{ "version": "1.0", "updated_at": "2026-10-15T08:41:33Z", "total_cases": 1274, "cases": [ { "id": "worker-ledger-freedom-dashboard", "title": "打工人小账本", "title_en": "Worker Ledger & Freedom Dashboard", "subtitle": "把工资、工时、支出和自由目标装进一个在线手账", "subtitle_en": "Wages, hours, spending, and freedom goals in one online journal", "description": "工资看起来还行，却很难回答自己一小时到底值多少钱、最近的钱花到哪里、离换一种活法还有多远；更别说好不容易记了一个月，换台设备打开又是空的。\n\n这个案例把真实时薪、10 秒记账、月度结余、自由基金、安全垫和情景模拟装进一个奶油薄荷手账风的在线工作台。它能把每笔支出换算成工作时间，首页即时展示最近 4 笔记录，用原生 SVG 绘制存款曲线；流水和个人设置存在资料库的在线数据表里，记完即时同步，换设备或换浏览器打开数据都在。\n\n适合：想做个人财务复盘的上班族、关注时间价值的自由职业者、需要多设备随时同步的用户、希望快速搭建轻量个人工作台的开发者与设计师。", "description_en": "A decent-looking salary still does not answer three practical questions: what is one hour of your life really worth, where has your money gone lately, and how far are you from choosing a different way to live? And after a month of diligent logging, opening the ledger on another device often means starting from scratch.\n\nThis case packages real hourly wage analysis, ten-second expense logging, monthly surplus, a freedom fund, safety-buffer milestones, and scenario simulations into an online workbench with a cream-and-mint journal aesthetic. Every expense is converted into work time, the latest four entries appear instantly on the dashboard, and savings and target lines are drawn with native SVG. Transactions and personal settings live in online library tables, syncing the moment you log something, so the data is still there on any device or browser.\n\nBest for: employees reviewing personal finances, freelancers tracking the value of time, people who need reliable multi-device sync, and developers or designers building a lightweight personal dashboard.", "categories": [ "workspace", "productivity", "finance", "data", "dev" ], "tags": [ "个人工作台", "真实时薪", "记账工具", "自由基金", "在线存储" ], "tags_en": [ "personal workbench", "real hourly wage", "expense tracking", "freedom fund", "online storage" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "打工人小账本（在线数据表存储，多端同步）" } ], "external_url": "", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/rYvVKPCRrbfSKmvekyothv），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n做一个名为「打工人小账本」的个人工作台，包含：真实时薪计算器（综合到手月薪、发薪月数、工作成本、在场时间、通勤和加班，并展开公式）；10 秒记账（金额、分类、备注，保存后立刻换算成工作时间）；首页最近 4 笔记录；月度收入与固定、弹性支出总结；自由基金目标与安全垫；原生 SVG 存款曲线；通勤、加班、涨薪情景模拟。\n\n数据要真正存在线上、不能只存在浏览器本地：请把流水记录和个人设置（时薪参数、自由基金目标等）分别建成对应的资料库数据表，页面读写都直接走线上数据，记一笔、改一笔、删一笔都即时写回，换设备或换浏览器打开数据都在；页面加载先渲染出界面再拉数据，读取失败或写入失败要有明确提示并可重试，不要白屏也不要静默丢数据。\n\n视觉采用精致可爱的奶油薄荷手账风，配统一的内联 SVG 图标和小猪存钱罐插画；桌面用侧栏，手机用底部 Tab，主要控件都适合触屏操作。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/rYvVKPCRrbfSKmvekyothv) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nUse the library (Workbuddy 资料库) capabilities to build me the same thing, with data stored online and synced both ways. Reference workbench: https://workbuddy.link/p/rYvVKPCRrbfSKmvekyothvCfhWZM Build a personal workbench called \"Worker's Little Ledger\" that includes: a real hourly-wage calculator (take-home monthly pay, number of pay months, work costs, hours on site, commute and overtime, with the formula expanded); 10-second expense logging (amount, category, note, instantly converted into working time); the four most recent entries on the home screen; a monthly summary of income and fixed versus flexible spending; a freedom-fund goal with a safety cushion; a native SVG savings curve; and what-if simulations for commute, overtime, and raises. The data must genuinely live online rather than only in browser storage: create separate library data tables for the transaction records and for personal settings (wage parameters, freedom-fund goal, and so on), have the page read and write directly against the online data, and persist every entry, edit, and deletion immediately so everything is still there on another device or browser. Render the interface first and then load the data; surface clear, retryable messages when a read or write fails, with no blank screens and no silently dropped records. Use a refined, cute cream-and-mint journal aesthetic with consistent inline SVG icons and a piggy-bank illustration; a sidebar on desktop, a bottom tab bar on mobile, and touch-friendly controls throughout.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "description_en": "Builds custom digital workspaces for students, professionals, creators, parents, and lifestyle users, with responsive HTML, local data persistence, and one-click deployment.", "expertType": "agent", "source": "enterprise", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 1, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-01T09:16:52Z", "updated_at": "2026-08-01T10:30:57Z", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario": "个人工作台", "scenario_id": 122, "task_mode": "日常办公", "scenario_use": "", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 1 }, { "id": "learning-goal-workspace", "title": "学习目标管理台", "title_en": "Learning Goal Workspace", "subtitle": "把有终点、有总量的学习目标拆成每天可完成的进度", "subtitle_en": "Turn finite learning goals into daily progress you can finish", "description": "“背完 2000 个单词”或“读完一本 440 页的书”最怕只剩一个截止日，却不知道今天该做多少、现在是提前还是拖后；更别说记了半个月，换台设备打开又是空的。\n\n这个案例演示了一个在线的学习目标管理台：根据剩余量和剩余天数计算今日建议量，用完成率、连续天数和预计完成日跟踪目标；支持每周一个弹性休息日、补打卡、逾期双出口、WOOP 障碍预案、近 14 天投入图和带环比的周报复盘。目标、打卡记录和周复盘存在资料库的在线数据表里，打完卡即时同步，换设备或换浏览器打开数据都在。\n\n适合：备考学生、读书计划执行者、在线课程学习者、技能训练者，以及所有需要管理“有终点、有总量”目标的人。", "description_en": "Goals such as learning 2,000 words or finishing a 440-page book often have only a deadline, leaving you unsure how much to do today or whether you are ahead or behind — and after two weeks of diligent logging, opening the tracker on another device often means starting over.\n\nThis case demonstrates an online learning-goal workspace that calculates today's suggested amount from remaining work and days, then tracks completion rate, streak, and estimated finish date. It supports one flexible rest day per week, backdated check-ins, two recovery actions for overdue goals, WOOP-style obstacle plans, a 14-day effort chart, and weekly reviews with week-over-week comparisons. Goals, check-in records, and weekly reviews live in online library tables, syncing the moment you check in, so the data is still there on any device or browser.\n\nSuitable for: exam candidates, reading-plan users, online-course learners, skill trainees, and anyone managing a finite goal with a measurable total.", "categories": [ "workspace", "dev" ], "tags": [ "目标管理", "学习计划", "进度追踪", "周报复盘", "在线存储" ], "tags_en": [ "Goal Management", "Study Planning", "Progress Tracking", "Weekly Review", "Online Storage" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "学习目标管理台（在线数据表存储，多端同步）" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/k9lbszKsR1DJh1rmiKUhcv），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n做一个学习目标管理台，专门管理有终点、有总量的学习目标，比如“背完 2000 个单词”或“读完一本 440 页的书”，不是每天喝水那种无限期习惯打卡。不需要我提供资料。\n\n每个目标包含名称、单位、总量、截止日、配色，以及两个选填字段：最容易拦住我的障碍和“如果它出现，我就……”的对策。每天自动计算今日建议量=剩余量÷剩余天数，打卡时预填但允许修改。每个目标显示完成率、连续打卡天数、预计完成日；预计完成日按近 7 天平均速度推算，与截止日比较后显示提前或拖后，样本不足时显示“暂无推算”，不要编数字。\n\n连续天数允许每周从周一开始自动使用 1 个休息日：本周第一次漏打卡不中断，第二次才断。昨日漏打卡的目标今天自动置顶，分别用温和黄卡或红卡解释休息日和滚入今日；逾期目标显示天数，并提供“立即打卡”和“调整计划”两个出口。卡片展示用户预案。一天可记录多次，支持选填分钟数和补记最近 6 天，补记带“补”标并按真实日期参与连续天数和推算。\n\n页面分四个 Tab：\n1. 今日：首屏显示新建目标按钮、今天要处理、可直接打卡的目标卡，列表底部再保留醒目的新建入口；支持编辑和二次确认删除。\n2. 看板：每个目标显示进度环、三个核心数字、预案和最近 10 条可删除记录；再用手写 SVG 绘制近 14 天投入分钟的分目标堆叠柱状图，未填分钟的记录不硬凑。\n3. 周报：按周一到周日汇总投入量、打卡天数和分钟数，附上周同口径环比；可查看历史周，并保存保持、问题、尝试、下周预案四栏；一键生成包含环比和四栏内容的周报文本。\n4. 我的：显示在线同步状态与最近一次同步时间、手动重新同步、清空示例、输入“清空”后清空全部，以及添加到手机主屏幕的三步说明。\n\n预置背英语单词、读《人类简史》、Python 入门课 3 个带示例标签的目标和几天记录，覆盖补记、休息日、连续漏打、障碍预案等状态。数据已在线保存，不需要提醒用户导出备份。\n\n数据要真正存在线上、不能只存在浏览器本地：请把学习目标、每日打卡记录、每周复盘分别建成对应的资料库数据表，页面读写都直接走线上数据，新建目标、打卡、补记、删除记录、保存周复盘都即时写回，换设备或换浏览器打开数据都在；页面加载先渲染出界面再拉数据，读取失败或写入失败要有明确提示并可重试，不要白屏也不要静默丢数据。\n\n视觉使用冷白底、紫色渐变、绿色成功、红色逾期、琥珀色休息日；目标达成时有彩带动效；图标全部使用内联 SVG，不用 emoji。手机底部四 Tab 和单列布局，平板顶部横向导航，电脑左侧固定边栏与多栏布局。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/k9lbszKsR1DJh1rmiKUhcv) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nUse the library (Workbuddy 资料库) capabilities to build me the same thing, with data stored online and synced both ways. Reference workbench: https://workbuddy.link/p/k9lbszKsR1DJh1rmiKUhcv1UnJ3K Build a learning-goal workspace specifically for finite, measurable goals such as learning 2,000 words or finishing a 440-page book, not endless habit tracking such as drinking water. No user-provided materials are needed. Each goal needs a name, unit, total amount, deadline, color, and two optional fields: the most likely obstacle and an If it happens, I will response. Calculate today's suggested amount as remaining amount divided by remaining days and prefill it during check-in while allowing edits. Show completion rate, streak, and estimated finish date. Estimate the finish date from the most recent seven-day average and compare it with the deadline to show days ahead or behind. When there is not enough activity, show No estimate instead of inventing a number. Allow one automatic rest day per Monday-to-Sunday week: the first missed day does not break the streak, while the second does. Put goals missed yesterday at the top today, using a gentle amber card when the rest day was used and a red card when work rolls into today. Show overdue days and provide Check in now and Adjust plan actions. Display the user's obstacle response on these cards. Allow multiple entries per day, optional minutes, and backdated entries within the previous six days. Mark backdated entries and use their real dates in streak and forecast calculations. Use four tabs. 1) Today: show New Goal in the first screen, a Today to Handle section, directly editable check-in cards, another prominent New Goal action at the bottom, plus edit and two-step delete. 2) Dashboard: show a progress ring, three key numbers, obstacle response, and the latest ten deletable records for each goal, plus a hand-written SVG 14-day stacked chart of minutes by goal without inventing minutes for entries that omitted them. 3) Weekly Review: summarize amount, active days, and minutes from Monday to Sunday, compare with the previous week using the same method, browse historical weeks, store Keep, Problem, Try, and Next-week response fields, and generate copyable review text containing comparisons and all four fields. 4) Me: export JSON, import after overwrite confirmation, migrate old backups, clear samples, clear all only after typing Clear, and show three steps for adding the page to the phone home screen. Include three sample goals—English vocabulary, reading Sapiens, and a Python introductory course—with sample labels and several days of records covering backdated entries, a rest day, consecutive misses, and an obstacle response. After 20 check-in records, show a gentle export-ready backup banner at the top. Use a cool-white background, purple gradient, green for success, red for overdue, and amber for rest days. Celebrate completed goals with confetti. Use inline SVG icons and no emoji. Use four bottom tabs and one column on mobile, horizontal top navigation on tablet, and a fixed left sidebar with multi-column content on desktop. The data must genuinely live online rather than only in browser storage: create separate library data tables for the goals, the daily check-in records, and the weekly reviews; have the page read and write directly against the online data, and persist every new goal, check-in, backdated entry, deleted record, and saved weekly review immediately so everything is still there on another device or browser. Render the interface first and then load the data; surface clear, retryable messages when a read or write fails, with no blank screens and no silently dropped records.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 2, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-01T18:50:21+08:00", "updated_at": "2026-08-01T18:50:21+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "preview_en": "output-en.html", "cover_image_en": "cover-en.png", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 5 }, { "id": "life-all-in-one-workbench", "title": "生活全能工作台", "title_en": "All-in-One Life Workbench", "subtitle": "把记账、习惯、健康、日程与精神生活装进一个在线工作台", "subtitle_en": "Manage money, habits, health, plans, and culture in one online hub", "description": "记账用一个 App、打卡用一个 App、日程和购物清单又散在别处，最烦的不是事情多，而是每天都要在不同工具之间来回切换；好不容易记了一个月，换台设备打开又是空的。\n\n这个案例把日常管理收进一个高级但易上手的在线工作台：包含今日概览、记账预算与消费结构、可自定义习惯和 30 天热力图、减脂趋势与周计划、日程统筹、待买清单、书影音收藏。数据存在资料库的在线数据表里，页面读写直接走线上、改完即时同步，换设备或换浏览器打开数据都在。\n\n适合：想统一管理个人生活的人；需要多设备随时同步的用户；希望快速搭个人仪表盘的产品与设计同学；想做数据驱动的在线单页应用的开发者。", "description_en": "Using one app for expenses, another for habits, and separate tools for schedules and shopping lists creates more friction than the tasks themselves. The real pain is constantly switching between scattered systems, and then finding an empty slate when you open things on another device.\n\nThis case brings everyday management into one polished yet approachable online workbench: a daily overview, budgeting and spending breakdowns, customizable habits with a 30-day heatmap, weight-loss trends and weekly plans, schedules, shopping lists, and a media library. Data lives in online library tables, so the page reads and writes against the cloud, syncs the moment you make a change, and is still there on any device or browser.\n\nGreat for: people who want one place to manage daily life; users who need reliable multi-device sync; product and design teams prototyping personal dashboards; developers building data-driven single-page apps.", "categories": [ "workspace", "productivity", "creation" ], "tags": [ "生活管理", "个人工作台", "在线存储", "双向同步", "习惯与健康" ], "tags_en": [ "life management", "personal workbench", "online storage", "two-way sync", "habits and health" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "生活全能工作台（在线数据表存储，多端同步）" } ], "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 3, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "usage_count": 0, "created_at": "2026-08-01T19:17:22+08:00", "updated_at": "2026-08-01T22:08:10+08:00", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/JlDtfA6n49ote46Jv3lHOn），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n做一个集成式的生活工作台，六个模块：今日概览；记账理财（收支记录、预算、月度对比、分类筛选、消费结构）；习惯健康（可自行新增删除习惯，支持勾选、计数、数值三种打卡方式和 30 天热力图）；减脂健身（体重体脂、7 日均线、BMI、目标进度、热量缺口、可自定义周计划）；日程统筹；待买清单；书影音收藏（状态、星级、短评、封面墙与列表两种视图、年度统计）。\n\n数据要真正存在线上、不能只存在浏览器本地：请按模块建好对应的资料库数据表，页面读写都直接走线上数据，新增、修改、删除即时写回，换设备或换浏览器打开数据都在；页面加载先渲染出界面再拉数据，读取失败或写入失败要有明确提示并可重试，不要白屏也不要静默丢数据。\n\n视觉保持克制、高级、有生活感，手机和电脑都能顺手用，避免千篇一律的模板化卡片风。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/JlDtfA6n49ote46Jv3lHOn) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nUse the library (Workbuddy 资料库) capabilities to build me the same thing, with data stored online and synced both ways. Reference workbench: https://workbuddy.link/p/JlDtfA6n49ote46Jv3lHOnDlxxvE Build an integrated life-management workbench with six modules: a daily overview; personal finance (income and expense records, budgets, monthly comparison, category filters, spending breakdown); habits and health (add or remove habits freely, with check, counter, and numeric tracking plus a 30-day heatmap); weight-loss and fitness (weight and body fat, 7-day average, BMI, goal progress, calorie deficit, customizable weekly plans); schedule planning; a shopping list; and a books/film/music library (status, star rating, short review, cover-wall and list views, annual statistics). The data must genuinely live online rather than only in browser storage: create the matching library data tables for each module, have the page read and write directly against the online data, and persist every add, edit, and delete immediately so everything is still there on another device or browser. Render the interface first and then load the data; surface clear, retryable messages when a read or write fails, with no blank screens and no silently dropped records. Keep the visual direction restrained, polished, and warm, comfortable on both phone and desktop, and avoid a generic AI-card template look.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 7 }, { "id": "home-renovation-workbench", "title": "装修工作台", "title_en": "Home Renovation Workbench", "subtitle": "施工进度、预算超支、灵感收藏一页管理，手机电脑都能用", "subtitle_en": "Track renovation progress, budget overruns, and inspiration in one page — works on mobile and desktop", "description": "装修最烦的不是花钱，是花了多少钱、进度到哪了、当初收藏的那些图长啥样，全都散落在好几个 APP 和聊天记录里，装修到一半根本理不清。\n\n这个案例演示了一套接资料库在线存储的装修工作台：首页顶部直接列出今天要处理的逾期和待办事项，八大施工阶段配甘特图和真实起止日期自动汇总进度；预算按项目名称、分类、施工阶段记明细，超支项会在首页和预算页高亮提醒；灵感库支持上传参考图、按风格筛选、写笔记。页面和资料库表格之间是双向绑定的——页面里的操作实时写回表格，表格里改的数据页面也能读到，未登录或访问不到资料库时自动降级为本机浏览器存储；支持导出 JSON 备份和导出 CSV 用 Excel 复盘，手机电脑都能用。\n\n适合：正在装修或即将装修的业主、需要盯多个施工方进度的人、想把预算和灵感都放在一个地方管理的人、不想为了记账装十几个 APP 的人。", "description_en": "Renovation stress rarely comes from the spending itself — it's that the budget, progress, and all those inspiration photos you saved end up scattered across a dozen apps and chat threads, impossible to track once the project is halfway done.\n\nThis case demonstrates a renovation workbench backed by an online database in the knowledge base: the homepage surfaces today's overdue and due items right at the top; eight construction phases are tracked with a Gantt chart and real start/finish dates that roll up into overall progress automatically; budget line items are recorded by project name, category, and phase, with overspending highlighted on both the homepage and budget page; an inspiration library lets you upload reference photos, filter by style, and jot down notes. The page and the database table are bidirectionally bound — actions on the page write back to the table in real time, and edits made in the table show up on the page too; it gracefully falls back to local browser storage if the knowledge base isn't reachable. JSON backup export and CSV export for Excel review are also included — it works on both mobile and desktop.\n\nBest for: homeowners renovating or about to renovate, anyone juggling progress across multiple contractors, people who want budget and inspiration in one place, and anyone who doesn't want to install a dozen apps just to track renovation spending.", "categories": [ "workspace", "productivity" ], "tags": [ "装修工作台", "施工进度", "预算管理", "灵感收藏", "资料库同步" ], "tags_en": [ "Renovation Workbench", "Construction Progress", "Budget Tracking", "Inspiration Board", "Database Sync" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整装修工作台单体 HTML" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/RQw9kjms5vngZMMGEwT8Zs），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：帮我搭建一个装修工作台，手机和电脑都能用，包含首页概览、施工进度、待办清单、预算管理、灵感库五个页面。首页顶部要有「今天要处理」区域，自动列出逾期和今天到期的事项，逾期项标红并给一键处理入口；下方是工程进度、实际花销、剩余预算、待处理事项四张数据卡，以及装修档案（风格、面积、户型、总预算、装修方式、开工日期、倒计时）。施工进度按设计、拆改、水电、泥木、油漆、安装、软装、完工八个阶段管理，每个阶段下面是具体施工小项，记录真实开始日和真实完成日（不是预计工期），阶段的整体进度和起止时间由小项的真实日期自动汇总，用甘特图展示各阶段时间线，今天所在位置要有一条竖线标出来。待办清单和施工小项是同一份数据：挂了施工阶段的就是工程任务，没挂阶段的是日常待办（比如「联系商家确认到货时间」），两处都能勾选完成、编辑、按优先级和分类筛选。预算管理按项目名称、预算分类、所属施工阶段记录每一笔明细（预算金额、实际花销、备注），自动算出总预算、已花销、结余、执行率，超支的项目要在首页和预算页都高亮出来，点击能看到超支明细列表；预算数据要能按分类和阶段两种维度切换查看。灵感库支持上传参考图片（自动压缩，最多存 30 张）、写风格标签和笔记，可以按风格筛选、标收藏；上传时能根据描述文字自动猜测风格标签。技术要求：单个完整 HTML 文件，CSS、JavaScript、SVG 图标和图表全部内联，零外部依赖；数据要接资料库 database 在线存储，页面和表格之间建立双向关联绑定——页面里的操作（勾选待办、编辑预算、上传灵感图）实时写回对应表格，表格里手动改的数据页面刷新后也要能读到最新结果；未登录或访问不到资料库时自动降级为 localStorage 本地模式，功能不缺；首屏必须有导出 JSON 备份和导出 CSV（可用 Excel 打开编辑再导入）两个入口，清空数据需要二次确认；首次打开要有几条示例数据（含施工进度、预算明细、灵感图），示例数据里要有 1 条逾期事项；手机端用底部 Tab 导航，电脑端用左侧栏导航，按钮点击区域不小于 44px，输入框字号不小于 16px 避免 iOS 自动放大。整体风格走沉稳的中古/侘寂调，暖色调、圆角卡片，具体配色和字体你可以自己发挥，不需要照搬某个特定风格库。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the knowledge base (资料库) capability to replicate this directly (https://workbuddy.link/p/RQw9kjms5vngZMMGEwT8Zs), with online data storage and bidirectional sync. Prefer the one-click replication path; if that capability is unavailable, build it as follows: Build me a home renovation workbench that works on both mobile and desktop, with five pages: Home, Construction Progress, To-dos, Budget, and Inspiration. The top of the homepage needs a \"Today's Focus\" section that automatically lists overdue and due-today items, with overdue ones highlighted in red and a one-click resolve action; below that, four data cards for construction progress, actual spending, remaining budget, and pending items, plus a renovation profile (style, floor area, layout, total budget, contract type, start date, countdown). Construction progress is managed across eight phases — design, demolition, plumbing/electrical, masonry, painting, installation, furnishing, and move-in — each phase containing specific line items that record real start and finish dates (not estimated durations); each phase's overall progress and date range should be automatically rolled up from its line items' real dates, displayed as a Gantt chart with a vertical line marking today's position. To-dos and construction line items share the same underlying data: items tied to a construction phase are project tasks, items without a phase are daily to-dos (e.g. \"confirm delivery time with vendor\"); both views should support checking off completion, editing, and filtering by priority and category. Budget tracking records each line item by project name, budget category, and construction phase (budgeted amount, actual spend, notes), automatically calculating total budget, amount spent, remaining balance, and execution rate; overspent items should be highlighted on both the homepage and budget page, with a click-through to see the list of overspent details; budget data should be viewable by both category and phase. The inspiration library supports uploading reference images (auto-compressed, up to 30 photos), tagging with style and notes, filtering by style, and marking favorites; when uploading, it should auto-guess a style tag based on the description text. Technical requirements: single complete HTML file with all CSS, JavaScript, SVG icons and charts inlined, zero external dependencies; data should be backed by an online database table in the knowledge base, with a bidirectional binding between the page and the table — actions on the page (checking off a to-do, editing a budget line, uploading an inspiration photo) write back to the table in real time, and edits made directly in the table should be reflected on the page after a refresh; if the knowledge base is unavailable or the user isn't logged in, gracefully fall back to local-only mode (localStorage) with no loss of functionality; the first screen must include both a \"Export JSON backup\" and \"Export CSV\" (editable in Excel and re-importable) entry point, with double confirmation required to clear data; on first open there should be a few sample records (including progress, budget items, and inspiration images) with one sample item deliberately overdue; use a bottom tab bar on mobile and a left sidebar on desktop, keep tap targets at least 44px and input font size at least 16px to avoid iOS auto-zoom. Overall style should lean toward a calm mid-century/wabi-sabi mood with warm tones and rounded cards — feel free to design the exact palette and typography yourself rather than copying a specific style library.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 4, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-12T15:00:00+08:00", "updated_at": "2026-08-12T15:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "shoufu-ledger", "title": "首富的记账本", "title_en": "Billionaire's Ledger", "subtitle": "一个月花光十个亿的记账工作台，轮盘抽花样、越花越有钱", "subtitle_en": "Spend a billion in 30 days — spin the wheel, trigger income storylines, sync to shared CSV", "description": "想体验「首富的钱花不完」？这个工作台让你和朋友们一起挑战一个月花光十个亿：转大轮盘决定花钱花样（包岛/包机/超跑车库…），每笔记账自动落盘到共享 CSV 表格，还有 40% 概率触发「越花越有钱」的意外回款剧情。适合聚会互动、直播演示、或者纯粹过过花钱的瘾。\n\n产物包含：首页进度看板（进度环 + 四宫格 + 支出/回款饼图）、记一笔表单（昵称 + 金额 + 分类 + 备注）、花钱大轮盘（16 个富豪玩法 + 朋友圈九图生成）、消费明细（筛选 + 甜甜圈图 + 冠军卡）。数据双向同步到资料库 database，多人共享同一张流水表。\n\n适合人群：想做趣味互动工作台的人、直播主播、活动策划、或者想体验「钱多到花不完」的人。", "description_en": "Ever wanted to spend a billionaire's money? This workspace lets you and friends race to blow through 1 billion in 30 days: spin a wheel for extravagant spending ideas (private islands, jets, supercar fleets...), every transaction syncs to a shared CSV, and there's a 40% chance of triggering \"the more you spend, the richer you get\" income storylines. Great for parties, livestreams, or just living the dream.\n\nIncludes: home dashboard (progress ring + stats + pie charts), add-entry form, spending wheel (16 ideas + social feed generation), and transaction log with filters. Data syncs bidirectionally to a shared database table.\n\nTarget audience: anyone wanting a fun interactive workspace, livestreamers, event planners, or those curious about the billionaire lifestyle.", "categories": [ "workspace", "finance" ], "tags": [ "记账工作台", "趣味互动", "轮盘抽奖", "多人共享", "CSV同步" ], "tags_en": [ "expense tracker", "interactive", "wheel spinner", "multiplayer", "CSV sync" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/NdX6fnH6IxEXbzBI03oh4F），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：主题：一个月和亲朋好友一起花光首富的十个亿，他不会破产——花得越狠，钱回来得越快。3 个模块：首页看板：败家进度环 + 四宫格（已挥霍/意外回款/还能造/日均得花）+ 支出/回款结构甜甜圈图 + 右侧最近记录时间线 记一笔：昵称 → 金额（带快捷按钮）→ 分类（14 种富豪消费分类 + 自定义，每种配马卡龙色系和 SVG 图标）→ 备注；提交后弹彩纸成功动画 + 第 N 笔 + 时间戳 花钱大轮盘：16 个真实富豪玩法（包岛/包机/超跑/游艇/私人烟花/买球队等，金额 50~800 万），转盘 4 秒停下后生成朋友圈九图 + 「记进账本」直接落账 回款剧情：每笔支出 40% 概率触发意外回款（16 个剧情事件，回款 = 支出 × 0.3~1.5 倍），带保底：连续 2 笔不给第 3 笔必给。数据接资料库 database：每笔收支落 CSV 表格（项目/收支/金额/分类/记录人/备注/记账时间），多人共享，15 秒轮询，页面和表格双向同步。消费明细支持滚动懒加载查看全部记录。未登录自动降级本地模式。界面：马卡龙柔和配色、1px 描边圆角卡片、 SVG 图标和图表（不用 emoji/Chart.js）。桌面三列布局，移动端底部 Tab 栏 + 首页只留任务卡。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/NdX6fnH6IxEXbzBI03oh4F) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch: Theme: spend 1 billion in 30 days with friends. He never goes broke — the more you spend, the faster money comes back. 3 modules: Home dashboard: progress ring + 4-stat grid (spent/income/remaining/daily) + expense/income donut charts + activity timeline. Add entry: nickname → amount (quick buttons) → category (14 billionaire types + custom, each with macaron color + SVG icon) → note; confetti success animation with streak count. Spending wheel: 16 real billionaire activities (islands/jets/supercars/yachts etc, 50万~800万), 4s spin, 9-grid social feed images, record button logs directly. Income storyline: 40% chance per expense (16 events, amount = expense × 0.3~1.5), pity: 2 misses → guaranteed. Data: Library database table (project/type/amount/category/person/note/time), multiplayer, 15s polling, bidirectional CSV sync, scroll lazy-loading. Anonymous → localStorage fallback. Visual: macaron palette, 1px borders, inline SVG icons & charts (no emoji/Chart.js). Desktop 3-column, mobile bottom tabs + home shows only task card.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "description": "个人数字工作台搭建专家，擅长为不同人群定制专属网页版工作台", "description_en": "Digital workspace builder — crafts personalized HTML workspaces for different audiences", "expertType": "agent", "source": "enterprise", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 5, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "个人工作台", "scenario_id": 122, "task_mode": "日常办公", "usage_count": 0, "artifacts": [ { "file": "output.html", "label": "首富的记账本 HTML 工作台" } ], "created_at": "2026-08-12T13:00:00+08:00", "updated_at": "2026-08-12T13:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "fat-loss-tracker", "title": "轻盈计划 · 减脂健身追踪台", "title_en": "LeanPlan · Fat-Loss Tracker", "subtitle": "单文件中英双语离线减脂追踪台，体重体脂趋势一目了然", "subtitle_en": "Single-file bilingual offline fat-loss tracker — weight & body-fat trends at a glance", "description": "减脂最头疼的就是每天记了体重却不知道趋势走向，热量吃多吃少全靠猜。\n\n这个案例产出了一个完整的单文件 HTML 减脂健身追踪台：每日记录体重和体脂后自动生成 Canvas 趋势折线图与 7 日移动平均线；用 Mifflin-St Jeor 公式估算个人 BMR/TDEE 并按 500–750 kcal 热量缺口给出饮食目标提醒；包含周训练计划打卡、习惯打卡、目标进度环、BMI 计算、数据导出 Excel/CSV 与 JSON 备份导入导出。视觉采用暖灰纸张底 + 深森林绿主色 + 陶土橙点缀的 Editorial 风格，开机自动用你的名字打招呼，支持自定义主题色（不改字号），手机电脑都能打开，完全离线可用。页面内置中英双语 I18N，默认中文，访问 ?lang=en 即切换为完整英文界面。\n\n适合：正在减脂塑形并希望系统化记录进度的个人用户；需要离线可用的健康追踪工具的用户；想看一个「单文件 HTML + 原生 Canvas 图表」完整实现的前端开发者。", "description_en": "The hardest part of fat loss is logging your weight every day without knowing where the trend is heading — and guessing whether you're eating too much or too little.\n\nThis case delivers a complete single-file HTML fat-loss fitness tracker: after daily weight and body-fat entries, it auto-generates Canvas trend line charts with a 7-day moving average; uses the Mifflin-St Jeor formula to estimate personal BMR/TDEE and provides diet target reminders based on a 500–750 kcal deficit; includes weekly workout plan check-ins, habit tracking, goal progress rings, BMI calculation, Excel/CSV export and JSON backup/import. The visual design uses warm gray paper + deep forest green primary + terracotta orange accents in an editorial style. It greets you by name on launch, supports custom theme colors (without changing font size), works on both phone and desktop, and runs fully offline. It includes built-in Chinese and English I18N: Chinese by default, with the full English interface available via ?lang=en.\n\nSuitable for: individuals who want systematic progress tracking during fat loss; users who need an offline-capable health tracker; frontend developers interested in a complete \"single-file HTML + native Canvas chart\" implementation.", "categories": [ "workspace", "data", "productivity" ], "tags": [ "减脂追踪", "体重管理", "趋势图表", "中英双语", "离线HTML" ], "tags_en": [ "Fat-loss tracking", "Weight management", "Trend charts", "Bilingual I18N", "Offline HTML" ], "artifact_type": "html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/0v5TZ9MYkGDzrXY51ddLEA），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个个人减脂健身追踪台，要求：所有 CSS/JS/可离线打开、手机和电脑都能用。功能包括：① 每日记录体重与体脂（含示例数据、可清空）；② 体重/体脂趋势折线图 + 7 日移动平均；③ 目标进度（当前体重 → 目标体重）与 BMI；④ 周训练计划打卡与习惯打卡；⑤ 用 Mifflin-St Jeor 公式估算 BMR/TDEE，并按每日 500–750 kcal 热量缺口给出饮食与蛋白/碳水/脂肪目标提醒；⑥ 顶部每日问候与下一步建议卡片；⑦ 数据云端同步（资料库 database 多表关联：身体记录、饮食记录、训练计划、训练日志、习惯打卡、每周复盘、日记）；⑧ 自定义主题色与按钮色（不改字号）。视觉风格：暖灰纸张底 + 深森林绿主色 + 陶土橙点缀的 Editorial 风，开机用我的名字打招呼。请用原生 Canvas 自绘图表，不要引用任何外部 CDN 或字体。\n\n交付：资料库在线工作台——在线 HTML 页面 + 8 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/0v5TZ9MYkGDzrXY51ddLEA) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a personal fat-loss tracker: works offline on mobile and desktop. Features: ① daily weight & body-fat logging; ② weight/body-fat trend line chart + 7-day moving average; ③ goal progress & BMI; ④ weekly workout plan check-in & habit tracking; ⑤ Mifflin-St Jeor BMR/TDEE estimation with 500–750 kcal deficit diet targets (protein/carb/fat); ⑥ daily greeting & next-step suggestion card; ⑦ cloud data sync via Library database (multi-table: body records, meals, workout plans, workout logs, habits, weekly reviews, diary); ⑧ customizable theme colors. Visual: warm gray paper + deep forest green + terracotta orange editorial style. Use native Canvas charts, no external CDN.\n\nDeliverable: an online Library workbench — an online HTML page plus 8 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 6, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "usage_count": 0, "created_at": "2026-08-01T11:26:59Z", "updated_at": "2026-08-01T14:23:34Z", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "减脂健身追踪台（单文件中英双语 HTML）" } ], "preview_en": "output.html", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "aihot-daily-briefing", "title": "今日 AI 日报一键生成", "title_en": "AI Daily Briefing Generator", "subtitle": "一句话生成可交互 HTML AI 晨报，把当日 AI 精选整理成五版块新闻仪表盘", "subtitle_en": "Generate an interactive HTML AI morning briefing that turns today's AI highlights into a 5-section news dashboard", "description": "调用 AI HOT 公开 API 拉取最新一期日报，按「模型发布 / 产品发布 / 行业动态 / 论文研究 / 技巧与观点」五大固定版块整理成响应式 HTML 晨报仪表盘。页面包含 Hero 数据概览、版块锚点导航、资讯卡片网格与原文跳转，适合直接投屏晨会、发公司群、做公众号选题池。无需配置 API Key，触发关键词：AI 日报 / 今天 AI 圈 / AI HOT。\n\n适合：产品经理、技术 Leader、AI 创业者、内容创作者、任何想快速了解当天 AI 圈发生了什么的人", "description_en": "Calls the AI HOT public API to fetch the latest daily briefing, then turns it into a responsive HTML morning dashboard organized into five sections: Model Releases, Product Launches, Industry News, Research Papers, Tips & Opinions. Includes a hero overview, section anchors, news cards, and source links — ready for standups, team sharing, or newsletter planning. No API key needed.\n\nBest for: product managers, tech leads, AI entrepreneurs, content creators, anyone wanting a quick daily AI digest", "categories": [ "search", "office" ], "tags": [ "AI 资讯", "日报", "晨会素材", "AI HOT", "公众号" ], "tags_en": [ "AI News", "Daily Briefing", "Morning Standup", "AI HOT", "Newsletter" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我想要今天的 AI 日报，请生成一个单文件 HTML 晨报仪表盘。\n\n要求：\n1. 调用 aihot skill 的日报接口拉数据；若当日尚未生成则回退到最近一期\n2. 按五个固定版块分组：模型发布/更新、产品发布/更新、行业动态、论文研究、技巧与观点\n3. 全局连续编号（不在版块内重新计数）\n4. 顶部 Hero 展示日期、总条数与五版块统计；中部锚点导航；正文用响应式卡片网格\n5. 每张卡片含：序号、标题、来源 chip、≤60 字中文摘要、原文跳转（target=\"_blank\" rel=\"noopener noreferrer\"）\n6. 时间转成北京时间人话格式，不展示 ISO 字符串\n7. 文末标注总条数与数据源\n8. 纯 HTML/CSS/JS 单文件，样式与脚本内联，无外部资源；响应式布局", "prompt_en": "I want today's AI daily briefing. Please generate a single-file HTML morning-briefing dashboard.\n\nRequirements:\n1. Call the aihot skill's daily-briefing API to pull data; if today's edition isn't ready yet, fall back to the most recent one\n2. Group into five fixed sections: Model Releases/Updates, Product Releases/Updates, Industry News, Papers & Research, Tips & Opinions\n3. Continuous global numbering (do not restart within sections)\n4. A top Hero showing the date, total count, and the five-section stats; anchor navigation in the middle; a responsive card grid for the body\n5. Each card includes: sequence number, title, source chip, a Chinese summary <=60 chars, and a source link (target=\"_blank\" rel=\"noopener noreferrer\")\n6. Convert times into human-friendly Beijing time; do not show raw ISO strings\n7. Note the total count and data source at the bottom\n8. A single pure HTML/CSS/JS file with inlined styles and scripts, no external resources; responsive layout", "skills": [ { "id": "aihot", "name": "AI HOT", "description": "AI 资讯聚合，提供日报、论文、公司动态追踪", "skill_id": "skill_2053079401717506048" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 7, "is_featured": true, "featured_order": 3, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-09T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "AI 日报 HTML 看板" } ], "is_recommended": true, "recommended_order": 10 }, { "id": "team-okr-radar-chart", "title": "团队 OKR 雷达图", "title_en": "Team OKR Radar Chart", "subtitle": "多维度目标达成可视化，一图看清团队健康度", "subtitle_en": "Multi-dimensional OKR visualization — see team health at a glance", "description": "每次 OKR 复盘只盯着完成百分比，却看不出团队能力短板在哪？六七个维度的表现高低，光靠数字表格根本比不出全貌，更别提跨团队横向对比了。\n\n这个案例用图表可视化技能生成了一张交互式团队 OKR 雷达图：6 个维度（产品交付、技术质量、用户增长、团队协作、创新能力、客户满意度）同时呈现目标值与实际值，两层多边形叠加一眼看出差距；支持产品团队、技术团队、设计团队三组数据一键切换；侧边栏自动汇总整体完成率、最强/最弱维度、环比趋势；颜色编码直观标记超额达成（绿）、接近目标（黄）、明显不足（红），纯 HTML/CSS/SVG 实现，响应式布局。\n\n适合：团队 Leader、HR、项目经理、CTO，以及任何需要在周会/月会/季度复盘中快速呈现团队多维表现的人", "description_en": "Tired of OKR reviews that only show completion percentages without revealing where your team's capability gaps actually are? Comparing performance across six or seven dimensions with spreadsheets alone never tells the full story — let alone cross-team benchmarking.\n\nThis case uses the chart-visualization skill to generate an interactive Team OKR Radar Chart: 6 dimensions (Product Delivery, Tech Quality, User Growth, Collaboration, Innovation, Customer Satisfaction) displaying both target and actual values as overlapping polygons for instant gap analysis; toggle between Product, Engineering, and Design teams with one click; a side panel auto-summarizes overall completion rate, strongest/weakest dimensions, and quarter-over-quarter trends; color-coded indicators mark exceeding target (green), close (yellow), and significantly below (red). Pure HTML/CSS/SVG, fully responsive.\n\nIdeal for: Team leads, HR, project managers, CTOs, and anyone who needs to present multi-dimensional team performance in weekly/monthly/quarterly reviews", "categories": [ "data", "office" ], "tags": [ "OKR", "雷达图", "团队管理", "数据可视化", "绩效分析" ], "tags_en": [ "OKR", "Radar Chart", "Team Management", "Data Visualization", "Performance Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个团队 OKR 雷达图页面（单文件 HTML，内联无外部依赖），响应式。\n\n展示 6 个维度（目标值 vs 实际值）：产品交付 90/85、技术质量 85/92、用户增长 80/68、团队协作 85/88、创新能力 75/72、客户满意度 90/86。\n目标值用虚线多边形、实际值用半透明填充多边形叠加；支持产品/技术/设计三组团队切换；侧栏显示整体完成率、最强/最弱维度、环比趋势；达成情况按超额/接近/不足用颜色区分。", "prompt_en": "Please generate a team OKR radar chart page (single-file HTML, inline with no external dependencies), responsive.\n\nDisplay 6 dimensions (target value vs. actual value): Product Delivery 90/85, Technical Quality 85/92, User Growth 80/68, Team Collaboration 85/88, Innovation 75/72, Customer Satisfaction 90/86.\nUse a dashed polygon for the target values and a semi-transparent filled polygon overlaid for the actual values; support switching among three teams: product / engineering / design; the sidebar shows overall completion rate, strongest/weakest dimensions, and month-over-month trend; distinguish achievement status by color according to exceeded / close / insufficient.", "artifacts": [ { "file": "output.html", "label": "OKR雷达分析图" } ], "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表", "description": "使用 Mermaid 语法生成流程图、时序图、甘特图等可视化图表", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 8, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-22T09:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "dali-guesthouse-search", "title": "大理古城民宿推荐", "title_en": "Dali Old Town Guesthouse Recommendations", "subtitle": "去哪儿+携程搜大理古城3人民宿，带院子和苍山景观的精选房源", "subtitle_en": "Search Qunar & Ctrip for 3-guest guesthouses in Dali Old Town with courtyards and mountain views", "description": "去大理旅游，住民宿比住酒店更有感觉，但平台上选项太多，好评多的不一定有特色，便宜的又怕条件差。\n\n这个案例演示了如何在携程、去哪儿搜出符合条件的大理古城民宿：指定人数、最低评分、区域偏好，AI 返回价格从低到高的房源列表，并自动标出带院子、有苍山景观的特色房源，附上古城内/洱海边/双廊三区域住宿特点对比。\n\n适合：计划云南自由行的旅行者、偏好民宿体验的用户、想找性价比住宿的3人出行小团体", "description_en": "Staying in a guesthouse in Dali beats a hotel for atmosphere — but there are hundreds of options and filtering through them takes forever.\n\nThis case shows how to search for Dali guesthouses via Qunar and Ctrip in one step: specify guest count, minimum rating, and location preference, and AI returns a price-sorted list with courtyard and mountain-view properties highlighted, plus a comparison of three main areas: Old Town, Erhai Lake, and Shuanglang.\n\nBest for: Yunnan self-guided travelers, guesthouse enthusiasts, groups of 3 looking for value accommodation", "categories": [ "travel" ], "tags": [ "民宿", "大理", "云南旅行" ], "tags_en": [ "Guesthouse", "Dali", "Yunnan Travel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "民宿推荐交互页面" } ], "skills": [ { "id": "travel-cn", "name": "travel-cn", "description": "聚合去哪儿、携程、飞猪，查机票、酒店、火车票、景点门票", "description_en": "Search flights, hotels and train tickets via Qunar, Ctrip and Fliggy", "skill_id": "skill_2053083000434532352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "帮我搜大理古城3人民宿，要求评分4.5以上，优先推荐带院子和苍山景观的房源，按价格从低到高排列，并对比古城内、洱海边、双廊三个区域的住宿特点。", "prompt_en": "Help me search for guesthouses in Dali Ancient Town for 3 people, requiring a rating above 4.5, prioritizing listings with a courtyard and views of Cangshan Mountain, sorted by price from low to high. Also compare the lodging characteristics of three areas: inside the Ancient Town, by Erhai Lake, and in Shuanglang.", "source": "official", "sort_order": 9, "featured_order": 0, "quality_score": 90, "usage_count": 0, "is_featured": false, "created_at": "2026-04-28T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "realtime-whiteboard-prototype", "title": "实时协作白板原型", "title_en": "Real-Time Collaborative Whiteboard Prototype", "subtitle": "便签拖拽 × 画笔绘制 × 连线标注，类 Miro 协作白板一键生成", "subtitle_en": "Sticky notes, freehand drawing & connectors — a Miro-like whiteboard in one click", "description": "团队头脑风暴想用白板，但 Miro/FigJam 要付费、内网又访问不了，临时搞个协作界面更是无从下手。这个案例演示了一个类 Miro 的协作白板原型：支持便签创建与拖拽、画笔自由绘制、元素间连线、无限画布缩放平移，整体交互流畅，设计风格干净现代。\n\n适合：产品经理做协作功能原型验证、前端工程师学习 Canvas/SVG 交互开发、SaaS 团队快速搭建白板 Demo。", "description_en": "Teams need a whiteboard for brainstorming, but Miro/FigJam require paid plans and are often inaccessible on internal networks — building a quick collaborative interface from scratch feels daunting. This case demonstrates a Miro-like collaborative whiteboard prototype: sticky note creation & dragging, freehand drawing, connector lines between elements, infinite canvas with zoom & pan, smooth interactions, and a clean modern design.\n\nIdeal for: product managers validating collaboration features, frontend engineers learning Canvas/SVG interaction development, SaaS teams rapidly building whiteboard demos.", "categories": [ "creation", "dev" ], "tags": [ "协作白板", "Canvas交互", "原型设计", "团队协作" ], "tags_en": [ "Collaborative Whiteboard", "Canvas Interaction", "Prototyping", "Team Collaboration" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "协作白板原型页面" } ], "prompt": "请创建一个类似 Miro 的实时协作白板原型页面（单文件 HTML，内联无外部依赖，用 Canvas 或 SVG 绘制），深色顶栏 + 白色画布 + 浅灰网格，干净现代。\n\n功能：1) 无限画布，鼠标拖拽平移 + 滚轮缩放；2) 便签（点击添加、拖拽移动、双击编辑、右键删除、多色切换）；3) 画笔（自由绘制，可调粗细颜色）；4) 连线工具；5) 选择工具（选中显示操作手柄）；6) 顶部工具栏；7) 右下角迷你地图。", "prompt_en": "Please create a Miro-like real-time collaborative whiteboard prototype page (single-file HTML, inlined with no external dependencies, drawn with Canvas or SVG), with a dark top bar + white canvas + light-gray grid, clean and modern.\n\nFeatures: 1) Infinite canvas with mouse-drag panning + scroll-wheel zoom; 2) Sticky notes (click to add, drag to move, double-click to edit, right-click to delete, multi-color switching); 3) Pen (freehand drawing with adjustable thickness and color); 4) Connector tool; 5) Selection tool (shows manipulation handles when selected); 6) Top toolbar; 7) Mini-map in the bottom-right corner.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "全栈前端开发，融合高端 UI 设计与动画效果", "skill_id": "skill_2052764137292271616" }, { "id": "impeccable", "name": "高品质前端设计", "description": "高品质 UI/UX 设计工具集，避免泛 AI 审美", "skill_id": "skill_2053082862415904768" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 10, "is_featured": true, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-05-02T14:37:22Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 2, "is_recommended": true, "recommended_order": 8 }, { "id": "saas-contract-comparison-report", "title": "SaaS 服务协议差异报告", "title_en": "SaaS Agreement Difference Report", "subtitle": "真实对比 8 处差异，聚焦数据、服务与责任变更", "subtitle_en": "A real eight-point comparison focused on data, service, and liability changes.", "description": "合同来回修订后，最头疼的是核心权利义务到底改了什么。\n\n本案例真实调用腾讯电子签合同对比能力，对两版 SaaS 服务协议输出 8 处差异、分类摘要，以及模拟 Word 修订体验的原文差异对照与本地下载文件。\n\n适合：SaaS 商务负责人、法务、采购人员、参与版本谈判的业务团队。", "description_en": "Contract revisions often make it hard to tell which core rights and obligations changed.\n\nThis case uses Tencent eSign contract comparison to show eight real differences, a classified summary, a Word-style tracked-change comparison of source clauses, and bundled local downloads.\n\nSuitable for: SaaS business leads, legal teams, procurement staff, and negotiation teams.", "categories": [ "business" ], "tags": [ "合同对比", "SaaS", "版本谈判" ], "tags_en": [ "Contract comparison", "SaaS", "Version negotiation" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "", "prompt": "请调用腾讯电子签合同法务专家的合同对比能力，对比以下两份合同：原版：[上传原版合同]；新版：[上传新版合同]。请按新增、修改、删除分类列出实质性变更，重点分析价款、服务范围、数据安全、知识产权、责任限制、违约责任及争议解决条款，并给出在线对比页面和可下载的差异明细。", "prompt_en": "Please use the Tencent eSign Contract Legal Expert comparison capability to compare two contracts: original: [upload original contract]; revised: [upload revised contract]. Classify substantive changes as additions, modifications, and deletions. Focus on price, service scope, data security, intellectual property, liability caps, breach, and dispute resolution, and provide an online comparison page and downloadable difference details.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "面向合同起草、审查、对比、签署与法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 10, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T21:46:05+08:00", "updated_at": "2026-08-13T13:39:33+08:00", "artifacts": [ { "file": "output.html", "label": "互动灵感卡片 HTML" }, { "file": "SaaS服务协议对比批注报告.pdf", "label": "对比批注报告 PDF" }, { "file": "SaaS服务协议对比差异明细.xlsx", "label": "对比差异明细 Excel" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "sancai-duck-incense-lineart", "title": "三彩鸭形香薰 · 线稿提取", "title_en": "Sancai Duck Incense Burner · Line-Art Extraction", "subtitle": "一张文物照，秒变可编辑的矢量单线稿", "subtitle_en": "Turn a relic photo into an editable single-line vector sketch", "description": "想拿文物做二创，却卡在第一步——照片没法直接编辑，手动描线又慢又不准。\n\n这个案例把「明成化素三彩鸭形香薰」的照片交给专家，识别主体造型、结构转折与关键纹样，输出干净精简、每笔都是可编辑贝塞尔路径的透明背景 SVG 单线稿，可直接在 Illustrator / Figma / Inkscape 里调锚点、改线宽。\n\n适合：文创设计师、文博 IP 开发者、插画与视觉从业者、博物馆数字化团队", "description_en": "This case hands a photo of the Chenghua Sancai duck incense burner to the expert, which recognizes the silhouette, structural transitions and key motifs, then outputs a clean transparent-background SVG single-line sketch where every stroke is an editable Bezier path — ready to tweak in Illustrator / Figma / Inkscape.\n\nGood for: Cultural-creative designers, Museum IP developers, Illustrators & visual creators, Museum digitalization teams", "categories": [ "creation" ], "tags": [ "文物矢量", "SVG线稿", "纹样提取", "文创设计" ], "tags_en": [ "Relic Vector", "SVG Line Art", "Motif Extraction", "Cultural Creative" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据这张文物图片，准确识别文物主体的整体造型、朝向、比例、结构转折与关键装饰特征，将其提炼为干净、精简、可编辑的 SVG 矢量单线稿。严格保持原文物的外形与辨识度，不得改变比例、重构造型、随意补全被遮挡区域或添加原图不存在的纹样；去除背景、底座以外的环境元素、摄影光影、材质色彩、锈迹、污渍、裂纹、颗粒和过于细碎的表面纹理，只保留主体的重要外轮廓、结构分界及少量具有识别作用的内部纹样。\n\n线稿必须采用真正的中心线矢量结构：每一笔均为单条开放或闭合的贝塞尔路径，由多个可编辑锚点连接，统一使用黑色描边、统一线宽、圆角转折和圆形端点，所有路径均设置为无填充。优先以连续、流畅的长曲线表现轮廓，主动合并方向一致的短线，减少断点、毛刺、小疙瘩、重复线、交叠线和不必要的锚点；曲线应自然顺滑，锚点数量精简但足以准确还原造型。\n\n最终输出透明背景的标准 SVG 文件，画面居中并保留适当留白。SVG 内不得嵌入原始位图、白色背景矩形、蒙版式伪矢量或由内外两圈边界构成的黑色填充轮廓；所有可见线条必须能够在 Illustrator、Figma、Inkscape 等软件中被单独选中、编辑贝塞尔锚点，并可统一调整描边颜色与线宽。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-vector-element-expert", "name": "文物矢量元素提取专家", "description": "将文物图像精准转译为可编辑的 SVG 矢量线稿、平面色块、轮廓线与代表色色卡", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uzh9tCiVNo2A" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 10, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "wes-anderson-short-film", "title": "韦斯·安德森风格短片", "title_en": "Wes Anderson–Style Short Film", "subtitle": "用对称构图和复古配色讲完一场礼宾员的秘密交接", "subtitle_en": "A secret handover told through symmetry and vintage color", "description": "有明确故事点子，却很难把美术风格、镜头节奏和声音统一成一支完整短片。\n\n案例以粉色老旅店为舞台，让年迈礼宾员在值夜时发现一把金色钥匙，并完成一场克制又荒诞的秘密交接；画面强调正面机位、中心对称、章节字幕、模型式横移和复古暖色。\n\n适合：品牌故事、酒店与文旅内容、风格化剧情短片，以及希望从故事梗概直接走到成片的创作者。", "description_en": "A clear story idea is often not enough to keep art direction, pacing, and sound coherent across a finished short film.\n\nThis case takes place in an old pink hotel, where an elderly night concierge discovers a golden key and completes a restrained, slightly absurd secret handover. Frontal framing, centered symmetry, chapter cards, model-like lateral moves, and warm vintage colors shape the film.\n\nBest for brand stories, hospitality and travel content, stylized narrative shorts, and creators who want to turn a premise into a finished film.", "categories": [ "creation" ], "tags": [ "电影感", "对称构图", "复古短片" ], "tags_en": [ "cinematic", "symmetry", "vintage short" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「电影感创作」专家调用 Flova 的「韦斯·安德森风格短片」Skill，制作一支约 1 分钟、16:9 的原创叙事短片《粉色旅店的金钥匙》。故事发生在一间粉色老旅店：年迈礼宾员穿酒红制服值夜，听见铜铃后从木质钥匙墙取下一把装饰繁复的金钥匙，交给一位始终不露正脸的客人；客人离开后，礼宾员发现钥匙又回到了原位。请先给出全局规格和 3 个章节的叙事结构，确认后设计综合色板、角色与场景、关键视觉和完整分镜；每个关键阶段都停下让我确认，再生成镜头、旁白或对白、环境音、复古室内乐并完成组装。视觉重点是正面机位、中心对称、克制表演、模型式横移、粉红与酒红配色；不要使用现成影视角色、品牌或受保护标识。", "prompt_en": "Ask the “Flova Cinematic Director” expert to invoke Flova’s “Wes Anderson–Style Short Film” Skill and create an original 16:9 narrative short of about one minute titled The Golden Key at the Pink Hotel. In an old pink hotel, an elderly concierge in a burgundy uniform hears a brass bell, removes an ornate golden key from a wooden key wall, and hands it to a guest whose face is never shown. After the guest leaves, the concierge discovers that the key has returned to its original hook. First present the global specification and a three-chapter narrative structure. After approval, design the unified color palette, characters and locations, key visuals, and full storyboard. Pause for my approval at every major stage before generating shots, narration or dialogue, ambience, vintage chamber music, and the final assembly. Emphasize frontal framing, centered symmetry, restrained acting, model-like lateral camera moves, and pink-and-burgundy colors. Do not use existing film characters, brands, or protected marks.", "skills": [], "experts": [ { "id": "flova-cinematic-director", "name": "电影感创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及电影感成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e0b061a92ad1", "expert_id": "ex_oQGh7uNdnrp6" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 10, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "deepseek-founder-quotes-sketch", "title": "梁文锋投资者交流会金句手绘", "title_en": "Liang Wenfeng Investor Q&A · Quotes Hand-Drawn", "subtitle": "把一场访谈的 18 句金句，做成铅笔手绘 + 语境解读的极简长图", "subtitle_en": "Turn 18 quotes from an interview into a minimalist longform of pencil sketches and context", "description": "读完一篇长访谈，最想留下的其实是那几句戳中人的话——但干巴巴摘录金句既没氛围也不好看，想配图又怕做成廉价拼贴。\n\n这个案例演示了如何把一场约 3 小时 44 分钟的创始人投资者交流会，提炼成 18 句金句，每句配一幅由 AI 依据金句意象生成的铅笔素描手绘图（双手捧种子、巨浪前静立的小人、崩塌的城墙与芯片新芽……），再补上回到原话的一段语境解读，最终拼成一份极简留白、衬线大字、红色引导线的单文件 HTML 长图，可离线打开、直接转发。\n\n适合：想把访谈/播客/长文金句做成可传播长图的内容创作者、公众号与新媒体编辑、知识博主、做读书笔记与观点卡片的深度阅读者。", "description_en": "After finishing a long interview, what you really want to keep are the few lines that hit home — but a dry list of quotes looks flat, and adding images risks cheap collages.\n\nThis case shows how to distill a ~3h44m founder investor Q&A into 18 key quotes, pair each with an AI-generated pencil-sketch illustration drawn from the quote's imagery (cupped hands holding a sprout, a small figure standing calm before a giant wave, a crumbling wall beside a chip sprout...), add a paragraph of original context, and assemble everything into a single-file HTML longform with minimalist whitespace, large serif type and a red accent rule — offline-ready and shareable.\n\nGreat for: content creators turning interviews/podcasts/long reads into shareable longforms, WeChat and new-media editors, knowledge bloggers, and deep readers making note cards.", "categories": [ "creation", "education" ], "tags": [ "金句手绘", "信息图", "内容创作", "长图排版", "AI 配图" ], "tags_en": [ "quote sketch", "infographic", "content creation", "longform layout", "AI illustration" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "梁文锋投资者交流会金句手绘" } ], "prompt": "我有一篇长访谈/播客的文字稿（见下方），帮我做成一份「金句 · 手绘 · 解读」的极简长图 HTML。\n\n第一步，通读全文，提炼出 15-18 句最有代表性的金句，每句给出：编号、所属主题（如「关于初衷」「关于克制」）、金句原话（尽量用原文）、以及一段回到语境的解读（说明这句话的背景和为什么重要）。\n\n第二步，为每一句金句用手绘信息图能力生成一幅配图：铅笔素描风、纯白纸背景、大量留白、单一诗意意象、黑白灰、无文字无颜色（文字交给排版负责），为每句金句设计一个贴切的视觉隐喻（如「舍弃芝麻换西瓜」画成手越过芝麻伸向远处西瓜）。竖版 1024x1536。\n\n第三步，组装成单文件 HTML 长图：整体极简、大量留白、衬线大字、浅色主题，卡片式图文分区（配图 + 编号主题 + 大字金句 + 解读），开头一个含标题/副标题/来源信息的封面区。真产品约束：单文件、所有图片以 base64 内联、无任何外部资源引用、可离线打开。\n\n访谈文字稿：[在此粘贴你的访谈/播客/长文文字稿]", "prompt_en": "I have a transcript of a long interview/podcast (below). Help me turn it into a minimalist \"Quotes · Sketch · Context\" longform HTML.\n\nStep 1: Read the whole transcript and distill 15-18 of the most representative quotes. For each, give: a number, a theme (e.g. \"On the original intent\", \"On restraint\"), the original quote (keep the original wording), and a paragraph of context explaining the background and why it matters.\n\nStep 2: For each quote, use the hand-drawn infographic capability to generate one illustration: pencil-sketch style, pure white paper background, abundant whitespace, a single poetic image, monochrome grayscale, no text and no color (text is handled by layout). Design a fitting visual metaphor per quote (e.g. \"give up sesame for the watermelon\" as a hand passing sesame seeds and reaching for a distant watermelon). Portrait 1024x1536.\n\nStep 3: Assemble into a single-file HTML longform: overall minimalist, abundant whitespace, large serif type, light theme, card-based image/text sections (illustration + numbered theme + large quote + context), starting with a cover area containing title/subtitle/source info. Real product constraints: single file, all images inlined as base64, no external resource references, openable offline.\n\nTranscript: [paste your interview/podcast/long-read transcript here]", "skills": [ { "id": "infographic-maker", "name": "Infographic Maker", "description": "把文章、概念和数据提炼成手绘卡通信息图", "skill_id": "skill_2053082171331530752" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 11, "is_featured": true, "featured_order": 1, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-24T00:38:00+08:00", "updated_at": "2026-07-24T01:43:43+08:00", "is_recommended": true, "recommended_order": 3 }, { "id": "pricing-comparison-page", "title": "产品定价对比页", "title_en": "Product Pricing Comparison Page", "subtitle": "SaaS 定价页一键生成，套餐对比清晰直观", "subtitle_en": "Generate a SaaS pricing page with clear plan comparison in one click", "description": "设计一个能提升转化率的定价页是 SaaS 产品最头疼的难题之一：套餐之间的差异要一目了然，年付月付切换要丝滑自然，功能对比矩阵要既全面又不让人眼花缭乱，还要兼顾移动端适配和视觉吸引力。\n\n这个案例用 frontend-dev 技能为虚拟项目管理工具「ProjectFlow」生成了一套专业级定价页：4 张套餐卡片并排展示（Free / Starter / Pro / Enterprise），支持月付/年付切换（年付享 8 折），Pro 套餐带「最受欢迎」标签，每张卡片包含价格、功能清单（勾叉标记）和行动按钮；下方是 20+ 项功能的详细对比矩阵表（按核心功能/协作/数据分析/客户支持分组），8 道常见 FAQ 手风琴折叠，底部 CTA 横幅和完整页脚。现代设计风格，玻璃态卡片，渐变 CTA 按钮，单文件 HTML，完全响应式。\n\n适合：SaaS 创始人、产品经理、市场营销团队、需要快速搭建定价页的任何人", "description_en": "Designing a pricing page that converts is one of the toughest challenges for SaaS products: plan differences must be instantly clear, monthly/annual billing toggling should feel seamless, the feature comparison matrix needs to be comprehensive yet scannable, and everything must be mobile-friendly and visually compelling.\n\nThis case uses the frontend-dev skill to generate a professional pricing page for the fictional project management tool 'ProjectFlow': 4 plan cards side by side (Free / Starter / Pro / Enterprise) with monthly/annual billing toggle (20% annual discount), a 'Most Popular' badge on Pro, each card showing price, feature list with checkmarks/crosses, and CTA buttons; below is a detailed 20+ feature comparison matrix grouped by Core / Collaboration / Analytics / Support, an 8-question FAQ accordion, a bottom CTA banner, and a complete footer. Modern glassmorphism design, gradient CTA buttons, single HTML file, fully responsive.\n\nIdeal for: SaaS founders, product managers, marketing teams, anyone who needs a quick pricing page", "categories": [ "creation", "business" ], "tags": [ "定价页", "SaaS", "对比表", "转化", "落地页" ], "tags_en": [ "Pricing Page", "SaaS", "Comparison", "Conversion", "Landing Page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为 SaaS 项目管理工具「ProjectFlow」设计一个价格对比页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，现代设计、响应式。\n\n1. 页头：Logo + 导航（功能/价格/关于我们）\n2. 计费切换：月付 / 年付（年付省 20%）\n3. 四档价格卡片：免费版（¥0，3 项目 1 用户）、入门版（¥29/月，10 项目 5 用户）、专业版（¥79/月，不限项目 20 用户，标「最受欢迎」）、企业版（定制）。每张含方案名/价格/描述/功能勾选/CTA\n4. 功能对比矩阵：20+ 项功能 × 4 方案，按类别分组（核心/协作/数据分析/技术支持），用勾/叉/数值标识\n5. FAQ：8 个计费相关问题折叠\n6. 底部 CTA 横幅\n7. 页脚", "prompt_en": "Please design a pricing comparison page for the SaaS project management tool \"ProjectFlow\" (single-file HTML, inline with no external dependencies), with the UI text in Simplified Chinese, modern design, responsive.\n\n1. Header: Logo + navigation (Features / Pricing / About Us)\n2. Billing toggle: monthly / annual (annual saves 20%)\n3. Four pricing tiers: Free (¥0, 3 projects 1 user), Starter (¥29/month, 10 projects 5 users), Pro (¥79/month, unlimited projects 20 users, marked \"Most Popular\"), Enterprise (custom). Each includes plan name / price / description / feature checklist / CTA\n4. Feature comparison matrix: 20+ features × 4 plans, grouped by category (Core / Collaboration / Analytics / Support), marked with checks/crosses/values\n5. FAQ: 8 collapsible billing-related questions\n6. Bottom CTA banner\n7. Footer", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "description_en": "Frontend development with AI media and animations", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 12, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-15T09:13:40Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "定价对比页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "baby-learning-workspace", "title": "4岁宝宝学习乐园工作台", "title_en": "Learning Playground Workspace for 4-Year-Olds", "subtitle": "字母·古诗·数字·逻辑·闯关 五合一趣味启蒙", "subtitle_en": "Alphabet · Poems · Numbers · Logic · Adventure – 5-in-1 Fun Learning", "description": "幼儿早教资源分散、互动性差，家长不知道今天该教什么。这个工作台把 26 个字母认读、经典古诗朗读、10 以内加减法、逻辑找规律、综合闯关游戏整合为一个彩虹糖果色的 H5 页面，孩子打开就能自主学、家长无需盯。适合 3-6 岁幼儿日常启蒙，iPad/手机/电脑均可使用。", "description_en": "Early childhood learning resources are scattered and lack interactivity. This workspace combines alphabet recognition, classic Chinese poem reading, basic arithmetic, pattern-finding logic, and adventure mini-games into one colorful HTML5 page. Kids can learn independently — no parental hovering needed. Perfect for 3-6 year olds on iPad, phone, or desktop.", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/aV7d15ja4vbZJTJzfXgJQb），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我给 4 岁小朋友做一个学习工作台「宝贝学习乐园」。包含 5 个模块：\n1. 字母乐园：26 个英文字母卡片点击发音（Web Speech API），大小写配对游戏\n2. 古诗花园：8 首经典幼儿古诗带拼音标注，点击朗读\n3. 数字王国：0-20 数字认知 + 10 以内加减法随机出题 + 数数小游戏\n4. 逻辑挑战：找规律/图形配对/排序题，每题 3-4 个选项\n5. 闯关冒险：综合出题闯关 + 星星奖励 + 成就徽章墙\n\n配色用彩虹糖果色（粉蓝黄绿紫），风格活泼可爱。PC 侧边栏导航 + 移动端底部 Tab。数据存 localStorage，闯关进度和星星奖励持久化。按钮 ≥ 44px 方便小手点击。完成后部署上线，手机扫码即可使用。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/aV7d15ja4vbZJTJzfXgJQb) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a learning workspace called 'Baby Learning Playground' for a 4-year-old. Include 5 modules:\n1. Alphabet Park: 26 letter cards with tap-to-speak (Web Speech API), uppercase-lowercase matching game\n2. Poem Garden: 8 classic Chinese poems with pinyin annotation and read-aloud\n3. Number Kingdom: 0-20 number recognition + arithmetic within 10 + counting game\n4. Logic Challenge: pattern finding / shape matching / sorting, 3-4 options each\n5. Adventure Quest: mixed-topic quizzes + star rewards + badge wall Use rainbow candy colors (pink, teal, yellow, green, purple). Lively and cute style. PC sidebar nav + mobile bottom tabs. Data in localStorage with persistent progress. Buttons ≥ 44px for small hands. Deploy online for mobile scanning.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "artifact_type": "html", "preview": "baby-learning.html", "artifacts": [ { "file": "baby-learning.html", "label": "宝贝学习乐园工作台", "label_en": "Baby Learning Playground Workspace" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace", "dev", "education" ], "tags": [ "儿童教育", "启蒙学习", "工作台", "响应式", "闯关游戏" ], "tags_en": [ "Kids Education", "Early Learning", "Workspace", "Responsive", "Adventure Game" ], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 13, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-01T07:08:38Z", "updated_at": "2026-08-01T07:46:07Z", "preview_en": "baby-learning.html", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 2 }, { "id": "contract-management-ledger-report", "title": "2026 年 H1 合同梳理报告", "title_en": "H1 2026 Contract Ledger Report", "subtitle": "以虚拟合同台账演示签约结构、到期预警与履约风险梳理", "subtitle_en": "A virtual contract ledger demonstrating signing mix, expiry alerts, and performance-risk review.", "description": "合同散落在多个文件夹时，最烦的是不知道哪些该续签、哪些待履约。\n\n本案例使用虚拟合同数据演示一份可交互的管理台账：按类型、相对方、金额、期限和履约状态聚合，并标记 30 日内到期与风险合同。\n\n适合：合同管理员、业务运营、采购与销售支持团队。", "description_en": "When contracts are scattered across folders, it is hard to know what needs renewal or performance follow-up.\n\nThis case uses virtual contract data to demonstrate an interactive ledger that groups by type, counterparty, amount, term, and status, and flags contracts expiring within 30 days and risks.\n\nSuitable for: contract administrators, business operations, and procurement or sales support teams.", "categories": [ "business" ], "tags": [ "合同台账", "到期预警", "履约管理" ], "tags_en": [ "Contract ledger", "Expiry alerts", "Performance management" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "", "prompt": "帮我梳理合同并建立合同管理台账：按合同类型、业务属性、合作相对方、合同金额、合同期限和履约状态分类；汇总各类合同的数量与金额，并重点标记 30 日内到期、待履约及存在风险提示的合同。请列明每份合同的合同名称、签约主体、关键日期和风险提示。", "prompt_en": "Help me organize contracts and build a contract management ledger: classify by contract type, business attribute, counterparty, amount, term, and performance status; summarize counts and amounts by category; flag contracts expiring within 30 days, pending performance, and risk alerts. List contract name, signing parties, key dates, and risk notes for each contract.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "面向合同起草、审查、对比、签署与法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 20, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T21:46:05+08:00", "updated_at": "2026-08-12T22:57:05+08:00", "artifacts": [ { "file": "output.html", "label": "互动灵感卡片 HTML" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "pomegranate-floral-color-block", "title": "石榴花鸟纹 · 矢量色块提取", "title_en": "Pomegranate Floral-Bird Motif · Vector Color-Block", "subtitle": "照片质感一键转成扁平化双色矢量色块稿", "subtitle_en": "Convert photographic texture into a flat two-tone vector color-block", "description": "老纹样上有墨色晕染、光影和做旧痕迹，想做成干净的矢量素材总是带一堆杂色噪点。\n\n案例完整保留原纹样的构图、走向与每个花头枝叶的形态，只做媒介转换——统一为深靛蓝主体 + 瓷白底的纯平涂双色色块稿，边缘平滑闭合、保留花心与镂空负形，可直接用于矢量化描摹。\n\n适合：纹样设计师、文创图案开发者、丝网印刷/图标设计师、品牌视觉团队", "description_en": "This case fully preserves the original composition, flow and every floral/leaf shape, doing only a medium conversion — a pure flat two-tone block (deep indigo subject on porcelain white), with smooth closed edges and preserved negative space, ready for vector tracing.\n\nGood for: Pattern designers, Cultural motif developers, Screen-print & icon designers, Brand visual teams", "categories": [ "creation" ], "tags": [ "纹样提取", "矢量色块", "扁平化", "青花蓝" ], "tags_en": [ "Motif Extraction", "Vector Color Block", "Flat Design", "Porcelain Blue" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据这张图片中的纹样，生成一张「平面化矢量剪影色块稿」。\n第一优先级：完全保留原图的构图、走向、疏密关系，以及每一个花头、枝叶、卷草的形态、比例与位置——不做任何再设计、不增删元素、不改变布局。只做媒介转换，不改动纹样本身。\n把原有的照片质感、墨色晕染、笔触、光影、立体感、做旧痕迹，全部转换为纯粹的平涂色块：\n纹样主体用单一实色平涂（统一为深靛蓝 #1F4E79），色块内部完全均匀，没有任何渐变、阴影、高光、纹理、颗粒、笔触或裂纹；\n背景是纯净的瓷白 #F7F4EE，整张画面只有两个颜色——主色与底色，不存在任何中间色调和过渡色；\n所有轮廓边缘平滑、连续、闭合，是干净利落的曲线，没有毛边、锯齿、断线或飞白；\n保留纹样内部的主要白色负形——花瓣之间的分隔线、花心、镂空——层次靠留白来形成，白色分隔线宽度要均匀且清晰可见，不要细到消失；\n同时消除所有细碎噪点、孤立小点、过细的杂线和不必要的断口，让每一个色块都是完整饱满、边界清晰的整块形状。\n整体观感：像一张专业绘制的矢量插画、丝网印刷色块稿或扁平化图标素材——干净、规整、高对比、边缘锐利，可以直接用于矢量化描摹。\n负面提示词：不要渐变，不要阴影，不要立体感，不要纹理，不要噪点，不要照片质感，不要裂纹做旧，不要第三种颜色，不要重新设计纹样，不要改变构图，不要虚线描边，不要模糊边缘，不要水印和文字。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-vector-element-expert", "name": "文物矢量元素提取专家", "description": "将文物图像精准转译为可编辑的 SVG 矢量线稿、平面色块、轮廓线与代表色色卡", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uzh9tCiVNo2A" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 20, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "shinkai-anime-film", "title": "新海诚动画电影风格", "title_en": "Shinkai Anime Film Style", "subtitle": "用天空、天气与光影讲述一段跨越距离的青春重逢", "subtitle_en": "A youthful reunion shaped by sky, weather, and light", "description": "青春情感故事容易只剩台词，缺少能承载距离感和重逢情绪的视觉线索。\n\n案例讲述分隔两地的少年与少女在夏末天台重逢：落日、电车、云层、流星雨和银河依次推动情绪，最终两人在星光下确认彼此。\n\n适合：校园故事、青春告白、动画预告片、歌曲视觉化和情绪型品牌短片。", "description_en": "Youthful emotional stories can become dialogue-heavy when they lack visual motifs for distance and reunion.\n\nThis case follows a boy and girl separated by distance who reunite on a rooftop at the end of summer. Sunset, trains, clouds, a meteor shower, and the Milky Way carry the emotional progression until they finally acknowledge each other under the stars.\n\nBest for school stories, youthful confessions, anime trailers, music visualization, and emotionally driven brand films.", "categories": [ "creation" ], "tags": [ "动漫电影", "青春情感", "星空" ], "tags_en": [ "anime film", "youth emotion", "starry sky" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「新海诚动画电影风格」Skill，制作一支约 60 秒、16:9 的原创青春动画短片《星河下的重逢》。男生住在海边小城，女生刚从东京返回；两人曾约定在夏末登上学校天台看流星雨，却因暴雨和停运电车几乎错过。请用落日电车、雨后积水倒影、被风吹动的校服、云层裂开的强光和银河流星作为贯穿视觉，结尾让两人在天台相向奔跑后安静对视。先确认故事规格和情绪曲线，再设计分镜、绑定角色与场景资产、生成关键帧；关键视觉与完整分镜确认后再逐镜生成，并完成环境音、钢琴弦乐和最终组装。保持角色外貌、服装和天气连续，不使用任何现成动画角色或片名标识。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “Shinkai Anime Film Style” Skill and create an original 16:9 youth anime short of about 60 seconds titled Reunion Beneath the Milky Way. A boy lives in a seaside town and a girl has just returned from Tokyo. They promised to watch the meteor shower from their school rooftop at summer’s end, but heavy rain and suspended trains nearly keep them apart. Use sunset trains, reflections in rain puddles, school uniforms moving in the wind, intense light breaking through clouds, and the Milky Way as recurring visual motifs. End with the two running toward each other and then quietly meeting each other’s gaze on the rooftop. Confirm the story specification and emotional curve first, then design the storyboard, bind character and location assets, and generate keyframes. Generate shots only after the key visuals and complete storyboard are approved, then finish ambience, piano-and-strings music, and final assembly. Keep faces, costumes, and weather continuity consistent, and use no existing anime characters or title marks.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 20, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-task-data-analysis-report", "title": "2026 年合同任务数据分析报告", "title_en": "2026 Contract Task Data Analysis Report", "subtitle": "以虚拟任务数据演示任务结构、趋势与处理效率分析", "subtitle_en": "Virtual task data demonstrating task mix, trends, and processing efficiency analysis.", "description": "合同任务越多，越难判断团队的工作重心和积压点。\n\n本案例使用虚拟任务数据，按起草、审查、法律咨询、对比及其他任务展示最近 90 天与累计统计、月度趋势和最新处理时间。\n\n适合：法务管理者、合同运营、需要复盘服务效率的团队负责人。", "description_en": "As contract tasks grow, it becomes hard to see the team’s focus and backlog.\n\nThis case uses virtual task data to show 90-day and cumulative statistics, monthly trends, and latest processing times across drafting, review, legal consultation, comparison, and other tasks.\n\nSuitable for: legal managers, contract operations, and team leads reviewing service efficiency.", "categories": [ "business" ], "tags": [ "任务分析", "合同运营", "数据看板" ], "tags_en": [ "Task analytics", "Contract operations", "Data dashboard" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "", "prompt": "帮我分析合同任务数据，按合同起草、合同审查、法律咨询、合同对比、其他任务五类进行统计。请分别展示最近 90 天及累计的任务数量、占比、月度趋势和最近处理时间；同时按任务创建时间从新到旧列出任务明细，包括任务名称、任务类型、创建时间、处理状态、涉及的合同类型。", "prompt_en": "Help me analyze contract task data by five categories: contract drafting, contract review, legal consultation, contract comparison, and other tasks. Show 90-day and cumulative task counts, proportions, monthly trends, and latest processing times. List task details from newest to oldest, including task name, type, creation time, status, and contract type.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "面向合同起草、审查、对比、签署与法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 30, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T21:46:05+08:00", "updated_at": "2026-08-12T22:57:05+08:00", "artifacts": [ { "file": "output.html", "label": "互动灵感卡片 HTML" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "nolan-director-style", "title": "诺兰导演风格", "title_en": "Nolan Director Style", "subtitle": "以三条不同速度的时间线交汇成一场原创战时撤离", "subtitle_en": "Three time scales converge in an original wartime evacuation", "description": "复杂叙事常因时间线不清而失去张力，宏大场面也容易变成互不关联的镜头堆叠。\n\n案例用海滩上一小时、海上一天、空中十分钟三条时间线，讲述虚构战争中的撤离行动；飞行员、年轻士兵和民船船长的选择在高潮处同步交汇。\n\n适合：悬疑预告、战争与灾难叙事、品牌史诗片，以及需要非线性结构的概念短片。", "description_en": "Complex stories lose tension when their timelines are unclear, while spectacle can collapse into a pile of unrelated shots.\n\nThis case uses three time scales—one hour on a beach, one day at sea, and ten minutes in the air—to portray an evacuation in a fictional war. The choices of a pilot, a young soldier, and a civilian boat captain converge at the climax.\n\nBest for suspense trailers, war and disaster narratives, epic brand films, and concept shorts that need nonlinear structure.", "categories": [ "creation" ], "tags": [ "非线性叙事", "史诗电影", "多时间线" ], "tags_en": [ "nonlinear narrative", "epic cinema", "multiple timelines" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「电影感创作」专家调用 Flova 的「诺兰导演风格」Skill，制作一支约 90 秒、21:9 的原创战时撤离短片《海陆空三重时钟》。设定为完全虚构的海岛战争：海滩线覆盖一小时，一名年轻通信兵寻找仍能工作的信号灯；海上线覆盖一天，一位民船船长穿过风暴接近撤离点；空中线覆盖十分钟，一名飞行员在燃料将尽时保护船队。三条时间线在信号灯亮起、民船入港和战机掠海的同一瞬间汇合。请先呈现故事分析与时间结构选项，确认后建立视觉圣经、角色系统、规格、关键视觉和完整分镜；各阶段停下确认，镜头通过后再生成旁白、音效、递进式管弦配乐并组装。避免复刻任何现有电影镜头、人物、台词或标识。", "prompt_en": "Ask the “Flova Cinematic Director” expert to invoke Flova’s “Nolan Director Style” Skill and create an original 21:9 wartime evacuation short of about 90 seconds titled Evacuation: Three Clocks Across Sea, Land, and Air. Set it in a completely fictional island war. The beach timeline spans one hour as a young signals soldier searches for a working beacon; the sea timeline spans one day as a civilian boat captain crosses a storm toward the evacuation point; the air timeline spans ten minutes as a pilot protects the convoy with fuel running out. The timelines converge at the same instant when the beacon lights, the civilian boat enters the harbor, and the aircraft skims the sea. Present the story analysis and time-structure options first. After approval, build the visual bible, character system, specification, key visuals, and full storyboard. Pause for approval at each stage, and only then generate narration, sound effects, escalating orchestral music, and the final edit. Do not recreate any existing film shots, characters, dialogue, or marks.", "skills": [], "experts": [ { "id": "flova-cinematic-director", "name": "电影感创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及电影感成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e0b061a92ad1", "expert_id": "ex_oQGh7uNdnrp6" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 30, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tang-sancai-horse-contour", "title": "唐三彩马 · 轮廓提取", "title_en": "Tang Sancai Horse · Contour Extraction", "subtitle": "白描式单线轮廓，保留镂空与负空间", "subtitle_en": "Gongbi-style single-line contour with negative space preserved", "description": "立体文物想抽成一根干净的轮廓线，AI 常常画成内外双线，还把腿间空隙填实。\n\n案例把唐三彩马从背景完整分离，保留姿态、比例与四腿间的透空负形，用纯黑单线勾勒外轮廓与关键结构分界，风格接近工笔白描，输出可继续编辑的透明背景矢量线稿。\n\n适合：文创设计师、文物插画师、文博 IP 开发者、视觉设计团队", "description_en": "This case cleanly separates the Tang Sancai horse from its background, keeps its pose, proportions and the see-through gaps between legs, and outlines the silhouette and key structural divisions in a single black stroke — a Gongbi-style, editable transparent-background vector sketch.\n\nGood for: Cultural-creative designers, Relic illustrators, Museum IP developers, Visual design teams", "categories": [ "creation" ], "tags": [ "轮廓提取", "白描线稿", "文物矢量", "SVG" ], "tags_en": [ "Contour Extraction", "Gongbi Line Art", "Relic Vector", "SVG" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据这张图片，提取画面主体的 SVG 矢量轮廓线。\n主体分割：将主体（含底座）从背景中完整分离，保留其造型、姿态、比例与构图；主体的镂空与负空间（如四腿之间的空隙、把手内圈）必须保留为透空，不可填实；主体在画面边缘被裁切处保持裁切。\n线条要求：纯黑 #000000 单线勾勒外轮廓，线条连续、流畅、粗细均匀、转折圆滑；同一边缘只画一根线，避免内外双线；风格接近白描 / 工笔勾线。\n取舍：以「外轮廓 + 关键结构分界」为主（如颈背线、腿的分界、底座边缘）；舍去釉色、色斑、光影渐变、材质颗粒与噪点。\n背景与输出：透明背景，无填充色、无阴影、无渐变、无外框；输出为可在 Illustrator / Figma / Inkscape 中继续编辑的矢量轮廓线稿，视觉干净、结构可辨。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-vector-element-expert", "name": "文物矢量元素提取专家", "description": "将文物图像精准转译为可编辑的 SVG 矢量线稿、平面色块、轮廓线与代表色色卡", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uzh9tCiVNo2A" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 30, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-renewal-performance-alert-report", "title": "2026 年合同续签与履约预警报告", "title_en": "2026 Renewal and Performance Alert Report", "subtitle": "以虚拟数据识别续签窗口、待办与履约异常", "subtitle_en": "Virtual data identifying renewal windows, action items, and performance exceptions.", "description": "续签窗口和履约节点一旦遗漏，往往直接影响收入与合作关系。\n\n本案例以虚拟合同池展示续签优先级、到期日历、待办事项及履约异常，并支持补充本地文件后刷新报告。\n\n适合：客户成功、销售运营、采购与合同管理团队。", "description_en": "Missing a renewal window or performance milestone can directly affect revenue and relationships.\n\nThis case uses a virtual contract pool to show renewal priorities, an expiry calendar, action items, and performance exceptions, with support for refreshing the report after adding local files.\n\nSuitable for: customer success, sales operations, procurement, and contract management teams.", "categories": [ "business" ], "tags": [ "续签管理", "履约预警", "合同运营" ], "tags_en": [ "Renewal management", "Performance alerts", "Contract operations" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "", "prompt": "请基于我的合同文件建立续签与履约预警报告：识别未来 90 天内到期或可进入续签谈判的合同，列出合同主体、金额、到期日、自动续期条款、当前履约状态和下一步建议；对已逾期付款、验收待确认或续签窗口不足 30 天的合同标记优先级。", "prompt_en": "Please build a renewal and performance alert report from my contract files: identify contracts expiring or entering renewal negotiations within 90 days; list parties, amount, expiry date, auto-renewal clauses, current performance status, and next steps; prioritize overdue payment, pending acceptance, or contracts with a renewal window under 30 days.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "面向合同起草、审查、对比、签署与法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 40, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T21:46:05+08:00", "updated_at": "2026-08-12T22:57:05+08:00", "artifacts": [ { "file": "output.html", "label": "互动灵感卡片 HTML" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "script-to-video", "title": "剧本生视频", "title_en": "Script to Video", "subtitle": "把室内悬疑剧本拆成光线、动作和声音都连续的成片", "subtitle_en": "A contained thriller turned into a visually and sonically continuous film", "description": "现成剧本直接做视频时，人物、空间和悬念节奏很容易在镜头之间断裂。\n\n案例把一段“独居女子半夜发现门外有人”的剧本拆成卧室、走廊与门口三处空间，用闹钟、手机光、影子和门锁声逐步提升压力。\n\n适合：悬疑短片、短剧样片、剧本可视化、导演提案和低场景成本的剧情内容。", "description_en": "When an existing script goes straight to video, character, spatial, and suspense continuity can break between shots.\n\nThis case divides a story about a woman who senses someone outside her door at night across a bedroom, hallway, and doorway, escalating pressure through an alarm clock, phone light, shadow, and lock sounds.\n\nBest for suspense shorts, drama samples, script visualization, directing pitches, and narrative content with limited locations.", "categories": [ "creation" ], "tags": [ "剧本可视化", "悬疑短片", "连续性" ], "tags_en": [ "script visualization", "thriller short", "continuity" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「电影感创作」专家调用 Flova 的「剧本生视频（需上传剧本）」Skill，把我上传的剧本《凌晨三点的门外人》制作成约 75 秒、16:9 的写实悬疑短片。故事限定在一间卧室和门外走廊：年轻女子被 03:00 的闹钟惊醒，手机显示三条来自未知号码的“别开门”；走廊感应灯亮起，门缝下出现停住的人影，她拿起手电靠近，门锁缓慢转动，最终发现门外只留着一张写有她姓名的旧照片。请先分析剧本并确认全局规格，再给出角色、场景、道具与完整分镜；在元素图和逐镜视频提示词草案阶段都停下让我确认，批准后再生成镜头、门锁与脚步音效、低频氛围音乐并组装。空间方位、人物睡衣和手机必须跨镜一致，不擅自增加角色或改写结局。", "prompt_en": "Ask the “Flova Cinematic Director” expert to invoke Flova’s “Script to Video (Script Upload Required)” Skill and turn my uploaded script The Figure Outside at 3 A.M. into a realistic 16:9 suspense short of about 75 seconds. The story is confined to a bedroom and the hallway outside. A young woman wakes to a 03:00 alarm and finds three messages from an unknown number saying, “Do not open the door.” The hallway sensor light switches on, a motionless shadow appears beneath the door, she takes a flashlight and approaches, the lock slowly turns, and she ultimately finds only an old photograph bearing her name. Analyze the script and confirm the global specification first, then present the characters, locations, props, and complete storyboard. Pause for my approval both after the element images and after the per-shot video prompt drafts. Once approved, generate the shots, lock and footstep effects, low-frequency atmospheric music, and final assembly. Keep spatial orientation, pajamas, and phone consistent across shots, and do not add characters or change the ending.", "skills": [], "experts": [ { "id": "flova-cinematic-director", "name": "电影感创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及电影感成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e0b061a92ad1", "expert_id": "ex_oQGh7uNdnrp6" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 40, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tang-tribute-horse-palette", "title": "唐献马图 · 色卡提取", "title_en": "Tang Tribute-Horse Painting · Color Palette", "subtitle": "原图 + 5 色代表色卡，一张专业配色分析图", "subtitle_en": "Original + 5 representative swatches in one pro palette sheet", "description": "想复用古画的配色做设计，却只能靠肉眼猜色号，取出来还老是几个相近的灰棕色。\n\n案例保留原始文物图不重绘不调色，识别本体后提取 5 种真实存在、彼此差异明显的代表色，用左图右卡的极简编辑版式呈现，每个色块标注精确 HEX 色值，输出高清配色分析图。\n\n适合：平面/品牌设计师、文创配色开发者、插画师、博物馆视觉团队", "description_en": "This case keeps the original relic image intact (no repaint, no recolor), identifies the subject and extracts 5 real, clearly distinct representative colors, laid out in a minimal editorial style (image left, swatches right) with precise HEX values on each block — a high-res palette analysis sheet.\n\nGood for: Graphic & brand designers, Cultural-creative colorists, Illustrators, Museum visual teams", "categories": [ "creation" ], "tags": [ "配色提取", "色卡", "HEX色值", "文物配色" ], "tags_en": [ "Color Extraction", "Palette", "HEX Values", "Relic Palette" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据我上传的文物图片，生成一张“文物原图＋代表色色卡”的专业配色分析图。完整保留原始图片，不重绘、不调色、不改变文物的造型、比例、纹样、光影与历史质感；先识别文物或作品本体，再提取5种真实存在、具有视觉代表性且彼此差异明显的颜色。若为白底文物图，应排除摄影背景、投影与高光；若为壁画、古画或织物等满幅作品，作品底色可以参与取色。色彩选择需兼顾主色、浅色基底、深色线描或材质色，以及具有文化识别度的彩绘或点缀色，避免提取裂纹、污点、压缩噪点及多个过度相近的米色、灰色或棕色。\n使用简洁的编辑设计版式：纯白色画布，左侧约占画面宽度60%—65%，完整展示原始文物图片；右侧约占35%—40%，纵向排列5个等高色块。原图框与色块均采用统一的小圆角，画面四周留白一致，色块间距均匀，不添加标题、说明、边框、阴影、Logo或水印。每个色块右侧居中标注对应的六位大写HEX色值，文字不带“#”；浅色块使用黑色无衬线字体，深色块使用白色字体。色块填充颜色必须与显示的HEX数值完全一致，最终输出高清PNG图。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-vector-element-expert", "name": "文物矢量元素提取专家", "description": "将文物图像精准转译为可编辑的 SVG 矢量线稿、平面色块、轮廓线与代表色色卡", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uzh9tCiVNo2A" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 40, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-review-risk-profile-report", "title": "2026 年合同审查风险画像报告", "title_en": "2026 Contract Review Risk Profile Report", "subtitle": "以虚拟审查任务演示风险分布、整改优先级与闭环跟踪", "subtitle_en": "Virtual review tasks demonstrating risk distribution, remediation priorities, and closed-loop tracking.", "description": "审查报告多了以后，单份风险很清楚，整体风险画像却常常看不见。\n\n本案例以虚拟审查数据聚合高、中、低风险，识别数据安全、责任限制和付款验收等高频问题，并给出整改看板。\n\n适合：法务负责人、合规团队、需要管理审查质量的业务负责人。", "description_en": "When review reports pile up, individual risks are clear but the overall risk profile is often invisible.\n\nThis case aggregates virtual review data by high, medium, and low risk, identifies recurring issues in data security, liability caps, and payment acceptance, and provides a remediation dashboard.\n\nSuitable for: legal leads, compliance teams, and business owners managing review quality.", "categories": [ "business" ], "tags": [ "合同审查", "风险画像", "整改闭环" ], "tags_en": [ "Contract review", "Risk profile", "Remediation tracking" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "", "prompt": "请汇总我近期的合同审查结果，形成合同风险画像报告：按高、中、低风险统计风险数量、涉及合同类型、责任部门和整改状态；重点归纳数据安全、知识产权、付款验收、责任限制、违约责任和争议解决等高频问题，并按影响程度和紧急程度给出整改优先级及后续审查建议。", "prompt_en": "Please consolidate my recent contract review results into a contract risk profile report: count high, medium, and low risks by contract type, responsible department, and remediation status; summarize recurring issues in data security, IP, payment acceptance, liability caps, breach, and dispute resolution; provide remediation priorities and follow-up review suggestions by impact and urgency.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "面向合同起草、审查、对比、签署与法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 50, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T21:46:05+08:00", "updated_at": "2026-08-12T22:57:05+08:00", "artifacts": [ { "file": "output.html", "label": "互动灵感卡片 HTML" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "drama-short-voice-reference", "title": "剧情短片(音色参考)", "title_en": "Drama Short with Voice Reference", "subtitle": "让双人对峙短剧保留角色音色与情绪推进", "subtitle_en": "A two-man confrontation that preserves voice identity and emotional escalation", "description": "剧情短片即使画面成立，也常因角色音色漂移、口型与情绪不匹配而失真。\n\n案例在寒冷空仓库里安排两名旧友围绕一份证据对峙，从点烟、拔枪到放下武器和拥抱，利用各自音色参考维持身份感。\n\n适合：双人短剧、角色对白样片、悬疑片段和需要指定演员音色的剧情内容。", "description_en": "Even strong drama footage can feel false when character voices drift or lip sync and emotion do not match.\n\nThis case stages a confrontation between two old friends in a cold empty warehouse, moving from a lit cigarette and drawn gun to a lowered weapon and embrace while preserving each character’s voice reference.\n\nBest for two-person dramas, dialogue samples, suspense scenes, and narratives that need authorized actor voice references.", "categories": [ "creation" ], "tags": [ "对白短剧", "音色参考", "情绪表演" ], "tags_en": [ "dialogue drama", "voice reference", "emotional acting" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「电影感创作」专家调用 Flova 的「剧情短片(音色参考)」Skill，制作一支约 60 秒、16:9 的双人现实主义短片《仓库里的最后一支烟》。我会上传剧本，并分别提供角色 A 与角色 B 的合法授权音色样本。场景是一间冬夜空仓库：A 靠窗点烟，质问 B 为何藏起关键证据；B 掏枪但手在发抖，A 说出“我来不是为了抓你”，B 最终放下枪，两人短暂拥抱后各自离开。请先审核并确认剧本，再建立规格、分镜、人物和场景资产；分别制作并让我确认两位角色的声音锚点，随后生成镜头、对白、环境声与配乐。每个关键阶段暂停，确保音色归属、口型、台词先后和情绪强度准确，不复制未经授权的真人声音。", "prompt_en": "Ask the “Flova Cinematic Director” expert to invoke Flova’s “Drama Short with Voice Reference” Skill and create a realistic 16:9 two-person short of about 60 seconds titled The Last Cigarette in the Warehouse. I will upload the script and legally authorized voice samples for Character A and Character B. In an empty warehouse on a winter night, A lights a cigarette by the window and asks why B hid the crucial evidence. B draws a gun with a trembling hand. A says, “I did not come to arrest you.” B lowers the weapon, they briefly embrace, and then leave separately. Review and approve the script first, then establish the specification, storyboard, character assets, and location assets. Create separate voice anchors for both characters and let me approve them before generating shots, dialogue, ambience, and music. Pause at every major stage and keep voice ownership, lip sync, line order, and emotional intensity accurate. Do not imitate any person’s voice without authorization.", "skills": [], "experts": [ { "id": "flova-cinematic-director", "name": "电影感创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及电影感成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e0b061a92ad1", "expert_id": "ex_oQGh7uNdnrp6" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 50, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "red-pottery-beast-toy-concept", "title": "红陶兽形器 · 玩具概念主图", "title_en": "Red Pottery Beast Vessel · Toy Concept Render", "subtitle": "把史前红陶兽形器变成枫木儿童拖拉玩具", "subtitle_en": "Turn a prehistoric beast vessel into a maple wooden pull-toy", "description": "文物开发文创，最难的是把古器造型转译成真正能生产、又保留辨识度的现代产品。\n\n案例以红陶兽形器为原型，保留圆润腹部、仰起兽首、拱形提梁与四足关系，转译为一款温润枫木、可拖拉的儿童玩具概念图——四足化作滚轮、兽口穿出棉绳木球，纯白背景棚拍级产品渲染，结构合理、无尖锐边角。\n\n适合：文创产品经理、工业/玩具设计师、博物馆文创团队、IP 授权开发方", "description_en": "Using the beast vessel as prototype, this case keeps the round belly, raised head, arched handle and four-leg relationship, translating it into a warm maple-wood pull-toy concept — legs become wheels, a cotton cord and wooden ball run through the mouth — rendered as studio-grade product photography on pure white, feasible and free of sharp edges.\n\nGood for: Cultural product managers, Industrial & toy designers, Museum creative teams, IP licensing developers", "categories": [ "creation" ], "tags": [ "文创概念", "产品渲染", "玩具设计", "文物活化" ], "tags_en": [ "Product Concept", "Product Render", "Toy Design", "Relic Revival" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据我上传的红陶兽形器文物原图，将其转译为一款儿童枫木拖拉玩具产品概念图。严格保留原文物最具辨识度的整体比例与关键结构：圆润饱满的兽形腹部、前端仰起并张嘴的兽首、两只竖耳、背部大型拱形提梁、后背圆筒形器口、上翘短尾和四足关系；不得无理由压扁、拉长或改变主体轮廓，仅允许基于儿童玩具安全性进行适度圆角化和轻微可爱化处理。将四足自然转化为四组短腿与圆形滚轮，兽口中穿出一根乳白色棉绳，绳端连接圆形木球，形成明确可用的拖拉结构。产品采用温润浅色枫木制作，保留细腻自然的木纹；腹部、拱形提梁、器口局部和双耳内侧使用低饱和红陶色涂装，颈部与轮缘加入少量深棕色细线作为层次点缀，整体配色来源于原文物的红陶色，不增加无关纹样。产品结构简洁、圆润、真实可生产，连接关系合理，无尖锐边角。画面采用正面偏左的四分之三产品视角，完整展示主体、四个轮子、拉绳和木球，单一产品居中，干净纯白背景，柔和棚拍光线，自然接触阴影，高端工业设计产品摄影与精细3D产品渲染质感，高清细节，无人物、无手、无场景道具、无文字、无标识、无拼图、无多视图，不生成陶瓷材质，不改变为其他动物。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-concept-expert", "name": "文创品概念设计专家", "description": "将文物特征转译为可落地的文创产品概念图，输出白底/棚拍/三视图/场景渲染", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3b6387f5c208", "expert_id": "ex_Pj5SOQeOdh2H" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 50, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "painted-pottery-ding-studio-shot", "title": "彩陶鼎 · 摆件棚拍图", "title_en": "Painted Pottery Ding · Studio Shot", "subtitle": "文创摆件秒变高级极简棚拍产品图", "subtitle_en": "Instantly stage a creative object as a premium minimalist studio shot", "description": "拍好一张干净又高级的文创产品图，要租棚、布光、修图，成本高还慢。\n\n案例锁定文创摆件的品类、比例与材质不做任何重设计，生成一张 4:5 竖版、主体严格居中的高级极简棚拍图——同色相低饱和背景、包裹式柔光、浅而宽的接触阴影，呈现克制细腻的编辑型商业静物质感。\n\n适合：文创电商运营、产品摄影师、品牌视觉设计师、博物馆商店团队", "description_en": "This case locks the object's category, proportion and material without any redesign, producing a 4:5 portrait, strictly centered high-end minimalist studio shot — same-hue low-saturation backdrop, wrapping soft light and a shallow wide contact shadow, delivering restrained editorial still-life quality.\n\nGood for: Cultural e-commerce ops, Product photographers, Brand visual designers, Museum shop teams", "categories": [ "creation" ], "tags": [ "产品棚拍", "文创摄影", "极简", "4:5竖版" ], "tags_en": [ "Studio Shot", "Product Photo", "Minimalism", "4:5 Portrait" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据这张产品或文创品图片，先识别并锁定主体的品类、外形、比例、朝向、结构、材质、颜色、纹样、拼接关系与所有辨识性细节，再制作一张竖版 4:5 的高级极简棚拍图。主体必须水平严格居中、端正稳定、尺度适中，保持原始观看角度，不得重新设计、拉伸、压扁、旋转、裁切、增减结构或改变材质与颜色。\n\n背景必须完全空无元素，仅保留墙面与地面连续衔接的无缝哑光棚拍空间；禁止植物、花卉、石头、几何体、台座、布料、烟雾、图形、暗纹、文字、标志和水印。先分析主体覆盖面积最大的主色或材质底色，再将该颜色转化为更浅、低饱和的同色相或邻近色作为背景；若主体为深色则提高背景明度，若主体为浅色则适度降低背景明度，始终保持柔和但清晰的轮廓分离。\n\n使用大面积柔光箱或柔光布形成包裹式散射光，降低方向性和反差，使高光宽阔平滑、没有集中光斑；脚底或底部保留轻微接触阴影，外围投影浅、宽、边缘充分羽化并自然消融。整体呈现真实、克制、细腻的编辑型商业静物摄影质感，避免硬边投影、强聚光、夸张轮廓光、强镜面塑料感、普通纯白电商图和廉价 CGI 效果。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-concept-expert", "name": "文创品概念设计专家", "description": "将文物特征转译为可落地的文创产品概念图，输出白底/棚拍/三视图/场景渲染", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3b6387f5c208", "expert_id": "ex_Pj5SOQeOdh2H" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 60, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "story-driven-video", "title": "故事驱动型视频", "title_en": "Story-Driven Video", "subtitle": "从一个荒诞包裹展开温柔又古怪的完整故事", "subtitle_en": "A tender, offbeat story unfolds from an absurd package", "description": "有趣的故事设定如果缺少角色与道具连续性，很难撑起两分钟以上的叙事视频。\n\n案例让独居青年收到一只写着“给你”的罐头，打开后一位陌生女孩出现；两人从做饭、吃饭到整理花园，最终以她在清晨消失留下空罐收尾。\n\n适合：轻奇幻剧情、独立短片、故事广告、寓言内容和需要完整起承转合的视频。", "description_en": "An imaginative premise cannot sustain a two-minute narrative without consistent characters, props, and locations.\n\nA lonely man receives a can labeled “For you,” and a mysterious woman appears when he opens it. They cook, eat, and restore a garden before she vanishes at dawn, leaving the empty can behind.\n\nBest for light fantasy, independent shorts, narrative advertising, modern fables, and videos that need a complete dramatic arc.", "categories": [ "creation" ], "tags": [ "故事短片", "轻奇幻", "角色连续性" ], "tags_en": [ "narrative short", "light fantasy", "character continuity" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「全能视频创作」专家调用 Flova 的「故事驱动型视频」Skill，制作一支约 2 分钟、16:9 的轻奇幻短片《罐头女友》。独居青年收到一个没有寄件人的纸箱，里面只有一只写着“For you”的银色罐头；他打开罐头后，一位穿淡黄色连衣裙的陌生女孩出现在厨房。两人一起做饭、吃饭、修剪荒废花园，关系从戒备转为默契；第二天清晨女孩消失，桌上只剩空罐和一朵新开的花。请先确认规格和完整故事板，再绑定纸箱、罐头、男女主、厨房与花园资产；元素图、关键帧、分镜视频和音频各阶段都停下确认。用温暖自然光、细微幽默和克制对白推进，确保人物服装、罐头外观与房屋空间连续，最后完成环境音、轻音乐和时间线组装。", "prompt_en": "Ask the “Flova Video Director” expert to invoke Flova’s “Story-Driven Video” Skill and create a 16:9 light-fantasy short of about two minutes titled The Canned Girlfriend. A lonely young man receives a parcel with no sender. Inside is a single silver can labeled “For you.” When he opens it, a stranger in a pale-yellow dress appears in his kitchen. They cook, eat, and restore his neglected garden together, gradually moving from suspicion to quiet understanding. The next morning she is gone, leaving only the empty can and a newly opened flower. Confirm the specification and complete storyboard first, then bind assets for the parcel, can, two leads, kitchen, and garden. Pause for approval after element images, keyframes, shot videos, and audio. Use warm natural light, subtle humor, and restrained dialogue. Keep costumes, the can’s appearance, and the house layout consistent, then complete ambience, light music, and the final timeline assembly.", "skills": [], "experts": [ { "id": "flova-video-director", "name": "全能视频创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及完整成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "4adc3d692804", "expert_id": "ex_6JBgKuGMsg6Y" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 60, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "gta6-life-demo", "title": "GTA 6 风格演示", "title_en": "GTA 6 Style Demo", "subtitle": "把人生选择做成原创开放世界城市预告", "subtitle_en": "Life choices become an original open-world city trailer", "description": "个人故事想做出开放世界预告感时，容易误用现成游戏镜头、UI 或品牌元素。\n\n案例在原创热带海滨城市“霓虹海湾”中展开：两位朋友驾车穿过日落大道、夜店区和暴雨街道，因一只装错的旅行箱卷入追逐，最终在黎明海滩作出选择。\n\n适合：人生故事包装、城市文化短片、游戏化社交内容和高能量角色预告。", "description_en": "Open-world styling for a personal story can easily misuse existing game footage, UI, or brand elements.\n\nSet in the original tropical city Neon Bay, this case follows two friends driving through sunset boulevards, nightlife, and rain after a misplaced suitcase draws them into a pursuit, ending with a choice on the beach at dawn.\n\nBest for personal-story packaging, city culture films, game-like social content, and energetic character trailers.", "categories": [ "creation" ], "tags": [ "开放世界", "城市预告", "霓虹夜景" ], "tags_en": [ "open world", "city trailer", "neon night" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「GTA 6 风格演示（玩转我的人生）」Skill，制作一支 45 秒、16:9 的原创开放世界城市预告《霓虹海湾的一夜》。两位二十多岁的朋友驾驶一辆无品牌红色跑车穿过热带海滨城市：日落棕榈大道、拥挤夜店区、暴雨高架桥和黎明海滩依次出现；他们因后备箱里一只装错的蓝色旅行箱被陌生车辆追逐，最后把箱子留在海滩并开向日出。请先给出两个原创人物方案和全局规格，确认后设计角色图与完整分镜；开场与角色资产通过后再生成镜头、城市环境音和节奏音乐。只借鉴开放世界犯罪城市预告的视觉语法，不得出现任何官方游戏镜头、Logo、UI、商标、车牌、角色或台词。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “GTA 6 Style Demo (Play My Life)” Skill and create an original 16:9 open-world city trailer of 45 seconds titled One Night in Neon Bay. Two friends in their twenties drive an unbranded red sports car through a tropical seaside city: a sunset palm boulevard, a crowded nightclub district, a rain-soaked overpass, and a beach at dawn. A blue suitcase placed in the wrong trunk draws them into a pursuit by an unknown vehicle. They leave the suitcase on the beach and drive toward sunrise. Present two original character options and the global specification first. After approval, design character images and the complete storyboard, and generate shots, urban ambience, and rhythmic music only after the opening and character assets are approved. Borrow only the visual grammar of an open-world crime-city trailer. Do not include official game footage, logos, UI, trademarks, license plates, characters, or dialogue.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 70, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sancai-duck-ewer-notebook-threeview", "title": "三彩鸭形水注 · 手账本三视图", "title_en": "Sancai Duck Ewer · Notebook Three-View", "subtitle": "正面 / 侧面 / 背面并排的产品概念三视图", "subtitle_en": "Front / side / back product concept in one three-view sheet", "description": "做文创产品提案，光一张主图说不清结构，客户总要问侧面、背面长什么样。\n\n案例把印有鸭形纹样的文创手账本生成 3:2 横版三视图，正面、正侧面、背面从左到右等高并排，共用同一基线，准确保留圆角硬壳、书脊厚度、纸纹质感与低饱和复古印刷色调，纯白背景产品渲染质感。\n\n适合：文创产品经理、工业设计师、文创提案设计师、品牌开发团队", "description_en": "This case renders a duck-motif creative notebook as a 3:2 three-view — front, side and back arranged left to right at equal height on a shared baseline — faithfully keeping the rounded hardcover, spine thickness, paper texture and low-saturation vintage print tone on a clean white background.\n\nGood for: Cultural product managers, Industrial designers, Proposal designers, Brand development teams", "categories": [ "creation" ], "tags": [ "三视图", "产品概念", "文创笔记本", "3:2横版" ], "tags_en": [ "Three-View", "Product Concept", "Creative Notebook", "3:2 Landscape" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据这张文创产品图片，生成一张横构图 3:2 的产品概念三视图，包含正面、正侧面、背面，三者从左到右并排展示，置于干净纯白背景中。参考图是唯一的产品依据，必须准确保留产品本身的品类、外形比例、圆角硬壳笔记本形态、左侧书脊厚度、封面边缘结构、米白色纸纹质感、复古颗粒、蓝色水纹、黄褐色鸭形纹样及整体低饱和复古印刷色调；不要把它理解成鸭形瓷器本体，而是印有鸭形纹样的文创笔记本。\n画面采用高品质商业产品摄影/产品设计渲染效果，三视图等比例、等高度、共用同一水平基线，正面展示完整封面图案，正侧面展示笔记本厚度和书脊结构，背面延续同一材质、纹理和图案语言。柔和棚拍散射光，轻微自然投影，不添加任何文字、标签、箭头、尺寸线、logo、水印、道具、包装、桌面元素或彩色背景。不得改变产品形状、颜色、材质、比例、纹样、厚度和圆角结构，不得拉伸、压扁、重新设计、卡通化或替换为其他品类。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-concept-expert", "name": "文创品概念设计专家", "description": "将文物特征转译为可落地的文创产品概念图，输出白底/棚拍/三视图/场景渲染", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3b6387f5c208", "expert_id": "ex_Pj5SOQeOdh2H" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 70, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "sancai-duck-incense-nightlight-scene", "title": "三彩鸭形香薰 · 夜灯场景图", "title_en": "Sancai Duck Incense Burner · Ambient Scene", "subtitle": "新中式轻奢场景，给文创品一张氛围海报", "subtitle_en": "A neo-Chinese luxe scene poster for your creative product", "description": "产品单图放上架总显得平，缺一张能讲氛围、能当主视觉的场景大片。\n\n案例保留同一件文创品的品类、比例与纹样，将其置于黑色火山岩上，配太湖石花器与白色兰花，远景商业视角、大面积留白、低饱和含灰渐变背景，加胶片颗粒与柔焦投影，呈现新中式、轻奢、克制的商业产品摄影海报。\n\n适合：文创品牌运营、产品视觉设计师、电商主视觉团队、博物馆文创团队", "description_en": "This case keeps the product's category, proportion and motif, placing it on black volcanic rock with a Taihu-stone vase and white orchid — a distant commercial angle, generous negative space, low-saturation gray-tinted gradient backdrop, film grain and soft-focus shadows — a neo-Chinese, understated luxe product poster.\n\nGood for: Cultural brand ops, Product visual designers, E-commerce KV teams, Museum creative teams", "categories": [ "creation" ], "tags": [ "场景图", "新中式", "产品海报", "文创摄影" ], "tags_en": [ "Scene Render", "Neo-Chinese", "Product Poster", "Cultural Photo" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请基于我上传的产品参考图，生成一张新中式轻奢商业产品摄影海报。主体必须是图中同一件产品，完整保留它的品类、轮廓比例、关键结构、颜色分区、材质和纹样，不得改品类、改色、增减结构或重绘纹样。\n场景：产品稳定摆放在一块造型自然的黑色火山岩上，一侧配一件低矮的太湖石花器，插一枝带少量花苞的白色兰花，再加几块小石头。所有道具侧置低置，不遮挡产品。不要窗格、屏风、建筑、案几。\n构图：远景商业摄影视角，极简，大面积留白。产品完整居中、不裁切、不贴边，连基座约占画面高度百分之三十五到四十五，上半部留干净背景。真实透视、真实接触阴影，不悬浮。\n背景：低饱和含灰渐变，明度中低，色相跟随产品主色选择协调的传统色，与产品拉开明暗或冷暖差以保护轮廓。不用砖红。加细腻胶片颗粒和避开主体的枝叶柔焦投影。\n摄影：真实高端商业静物摄影，柔和有层次的全局布光，准确表现材质、纹理、光泽、透明度和清晰边缘，背景自然虚化。气质为新中式、轻奢、极简、复古胶片、禅意、克制。\n禁止：文字、标识、水印、边框、插画、卡通、塑料玩具感、过度装饰、悬浮、道具遮挡，以及与参考图不一致的图案。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-concept-expert", "name": "文创品概念设计专家", "description": "将文物特征转译为可落地的文创产品概念图，输出白底/棚拍/三视图/场景渲染", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3b6387f5c208", "expert_id": "ex_Pj5SOQeOdh2H" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 80, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "soft-3d-healing-animation", "title": "软萌 3D 治愈动画", "title_en": "Soft 3D Healing Animation", "subtitle": "一盏小灯照亮月夜森林，也安慰每个怕黑的伙伴", "subtitle_en": "A small lantern comforts every creature afraid of the dark", "description": "治愈动画容易只有可爱造型，却缺少清楚的小事件和可持续的情绪线。\n\n案例让红帽蘑菇孩子提着装满星光的灯穿过月夜森林，为迷路刺猬、睡不着的小鸟和害怕雷声的小兔依次送光，最后灯空了，整片森林却亮起来。\n\n适合：亲子内容、治愈系账号、睡前故事、角色 IP 概念和温柔品牌表达。", "description_en": "Healing animation often stops at cute character design instead of building a clear event and emotional progression.\n\nA red-capped mushroom child carries a lantern of starlight through a moonlit forest, helping a lost hedgehog, a sleepless bird, and a rabbit afraid of thunder before the whole forest begins to glow.\n\nBest for family content, comforting channels, bedtime stories, character-IP concepts, and gentle brand expression.", "categories": [ "creation" ], "tags": [ "软萌3D", "治愈动画", "森林童话" ], "tags_en": [ "soft 3D", "healing animation", "forest fairy tale" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「软萌 3D 治愈动画」Skill，制作一支约 60 秒、16:9 的原创 3D 动画《蘑菇孩子的星光灯》。主角是红色圆帽、米白身体、短圆四肢的蘑菇孩子，提着装有星光的木质小灯穿过月夜森林，依次为迷路刺猬指路、陪睡不着的小鸟安静下来、把一小团光留给怕雷声的兔子；灯最终变空，但蘑菇和树叶全部亮起微光。请先给 3 个可执行创意方向，我选定后再做 60 秒故事结构和分镜；随后依次确认主角图、场景/道具/首尾帧和开场 1–3 镜。全程固定角色比例、红帽与木灯造型，逐镜生成后完成森林环境音与轻柔音乐，不模仿任何动画工作室或现成 IP。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “Soft 3D Healing Animation” Skill and create an original 16:9 animated short of about 60 seconds titled The Mushroom Child’s Starlight Lantern. The protagonist is a mushroom child with a round red cap, cream body, and short rounded limbs, carrying a small wooden lantern filled with starlight through a moonlit forest. The child guides a lost hedgehog, sits with a sleepless bird until it calms down, and leaves a small glow with a rabbit afraid of thunder. The lantern ends up empty, but the mushrooms and leaves across the forest begin to glow. Present three executable creative directions first. After I select one, design the 60-second story structure and storyboard, then obtain approval for the hero image, locations, props, opening and ending frames, and opening shots 1–3 in sequence. Keep the character proportions, red cap, and wooden lantern fixed throughout. Generate each shot and finish with forest ambience and gentle music. Do not imitate any animation studio or existing IP.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 80, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ai-short-drama", "title": "AI 短剧一站式生成", "title_en": "AI Short Drama One-Stop Generation", "subtitle": "用 15 秒把合同、背叛与反转压进一场办公室对峙", "subtitle_en": "A contract, betrayal, and reversal compressed into one office confrontation", "description": "短剧开场需要快速建立冲突，但人物动机、道具位置和镜头连续性经常顾此失彼。\n\n案例在夜间办公室中安排一男一女围绕并购合同对峙：女主拒签、男主威胁停电，女主亮出已经录音的手机完成反转。\n\n适合：竖屏短剧、商业悬疑、剧情广告、投流素材和快速验证剧本钩子。", "description_en": "A short drama needs immediate conflict, but motivation, prop placement, and shot continuity are difficult to hold together.\n\nIn a late-night office, an executive refuses to sign a merger contract, her partner threatens her, and she reveals that her phone has recorded and uploaded the entire exchange.\n\nBest for short-form dramas, commercial suspense, narrative ads, performance marketing, and rapid hook testing.", "categories": [ "creation" ], "tags": [ "AI短剧", "办公室对峙", "剧情反转" ], "tags_en": [ "AI short drama", "office confrontation", "plot twist" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「AI 短剧一站式生成」Skill，把以下剧情制作成约 15 秒、16:9 的写实短剧《午夜办公室的签字》：深夜办公室里，女高管把并购合同推回去，说“数字不对，我不签”；男合伙人关掉一半灯光，低声说“今晚没人会替你作证”；女高管握紧拳头后把手机放到桌上，屏幕显示录音正在上传，她回答“我从进门就没打算让你替我作证”。请先确认已有素材、输出语言和全局规格，再建立人物、办公室、合同、手机的故事板与设定图；分镜预览、元素图和全部镜头生成后分别暂停确认，再完成对白、环境声与剪辑。严格保持人物服装、桌面方位、合同和手机位置连续，不补写新的结局。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “AI Short Drama One-Stop Generation” Skill and turn the following story into a realistic 16:9 drama of about 15 seconds titled The Midnight Signature. In a late-night office, a female executive pushes a merger contract back and says, “The numbers are wrong. I will not sign.” Her male partner switches off half the lights and says quietly, “No one will testify for you tonight.” She clenches her fist, places her phone on the desk, and reveals that the recording is uploading. She replies, “I never came in expecting you to testify for me.” Confirm available materials, output language, and the global specification first. Then build a storyboard and design images for both characters, the office, contract, and phone. Pause for approval after the storyboard preview, element images, and all generated shots, then complete dialogue, ambience, and editing. Keep costumes, desk orientation, contract, and phone positions continuous, and do not add a new ending.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 90, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sancai-duck-incense-poster", "title": "三彩鸭形香薰文创 · 海报", "title_en": "Sancai Duck Incense Cultural Product · Poster", "subtitle": "复古铜版画式博物馆文创品海报", "subtitle_en": "A vintage engraving-style museum editorial poster", "description": "文创品想出一张有档案感、有杂志质感的海报，排版和风格全靠反复试，太费劲。\n\n案例把鸭形香薰文创品转绘为精细铜版画式插图，暖象牙白旧纸背景 + 哑光赭金油墨，超大英文主标题「INCENSE / DUCK」与产品前后叠压，左下角配中英文产品档案信息区，呈现独立杂志与博物馆文创档案气质的 2:3 竖版海报。\n\n适合：文创品牌营销、视觉/平面设计师、博物馆展陈团队、文创电商推广", "description_en": "This case redraws the duck incense product as a fine engraving-style illustration on warm ivory old paper with matte ochre-gold ink, an oversized English title 'INCENSE / DUCK' layered with the product, and a bilingual archive info block at bottom-left — a 2:3 poster with indie-magazine and museum-archive character.\n\nGood for: Cultural brand marketers, Visual & graphic designers, Museum exhibition teams, Cultural e-commerce", "categories": [ "creation" ], "tags": [ "文创海报", "铜版画", "博物馆海报设计", "超大标题" ], "tags_en": [ "Cultural Poster", "Engraving", "Museum Editorial", "Oversized Title" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据参考图里的文创品，生成一张2:3竖版复古博物馆编辑海报。该文创品是一款以“明成化素三彩鸭形香薰”为设计原型的文创产品，并非原文物；必须准确保留产品的鸭形轮廓、张开的鸭喙、圆形眼睛、修长弯曲的颈部、圆润身体、上翘尾部、背部香插、层叠羽毛纹样、双足及带镂空孔的方形底座，保持原有造型、比例、朝向、金色材质和关键结构，不得将其改成普通鸟类、其他动物、古代原始文物或不同品类的产品。保留点燃的线香及自然向上飘散的烟雾，仅改变画面风格与排版，不重新设计产品。\n\n将产品转绘为精细的古典铜版画式插图，以细密排线、蚀刻线条和克制的明暗层次刻画黄铜质感、羽毛纹样和结构细节。画面使用暖象牙白无涂布旧纸背景，主体和主标题统一采用哑光赭金色油墨，辅以少量深暖褐色表现阴影及小字；加入细腻网点、轻微油墨磨损、不均匀印压和自然纸张纤维，保持鸭喙、眼睛、羽纹、香插和底座清晰。避免写实摄影感、现代3D渲染、强金属反光、渐变、发光和假烫金。\n\n采用不对称的“超大标题＋主体叠压”构图。将鸭形香薰文创品放大置于画面中部偏右及下方，占画面约65%—75%，底座接近下边缘，鸭头伸向左上方；右上方保留较大留白，让烟雾自然向上延伸。画面左侧排版两行超大英文主标题：第一行“INCENSE”，第二行“DUCK”。标题必须全部使用大写字母并准确拼写，采用极粗、高挑、紧缩的窄体无衬线字体，部分字母可被左侧画幅裁切；标题置于产品后方，让鸭头、颈部与身体覆盖部分字母，形成前后层次，但不得遮挡产品的关键识别结构。\n\n画面左下角设置小型文创产品档案信息区，准确排版以下内容：\n\n中文名称：“明成化素三彩鸭形香薰文创品”\n英文标签：“CULTURAL PRODUCT STUDY”\n细节标签：“FORM & MATERIAL”\n编号标签：“ARCHIVE 01”\n\n“CULTURAL PRODUCT STUDY”下方加入一条纤细横线和小型十字定位标记；“FORM & MATERIAL”下方放置一个方形局部细节图，真实截取产品身上的羽毛纹样，展示鳞羽与弧形翼纹，四角加入细小裁切线；“ARCHIVE 01”下方再次设置纤细横线与定位标记。中文名称使用清晰、克制的小号黑体，英文标签使用深暖褐色窄体无衬线字体，整体严格对齐、间距疏朗。\n\n除“INCENSE DUCK”“明成化素三彩鸭形香薰文创品”“CULTURAL PRODUCT STUDY”“FORM & MATERIAL”“ARCHIVE 01”外，不生成其他文字或乱码。整体呈现独立文化杂志、博物馆文创档案与复古印刷海报气质，但必须明确表现为文创产品视觉研究，不得将产品冒充为馆藏原件，也不要虚构年代、博物馆、出土信息或文物编号。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-promo-expert", "name": "文创宣传视觉设计专家", "description": "将定稿文创设计转化为海报、明信片、收藏小卡、易拉宝等宣传物料", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c6229d08eb8d", "expert_id": "ex_zaYhWoDF5zrp" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 90, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "21day-early-sleep-workbench", "title": "黎贝卡的21天早睡打卡", "title_en": "21-Day Early Sleep Challenge", "subtitle": "21天12:30前睡就去香港两天一夜！", "subtitle_en": "Sleep before 12:30 for 21 days and win a Hong Kong 2-day trip", "description": "想养成早睡习惯却总是三天打鱼两天晒网，最缺的就是一个看得见的进度和一份值得期待的奖励。\n\n这个案例展示了一个专属于自己的「21天早睡打卡」工作台：日历里每天点一下记录早睡，自动累成连续天数，连续 21 天达成后弹出庆祝提示。配套还有每月主线任务（默认「早睡&新书完成」，可编辑）、每日待办清单、灵感记录区（文字随笔 + 穿搭拍照，按日期归档）、八段锦21天打卡，以及一个能本地分析随笔主题和情绪、给出选题方向的「内容创作专家」。所有数据落在资料库的结构化表里，手机电脑同步，不丢不换设备。\n\n适合：想认真培养一次早睡习惯的人、喜欢给生活留个进度条的人、需要把待办和灵感收在一个干净地方的人、想要一个属于自己的生活工作台的人。", "description_en": "Most early-sleep resolutions die within a week because there's no visible streak and no reward worth waiting for.\n\nThis case shows a personal 21-day early-sleep challenge workbench: tap a calendar day to log sleep before 12:30, watch the streak build, and get a celebration alert after 21 consecutive days. It also comes with a monthly main-mission card (default \"Sleep Early & Finish New Books\", editable), a daily to-do list, an inspiration journal (text notes + outfit photos, archived by date), a 21-day Baduanjin check-in, and a content-creation expert that analyzes your notes' themes and mood locally to suggest topics. All data lives in structured tables in WorkBuddy Library, so it syncs across devices and never gets lost.\n\nBest for: anyone determined to actually fix their bedtime once, people who like a visible progress bar in life, those who want to-do and inspiration in one calm place, and anyone who'd like a workbench that's truly their own.", "categories": [ "workspace" ], "tags": [ "习惯打卡", "早睡挑战", "生活工作台", "数据同步", "个人管理" ], "tags_en": [ "Habit Tracking", "Early Sleep", "Life Workbench", "Data Sync", "Personal Management" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "21天早睡打卡工作台（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://www.workbuddy.cn/space/d/6Qvqouwws4JA5na8z6rLzR），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n帮我做一个「21天早睡打卡」生活工作台，做成单个完整 HTML 文件。\n\n包含：① 顶部本月主线任务（默认「早睡&新书完成」，右上角小编辑按钮弹窗修改、按月回顾）；② 每日待办（可添加/打勾/删除，首次预置「走7000步 / 八段锦 / 睡前10分钟手帐 / 喝1000ml水」）；③ 灵感记录（文字随笔 + 穿搭拍照上传压缩，按日期归档时间线）；④ 21天早睡打卡（日历可点击、连续天数自动累加，副标题「21天12:30前睡就去香港两天一夜！」，21天达成弹提示）；⑤ 八段锦21天打卡（同款日历）；⑥ 内容创作专家（本地分析随笔主题/情绪，给选题方向）。\n\n主题「更广阔 · 更自由 · 更轻松」，视觉走低饱和治愈系：暖粉薄荷渐变背景 + 墨绿主色 + 暖珊瑚强调，卡片圆角毛玻璃。制作完成后：① 将 HTML 文件存入我的个人资料库，并创建资料库子数据表（主线任务/待办/灵感/睡眠/八段锦5张表，跨设备同步）绑定数据；② 将 HTML 发布为公开链接并返回链接给我CSS/JS 无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 5 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/6Qvqouwws4JA5na8z6rLzR, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a \"21-Day Early Sleep Challenge\" life workbench. Deliver a single complete HTML file. Include: ① a top monthly main-mission area (default \"Sleep Early & Finish New Books\", with a small edit button that opens a modal for editing and month-by-month review); ② a daily to-do list (add / check off / delete, pre-seeded with \"Walk 7,000 steps / Baduanjin / 10-min journal before bed / Drink 1000ml water\"); ③ an inspiration journal (text notes + outfit photo uploads with local compression, archived by date on a timeline); ④ a 21-day early-sleep check-in (clickable calendar with auto streak, subtitle \"Sleep before 12:30 for 21 days and win a Hong Kong 2-day trip\", with a celebration alert at 21 days); ⑤ a 21-day Baduanjin check-in (same calendar style); ⑥ a content-creation expert that analyzes your notes' themes and mood locally and suggests topics. Theme: \"Broader · Freer · Lighter\", with a low-saturation healing look — warm pink + mint gradient background, dark green primary, warm coral accents, rounded frosted-glass cards. When done: ① save the HTML file to my personal WorkBuddy Library and create bound Library child tables (main mission / to-do / inspiration / sleep / Baduanjin, synced across devices) for its data; ② publish the HTML to a public link and return the link to me. Fall back to localStorage when the SDK is unavailable. Inline all CSS/JS with no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus 5 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "黎贝卡", "name_en": "黎贝卡", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-14T10:30:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "3d-guoman-costume-drama", "title": "3D国漫古装精品短剧", "title_en": "Premium 3D Chinese Costume Drama", "subtitle": "在金殿对峙中完成将军、皇后与少年谋士的权力反转", "subtitle_en": "A palace confrontation reverses power among a general, empress, and young strategist", "description": "古装短剧角色多、场景复杂，首镜空间一旦不稳，后续站位和人物形象就会连续漂移。\n\n案例把故事锁在金殿：凯旋将军逼皇后交出玉玺，少年谋士亮出先帝密诏，皇后在众臣面前完成反制；3D 国漫质感贯穿角色、盔甲与宫殿。\n\n适合：古装权谋、国漫短剧、连续剧预告、角色概念样片和强冲突开场。", "description_en": "Historical dramas involve many characters and complex spaces, so an unstable opening shot can cause cascading continuity errors.\n\nIn a throne hall, a victorious general demands the imperial seal, a young strategist reveals a secret decree, and the empress reverses the balance of power in a consistent 3D Chinese-animation style.\n\nBest for costume intrigue, Chinese-animation dramas, serial trailers, character concept reels, and conflict-driven openings.", "categories": [ "creation" ], "tags": [ "3D国漫", "古装权谋", "宫殿对峙" ], "tags_en": [ "3D Chinese animation", "costume intrigue", "palace confrontation" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「3D国漫古装精品短剧」Skill，制作一支约 75 秒、固定 16:9 的 3D 古风短剧《金殿夺印》。故事开场即是凯旋将军带甲闯入金殿，要求皇后交出玉玺；站在阶下的少年谋士取出先帝密诏，揭示将军伪造军报，皇后顺势命禁军封殿。请先评估剧本钩子、人物动机和台词，再逐项让我确认修改建议；之后完成全局规格、故事板、角色三视图、宫殿四视图、玉玺与密诏资产。第一个镜头生成后必须暂停，让我核对王座、台阶、人物站位和光影；再让我选择一键生成或逐镜确认。角色服饰、铠甲、玉玺与空间关系全片固定，完成对白、环境声、配乐和组装。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “Premium 3D Chinese Costume Drama” Skill and create a fixed 16:9 3D historical-fantasy drama of about 75 seconds titled Seizing the Imperial Seal. Open with a victorious general entering the throne hall in armor and demanding that the empress surrender the imperial seal. A young strategist below the steps produces the late emperor’s secret decree, revealing that the general forged military reports, and the empress orders the imperial guards to seal the hall. Evaluate the script hook, character motivations, and dialogue first, and let me approve every proposed revision. Then complete the global specification, storyboard, character turnaround sheets, four-view palace reference, imperial seal, and decree assets. After generating the first shot, pause so I can inspect the throne, steps, blocking, and lighting, then let me choose one-click generation or shot-by-shot approval. Keep costumes, armor, seal, and spatial relationships fixed throughout, and complete dialogue, ambience, music, and assembly.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "a-stock-dragon-tiger-hotspot-daily", "title": "龙虎榜情绪日报", "title_en": "Dragon-Tiger Board Sentiment Daily", "subtitle": "收盘后一页看清：谁上榜、什么题材在炒、哪个行业最强", "subtitle_en": "After close, one page: who made the board, which themes are hot, which sector leads", "description": "收盘后想复盘当天盘面，龙虎榜要一只只翻、题材要自己归纳、行业强弱得手动排——等你整理完，第二天都开盘了。这个案例把 2026-07-24 这一天的盘面自动做成一页情绪日报：84 家龙虎榜上榜股票按净买入排名（滨化股份净买入 1.78 亿登顶），同花顺当日强势股的题材归因做成热度榜（「兵装重组」「央企」「间接投资长鑫科技」领跑），再叠上全市场行业涨跌榜（美容护理领涨 +4.11%、保险领跌）。\n\n数据全部来自东方财富龙虎榜、同花顺热点、东财行业板块的真实公开接口，通过「A股全栈数据」skill 现场抓取，红涨绿跌、条形图直观。\n\n适合：每天要做收盘复盘的短线交易者、跟踪题材轮动的游资研究者、写盘后点评的财经博主、需要快速掌握当日资金动向的投研人员。", "description_en": "After the close, reviewing the day means flipping through the dragon-tiger board one by one, summarizing themes yourself and manually ranking sector strength — by the time you're done, the next session has already opened. This case auto-builds a one-page sentiment daily for 2026-07-24: 84 dragon-tiger board stocks ranked by net buying (Binhua Co. tops it with ¥178M net buy), THS strong-stock theme attributions turned into a heat ranking ('defense-asset restructuring', 'SOE', 'indirect CXMT investment' leading), plus a market-wide sector gain/loss board (Beauty & Care leads +4.11%, Insurance lags).\n\nAll data comes from the real public APIs of EastMoney's dragon-tiger board, THS hotspots and EastMoney sector boards, pulled live via the A-Stock Data toolkit, with up=red/down=green coloring and intuitive bar charts.\n\nIdeal for: short-term traders doing daily post-close reviews, hot-money researchers tracking theme rotation, finance bloggers writing after-hours commentary, and analysts needing a fast read on the day's capital moves.", "categories": [ "finance", "search" ], "tags": [ "龙虎榜", "题材归因", "行业轮动", "盘后复盘", "市场情绪" ], "tags_en": [ "Dragon-Tiger Board", "Theme Attribution", "Sector Rotation", "Post-Close Review", "Market Sentiment" ], "artifact_type": "html", "prompt": "帮我用「A股全栈数据」skill 做一份指定交易日（如 2026-07-24）的 A股「龙虎榜+题材热点+行业轮动」市场情绪日报，做成单页 HTML，数据必须真实抓取并注明来源。\n\n请查询并整合这些真实数据：①全市场龙虎榜当日全部上榜股票（代码、名称、涨跌幅、净买入额、上榜原因、换手率），按净买入降序；②同花顺当日强势股的题材归因（reason 标签），把题材按「+」拆开做词频统计，得出当日 TOP 热门题材及出现只数；③东财全市场行业板块涨跌幅排名（领涨/领跌行业、上涨下跌家数、领涨股）。\n\n顶部放当日概览（上榜家数、净买入合计、最热题材、领涨行业）。做成单文件 HTML、资源全内联可离线打开，暗色盘面复盘风、信息密度高，用内联 SVG 画题材热度条形图、行业涨跌榜、龙虎榜净买入排名。界面简体中文，涨用红色、跌用绿色，金额用¥和亿/万单位。页脚注明数据来源（东方财富龙虎榜/同花顺热点/东财行业板块）与交易日期。", "prompt_en": "Use the A-Stock Data toolkit to build a single-page HTML A-share 'dragon-tiger board + hot themes + sector rotation' market-sentiment daily for a given trading day (e.g. 2026-07-24); all data must be pulled live with sources cited.\n\nQuery and integrate this real data: (1) all dragon-tiger board stocks that day (code, name, change%, net buy amount, listing reason, turnover), sorted by net buying descending; (2) THS strong-stock theme attributions (reason tags), split by '+' and frequency-counted to get the day's TOP hot themes and stock counts; (3) EastMoney market-wide sector gain/loss ranking (top/bottom sectors, advancers/decliners, leading stock).\n\nPut a daily overview at top (board count, total net buying, hottest theme, leading sector). Deliver a single-file HTML with all resources inlined and offline-openable, dark trading-desk review style, high information density, using inline SVG for the theme-heat bar chart, sector board and dragon-tiger net-buy ranking. UI in Simplified Chinese, up=red and down=green, amounts in ¥ with 100M/10K units. Footer must cite sources (EastMoney dragon-tiger / THS hotspots / EastMoney sectors) and the trading date.", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "龙虎榜+题材+行业市场情绪日报" } ], "skills": [ { "id": "a-stock-data", "name": "A股全栈数据", "description": "A 股行情、研报、资金流、公告与财报查询工具包", "description_en": "A-share data toolkit for quotes, reports, funds flow, announcements, news, and financial analysis", "skill_id": "skill_2059464154913587200" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-07-27T11:30:00Z", "updated_at": "2026-07-27T11:30:00Z", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "a-stock-lockup-expiry-radar", "title": "限售解禁预警雷达", "title_en": "Lockup Expiry Warning Radar", "subtitle": "提前 90 天扫出解禁抛压，哪只压力最大一眼看清", "subtitle_en": "Scan 90 days ahead for unlock selling pressure, spot the biggest at a glance", "description": "限售股解禁是A股绕不开的抛压来源，可解禁公告藏在各家披露里，散户往往等股价砸下来才反应过来。这个案例把未来90天（基准日 2026-07-24 起）全市场解禁排期一次性扫成一张预警雷达：415 批、涉及 390 只个股、合计解禁市值约 6010 亿元；按周分布画出时间轴，10-19 当周 1268 亿元的解禁高峰被高亮预警；重点压力榜里，中船特气以 944 亿元解禁市值居首。\n\n数据全部来自东方财富数据中心的限售解禁与分红真实接口，通过「A股全栈数据」skill 现场抓取，橙红警示配色、时间轴与条形图俱全。\n\n适合：想提前规避解禁风险的持股者、做事件驱动策略的交易者、跟踪个股筹码变化的研究者、需要给持仓做风控排查的投顾。", "description_en": "Lockup-share unlocking is an unavoidable source of A-share selling pressure, yet the unlock schedules are buried across scattered disclosures — retail investors often only react after the price has dropped. This case scans the entire market's unlock schedule for the next 90 days (from base date 2026-07-24) into one warning radar: 415 batches across 390 stocks, ~¥601B total unlock value; a weekly-distribution timeline highlights the ¥126.8B peak in the week of Oct 19; and on the pressure ranking, China Shipbuilding Special Gas tops it with ¥94.5B of unlock value.\n\nAll data comes from EastMoney Data Center's real lockup-expiry and dividend APIs, pulled live via the A-Stock Data toolkit, with amber-red warning colors and full timeline and bar charts.\n\nIdeal for: holders wanting to dodge unlock risk early, event-driven traders, researchers tracking share-structure changes, and advisors running risk checks on portfolios.", "categories": [ "finance", "data" ], "tags": [ "限售解禁", "解禁预警", "抛压风险", "事件驱动", "风控排查" ], "tags_en": [ "Lockup Expiry", "Unlock Warning", "Selling Pressure", "Event-Driven", "Risk Screening" ], "artifact_type": "html", "prompt": "帮我用「A股全栈数据」skill 做一份 A股「未来90天限售解禁预警雷达」单页 HTML（基准日如 2026-07-24），数据必须真实抓取并注明来源。\n\n请查询并整合这些真实数据：①全市场未来90天待解禁排期（东财数据中心 RPT_LIFT_STAGE，筛选解禁日期在未来90天窗口内），字段含解禁日期、股票代码/简称、解禁类型、解禁股数、解禁比例、解禁市值；②按解禁市值降序取 TOP 20~30 只作为重点预警名单；③对压力最大的几只补充实时行情估算解禁市值占流通盘比例；④挑 2~3 只代表股拉历史解禁记录和最近分红做详情展示；⑤统计未来90天按周的解禁批次与市值分布，找出解禁高峰周。\n\n顶部放概览（待解禁批次、涉及个股数、合计解禁市值、最近高峰）。做成单文件 HTML、资源全内联可离线打开，暗色风控预警雷达风（橙红警示色+青色信息色）、信息密度高，用内联 SVG 画未来90天按周解禁分布柱状图、解禁市值TOP条形图、解禁类型占比。金额用¥和亿/万单位，股数用万股/亿股，临近或高比例解禁标红预警。界面简体中文。页脚注明数据来源（东方财富数据中心）与基准日，并注明「非投资建议，解禁不必然导致下跌」。", "prompt_en": "Use the A-Stock Data toolkit to build a single-page HTML A-share 'next-90-day lockup expiry warning radar' (base date e.g. 2026-07-24); all data must be pulled live with sources cited.\n\nQuery and integrate this real data: (1) the market-wide unlock schedule for the next 90 days (EastMoney Data Center RPT_LIFT_STAGE, filtering free-date within the 90-day window), with fields for unlock date, stock code/name, unlock type, unlock shares, unlock ratio and unlock market value; (2) take the TOP 20-30 by unlock value as the key watchlist; (3) for the highest-pressure names, add live quotes to estimate unlock value as a share of float; (4) pick 2-3 representative stocks and pull their historical unlock records and recent dividends for detail cards; (5) tally the next 90 days' weekly unlock batches and value distribution to find the peak week.\n\nPut an overview at top (pending batches, stocks involved, total unlock value, nearest peak). Deliver a single-file HTML with all resources inlined and offline-openable, dark risk-control warning-radar style (amber-red warning + cyan info), high information density, using inline SVG for the weekly unlock-distribution bar chart, unlock-value TOP bar chart and unlock-type breakdown. Amounts in ¥ with 100M/10K units, shares in 10K/100M shares, flag imminent or high-ratio unlocks in red. UI in Simplified Chinese. Footer must cite the source (EastMoney Data Center) and base date, noting 'not investment advice; an unlock does not necessarily cause a decline'.", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "未来90天限售解禁预警雷达" } ], "skills": [ { "id": "a-stock-data", "name": "A股全栈数据", "description": "A 股行情、研报、资金流、公告与财报查询工具包", "description_en": "A-share data toolkit for quotes, reports, funds flow, announcements, news, and financial analysis", "skill_id": "skill_2059464154913587200" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-07-27T11:30:00Z", "updated_at": "2026-07-27T11:30:00Z", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "a-stock-moutai-valuation-checkup", "title": "贵州茅台估值体检", "title_en": "Kweichow Moutai Valuation Checkup", "subtitle": "一页看清贵不贵、机构怎么看、主力在进还是出", "subtitle_en": "One page: is it pricey, what do analysts think, is smart money in or out", "description": "想看一只票到底贵不贵，得东翻行情、西查财报、再翻研报和资金流，普通人根本凑不齐这些数据。这个案例把贵州茅台一次性做成全维体检单页：实时价¥1287、PE 19.5倍、PB 6.9倍开局，接上46家机构一致预期EPS算出前向PE和PEG，再叠上近120日主力资金净流入、融资余额、股东户数变化和最新研报评级，最后给出「贵不贵、能不能拿」的综合判断。\n\n所有数字都来自腾讯财经、东方财富、同花顺、新浪的真实公开数据，通过「A股全栈数据」skill 现场抓取，红涨绿跌、图表俱全。\n\n适合：持有或想买核心资产的价值投资者、想快速判断个股估值水位的散户、需要给客户出个股体检的投顾、做投研笔记的财经内容创作者。", "description_en": "To judge whether a stock is pricey, you have to dig through quotes, financials, research reports and fund flows — data an ordinary investor can never fully assemble. This case turns Kweichow Moutai into a one-page full-dimension checkup: starting from a live price of ¥1287, PE 19.5x and PB 6.9x, it adds consensus EPS from 46 institutions to derive forward PE and PEG, layers on 120-day net main-force inflows, margin balance, shareholder-count changes and the latest analyst ratings, and ends with an overall 'pricey or not, hold or not' verdict.\n\nEvery number is real public data from Tencent Finance, EastMoney, THS and Sina, pulled live via the A-Stock Data toolkit, with China-convention up=red/down=green coloring and full charts.\n\nIdeal for: value investors holding or eyeing blue chips, retail traders wanting a quick read on valuation levels, advisors producing stock checkups for clients, and finance content creators writing research notes.", "categories": [ "finance", "data" ], "tags": [ "贵州茅台", "估值分析", "一致预期", "PEG", "资金流向" ], "tags_en": [ "Kweichow Moutai", "Valuation", "Consensus Forecast", "PEG", "Capital Flow" ], "artifact_type": "html", "prompt": "帮我用「A股全栈数据」skill 给贵州茅台(600519)做一份全维估值体检的单页 HTML 报告，数据必须真实抓取并注明来源。\n\n请查询并整合这些真实数据：①实时行情（现价、PE(TTM)、PB、总市值/流通市值、换手率、涨跌幅）；②同花顺机构一致预期EPS（各年度均值、覆盖机构数），据此算前向PE、EPS增速CAGR、PEG，并判断当前PE消化到30倍需要几年；③最近120日主力资金净流入（日级）并统计近20日累计；④融资余额趋势；⑤股东户数变化（判断筹码集中还是分散）；⑥最近8条研报的机构、评级和标题；⑦分红送转历史。\n\n最后给出「贵不贵、筹码稳不稳、能不能拿」的综合判断（注明非投资建议）。做成单文件 HTML、资源全内联可离线打开，暗色科技投研风、信息密度高，用内联 SVG 画价格走势和资金流柱状图。界面简体中文，涨用红色、跌用绿色，金额用¥和亿/万单位。页脚注明数据来源（腾讯财经/东方财富/同花顺/新浪）与数据日期。", "prompt_en": "Use the A-Stock Data toolkit to build a single-page HTML full-dimension valuation checkup for Kweichow Moutai (600519); all data must be pulled live with sources cited.\n\nQuery and integrate this real data: (1) live quote (price, PE(TTM), PB, total/float market cap, turnover, change%); (2) THS consensus EPS by year and analyst coverage count, then compute forward PE, EPS CAGR, PEG, and how many years to digest the current PE down to 30x; (3) 120-day daily net main-force inflow and the 20-day cumulative; (4) margin balance trend; (5) shareholder-count change (concentrating or dispersing); (6) the latest 8 research reports with institution, rating and title; (7) dividend history.\n\nEnd with an overall verdict on 'pricey or not, chips stable or not, hold or not' (note: not investment advice). Deliver a single-file HTML with all resources inlined and offline-openable, dark tech research style, high information density, using inline SVG for the price trend and capital-flow bar charts. UI in Simplified Chinese, up=red and down=green, amounts in ¥ with 100M/10K units. Footer must cite sources (Tencent Finance / EastMoney / THS / Sina) and the data date.", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "贵州茅台全维估值体检报告" } ], "skills": [ { "id": "a-stock-data", "name": "A股全栈数据", "description": "A 股行情、研报、资金流、公告与财报查询工具包", "description_en": "A-share data toolkit for quotes, reports, funds flow, announcements, news, and financial analysis", "skill_id": "skill_2059464154913587200" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-07-27T11:30:00Z", "updated_at": "2026-07-27T11:30:00Z", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "aa-ledger-split-workbench", "artifact_type": "html", "title": "旅游AA账本", "title_en": "Travel AA Ledger", "subtitle": "出行路上记下每一笔共同消费，多币种自动折算，自动算清谁该给谁转多少", "subtitle_en": "Log every shared cost as you travel, convert currencies on the fly and work out who pays whom", "description": "一起出去玩最容易伤感情的不是花了多少钱，而是最后算不清谁该给谁转多少。这个工作台把 AA 分账拆成三张数据表：账目表记下每一笔谁付的钱、谁参与分摊，成员表管住一起花钱的人，设置表管基准币种和常用币种；记一笔时外币会按汇率自动折算成基准币种再分账，分摊方式支持平均分摊和按人自定义，账目页能按付款人筛选、导出与导入 CSV，成员页直接给出每人垫付、应摊、结余和「最少转账」还款方案，待确认的账目不计入结算。内置 10 笔泰国旅行演示账目（4 名成员，含一笔人民币汇率折算和一笔自定义分摊）与 1 笔待确认记录，打开就能看到完整的结算结果。适合和朋友出行 AA、合租分摊、多人聚餐后算账的人。", "description_en": "On a group trip the thing that strains a friendship is rarely the money itself — it is not knowing who owes whom at the end. This workbench splits AA bookkeeping into three data tables: the entries table records who paid and who shares each cost, the members table holds everyone who spends together, and the settings table keeps the base currency and common currencies. While recording an expense, a foreign currency is converted at the saved rate before splitting; the split can be even or customised per person; the entries tab filters by payer and imports/exports CSV; the members tab shows each person's paid, share and balance plus a \"fewest transfers\" settlement plan, and pending entries stay out of the settlement. It ships with 10 demo entries from a Thailand trip (4 members, including one CNY conversion and one custom split) and 1 pending record, so the full settlement result is visible on first open. Great for friends travelling together, flatmates splitting rent and bills, or any group settling up after a shared meal.", "tags": [ "AA记账", "多人分账", "旅行账本", "在线数据表" ], "tags_en": [ "bill splitting", "group expenses", "travel ledger", "online tables" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/h6j527OFEyiZZVDl3tMZhy ，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「旅游AA账本」多人分账工作台，暖色纸质旅行账本风格（深墨绿 + 米白 + 珊瑚橙），移动端优先、PC 自适应，用三张数据表管住全部数据。\n\n数据表结构：\n- 账目表：项目（文本）、日期（日期）、类别（单选：餐饮/交通/购物/住宿/娱乐/医疗/其他）、付款人（文本，存成员姓名）、参与人（文本，多人以 / 连接）、金额（数字，2 位小数）、分摊方式（单选：平均分摊/自定义）、自定义分摊（文本，形如「Leo=600.00/Iris=300.00」）、状态（单选：已确认/待确认）、备注（文本）、币种（单选，形如「THB 泰铢」）、汇率（数字，6 位小数，表示 1 单位该币种折多少基准币种）。\n- 成员表：姓名（文本）、备注（文本）。\n- 设置表：基准币种（单选）、常用币种（多选）、已初始化（复选框），只保留一行记录。\n\n三个页签：\n- 记一笔：大号金额输入区，币种下拉取自设置表的常用币种；选到非基准币种时出现汇率输入框并实时显示折算金额；类别、付款人用单选胶囊，参与人用多选胶囊；分摊方式选「自定义」时按参与人展开每人承担金额的输入格，合计与实付金额不符时拒绝提交并提示；日期默认今天；点账目卡片的编辑进入编辑态（按钮变「保存修改」，可取消回到新增）。\n- 账目：顶部按付款人筛选的胶囊筛选条；卡片显示项目、金额（外币显示原币种，并在副行注明折算金额与汇率）、日期、类别、付款人、几人分、状态标签，以及平均分摊每人多少或自定义分摊明细；卡片带编辑与删除入口，删除前二次确认；提供导出 CSV 与导入 CSV（导入可勾选先清空云端账目）；新增、编辑、删除后列表、合计与结算建议立即刷新。\n- 成员：成员增删改名（改名时可选择同步更新历史账目里的姓名，删除成员不影响已记账目）；顶部显示已确认共同支出合计与待确认笔数；每人一张卡片显示垫付、应摊与结余，正负用不同颜色和进度条区分；下方「怎么结（最少转账）」按债权债务配对，给出笔数最少的转账建议。\n\n首次打开（设置表尚无初始化记录）自动弹出 3 步初始化向导：添加成员 → 选基准币种 → 勾选常用币种；成员少于 2 人时下一步前给出确认提示；右上角齿轮可随时重新打开设置向导改成员与币种。\n\n多人在线：所有读写都走资料库数据表，一个人改动后其他人的页面会自动刷新；按记录逐条保存，不做整表覆盖。\n\n交付：资料库在线工作台，PC 与移动端自适应。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/h6j527OFEyiZZVDl3tMZhy with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows:\n\nBuild a \"Travel AA Ledger\" shared-expense workbench in a warm paper travel-ledger style (deep teal + cream + coral), mobile-first and responsive on desktop, keeping all data in three database tables.\n\nTables:\n- Entries: item (text), date (date), category (select: Dining / Transport / Shopping / Stay / Fun / Health / Other), payer (text, member name), participants (text, names joined by /), amount (number, 2 decimals), split method (select: split evenly / custom), custom split (text, e.g. \"Leo=600.00/Iris=300.00\"), status (select: confirmed / pending), note (text), currency (select, e.g. \"THB Thai Baht\"), rate (number, 6 decimals, meaning how much base currency 1 unit of that currency is worth).\n- Members: name (text), note (text).\n- Settings: base currency (select), common currencies (multi-select), initialised (checkbox) — a single row only.\n\nThree tabs:\n- Add an expense: a large amount input; the currency dropdown is built from the common currencies in Settings; picking a non-base currency reveals a rate input and shows the converted amount live; category and payer are single-select chips, participants are multi-select chips; choosing the custom split method expands one amount input per participant and blocks submission with a hint when the total does not match the amount paid; the date defaults to today; tapping Edit on an entry card enters edit mode (the button becomes \"Save changes\" and can be cancelled back to a new entry).\n- Entries: a chip filter bar by payer; each card shows item, amount (foreign currency shows the original currency plus a sub-line with the converted amount and rate), date, category, payer, how many people split it, a status tag and either the per-person amount for an even split or the custom split details; cards offer edit and delete with a confirmation before deleting; CSV export and import are provided (import can clear cloud entries first); creating, editing or deleting refreshes the list, totals and settlement suggestions immediately.\n- Members: add, rename and delete members (renaming can also update that name in past entries; deleting a member keeps existing entries); the top shows the confirmed shared total and the number of pending entries; each person gets a card with paid, share and balance, with different colours and bars for positive and negative; below, \"How to settle up (fewest transfers)\" pairs creditors with debtors to propose the smallest number of transfers.\n\nOn first open (no initialised record in Settings) a 3-step setup wizard appears automatically: add members → pick the base currency → tick the common currencies; warn before continuing when fewer than 2 members exist; the gear icon in the header reopens the wizard any time to change members and currencies.\n\nMulti-user online: every read and write goes through the Library data tables, so when one person changes something the others refresh automatically; records are saved one by one rather than overwriting the whole table.\n\nDeliverable: an online Library workbench, responsive on desktop and mobile.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "旅游AA账本（自包含演示）" } ], "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "categories": [ "travel" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd9b7744a331", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "skill_id": null, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 98, "created_at": "2026-09-11T19:40:00+08:00", "updated_at": "2026-09-11T20:28:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "abnormal-order-action-list", "title": "异常订单处理清单", "title_en": "Abnormal Order Action List", "subtitle": "待退款、超时未发货与地址异常订单分级处理", "subtitle_en": "A practical Weimob playbook for abnormal order action list", "description": "做异常订单处理清单时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“异常订单处理清单”展开：汇总最近 7 天待退款、超时未发货和地址异常订单，按照紧急程度生成处理顺序、建议和回复话术。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、订单履约运营人员、经营分析团队。", "description_en": "The hardest part of abnormal order action list is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates abnormal order action list through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Order Fulfillment specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "异常订单", "待退款", "超时未发货", "地址异常", "处理话术" ], "tags_en": [ "Abnormal Order Action List", "Mall Operations", "Order Fulfillment", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我找出最近7天待退款、超时未发货和地址异常的订单，按紧急程度排序列出明细，并给出每笔订单的处理建议和回复话术。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a abnormal order action list covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "acid-base-neutralization", "title": "酸和碱的中和反应互动教学", "title_en": "酸和碱的中和反应互动教学", "subtitle": "初中化学 · 酸和碱的中和反应互动教学", "subtitle_en": "初中化学 · 酸和碱的中和反应互动教学", "description": "初三化学《酸和碱的中和反应》互动课件，用动画演示电离与中和反应的微观本质，并配方程式练习和生活应用小测。", "description_en": "初三化学《酸和碱的中和反应》互动课件，用动画演示电离与中和反应的微观本质，并配方程式练习和生活应用小测。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "acid-base-neutralization.html", "artifacts": [ { "file": "acid-base-neutralization.html", "label": "酸和碱的中和反应互动教学｜互动教学网页" } ], "prompt": "帮我做一个初三化学《酸和碱的中和反应》的互动教学网页，要用动画演示电离和 H⁺ 与 OH⁻ 的结合过程。", "prompt_en": "帮我做一个初三化学《酸和碱的中和反应》的互动教学网页，要用动画演示电离和 H⁺ 与 OH⁻ 的结合过程。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "active-product-sales-status-check", "title": "在售商品状态与动销体检", "title_en": "Active Product Status and Sell-Through Check", "subtitle": "上架时间、近30天销量与库存状态诊断", "subtitle_en": "A practical Weimob playbook for active product status and sell-through check", "description": "做在售商品状态与动销体检时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“在售商品状态与动销体检”展开：核对当前在售商品的上架时间、近 30 天销量和库存，区分新品观察、正常动销和需重点运营商品，不直接执行下架。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、商品运营运营人员、经营分析团队。", "description_en": "The hardest part of active product status and sell-through check is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates active product status and sell-through check through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Product Operations specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "在售商品", "上架时间", "商品动销", "库存状态", "运营建议" ], "tags_en": [ "Active Product Status and Sell-Through Check", "Mall Operations", "Product Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "核对当前在售商品的上架时间、近30天销量和库存，区分上架不足30天的新品、正常动销商品和需要重点运营的商品，并输出逐商品建议；不要执行下架。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a active product status and sell-through check covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "after-close-trade-review", "title": "盘后两笔交易复盘", "title_en": "Two-Trade Post-Close Review", "subtitle": "把一笔早卖的盈利单和一笔午后逆势单拆开看清楚", "subtitle_en": "Dissect one early-taken winner and one low-probability afternoon long", "description": "收盘后最头疼的不是看盈亏，而是不知道问题到底出在买点、卖点还是节奏上。这个案例把 2026-07-27 的两笔短线样例交易先用官方交割单解析脚本配对，再结合真实指数、分时和资金数据，拆出为什么北化股份拿不住、为什么宝丰能源不该接。\n\n页面里演示了市场背景、操作统计、逐笔打分、共性问题和明日关注，重点不是复述行情，而是把执行偏差落到可复现的结构判断上。所有内容都在单文件双语 HTML 里完成，右上角可切换中文和英文。\n\n适合：想把交割单复盘做细的短线交易者；需要训练卖点纪律的人；希望把“盘感”变成可回看框架的 A 股参与者。", "description_en": "The hardest part after the close is not reading PnL, but figuring out whether the real issue came from the entry, the exit, or the trading rhythm. This case first pairs two sample trades from 2026-07-27 with the official trade-record parser, then reviews them against real index, intraday and flow data to explain why NCIC was sold too early and why BF Energy should not have been bought in the afternoon.\n\nThe page demonstrates market context, session metrics, trade-by-trade scoring, recurring mistakes and the next-day watchlist. The goal is not to retell the tape, but to convert execution mistakes into a repeatable structure-based framework. Everything is delivered in a single bilingual HTML page with a top-right language switch.\n\nSuitable for: short-term traders who want deeper post-trade review; anyone training exit discipline; A-share participants who want to turn intuition into a reviewable framework.", "categories": [ "finance", "data" ], "tags": [ "交割单复盘", "盘后诊断", "短线交易", "执行纪律" ], "tags_en": [ "trade review", "post-close diagnosis", "short-term trading", "execution discipline" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请按 A股短线交易 的盘后诊断流程处理以下样例交割单，并保留这份输入作为复现依据：2026-07-27 09:35 买入 北化股份 002246 19.42 5000股；2026-07-27 10:15 卖出 北化股份 002246 19.47 5000股；2026-07-27 13:31 买入 宝丰能源 600989 23.04 4000股；2026-07-27 14:30 卖出 宝丰能源 600989 22.84 4000股。先用官方 scripts/parse_trade_record.py 解析并完成买卖配对，再用 WeStock 官方数据查询 2026-07-27 对应标的和大盘的真实分时、行情与资金信息，生成一份单文件、资源全内联的 HTML 复盘页面。内容请包含市场背景、操作统计、逐笔评分、共性问题、明日关注与交易心得，并把分析写具体，不要用示意数据。", "prompt_en": "Please follow the post-close review workflow of A股短线交易 and use the following sample trade record as the reproducible input: 2026-07-27 09:35 buy NCIC 002246 at 19.42 for 5000 shares; 2026-07-27 10:15 sell NCIC 002246 at 19.47 for 5000 shares; 2026-07-27 13:31 buy BF Energy 600989 at 23.04 for 4000 shares; 2026-07-27 14:30 sell BF Energy 600989 at 22.84 for 4000 shares. First parse it with the official scripts/parse_trade_record.py and complete the buy-sell pairing, then use official WeStock data queries to pull the real intraday, quote and fund-flow data for the traded symbols and the market on 2026-07-27, and generate a single-file HTML review page with all resources embedded. Include market context, session metrics, trade-by-trade scoring, recurring issues, next-day watch items and trading notes, and keep the analysis specific rather than using placeholder data.", "skills": [ { "id": "ashare-short-term-trading", "name": "A股短线交易", "description": "A股短线交易体系，支持交割单复盘分析、策略筛选选股、盘中关键时间点推送三大场景。", "skill_id": "skill_2076627819394322432" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "artifacts": [ { "file": "output.html", "label": "双语复盘页面" } ], "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:05:00+08:00", "updated_at": "2026-07-27T20:16:07+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "after-sales-refund-cause-review", "title": "售后退款原因复盘", "title_en": "After-Sales Refund Cause Review", "subtitle": "本月售后退款趋势与优化建议", "subtitle_en": "A practical Weimob playbook for after-sales refund cause review", "description": "做售后退款原因复盘时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“售后退款原因复盘”展开：汇总本月售后单和退款原因，识别退款集中的时间段及高售后商品，并提出跨商品、物流和客服环节的改进建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、订单履约运营人员、经营分析团队。", "description_en": "The hardest part of after-sales refund cause review is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates after-sales refund cause review through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Order Fulfillment specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "售后复盘", "退款原因", "高售后商品", "退款趋势", "服务优化" ], "tags_en": [ "After-Sales Refund Cause Review", "Mall Operations", "Order Fulfillment", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "汇总本月截至昨天的售后单和退款原因，识别退款集中的时间段及高售后商品，并给出商品、物流和客服方面的优化建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a after-sales refund cause review covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-batch-image-processing", "title": "Agent 批量图片处理工作流", "title_en": "Agent Batch Image Processing Workflow", "subtitle": "多 Agent 协作完成电商产品图批量裁剪、水印、缩略图与格式转换", "subtitle_en": "Multi-Agent Workflow for Batch Cropping, Watermarking, Thumbnailing & Format Conversion", "description": "电商运营每次上新都要处理几十甚至上百张产品图：裁剪为平台要求尺寸、加品牌水印、生成缩略图、转换为 WebP 以优化加载速度……逐张手动处理耗时巨大，且容易遗漏或出错，导致图片规格不一致、水印偏移等问题。\n\n这个案例演示了一个多 Agent 批量图片处理工作流：规划 Agent 分析图片清单并制定处理方案（裁剪规则、水印位置、命名规范）；处理 Agent 对每张图执行 800x800 正方形裁剪、右下角品牌水印叠加、200x200 缩略图生成和 WebP 格式转换；质检 Agent 逐张验证输出质量（尺寸/水印/文件大小）；汇总 Agent 生成包含处理前后对比和统计信息的可视化报告。\n\n适合：电商运营、视觉设计师、内容运营、品牌方营销团队、需要批量图片处理的任何场景。", "description_en": "Every product launch requires processing dozens of images: cropping to platform specs, adding brand watermarks, generating thumbnails, converting to WebP — manual processing is slow and error-prone.\n\nThis case demonstrates a multi-agent batch image processing workflow: a Planning Agent analyzes the image list and defines processing rules; a Processing Agent executes cropping, watermarking, thumbnail generation, and format conversion for each image; a QA Agent validates output quality; a Summary Agent generates a visual processing report with before/after comparisons.\n\nIdeal for: e-commerce operations, visual designers, content teams, brand marketing teams, and any batch image processing scenario.", "categories": [ "agent", "creation" ], "tags": [ "批量图片处理", "电商运营", "多Agent协作", "自动化工作流" ], "tags_en": [ "Batch Image Processing", "E-commerce Ops", "Multi-Agent", "Automated Workflow" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我有一批 20 张电商产品图需要批量处理。请设计一个 Agent 工作流：1) 规划 Agent 分析图片列表，制定处理方案（尺寸裁剪规则、水印位置、命名规范）；2) 处理 Agent 对每张图片执行：自动裁剪为 800x800 正方形、添加品牌水印「ShopFlow」在右下角、生成缩略图 200x200、转换为 WebP 格式；3) 质检 Agent 验证每张图的输出质量（尺寸/水印/文件大小）；4) 汇总 Agent 生成处理报告页面。产出一个 HTML 格式的批量处理报告，展示每张图的处理前后对比、状态和统计。", "prompt_en": "I have a batch of 20 e-commerce product images that need batch processing. Please design an Agent workflow: 1) A Planning Agent analyzes the image list and defines a processing plan (size/cropping rules, watermark position, naming conventions); 2) A Processing Agent performs the following on each image: auto-crop to an 800x800 square, add the brand watermark \"ShopFlow\" in the bottom-right corner, generate a 200x200 thumbnail, and convert to WebP format; 3) A QA Agent verifies the output quality of each image (dimensions / watermark / file size); 4) A Summary Agent generates a processing report page. Produce a batch-processing report in HTML format that shows the before/after comparison, status, and statistics for each image.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "批量图片处理报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ai-liquid-cooling-chain-report", "title": "AI液冷产业链深度分析", "title_en": "AI Liquid Cooling Chain Deep Dive", "subtitle": "从液冷系统到关键零部件，定位 AI 算力扩张中的瓶颈段。", "subtitle_en": "Trace the chokepoints from system integration to core components in the AI liquid-cooling stack.", "description": "AI 算力越堆越高，真正卡住数据中心扩容的往往不是芯片本身，而是散热、供电与交付链条。这个案例围绕 AI 液冷产业链，演示如何从需求端一路拆到液冷系统、冷板、CDU、快接头与管路环节。\n\n报告会基于真实公开资料，区分哪些公司只是主题映射，哪些环节更可能因为认证、良率和交付能力而持续拿走利润，并给出可跟踪的证据链与风险点。\n\n适合：想快速理解 AI 液冷逻辑的人；需要做赛道拆解与公司映射的投研人员；要建立跟踪指标的产业分析师。", "description_en": "As AI compute density keeps climbing, the real bottleneck for data-center expansion is often not the chip itself but the cooling, power, and delivery chain. This case shows how to break the stack down from end demand to liquid-cooling systems, cold plates, CDUs, quick disconnects, and fluid piping.\n\nUsing real public materials, the report separates simple theme exposure from the segments more likely to keep capturing profit because of certification, yield, and delivery capability, then builds a traceable evidence chain and risk map.\n\nBest for: users who want a fast understanding of AI liquid cooling; researchers doing sector breakdown and company mapping; industry analysts building tracking indicators.", "categories": [ "finance", "data", "business" ], "tags": [ "AI液冷", "产业链", "数据中心", "冷板", "液冷机柜" ], "tags_en": [ "AI liquid cooling", "industry chain", "data centers", "cold plates", "liquid-cooled racks" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "完整分析报告" } ], "prompt": "请围绕 AI 液冷产业链做一份深度分析。先用真实查询获取最新产业链、供需、关键公司、财务与公告/研报信息，并注明来源与日期；再按“全景拆解 → 瓶颈环节 → 公司映射 → 风险与跟踪”的路径输出一份单文件、可离线打开的 HTML 报告，正文对重点公司标注证据等级并附跟踪里程碑。", "prompt_en": "Create a deep-dive report on the AI liquid-cooling chain. First use real queries to gather the latest industry-chain, supply-demand, company, financial, and filing/research information with explicit sources and dates; then present a single-file offline HTML report following the path of full-stack breakdown → bottleneck segments → company mapping → risks and tracking, with evidence grades and tracking milestones for the key companies.", "skills": [ { "id": "industry-chain-analysis", "name": "产业链分析", "description": "产业链/赛道深度研究，支持全景拆解与瓶颈环节、金矿雷达、Serenity五维瓶颈验证。", "skill_id": "skill_2076627978773680128" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 87, "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T09:17:31Z", "updated_at": "2026-07-28T09:40:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ai-task-manager-workbench", "title": "AI朱朱任务管理台", "title_en": "AI ZhuZhu Task Manager", "subtitle": "论文、项目、求职三线并进的研究生任务工作台", "subtitle_en": "A grad-student workbench for papers, projects and job hunting in one place", "description": "研究生日常是论文阅读、科研项目、组会汇报、求职准备多线并行，任务散在备忘录和表格里，哪些今天该做、哪些卡住了、离截止还有几天，全靠脑子记。\n\n这个案例展示了一个「AI朱朱任务管理」个人工作台：首屏三张今日卡片（今日重点 / Deadline 倒计时 / 卡点提醒）一眼看清今天要处理什么；任务按论文阅读、项目推进、组会准备分类管理，每张卡片带优先级、进度条和卡点标注；进度看板按待开始/进行中/已完成三列归拢，还有月度和年度汇总统计；独立的求职管理页把从科研积累到实习投递的时间线节点铺开，每个节点带倒计时和行动项，点一下就能推进状态。数据落在资料库的两张子数据表里（任务管理 + 求职时间线），新增、编辑、标记完成即改即存、跨设备同步，离线时自动降级本地存储。\n\n适合：论文和项目并行的研究生、正在准备实习求职的本科生、想要一个把任务和求职时间线收在一起的个人工作台的人。", "description_en": "Grad school means juggling paper reading, research projects, group-meeting reports and job prep at once. Tasks scatter across memos and spreadsheets, and you track deadlines, priorities and blockers in your head.\n\nThis case shows an \"AI ZhuZhu Task Manager\" personal workbench: three cards up top (Today's Focus / Deadline Countdown / Blockers) tell you what needs attention today at a glance. Tasks are organized by type — paper reading, project work, group meeting, others — each card carrying priority, a progress bar and blocker notes. A kanban view groups everything into To Do / In Progress / Done, with monthly and yearly summary stats. A dedicated job-hunt page lays out the timeline from building research experience to internship applications, each milestone with a countdown and action items, advanceable with one tap. Data lives in two child tables in WorkBuddy Library (Tasks + Job-Hunt Timeline) — add, edit and mark done, saved instantly and synced across devices, with automatic localStorage fallback when offline.\n\nBest for: grad students juggling papers and projects, undergrads preparing for internship recruiting, and anyone who wants tasks and a job-hunt timeline in one personal workbench.", "categories": [ "workspace" ], "tags": [ "任务管理", "求职时间线", "学习工作台", "进度看板", "个人管理" ], "tags_en": [ "Task Management", "Job Hunt Timeline", "Study Workbench", "Kanban", "Personal Management" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "AI朱朱任务管理工作台（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/BC6raMitDwnqIRIL2NS8lP，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n做一个「AI朱朱任务管理」个人工作台，单个完整 HTML 文件。\n\n包含：① 首屏「今天要处理」三张卡片——今日重点、Deadline 倒计时（按截止日排序、剩余天数分色）、卡点提醒；② 任务管理（新增/编辑/删除/标记完成，字段含任务类型、状态、优先级、进度、开始/截止日期、卡点、备注、今日重点，类型分论文阅读/项目推进/组会准备/其他，表单未提交内容自动缓存）；③ 进度看板（待开始/进行中/已完成三列，附月度/年度汇总弹窗：总览、完成率、按类型统计、卡点统计、已完成清单）；④ 求职管理页（时间线节点卡片：阶段、节点名称、截止倒计时、行动项、备注，一键切换未开始/进行中/已完成状态，顶部汇总总节点/已完成/待完成/下一节点）。\n\n视觉：浅灰底 + 白色圆角卡片 + 靛蓝主色，移动端优先、PC 自适应，截止时间按紧急度分红/橙/绿。制作完成后：① 将 HTML 存入我的资料库，并创建两张资料库子数据表（任务管理/求职时间线）绑定数据、预置一批演示数据；② 将页面发布为公开链接并返回给我CSS/JS 无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use WorkBuddy Library to make me an instant replica of https://www.workbuddy.cn/space/d/BC6raMitDwnqIRIL2NS8lP with online data storage and two-way sync. Prefer the one-click replica flow; if that capability is unavailable, build it from scratch as follows:/n/nBuild an \"AI ZhuZhu Task Manager\" personal workbench as a single complete HTML file. Include: ① a top \"Today's Focus\" area with three cards — today's priorities, deadline countdown (sorted by due date, color-coded by urgency), and blocker alerts; ② task management (add / edit / delete / mark done, with fields for type, status, priority, progress, start/due dates, blocker, notes and today's-focus flag; types are paper reading / project work / group meeting / other; unsubmitted form input is auto-cached); ③ a kanban view (To Do / In Progress / Done columns, plus monthly and yearly summary modals: overview, completion rate, stats by type, blocker stats, completed list); ④ a job-hunt page (timeline milestone cards with stage, name, countdown, action items and notes, one-tap status cycling through Not Started / In Progress / Done, and a top summary of total / done / remaining / next milestone). Visuals: light gray background, white rounded cards, indigo primary color, mobile-first with PC adaptation, deadlines colored red/orange/green by urgency. When done: ① save the HTML to my WorkBuddy Library and create two bound Library child tables (Tasks / Job-Hunt Timeline) with a pre-seeded demo dataset; ② publish the page to a public link and return it to me. Fall back to localStorage when the SDK is unavailable. Inline all CSS/JS with no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "牛津 Kate 朱朱", "name_en": "Oxford Kate ZhuZhu", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-17T10:30:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "ai-teahouse-event-workbench", "artifact_type": "html", "title": "AI小茶桌·活动运营台", "title_en": "AI Tea Table Event Ops Workbench", "subtitle": "选题匹配 · 方案设计 · 会后复盘 · 推文海报，一桌一档全流程在线同步", "subtitle_en": "Topic matching, agenda design, wrap-up and tweets & posters — one table, one record, fully in sync online", "description": "组织 AI 分享活动最怕东一张表、西一个文档：选题靠群接龙，排期靠拍脑袋，复盘完就散。这个工作台把选题意愿、档期匹配、流程单、复盘金句和推文海报放进同一张活动表里，谁想听、什么时候开、物料写没写，一眼看清。演示数据是 5 期「AI 小茶桌」活动，从灵感池到已完成全覆盖。适合社群组织者、团队分享会主理人，以及任何想用一张表管住一场活动的人。", "description_en": "Running an AI sharing event usually means scattered sheets and docs: topics collected in group chats, schedules guessed, retrospectives forgotten. This workbench puts topic demand, schedule matching, run sheets, highlight quotes and tweet & poster copy into one event table — who wants to attend, when it starts, whether assets are ready, all at a glance. The demo data covers 5 sessions of an \"AI Tea Table\" event, from idea pool to completed. Great for community organizers, internal sharing-session hosts, and anyone who wants one table to run one event.", "tags": [ "活动运营", "复盘沉淀", "社群组织", "流程单", "文案生成" ], "tags_en": [ "event ops", "retrospective", "community", "run sheet", "copy generation" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/b8N9LDNFMfOimiulhlVLKr，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「AI 小茶桌」分享活动的运营工作台，茶绿色调，面向小型 AI 分享活动的组织者，用一张「活动管理」表管住全流程。表字段：主题、嘉宾、嘉宾简介、方向（多选：Agent实战/大模型应用/AI工具提效/行业落地/提示词与工程/其他）、状态（灵感池/意愿收集中/已排期/已完成）、想听人数、观众时间偏好与嘉宾可用时间（多选：工作日晚上/周五晚上/周末下午/周末晚上）、活动日期、推文状态、海报状态、复盘总结、金句亮点、待改进、活动方案、推文文案、海报文案。页面包含：顶部同步状态与刷新按钮；「今天要处理」待办提醒（逾期补复盘、临期查物料、高热度可排期）；四个页签：选题匹配（匹配度=时段契合60%+想听热度40%，时段点击投票 +1 想听，推荐开桌时段）、方案设计（按活动形式/人数规模/时长生成流程单并可保存）、会后复盘（复盘/金句/待改进失焦自动保存，经验沉淀墙，各期想听人数柱状图）、推文海报（自动生成报名推文与海报文案，可编辑保存，状态跟踪）。内置 5 期演示活动数据，支持新建、编辑、删除选题，所有改动实时写回数据表。PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/b8N9LDNFMfOimiulhlVLKr with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build an ops workbench for an \"AI Tea Table\" sharing event, in a tea-green tone, for organizers of small AI meetups, managing the whole flow with one \"Event Management\" table. Fields: topic, guest, guest bio, direction (multi-select: Agent practice / LLM apps / AI productivity / Industry adoption / Prompt engineering / Other), status (idea pool / gauging interest / scheduled / completed), interested count, audience time preference and guest availability (multi-select: weekday evening / Friday evening / weekend afternoon / weekend evening), event date, tweet status, poster status, retrospective, highlight quotes, to-improve, event plan, tweet copy, poster copy. The page includes: a header with sync status and refresh; a \"To handle today\" reminder list (overdue wrap-ups, near-term asset checks, high-demand topics ready to schedule); four tabs: Topic Match (score = slot fit 60% + demand heat 40%, tap-a-slot voting +1 interested, suggested slot), Plan Design (generate a run sheet from format / size / duration, saveable), Wrap-up (retrospective / quotes / to-improve auto-save on blur, learnings wall, per-session interest bar chart), Tweets & Posters (auto-generated signup tweet and poster copy, editable and saveable, status tracking). Ships with 5 demo sessions; supports creating, editing and deleting topics, with every change written back to the table in real time. responsive on PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "AI小茶桌·活动运营台（自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "康皓凯", "name_en": "Kang Haokai", "avatar": "", "badge": "community" }, "source": "community", "skill_id": null, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 95, "created_at": "2026-09-02T18:10:00+08:00", "updated_at": "2026-09-02T18:48:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "aihot-company-tracker", "title": "OpenAI 最近 7 天动态追踪", "title_en": "AI Company Weekly Tracker", "subtitle": "一句话追踪任意 AI 公司 / 关键词近 7 天动态，生成商业情报 HTML 看板", "subtitle_en": "Track any AI company or keyword over the past 7 days and generate a business-intelligence HTML dashboard", "description": "支持 OpenAI / Anthropic / Google / DeepSeek / 智谱 / 任意关键词。调用 AI HOT 的 items 接口进行全文搜索，覆盖中文标题、英文标题与中文摘要，并按 5 个 category 分组生成商业情报 HTML 看板。页面包含 KPI 指标、分类分布图、关键词高亮、分组动态流与原文跳转，比“刷 X 列表”更高效，比直接问 GPT 更新。\n\n适合：AI 从业者、投资人、产品经理、竞品分析师", "description_en": "Supports OpenAI / Anthropic / Google / DeepSeek / Zhipu / any keyword. Uses AI HOT full-text search across Chinese titles, English titles, and Chinese summaries, then generates a business-intelligence HTML dashboard grouped into 5 categories. Includes KPI cards, category distribution, keyword highlighting, grouped activity feeds, and source links — more efficient than scrolling X and fresher than asking GPT directly.\n\nBest for: AI practitioners, investors, product managers, competitive analysts", "categories": [ "search", "business" ], "tags": [ "公司动态", "OpenAI", "Anthropic", "AI HOT", "竞品追踪" ], "tags_en": [ "Company News", "OpenAI", "Anthropic", "AI HOT", "Competitive Intel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我想追踪某家 AI 公司最近 7 天的动态，请生成一个单文件 HTML 商业情报看板。\n\n要求：\n1. 调用 aihot skill 的 items 接口（mode=selected、q=关键词、since=7天前、take=30）做全文搜索\n2. 按 5 个 category 分组：模型发布/更新 → 产品发布/更新 → 行业动态 → 论文研究 → 技巧与观点（顺序固定）\n3. 全局连续编号（不在版块内重新计数）\n4. 顶部展示时间窗、命中总数、关键词与分类分布 KPI；正文按 category 分段展示动态流\n5. 每条含：序号、标题、来源、北京时间相对时间、≤160 字摘要、原文链接（target=\"_blank\" rel=\"noopener noreferrer\"）；高亮命中的关键词\n6. 来源原样保留，不要简化\n7. 纯 HTML/CSS/JS 单文件，样式与脚本内联，无外部资源；响应式布局", "prompt_en": "I want to track an AI company over the past 7 days. Please generate a single-file HTML business-intelligence dashboard.\n\nRequirements:\n1. Call the aihot skill's items interface (mode=selected, q=keyword, since=7 days ago, take=30) for full-text search\n2. Group by 5 categories: Model Releases/Updates -> Product Releases/Updates -> Industry News -> Papers & Research -> Tips & Opinions (fixed order)\n3. Continuous global numbering (do not restart within sections)\n4. Top shows the time window, total hits, keyword, and category-distribution KPIs; the body presents the activity feed in sections by category\n5. Each item includes: sequence number, title, source, Beijing-time relative timestamp, summary <=160 chars, and source link (target=\"_blank\" rel=\"noopener noreferrer\"); highlight the matched keyword\n6. Keep sources verbatim, do not simplify\n7. A single pure HTML/CSS/JS file with inlined styles and scripts, no external resources; responsive layout", "skills": [ { "id": "aihot", "name": "AI HOT", "description": "AI 资讯聚合，提供日报、论文、公司动态追踪", "skill_id": "skill_2053079401717506048" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-09T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "公司动态 HTML 情报看板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "aihot-weekly-papers", "title": "最近一周 AI 论文盘点", "title_en": "Weekly AI Paper Roundup", "subtitle": "拉取过去 7 天 aihot 精选 AI 论文，生成浅色科技蓝学术档案库式 HTML 周报", "subtitle_en": "Fetch the past 7 days of curated AI papers from aihot and turn them into a light tech-blue academic HTML archive", "description": "面向 AI 研究者 / 创业者 / 内容创作者：调用 AI HOT 的 items 接口，按 category=paper + since=过去 7 天拉取精选论文，按发布时间倒序生成可阅读、可跳转的 HTML 学术档案库。页面用浅色科技蓝主题（浅蓝白底 + 深蓝主文 + 科技蓝点缀）+ 时间轴 + 论文卡片 + 来源徽章和摘要区替代纯 Markdown 列表，省去刷 X、HuggingFace Daily Papers、arXiv 的时间成本。\n\n适合：AI 研究者、技术 Leader、创业者、内容创作者", "description_en": "For AI researchers, founders, and content creators: calls AI HOT's items API with category=paper and a 7-day window, then turns curated papers into a readable HTML academic archive in reverse chronological order. A light tech-blue theme (light-blue/white background + deep-blue text + tech-blue accents), timeline, paper cards, source badges, and summaries replace a plain Markdown list — saving you from scrolling X, HuggingFace Daily Papers, and arXiv.\n\nBest for: AI researchers, tech leads, founders, content creators", "categories": [ "search", "education" ], "tags": [ "AI 论文", "研究综述", "AI HOT", "arXiv", "HuggingFace" ], "tags_en": [ "AI Papers", "Research Roundup", "AI HOT", "arXiv", "HuggingFace" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我盘点最近 7 天的 AI 论文，并生成一个单文件 HTML 学术档案库。\n\n要求：\n1. 调用 aihot skill 的 items 接口，拉取最近 7 天的精选论文（约 50 篇）\n2. 按发布时间倒序，扁平编号列表，标题用中文版\n3. 页面风格：浅色科技蓝学术档案库风，左侧时间轴 + 右侧论文卡片，来源生成简洁徽章；顶部展示时间窗与论文总数\n4. 每篇卡片含：序号、中文标题、来源、相对时间、200 字以内摘要、原文跳转链接（新标签打开）\n5. 不在输出里暴露 API 端点路径与参数名等基础设施细节\n6. 纯 HTML/CSS/JS 单文件，样式与脚本内联，禁止外部资源；响应式 + 滚动渐入动画", "prompt_en": "Help me round up AI papers from the last 7 days and generate a single-file HTML academic archive.\n\nRequirements:\n1. Call the aihot skill's items interface to fetch curated papers from the last 7 days (about 50)\n2. Sort by publication time descending, as a flat numbered list; use the Chinese title\n3. Page style: a light tech-blue academic archive look, left timeline + right paper cards, concise source badges; show the time window and total paper count at the top\n4. Each card includes: number, Chinese title, source, relative time, an abstract within 200 characters, and a link to the original (opens in a new tab)\n5. Do not expose infrastructure details such as API endpoint paths or parameter names in the output\n6. A pure HTML/CSS/JS single file, styles and scripts inline, no external resources; responsive + scroll fade-in animations", "skills": [ { "id": "aihot", "name": "AI HOT", "description": "AI 资讯聚合，提供日报、论文、公司动态追踪", "skill_id": "skill_2053079401717506048" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-09T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "AI 论文 HTML 档案库" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-technical-triad-review", "title": "A股技术面三维复盘", "title_en": "A-share Three-Lens Technical Review", "subtitle": "上证看风险，国证2000看进攻，广度看情绪", "subtitle_en": "Use the SSE for risk, Guozheng 2000 for offense, and breadth for sentiment", "description": "很多人收盘后会看见指数反弹，却说不清这是趋势转强，还是下跌过程里的技术性修复。真正难的是把风险、进攻和情绪拆开看，而不是被一根阳线带着跑。\n\n这个案例演示了如何用 A 股收盘后的真实指数、宽度和板块数据，做一份技术面三维复盘：先判断上证指数的风险位置，再看国证2000是否代表进攻修复，最后用涨跌家数、涨停数和量能验证情绪热度，并给出三档建仓参考与次日观察清单。\n\n适合：需要收盘后快速判断仓位节奏的主动投资者、做指数与题材联动复盘的人、想把“感觉”变成结构化判断的 A 股研究者。", "description_en": "Many people see an after-close rebound but still can't tell whether it signals a genuine trend turn or just a technical bounce inside a down leg. The hard part is separating risk, offense and sentiment instead of being dragged around by one green candle.\n\nThis case shows how to turn real A-share closing data into a three-lens technical review: first assess the SSE risk position, then check whether Guozheng 2000 confirms offensive repair, and finally validate sentiment with breadth, limit-ups and turnover before outlining a three-tier deployment ladder and a next-session watchlist.\n\nIdeal for active investors managing close-to-close pace, anyone reviewing the interplay between indexes and themes, and A-share researchers who want to replace vague feeling with a structured framework.", "categories": [ "finance", "data" ], "tags": [ "A股复盘", "技术面", "市场宽度", "国证2000", "单文件HTML" ], "tags_en": [ "A-share Review", "Technicals", "Market Breadth", "Guozheng 2000", "Single-file HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "技术面三维复盘 HTML报告" } ], "prompt": "请基于指定交易日做一份 A 股技术面三维复盘单文件 HTML。先判断当天是否为交易日；若休市直接说明。若有交易数据，请围绕上证指数、国证2000和全市场广度，输出首屏结论、风险面、进攻面、情绪面、主线板块、三档建仓参考、次日观察清单与风险提示。所有资源内联、可离线打开，使用真实数据并在页尾注明统计日期。", "prompt_en": "Create a single-file HTML A-share three-lens technical review for a specified trade date. First confirm whether the day is a trading session; if the market is closed, state that directly. If data exists, structure the page around the SSE Composite, Guozheng 2000 and whole-market breadth, and output a TL;DR, risk lens, offense lens, sentiment lens, mainline sectors, a three-tier deployment ladder, a next-session watchlist and risk notes. Fully self-contained for offline use, and based on real data with the statistics date shown in the footer.", "skills": [ { "id": "a-share-daily-review", "name": "A股每日复盘", "description": "A股每日收盘后全维度复盘报告生成，支持五种分析风格，输出自包含HTML报告。", "skill_id": "skill_2076628582742429696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:12:45+08:00", "updated_at": "2026-07-27T20:12:45+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "audio-spectrum-visualizer", "title": "音频频谱可视化器", "title_en": "Audio Spectrum Visualizer", "subtitle": "基于 Web Audio API 的实时音频频谱动画展示", "subtitle_en": "Real-time audio spectrum animation powered by Web Audio API", "description": "想要创建炫酷的音乐可视化效果？本案例展示如何构建一个实时音频频谱可视化器，包含柱状频谱和环形波形两种展示模式，配合彩虹渐变色彩和流畅动画，打造专业级音频可视化体验。适合音乐应用开发者、前端动效爱好者和多媒体项目需求者。", "description_en": "Want to create stunning music visualization effects? This case demonstrates how to build a real-time audio spectrum visualizer with bar spectrum and circular waveform displays, featuring rainbow gradient colors and smooth animations for a professional audio visualization experience.", "categories": [ "creation" ], "tags": [ "音频可视化", "Web Audio", "Canvas动画", "频谱分析" ], "tags_en": [ "Audio Visualization", "Web Audio", "Canvas Animation", "Spectrum Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "音频频谱可视化器页面" } ], "prompt": "请创建一个实时音频频谱可视化器页面（单文件 HTML，内联无外部依赖），深色主题、响应式。\n\n要求：Canvas 绘制 60+ 频谱柱条（彩虹渐变色，平滑过渡）、中央环形波形、Play/Pause 控制。用模拟数据驱动动画（Math.sin + 随机噪声），无需真实音频输入。", "prompt_en": "Please create a real-time audio spectrum visualizer page (single-file HTML, inline with no external dependencies), dark theme, responsive.\n\nRequirements: Use Canvas to draw 60+ spectrum bars (rainbow gradient colors, smooth transitions), a central circular waveform, and Play/Pause controls. Drive the animation with simulated data (Math.sin + random noise) — no real audio input needed.", "skills": [ { "id": "songsee", "name": "音频可视化", "description": "从音频生成频谱图和可视化", "skill_id": "skill_2053082736246448128" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "authorization-letter", "title": "商标事务授权委托书", "title_en": "Trademark Authorization Letter", "subtitle": "用清晰授权边界推进商标事务办理", "subtitle_en": "Advance trademark matters with clear authorization boundaries.", "description": "办理商标事务时，授权范围过宽或期限不清，容易造成代理权限争议。\n\n本案例演示如何生成一份可编辑的《商标事务授权委托书》，将办理事项、期限、禁止转委托、资料保密和撤销机制集中写明。\n\n适合：品牌负责人、法务人员、商标权人、代理协作团队。", "description_en": "Trademark filings can create agency disputes when the authorization scope or duration is unclear.\n\nThis case demonstrates an editable Trademark Authorization Letter that consolidates authorized matters, term, no-subdelegation restrictions, confidentiality and revocation.\n\nSuitable for: brand owners, legal teams, trademark holders and agency collaboration teams.", "categories": [ "business" ], "tags": [ "授权委托书", "商标事务", "权限边界", "品牌法务" ], "tags_en": [ "Authorization Letter", "Trademark Matters", "Authority Boundary", "Brand Legal" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请起草一份商标事务授权委托书。委托人为[委托人名称]，受托人为[受托人姓名/机构]，就“[商标名称]”商标授权办理[申请、补正、答复、领取文书等事项]，授权期限至[日期]。请明确不得转委托、不得超越授权或处分商标权，约定资料保密、撤销方式、责任承担和争议解决。请输出可直接签署的中文 Word 文书。", "prompt_en": "Please draft a trademark authorization letter. The principal is [委托人名称] and the agent is [受托人姓名/机构]. Authorize the agent to handle [申请、补正、答复、领取文书等事项] related to the “[商标名称]” trademark through [日期]. Clearly prohibit subdelegation, exceeding authority or disposing of trademark rights, and specify confidentiality, revocation, liability and dispute resolution. Produce a signable Chinese Word instrument.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-11T21:04:09+08:00", "updated_at": "2026-08-12T13:37:31+08:00", "artifacts": [ { "file": "output.html", "label": "V3.0 互动合同起草页面" }, { "file": "商标事务授权委托书.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "auto-shopping-compare", "title": "自动搜索商品并比价", "title_en": "Automated Product Search & Price Comparison", "subtitle": "让浏览器自己搜商品、翻页、抓价格，一键生成横向比价看板", "subtitle_en": "Let the browser search, paginate, and extract prices to build a side-by-side comparison board", "description": "想买东西时挨个开网页、逐条比价格，翻来翻去还容易漏、记不住，纯靠人工又慢又累。\n\n这个案例让 QQ 浏览器自动化 Skill 替你跑完整条链路：打开电商站点、在搜索框输入商品关键词、点击搜索、翻页滚动，再从结果页结构化抓取多个商品的标题、价格和店铺，最后整理成一份可离线打开的 HTML 比价看板，让你一眼看清自动化流程、也能方便地横向比较各商品价格。\n\n适合：网购前做比价、运营做竞品价格盯盘、任何需要批量抓取搜索结果的场景。", "description_en": "Shopping usually means opening pages one by one and comparing prices manually—slow, tiring, and easy to miss items or forget numbers.\n\nThis case lets the QQ Browser Automation Skill run the whole loop for you: open an e-commerce site, type a product keyword into the search box, click search, scroll and paginate, then structurally extract multiple products' titles, prices, and shops from the results, and finally compile an offline-openable HTML comparison board that lets you see the automation flow at a glance and compare product prices side by side.\n\nFor: pre-purchase price comparison, competitor price monitoring for operations, and any scenario requiring bulk extraction of search results.", "categories": [ "search", "business" ], "tags": [ "浏览器自动化", "电商比价", "表单交互" ], "tags_en": [ "Browser Automation", "Price Comparison", "Form Interaction" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "自动搜索商品并比价" } ], "prompt": "用 QQ 浏览器自动化 skill 帮我做一次电商比价。打开一个不需要登录的购物/搜索站点（比如当当网 search.dangdang.com 或京东搜索），在搜索框里输入一个具体的商品关键词（例如“保温杯”“机械键盘”任选一个），点击搜索按钮进入结果页，再通过滚动和翻页加载更多结果。然后从搜索结果里真实抓取 6-8 个商品，逐条提取商品标题、价格、店铺/品牌，以及评价数等可对比字段，并记录数据来源站点、搜索关键词和抓取时间。请务必真实搜索抓取并注明来源，不要使用任何示意或模拟数据。\n\n最后把结果整理成一个单文件 HTML（保存为 output.html），资源全部内联、可离线打开。让人能一眼看懂 skill 是怎么一步步自动完成搜索抓取的，并能方便地横向比较各商品价格、看出哪个最便宜，风格你来把握。", "prompt_en": "Use the QQ Browser Automation skill to run an e-commerce price comparison for me. Open a shopping/search site that requires no login (e.g. Dangdang search.dangdang.com or JD search), type a specific product keyword into the search box (pick one such as \"vacuum flask\" or \"mechanical keyboard\"), click the search button to reach the results page, then load more results by scrolling and paginating. Then genuinely extract 6-8 products from the search results, pulling each product's title, price, shop/brand, and comparable fields like review count, and record the source site, search keyword, and capture time. Be sure to really search and scrape, and cite the source—do not use any illustrative or mock data.\n\nFinally compile the results into a single-file HTML saved as output.html, with all resources inlined and openable offline. It should make it easy to see step by step how the skill automated the search and extraction, and to compare products side by side and spot the cheapest one; the visual style is up to you.", "skills": [ { "id": "qqbrowser-skill", "name": "QQ浏览器自动化", "description": "面向 AI 代理的浏览器自动化技能，可导航网页、填表、点击、截图、提取数据", "skill_id": "skill_2070018474844057600" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "official", "usage_count": 0, "created_at": "2026-07-27T03:49:07Z", "updated_at": "2026-07-27T06:59:49Z", "is_recommended": false, "recommended_order": 0 }, { "id": "b2b-quotation", "title": "3分钟出一份专业含税到货报价单", "title_en": "B2B Quotation", "subtitle": "客户、品类、数量一填，含税到货价自动核算", "subtitle_en": "Fill in customer, category and quantity — tax-inclusive delivered prices calculated automatically", "description": "输入客户、品类和数量，自动核算含税单价、运费、账期，生成可直接发客户的正式报价单。", "description_en": "Enter the customer, product category and quantities, and it automatically works out tax-inclusive unit prices, freight and payment terms, then generates a formal quotation you can send straight to the customer.", "categories": [ "creation" ], "tags": [ "销售报价", "B2B贸易", "贸易流通", "报价单" ], "tags_en": [ "Quotation", "B2B Trade", "Trade & Distribution", "Sales Pricing" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "含税到货报价单（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易销售专家。请为我生成一份可直接发给客户的正式报价单，要求专业、含税、可落地。\n\n【客户信息】\n- 客户名称：{客户名称，如 江苏鼎盛建设工程有限公司}\n- 联系人：{联系人及电话}\n- 交货地点：{交货地，如 南京市江宁区项目工地}\n\n【报价明细】\n请对以下每一项，结合当前市场行情核算含税到货单价：\n{逐行填写：品名/规格/材质/数量。例如：\n- 螺纹钢 HRB400E Φ20，120吨\n- 高线盘螺 HPB300 Φ8，50吨}\n\n【报价要求】\n1. 调用【找钢行情MCP】取真实行情、查询【行业信息资产库】里这个行业的定价与议价要点，核算每项含13%增值税的到货单价；\n2. 计算货物小计、运费、价税合计，金额附大写；\n3. 输出报价说明：含税口径、报价有效期、付款方式（预付比例/尾款账期）、交货周期、材质单质保；\n4. 以整洁的表格 + 报价单版式呈现，可直接发给客户。\n\n请生成完整报价单。", "skills": [ { "id": "quotation-generation", "name": "报价单生成", "description": "TradeBuddy把“我要给客户报这批货”变成一张规范、专业、能直接发出去的报价单——该有的项一个不缺，排版体面，可导出。价格由你提供，系统负责算合计、排格式。", "skill_id": "skill_2091154967066689536" } ], "experts": [ { "id": "xiaoshou-expert", "name": "B2B贸易销售专家", "description": "懂B2B贸易的销售专家：发货盘、写文案、出报价、盯商机、经营客户、看业绩、谈价成单，助你多成单。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_42GwnCqPUv7a" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "bakery-receipt-sales-poster", "title": "麦子烘焙活动日报海报", "title_en": "Bakery Daily Sales Poster", "subtitle": "借一条收银小票，把当天的数字印在上面", "subtitle_en": "A shop receipt carries the day's numbers", "description": "店铺日报最常见的样子是几张宫格卡片加一串数字，做出来像 Word 排的，发到朋友圈没人停下来看。\n\n这个案例给虚构的烘焙店「麦子烘焙」做了一张竖版战报：一条长长的收银小票从画面上方垂下来，“412 个今日售出”印在小票上，条形码压在小票底端，旁边是木板上一只撒了粉的酵母面包，右下角留着二维码的位置。物件是从面包店自己的日常里找出来的，不套固定版式。\n\n适合每天或每场活动要发一张战报的小微商家，也适合想参考“从店里找一件物件当主角”的海报设计者。", "description_en": "Shop daily reports usually look like a few grid cards and a string of figures. They read as if typeset in Word, and nobody stops to look when they are posted to Moments.\n\nThis case makes a vertical sales poster for a fictional bakery, Maizi Bakery. A long shop receipt hangs from the top of the frame with “412 sold today” printed on it, a barcode at its end, a floured sourdough loaf on a wooden board beside it, and a spot reserved at the lower right for a QR code. The object comes from the bakery's own daily life rather than a fixed layout.\n\nUseful for small merchants who post a poster every day or every event, and for poster designers exploring how to pick one object from the shop as the hero.", "categories": [ "creation", "business" ], "tags": [ "战报海报", "烘焙", "微信小店", "店铺运营" ], "tags_en": [ "Sales Poster", "Bakery", "WeChat Shop", "Shop Operations" ], "artifact_type": "image", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "麦子烘焙活动日报海报（PNG）" } ], "cover_image": "cover-portrait.png", "cover_mode": "portrait", "prompt": "为烘焙店「麦子烘焙」做一张微信小店活动日报海报，竖版约 9:16，用于发朋友圈和店铺动态。数字如下：今日售出 412 个，今日出炉 6 炉，活动第 3 天，明早 7:00 第一炉；统计时段 10 月 14 日 00:00 到 21:00，数据来源微信小店后台。\n\n创意不要宫格、卡片、居中堆叠这类规整排版，从面包店自己的日常里借一件物件当主角：借一条收银小票。一条长长的小票卷从画面上方垂下，“412 个”和“今日售出”印在小票上，虚线下是“今日出炉 6 炉、第 3 天”，小票底端是条形码；旁边是木砧板上一只撒了面粉的酵母面包，右侧写“明早 7:00 第一炉”。画面右下角留一块白色方形空白放二维码，底部一行小字写数据来源和统计时段。\n\n整体奶油色底、焦糖棕色字，画面用写实棚拍质感，最大的数字最先被看见。\n\n战报数字先请商家确认；海报只呈现商家后台的真实数字，不编造销量或评价；二维码由程序贴在预留的空白处。\n\n交付：一张 PNG 竖版海报。", "prompt_en": "Create a WeChat Shop activity daily poster for a bakery, Maizi Bakery, in vertical, about 9:16, for Moments and shop updates. Figures: 412 items sold today, 6 bakes today, day 3 of the promotion, first bake tomorrow at 7:00; period 14 October 00:00 to 21:00, source: WeChat Shop back office.\n\nAvoid tidy grids, cards or centred stacks. Borrow one object from the bakery's own day as the hero: a shop receipt. A long receipt roll hangs from the top of the frame with “412” and “sold today” printed on it, “6 bakes today, day 3” under a dashed line, and a barcode at its end. Beside it is a floured sourdough loaf on a wooden board, with “First bake tomorrow at 7:00” on the right. Leave a white square at the lower right for a QR code and put one small line at the bottom for the source and period.\n\nCream background and caramel-brown type in a realistic studio-photo look, with the biggest number seen first.\n\nConfirm the merchant's figures first; the poster only shows real figures from the merchant's back office and never invents sales or reviews; the QR code is placed programmatically on the blank area reserved for it.\n\nDeliverable: one vertical PNG poster.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 97, "created_at": "2026-10-15T15:30:00+08:00", "updated_at": "2026-10-15T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "beauty-personal-care-bsr-analysis", "title": "亚马逊美容个护畅销榜竞品全景分析", "title_en": "Amazon Beauty & Personal Care BSR Competitive Landscape Analysis", "subtitle": "基于 Top 30 实时榜单的品牌格局、价格带机会与品类切入策略", "subtitle_en": "Brand landscape, price band opportunities, and category entry strategy based on Top 30 real-time BSR data", "description": "想做亚马逊美妆个护类目，最头疼的就是不知道哪个价格带还有机会、哪些品牌在统治市场。\n\n这个案例演示了如何只通过一个 Amazon 畅销榜链接，完成一次完整的品类竞品全景分析：自动抓取 Top 30 商品数据，分析品牌集中度、价格带分布、子品类结构，识别 K-Beauty 崛起趋势，并给出从价格带选择到 SKU 组合的完整切入策略。\n\n适合：亚马逊选品经理、跨境电商运营、品牌出海负责人、美妆个护品类创业者。", "description_en": "Entering Amazon's beauty & personal care category is daunting without knowing which price bands have room and which brands dominate.\n\nThis case demonstrates a complete competitive landscape analysis using just one Amazon Best Sellers URL: auto-extract Top 30 product data, analyze brand concentration, price band distribution, sub-category structure, identify the K-Beauty trend, and deliver a full entry strategy from price positioning to SKU matrix.\n\nFor: Amazon product selection managers, cross-border e-commerce operators, brand globalization leads, beauty & personal care entrepreneurs.", "categories": [ "business" ], "tags": [ "亚马逊", "选品分析", "竞品分析", "美妆个护", "BSR榜单" ], "tags_en": [ "Amazon", "Product Selection", "Competitive Analysis", "Beauty & Personal Care", "BSR Ranking" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.html", "prompt": "请走在线快速分析路径：抓取亚马逊 Beauty & Personal Care 畅销榜（https://www.amazon.com/Best-Sellers-Beauty-Personal-Care/zgbs/beauty），做品类竞品全景分析。\n\n要求：\n1. 提取 Top 30 商品的排名、品牌、ASIN、产品名称、价格\n2. 分析品牌竞争格局（集中度 CR3/CR5、品牌矩阵）\n3. 分析价格带分布与子品类结构\n4. 识别趋势品类（如 K-Beauty）\n5. 给出品类切入策略建议（价格带、SKU 组合、品牌定位）\n6. 生成含交互图表的自包含 HTML 报告", "prompt_en": "Use the online quick analysis path: fetch the Amazon Beauty & Personal Care Best Sellers page (https://www.amazon.com/Best-Sellers-Beauty-Personal-Care/zgbs/beauty) and perform a complete competitive landscape analysis.\n\nRequirements:\n1. Extract Top 30 products: rank, brand, ASIN, product name, price\n2. Analyze brand competitive landscape (CR3/CR5 concentration, brand matrix)\n3. Analyze price band distribution and sub-category structure\n4. Identify trending categories (e.g. K-Beauty)\n5. Deliver category entry strategy (price band, SKU matrix, brand positioning)\n6. Generate a self-contained HTML report with interactive charts", "skills": [ { "id": "amazon-category-selection-analysis", "name": "亚马逊品类竞品全景分析", "description": "亚马逊品类竞品全景分析，覆盖选品调研、品类切入评估、竞品格局诊断、价格带机会分析与VOC驱动产品开发", "skill_id": "skill_2085355112879824896" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "report.html", "label": "品类竞品全景分析 HTML 报告" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-04T13:56:40+08:00", "updated_at": "2026-08-04T13:56:40+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "behavioral-crowding-risk-radar", "title": "A股群体心理风险雷达", "title_en": "A-Share Behavioral Risk Radar", "subtitle": "用真实市场数据拆解拥挤交易、融资挤压与情绪偏差", "subtitle_en": "Decode crowded trades, leverage stress, and sentiment bias with real market data", "description": "最难的不是看见市场波动，而是分清这是一轮正常回撤，还是拥挤交易开始松动。这个案例用大盘走势研判 skill 的群体心理路径，把真实 A 股广度、板块轮动、融资变化、主力流向和热度线索放到一张风险雷达里，帮助快速识别 FOMO、代表性偏差、杠杆拥挤与风格背离。\n\n演示内容不是泛泛复盘，而是用 westock-data / westock-tool 的官方查询结果，生成一份可离线打开的单文件双语 HTML 报告，包含一句话结论、行为偏差热力图、关键证据板、风险分层卡片，以及尽量标注来源与日期的数据说明。\n\n适合：盘后做市场复盘的短线交易者、关注拥挤风险与行为金融的研究员、需要把复杂市场信号讲清楚的投资内容作者、想快速判断当前市场是进攻还是防守阶段的 A 股参与者。", "description_en": "The hard part is not seeing market volatility. It is telling a normal pullback from the moment when crowded trades start to loosen. This case uses the behavioral-finance path of the Market Trend Assessment skill to place real A-share breadth, sector rotation, margin changes, main-force flows, and heat signals into one risk radar, so users can quickly spot FOMO, representativeness bias, leverage crowding, and style divergence.\n\nThis demo is not a generic market recap. It uses official queries from westock-data and westock-tool to generate a single-file bilingual HTML report that works offline, including a one-line conclusion, a behavioral-bias heatmap, an evidence board, layered risk cards, and data notes with sources and dates wherever possible.\n\nSuitable for: short-term traders doing post-close review, researchers tracking crowding risk and behavioral finance, investment content creators who need to explain market signals clearly, and A-share participants who want a fast read on whether the market is in offense or defense mode.", "categories": [ "finance", "data", "search" ], "tags": [ "行为金融", "拥挤度", "融资风险", "市场情绪", "A股复盘" ], "tags_en": [ "Behavioral Finance", "Crowding", "Margin Risk", "Market Sentiment", "A-Share Review" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "群体心理风险雷达" } ], "prompt": "请用大盘走势研判 skill 做一份“群体心理风险雷达”A 股报告，重点不是泛泛复盘，而是用真实查询结果识别行为金融偏差、拥挤度/融资/热度与风险分层。先遵循 wb-finance-skill 的红线、时间口径与 HTML 规范，再通过 westock-data / westock-tool 的官方能力查询真实数据，至少覆盖市场涨跌分布与市场总览、上证/深成指/创业板指表现、板块排行、热搜股票、核心宏观指标、两融变化、主力 5 日净流入和连板高度。输出单文件 HTML，资源全内联，可离线打开。页面做成明亮、杂志化的分析板与风险热力图风格，包含一句话结论、行为偏差热力图、关键证据表、风险分层卡片和数据来源/日期说明，文末加入“不构成投资建议”的合规免责声明。", "prompt_en": "Use the Market Trend Assessment skill to produce an A-share report titled \"Behavioral Risk Radar\". The goal is not a generic market recap, but to identify behavioral-finance bias, crowding, margin stress, heat signals, and layered risk with real query results. First follow the red lines, time conventions, and HTML rules of wb-finance-skill, then query real data through the official capabilities of westock-data and westock-tool. At minimum, cover market breadth and market overview, performance of the SSE Composite / SZSE Component / ChiNext, sector ranking, hot stocks, core macro indicators, margin changes, 5-day main-force inflows, and the limit-up ladder. Output a single-file HTML report with all resources inlined so it can be opened offline. Use a bright, magazine-style analysis board and risk-heatmap direction. Include a one-line conclusion, a behavioral-bias heatmap, key evidence tables, layered risk cards, and notes for sources and dates. End the page with a compliance disclaimer stating that the report is not investment advice.", "skills": [ { "id": "market-trend-assessment", "name": "大盘走势研判", "description": "A股大盘走势研判，三套框架：短线情绪流、周线阶段论、行为金融。", "skill_id": "skill_2076627841600368640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-07-28T17:23:05.020265+08:00", "updated_at": "2026-07-28T17:50:00+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "beneficiary-photo-mask-multi", "title": "受助人多人合影脱敏", "title_en": "Group Photo Anonymization for Beneficiaries", "subtitle": "合影里逐张识别、批量贴合——集体活动照一次搞定", "subtitle_en": "Detect every face in the group, batch-anonymize — one pass, one deliverable", "description": "公益机构在整理活动现场照片、集体合影或项目影像资料时，常常被多人脸批量脱敏的繁琐工作压住——人脸要逐一识别、位置要逐张对齐、合照里容易漏掉几张脸、几十张照片处理下来常常半天就过去了，手动逐张修图既费时又容易出错。\n\n让「公益人脸照片脱敏专家」接住这个活：它自动扫描多人照片中的所有人脸区域，对每一张正脸逐一施加口罩贴纸脱敏，批量输出保留场景原貌、所有人脸均已脱敏处理的合影照片。\n\n适合：公益机构宣传人员、项目执行人员，以及需要批量处理活动现场多人合影照片的隐私合规场景。", "description_en": "Handling event photos, group shots, or campaign imagery means anonymizing many faces at once — spotting every face, aligning stickers one by one, catching every partial face in the crowd. Dozens of photos routinely eat half a day and still leave errors.\n\nLet the Beneficiary Photo Anonymization Expert take it: automatically scans group photos for every face region, applies a mask sticker to each frontal face, and batch-outputs anonymized group photos that keep the scene intact.\n\nFor: comms and project execution staff at nonprofits, and any team routinely processing group-shot compliance for events.", "categories": [ "creation", "business" ], "tags": [ "公益", "照片脱敏", "合影处理", "批量处理" ], "tags_en": [ "Public Welfare", "Photo Anonymization", "Group Photo", "Batch Processing" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我会给你上传几张照片，你帮我给照片上的人脸戴上口罩。", "prompt_en": "I'll upload a few photos — please put a mask on the faces in them.", "skills": [], "experts": [ { "id": "ex_OnNMGGf0L1nj", "name": "公益受助人照片脱敏专家", "description": "对受助人照片中的人脸自动脱敏（口罩贴纸），保护隐私同时保留场景", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "4a0517f0f1e8", "expert_id": "ex_OnNMGGf0L1nj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "多人合影脱敏样例（学校发放活动集体照）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "beneficiary-photo-mask-single", "title": "受助人单人照片脱敏", "title_en": "Single-Person Photo Anonymization for Beneficiaries", "subtitle": "识别人脸、贴合位置、自然脱敏——单张照片一步到位", "subtitle_en": "Detect the face, align the sticker, anonymize naturally — done in one step", "description": "公益机构在发布项目进展、编制年报或制作宣传素材时，常常被受助人照片的隐私合规问题绊住——正脸要识别、面部要精确定位、贴纸要自然贴合、输出还不能失真，手动在修图软件里逐张处理往往要花好几个小时，还总怕漏掉一张脸、贴歪一个位置。\n\n让「公益人脸照片脱敏专家」接住这个活：它自动扫描单人照片中的人脸区域，精准定位面部位置，对正脸施加口罩贴纸脱敏处理，输出一份既保护受助人隐私、又保留画面完整性的脱敏照片。\n\n适合：公益机构宣传人员、项目执行人员，以及需要快速合规处理受助人单人照片的隐私保护场景。", "description_en": "Publishing project updates, compiling annual reports, or preparing marketing materials means clearing beneficiary photo privacy — detecting frontal faces, locating them precisely, applying stickers that look natural, and keeping the image undistorted. Doing this manually in photo software eats hours and still risks a missed face or a crooked sticker.\n\nLet the Beneficiary Photo Anonymization Expert handle it: automatically scans a single-person photo, precisely locates the face, applies a mask sticker to the frontal face, and outputs an anonymized image that protects the beneficiary while preserving the scene.\n\nFor: comms and project execution staff at nonprofits, and anyone needing quick, compliant handling of single-person beneficiary photos.", "categories": [ "creation", "business" ], "tags": [ "公益", "照片脱敏", "隐私保护", "图片处理" ], "tags_en": [ "Public Welfare", "Photo Anonymization", "Privacy Protection", "Image Processing" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我会给你上传几张照片，你帮我给照片上的人脸戴上口罩。", "prompt_en": "I'll upload a few photos — please put a mask on the faces in them.", "skills": [], "experts": [ { "id": "ex_OnNMGGf0L1nj", "name": "公益受助人照片脱敏专家", "description": "对受助人照片中的人脸自动脱敏（口罩贴纸），保护隐私同时保留场景", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "4a0517f0f1e8", "expert_id": "ex_OnNMGGf0L1nj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "单人照片脱敏样例（受助儿童正脸贴花）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "best-selling-product-basket-size-analysis", "title": "热销商品连带件数分析", "title_en": "Best-Selling Product Basket-Size Analysis", "subtitle": "热销商品订单购买件数与连带销售分析", "subtitle_en": "A practical Weimob playbook for best-selling product basket-size analysis", "description": "做热销商品连带件数分析时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“热销商品连带件数分析”展开：找到近 30 天销量最高的商品，分析其订单购买件数分布，判断用户以单件购买还是多件连带购买为主。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、商品运营运营人员、经营分析团队。", "description_en": "The hardest part of best-selling product basket-size analysis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates best-selling product basket-size analysis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Product Operations specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "热销商品", "购买件数", "连带销售", "订单分析", "商品策略" ], "tags_en": [ "Best-Selling Product Basket-Size Analysis", "Mall Operations", "Product Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我找到近30天销量最高的商品，分析它的订单购买件数分布，看看用户是单买更多，还是经常一次购买多件。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a best-selling product basket-size analysis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "best-selling-product-sales-script", "title": "热销商品导购话术", "title_en": "Best-Selling Product Sales Script", "subtitle": "热销商品与促销活动的五阶段导购话术", "subtitle_en": "A practical Weimob playbook for best-selling product sales script", "description": "做热销商品导购话术时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“热销商品导购话术”展开：结合近期热销商品和促销活动生成覆盖迎宾、需求了解、卖点讲解、异议处理和成交引导的完整导购话术。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购任务与话术运营人员、经营分析团队。", "description_en": "The hardest part of best-selling product sales script is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates best-selling product sales script through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Tasks and Scripts specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "导购话术", "热销商品", "促销活动", "异议处理", "成交引导" ], "tags_en": [ "Best-Selling Product Sales Script", "Store and Sales Associate Operations", "Sales Associate Tasks and Scripts", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "针对近期热销商品和促销活动生成导购话术，覆盖迎宾、需求了解、卖点讲解、异议处理和成交引导。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a best-selling product sales script covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "bid-response-document-review-1", "title": "标书提交前自检与估分", "title_en": "Bid Response Review & Score Estimation", "subtitle": "全量抽取标书正文，逐项排查填写缺口，按评标办法重估得分", "subtitle_en": "Full-text extraction, gap check, and score re-estimation against the evaluation criteria", "description": "标书写完要提交了，最怕的就是有填写缺口或评分隐患没发现，等到评审被扣分甚至废标才后悔。这个案例演示了如何把一份半导体晶圆检测设备采购的投标响应文件和评标办法交给 AI：全量抽取正文（含表格、目录、声明），逐项核验商务符合性应答、目录页码、盖章声明、占位残留与修订痕迹，按评标办法核定评分口径并给出保守/基准/乐观三档估分，最后产出一份自包含 HTML 核查报告，把售后承诺漏答、业绩案例混同资格案例、团队配置缺失等风险点单列提示。\n\n适合：投标人提交前的标书自查、招投标评审前的疏漏排查、需要按评标办法估算得分的商务团队、以及负责标书合规审核的质量/法务同学。", "description_en": "The most annoying thing after finishing a bid document is submitting it with undetected gaps or scoring risks, only to lose points or get disqualified at evaluation. This case shows how to hand a semiconductor wafer inspection equipment bid response and its evaluation criteria to AI: extract the full text (including tables, TOC and declarations), verify compliance response tables, TOC page numbers, seals and placeholders one by one, re-estimate scores in conservative / baseline / optimistic scenarios per the evaluation criteria, and produce a self-contained HTML review report that flags missing after-sales commitments, performance cases mixed with qualification cases, and missing team configuration as standalone risks.\n\nPerfect for: bidders self-checking before submission, gap hunting before tender evaluation, business teams estimating scores per the evaluation criteria, and quality/legal staff reviewing bid compliance.", "categories": [ "productivity", "business" ], "tags": [ "标书核查", "投标自检", "估分", "评标办法", "文档审阅" ], "tags_en": [ "Bid review", "Tender self-check", "Score estimation", "Evaluation criteria", "Document review" ], "artifact_type": "html", "preview": "核查报告.html", "prompt": "对投标响应文件做提交前全面自检与估分，全程只读，不改动原文件。\n\n输入材料（放置在工作区内）：\n- 投标响应文件.docx：待核查的标书（含商务响应、技术方案、业绩案例、声明等章节与表格）\n- 评标办法.xlsx：评分规则（商务/技术/价格分值权重、业绩有效性、资质与质保要求）\n\n核查步骤：\n1. 全量抽取正文（含表格、目录、声明），逐项排查填写缺口：商务符合性应答表是否逐项作答、索引/目录页码是否填好、盖章/声明是否到位、有无「待补充 / TODO / Pg XX」之类占位残留；\n2. 检查售后本地化承诺等关键应答项是否漏答；\n3. 依据评标办法核定评分口径（商务/技术/价格权重、业绩有效性判定、资质与质保要求），重估得分并区分保守/基准/乐观三种情形，说明各自假设；\n4. 把「图片类证书只能文字引用核验、实际装订需人工目检」这类事项单列提示。\n\n输出：自包含 HTML 核查报告（内联样式、无外部资源、可离线打开），包含「核查概要 / 调整项核验 / 评分风险与建议 / 重估得分 / 待处理检查项」五个区块，保存为 核查报告.html。", "prompt_en": "Perform a full pre-submission self-check and score estimation on the bid response document. Read-only throughout; do not modify the original files.\n\nInputs (placed in the working directory):\n- 投标响应文件.docx (bid response document): the tender document to review (chapters and tables covering commercial response, technical proposal, performance cases, declaration, etc.)\n- 评标办法.xlsx (evaluation criteria): scoring rules (commercial/technical/price weights, performance validity, qualification and warranty requirements)\n\nReview steps:\n1. Extract the full text (including tables, TOC and declarations) and check every item for gaps: whether the commercial compliance response table is fully answered, whether index/TOC page numbers are filled in, whether seals/declarations are in place, and whether placeholder leftovers such as “待补充 / TODO / Pg XX” remain;\n2. Check whether key response items such as the after-sales localization commitment are missing;\n3. Verify the scoring criteria per the evaluation rules (commercial/technical/price weights, performance validity, qualification and warranty requirements), re-estimate scores in conservative / baseline / optimistic scenarios, and state the assumptions of each;\n4. List items like “image-based certificates can only be verified by text reference; actual binding requires manual visual inspection” as standalone notices.\n\nOutput: a self-contained HTML review report (inline styles, no external resources, opens offline) with five sections — “Overview / Item Verification / Scoring Risks & Suggestions / Re-estimated Scores / Pending Items” — saved as 核查报告.html.", "skills": [ { "id": "bid-response-document-review", "name": "投标响应文件审阅与复核", "description": "对 .docx 投标响应文件做只读检查与复核打分：全量抽取正文、按清单排查填写缺口（表格、页码、应答、盖章、目录、修订、占位符），依据评标办法核定评分标准并重估分数，产出结构化核查报告。适用于定稿前自查标书完整性、排查疏漏、估算得分的中文招投标场景。", "skill_id": "skill_2090718419809648640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-08-21T04:59:49Z", "updated_at": "2026-08-21T04:59:49Z", "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "核查报告.html", "label": "核查报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "bid-response-document-review-2", "title": "投标响应文件只读复核", "title_en": "Bid Response Document Review", "subtitle": "定稿前逐条核验供应商调整项，输出带证据定位的自包含 HTML 核查报告", "subtitle_en": "Verify every supplier adjustment item before finalizing and output a self-contained HTML report with evidence", "description": "标书定稿前最怕供应商说「改了」其实没改到位，逐页翻文件核对表格宽度、目录页码、应答逐项、盖章声明，又慢又容易漏。\n\n本案例把供应商回传的 4 项待办调整清单交给 AI 做只读核验：脚本全量抽取 docx 正文与表格结构，逐条验证「表格列宽是否超出页面可用宽度、Pg XX 占位是否残留、符合性应答是否逐项作答并引用资质 3.1/检验 3.2/业绩 4.1、盖章声明关键词出现次数与位置」，每条结论都附脚本扫描证据与正文定位，最终输出自包含 HTML 核查报告。\n\n适合：投标经办人、标书制作人员、参与投标的项目经理、需要复核供应商交稿质量的采购与商务同事。", "description_en": "The most annoying thing before a bid document is finalized: the supplier says \"fixed\" but nothing was actually fixed. Going page by page to check table widths, TOC page numbers, compliance answers, and seals is slow and easy to miss.\n\nThis case hands the supplier's 4-item adjustment checklist to AI for read-only verification: scripts fully extract the docx body and table structure, then verify item by item — whether table column widths exceed the usable page width, whether Pg XX placeholders remain, whether compliance responses are answered item by item with references to sections 3.1/3.2/4.1, and where seal/declaration keywords appear. Every conclusion carries script evidence and location, and the final output is a self-contained HTML review report.\n\nIdeal for: bid handlers, bid document preparers, project managers involved in bidding, and procurement or business colleagues who need to verify supplier deliverables.", "categories": [ "office", "productivity" ], "tags": [ "标书核查", "投标响应", "文档审阅", "合规检查", "HTML报告" ], "tags_en": [ "bid review", "bid response", "document review", "compliance check", "HTML report" ], "artifact_type": "html", "preview": "核查报告.html", "artifacts": [ { "file": "核查报告.html", "label": "投标响应文件核查报告（自包含 HTML）" } ], "prompt": "请对投标响应文件《投标响应文件_市政排水泵站设备采购_禾川机电.docx》做只读核验，不要改动原文件。逐条核验以下 4 项调整是否已处理到位：\n1. 符合性应答表的表格列宽是否已调整，合计没有超出页面正文可用宽度；\n2. 索引/目录页码是否已填入，没有残留 Pg XX 之类的占位；\n3. 符合性应答表是否逐项明确作答（例如写「完全满足」），商务部分是否引用了对应章节（资质 3.1、检验 3.2、业绩 4.1）；\n4. 盖章/声明是否在各节出现（统计声明关键词出现次数与位置）。\n请逐条给出「已处理 / 待处理 / 未处理」结论，并附正文定位或脚本扫描结果作为证据，最后整理成自包含 HTML 核查报告，保存到 ./output/核查报告.html。可替换为你自己的标书文件名与调整项清单后复用。", "prompt_en": "Perform a read-only review of the bid response document《投标响应文件_市政排水泵站设备采购_禾川机电.docx》— do not modify the original file. Verify item by item whether the following 4 adjustments have been handled:\n1. Whether the compliance response table column widths have been adjusted so the total does not exceed the usable body width of the page;\n2. Whether the index/TOC page numbers have been filled in, with no residual Pg XX placeholders;\n3. Whether every row of the compliance response table has an explicit answer (e.g. \"完全满足\"), and whether the business section references the corresponding chapters (qualification 3.1, inspection 3.2, track record 4.1);\n4. Whether seals/declarations appear in each section (count the occurrences and locations of declaration keywords).\nFor each item, give a \"已处理 / 待处理 / 未处理\" conclusion backed by body locations or script scan evidence, then compile everything into a self-contained HTML review report saved to ./output/核查报告.html. You can replace the file name and checklist with your own bid document to reuse it.", "skills": [ { "id": "bid-response-document-review", "name": "投标响应文件审阅与复核", "description": "对 .docx 投标响应文件做只读检查与复核打分：全量抽取正文、按清单排查填写缺口（表格、页码、应答、盖章、目录、修订、占位符），依据评标办法核定评分标准并重估分数，产出结构化核查报告。", "skill_id": "skill_2090718419809648640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-08-21T12:58:41+08:00", "updated_at": "2026-08-21T12:58:41+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "bid-response-document-review-3", "title": "Apache 2.0 合规商务应答核查", "title_en": "Apache 2.0 Compliance Bid Response Review", "subtitle": "一句「完全满足」的标书商务应答,如何按许可义务补齐承诺点、化解评分风险", "subtitle_en": "How to expand a one-line \"fully compliant\" bid answer into obligation-by-obligation commitments", "description": "投标最怕的就是商务条款只写了句「完全满足」,评标专家无从核验,评分风险极高。\n\n这个案例演示了当招标文件要求交付软件中的第三方开源组件符合 Apache License 2.0 时,如何按该许可的关键义务(版权声明保留、NOTICE 文件、再分发附许可文本、修改文件标注、无商标授权、专利条款、免责边界)逐条梳理商务应答必须覆盖的承诺点,指出只写「完全满足」的评分风险与废标隐患,并给出可直接套用的应答替换范文。报告按「核查概要 / 合规核验 / 评分风险与建议 / 待处理检查项」四区块输出,附开源组件清单与法务复核等待办项。\n\n适合:投标商务人员、标书编制与审查人员、法务合规岗、参与招投标项目管理的项目经理、需要快速自查标书合规性的团队。", "description_en": "Nothing is more nerve-wracking before bidding than a commercial response that only says \"fully compliant\" — evaluators cannot verify it and scoring risk runs high.\n\nThis case demonstrates how, when a tender requires third-party open-source components in the delivered software to comply with Apache License 2.0, to map each key obligation (copyright notice preservation, NOTICE file, attaching license text on redistribution, marking modified files, no trademark endorsement, patent terms, warranty disclaimer boundary) into the commitment points a bid response must cover, flag the scoring and disqualification risks of a bare \"fully compliant\" answer, and provide a ready-to-use replacement paragraph. The report follows four sections — Review Summary / Compliance Verification / Scoring Risks & Suggestions / Pending Check Items — with a component BOM and legal review checklist.\n\nIdeal for: bid commercial staff, tender document writers and reviewers, legal & compliance professionals, project managers involved in bidding, and teams that need a fast self-check of bid compliance.", "categories": [ "business", "productivity" ], "tags": [ "标书核查", "开源许可", "合规应答", "招投标", "Apache License" ], "tags_en": [ "bid review", "open-source license", "compliance response", "bidding", "Apache License" ], "artifact_type": "html", "preview": "核查报告.html", "artifacts": [ { "file": "核查报告.html", "label": "Apache License 2.0 合规商务应答核查报告" } ], "prompt": "我们投「智慧园区软件采购」项目,招标文件商务要求写:投标人须承诺交付软件中的第三方开源组件符合 Apache License 2.0。目前标书商务应答只有一句「完全满足」,还没有完整响应文件。\n\n请依据 Apache License 2.0 官方文本(apache.org/licenses/LICENSE-2.0)逐条核对许可关键义务(版权声明保留、NOTICE 文件、再分发附许可文本、修改文件标注、无商标授权、专利条款、免责与责任边界等),梳理商务应答至少应覆盖的承诺点清单,并指出只写「完全满足」的评分风险(含必须满足条款的废标隐患)和改法——给出可直接套用的逐条承诺应答范文。\n\n按「核查概要 / 合规核验 / 评分风险与建议 / 待处理检查项」四个区块输出自包含 HTML 报告到 ./output/核查报告.html:单文件、CSS 内联、浅色商务风,无需外部资源,可离线打开;待处理检查项需覆盖组件清单(BOM)核对、招标条款星标回查、NOTICE 交付物确认、法务复核等。", "prompt_en": "We are bidding on a \"Smart Park Software Procurement\" project. The commercial requirement in the tender documents states: bidders must commit that third-party open-source components in the delivered software comply with Apache License 2.0. Our commercial response currently only says \"fully compliant\" — there is no complete response document yet.\n\nPlease check the key obligations of the Apache License 2.0 official text (apache.org/licenses/LICENSE-2.0) one by one (copyright notice preservation, NOTICE file, attaching license text on redistribution, marking modified files, no trademark endorsement, patent terms, warranty disclaimer & liability boundaries, etc.), compile a checklist of commitment points the commercial response must cover, and point out the scoring risks of only writing \"fully compliant\" (including the disqualification risk if the clause is a mandatory \"★\" requirement) and how to fix it — provide a ready-to-use paragraph with item-by-item commitments.\n\nOutput a self-contained HTML report to ./output/核查报告.html with four sections — Review Summary / Compliance Verification / Scoring Risks & Suggestions / Pending Check Items: single file, inline CSS, light business style, no external resources, openable offline. Pending check items must cover component BOM verification, re-checking the tender clause's star mark and evaluation rules, NOTICE deliverable confirmation, and legal review.", "skills": [ { "id": "bid-response-document-review", "name": "投标响应文件审阅与复核", "description": "对 .docx 投标响应文件做只读检查与复核打分,产出结构化核查报告", "skill_id": "skill_2090718419809648640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-08-21T04:56:20Z", "updated_at": "2026-08-21T05:00:00Z", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "blogger-sentiment-compass", "title": "财经神奇宝贝大PK", "title_en": "Finance Card PK Showdown", "subtitle": "48 位财经博主神奇宝贝卡反向指标大PK", "subtitle_en": "A card-battle contrarian showdown of 48 finance bloggers", "description": "看财经博主喊多喊空，到底该信谁？这个工作台是一场财经神奇宝贝大PK：把关注的博主做成集换式卡片：反向命中率越高稀有度越高，翻卡看毒舌解读；情绪指南针把 48 人观点加权成 0-100 读数，越极端反向信号越强；大事件复盘回看每次行情节点谁踏空谁逃顶。适合想用群体情绪做反向参考的投资者，也适合任何想把\"一群人的观点\"变成可跟踪指标的人。", "description_en": "Finance bloggers shout bull and bear — who should you trust? This workbench is a finance card PK showdown: it turns tracked bloggers into collectible cards: higher contrarian hit rate means higher rarity; flip a card for a snarky read. A sentiment compass weighs 48 opinions into a 0-100 reading — the more extreme, the stronger the contrarian signal. Event reviews show who missed and who escaped at each market turn. For investors using crowd sentiment as a contrarian reference, and for anyone turning 'a crowd of opinions' into a trackable indicator.", "tags": [ "情绪反向指标", "集换式卡片", "行情复盘", "数据表驱动" ], "tags_en": [ "Contrarian sentiment", "Collectible cards", "Event review", "Database-driven" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "财经神奇宝贝大PK（双语自包含演示）" } ], "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/iSIVEnoVhjmwqP5NdKGpnT，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「财经神奇宝贝大PK」数据工作台，数据全部存资料库数据表（博主因子池 + 大事件复盘两张表，页面运行时读写，支持在表中加记录后页面刷新即出现新卡片）。功能：①宝贝卡视图——每位博主一张集换式卡片，展示赛道/今日情绪/强度/原话/反向命中率，稀有度按命中率分 SSR/Sr/R/N 四档配色区分（SSR 金色主题、SR 紫色），卡片可翻转看毒舌统计解读，可甄选最多 3 张宝贝卡（写回数据表）；②情绪指南针——全体博主情绪强度加权汇总为 0-100 仪表盘，附看多/看空/中性人数分布与极端区间反向提示；③大事件复盘——可展开的行情节点手风琴，含市场结果、群体情绪读数、减仓/抄底/被套人数、操作名单分组与复盘点评；④赛季榜——全量博主表格，支持搜索、方向/稀有度/情绪三重筛选和表头排序。页内提供「添加博主」「添加事件」表单和右键编辑/归档/删除：归档的博主不参与指南针、统计与榜单汇总，收进单独归档区。数据字段以演示数据为准（博主全部虚构化名，不指向真实人物），统计样本不足时显示「积累中」而不是编造数字。视觉走游戏化卡片风但保持克制：深色顶栏、圆角卡片、红涨绿跌、稀有度用边框与主题色区分不堆砌。\n\n交付：资料库在线工作台——在线 HTML 页面 + 4 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/iSIVEnoVhjmwqP5NdKGpnT, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a 'Finance Card PK Showdown' data workbench as a single HTML file, with all data in library databases (two tables: blogger factor pool + event reviews) read and written at runtime, so adding records in the table makes new cards appear after refresh. Features: 1) Treasure cards — one collectible card per blogger showing sector / today's sentiment / strength / quote / contrarian hit rate; rarity tiers SSR/SR/R/N by hit rate with distinct themes (SSR gold, SR purple); cards flip to reveal snarky stats; select up to 3 favorites (written back to the table). 2) Sentiment compass — weighted 0-100 gauge of all bloggers' sentiment strength, with bull/bear/neutral counts and an extreme-zone contrarian warning. 3) Event reviews — expandable market-turn accordions with market outcome, group reading, de-risk/bottom-call/trapped counts, grouped actor lists and commentary. 4) Season board — full table with search, direction/rarity/sentiment filters and sortable headers. In-page 'Add Blogger' and 'Add Event' forms plus right-click edit/archive/delete: archived bloggers are excluded from compass, stats and rankings, kept in a separate archive area. Fields follow the demo data (all bloggers are fictional aliases, no real persons); show 'Accruing' instead of fabricating numbers when samples are insufficient. Visual style: game-like cards but restrained — dark top bar, rounded cards, red-up/green-down, rarity via borders and theme colors without clutter.\n\nDeliverable: an online Library workbench — an online HTML page plus 4 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "温义飞", "name_en": "Wen Yifei", "avatar": "", "badge": "community" }, "source": "community", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "created_at": "2026-08-18T16:30:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "bmi-bureau-workbench", "title": "BMI管理局 · 局长日报", "title_en": "BMI Bureau · Director Daily", "subtitle": "游戏化健康打卡工作台 ｜ 打卡 · 三餐 · 勋章 · 局内商店", "subtitle_en": "Gamified health check-in workbench | Check-ins · Meals · Medals · Bureau Shop", "description": "体重管理坚持不下来？把打卡做成一份「局长日报」。本案例把减重这件事拟人化成一座 BMI 管理局：你是局长，每天提交一份体重日报就算「签批公文」，连续打卡从实习干事一路晋升到局长（1/7/21/50/100/200/365 天七级职级），勤勉打卡攒局内币，去局内商店兑换档案主题、职级边框和印章。四大板块：① 今日——打卡日报表单（体重/腰围/体脂率/睡眠/饮食状态/运动/饮水/心情）+ BMI 实测卡（当前 BMI、超重标签、较上次变化、距目标区间、连续天数、距下个职级）+ 今日状态摘要 + 三餐热量（拍照或快捷食物录入，自动汇总总摄入与目标对比）；② 档案——目标设置（目标体重/每日热量/BMI 区间/运动/饮水）+ 体重、BMI、热量三张趋势图（周/月切换）+ 目标进度与达标建议；③ 成就——职级档案、12 枚成就勋章、局内币与获取规则；④ 我的——局长档案、每日提醒时间设置、局内商店。数据存在资料库数据表（16 个字段按日期记录，双向同步，改动任一设备都能看到），内置 3 天示例数据开箱即用。适合想养成记录习惯的减重人群、喜欢游戏化激励的自律挑战者。", "description_en": "Struggling to stay consistent with weight management? Turn daily tracking into a 'Director's Daily Report'. This case personifies weight management as a playful 'BMI Bureau': you are the Director, and each daily report you file counts as signing an official document. Check in consecutively to rise through seven ranks from Intern Officer to Director (day 1/7/21/50/100/200/365), earn Bureau coins for diligence, and spend them in the Bureau Shop on file themes, rank frames and seals. Four tabs: ① Today — the daily report form (weight/waist/body fat/sleep/diet status/exercise/water/mood), a live BMI card (current BMI, zone tag, change vs last entry, distance to target zone, streak, days to next rank), a today-at-a-glance summary, and meal calories (photo or quick-pick food entry with automatic totals against your target); ② Records — goal settings (target weight/daily calories/BMI zone/workouts/water) plus weight, BMI and calorie trend charts (week/month) with progress and suggestions; ③ Achievements — rank profile, 12 achievement medals, and the coin balance with earning rules; ④ Profile — director profile, daily reminder time, and the Bureau Shop. Data lives in a library database table (16 fields recorded by date, two-way sync across devices) with 3 days of sample data so it works out of the box. Great for anyone building a tracking habit, or anyone who responds better to gamified incentives.", "categories": [ "workspace" ], "tags": [ "健康管理", "BMI", "打卡", "游戏化", "习惯养成" ], "tags_en": [ "Health", "BMI", "Daily Check-in", "Gamification", "Habit Building" ], "artifact_type": "html", "quality_score": 95, "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "BMI管理局 · 局长日报（双语自包含工作台）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/Mtiu6gm1yIU9PyK4H9af1z，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「BMI管理局 · 局长日报」 健康打卡工作台，把体重管理拟人化成一座「管理局」：用户是局长，每天提交体重日报即算签批公文。数据全部存资料库数据表（表名「BMI打卡日报」，16 个字段：日期/体重/身高/BMI/腰围/体脂率/运动/睡眠/早餐热量/午餐热量/晚餐热量/总热量/备注/饮食状态/饮水/心情），页面运行时按日期读写、双向同步，并内置 3 天示例数据（身高 172cm，体重 74.6/74.2/73.8kg 递减）。\n\n四大板块：\n1. 今日——顶部今日待办条（日报待提交/三餐未记录提醒，可跳转）；打卡卡：表单（体重/腰围/体脂率/睡眠/饮食状态 chips/是否运动/饮水量/心情 chips）提交得 10 局内币，已有记录则原地更新不重复发币；BMI 实测卡：当前 BMI 大数字+偏瘦/标准/超重/肥胖标签+较上次变化箭头+距目标区间+连续打卡天数+距下个职级还需天数；今日状态摘要（运动/睡眠/饮食/摄入四格）；三餐热量卡：早/午/晚三格，每格可拍照上传（存档+2 币）或从 41 种常见食物快捷选择/手动录入，自动汇总今日总摄入并与目标热量对比给出饮食状态建议；\n2. 档案——目标设置表单（目标体重/每日目标热量/BMI 上下限/每周运动次数/每日饮水目标）；体重/BMI/热量三张 SVG 趋势图（周/月切换）；目标进度条与达标情况建议文案；\n3. 成就——职级档案（实习干事 1 天/科员 7 天/副科长 21 天/科长 50 天/副处长 100 天/处长 200 天/局长 365 天，当前级高亮，晋升高职级弹出「任命书」公文弹窗）；12 枚成就勋章（首日入职/七日考勤/廿一坚守/百日筑基/一年全勤/BMI 达标/运动健将/水润之星/三餐有据/轻装前行/早睡局长/能量平衡，达成点亮）；局内币卡（余额+获取规则说明）；\n4. 我的——局长档案（身高可改/初始体重/入职日期/当前职级+头衔）；提醒设置（每日打卡提醒时间+三餐拍照提醒时间，浏览器通知权限说明）；局内商店（10 件商品：3 主题/4 边框/3 印章，局内币兑换，兑换后即用）；\n5. 底部导航四 tab，缺勤日可补签。\n\n游戏化规则：连续打卡额外奖励（每日最多 +20）、三餐拍照 +2 币/餐；职级与勋章由打卡数据自动计算。\n\n视觉走「档案卷宗」暖黄公文风（米黄底、墨绿主色、衬线标题、印章元素），手机竖版布局，所有 CSS/JS 无外部依赖，数据读写走资料库 SDK。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/Mtiu6gm1yIU9PyK4H9af1z, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build it as follows: Build 'BMI Bureau · Director Daily' as a health check-in workbench that personifies weight management as a 'Bureau': the user is the Director, and each daily report filed counts as signing an official document. All data lives in a library database table ('BMI Daily Reports', 16 fields: date/weight/height/BMI/waist/body fat/exercise/sleep/breakfast kcal/lunch kcal/dinner kcal/total kcal/notes/diet status/water/mood), read and written by date with two-way sync, seeded with 3 days of sample data (height 172cm, weight 74.6/74.2/73.8kg trending down). Four tabs:\n1. Today — a top to-do bar (report pending / meals not logged, with jump links); the check-in card: form (weight/waist/body fat/sleep/diet-status chips/exercise/water/mood chips) worth 10 Bureau coins on submit, with same-day edits updating in place without re-rewarding; the live BMI card: big current-BMI number with Underweight/Normal/Overweight/Obese tag, arrow change vs last entry, distance to target zone, check-in streak, and days to the next rank; a today-at-a-glance summary (exercise/sleep/diet/intake); the meal-calories card: breakfast/lunch/dinner cells, each accepting a photo upload (archive, +2 coins) or quick picks from 41 common foods / manual entry, auto-totalling daily intake against the calorie target with diet-status suggestions;\n2. Records — goal settings (target weight/daily calories/BMI bounds/workouts per week/daily water); three SVG trend charts for weight, BMI and calories (week/month toggle); progress bar with status suggestions;\n3. Achievements — rank profile (Intern Officer day 1 / Clerk 7 / Deputy Section Chief 21 / Section Chief 50 / Deputy Division Chief 100 / Division Chief 200 / Director 365, current rank highlighted, a 'Letter of Appointment' modal on promotion); 12 achievement medals (First Day, 7-Day Attendance, 21-Day Persistence, 100-Day Foundation, 365-Day Full Attendance, BMI On Target, Fitness Pro, Hydration Star, Well-Documented Meals, Lighter Forward, Early-Sleep Director, Energy Balance — lit when earned); a coins card (balance + earning rules);\n4. Profile — director profile (editable height, start weight, join date, current rank + title); reminder settings (daily check-in time + three meal-photo reminder times with browser-notification notes); the Bureau Shop (10 items: 3 themes / 4 frames / 3 seals, redeemable with coins, applied on redemption);\n5. Bottom four-tab navigation; missed days can be backfilled. Gamification: streak bonuses (up to +20/day), +2 coins per meal photo; ranks and medals are computed automatically from check-in data. Visual: a warm 'official file' parchment style (cream background, ink-green primary, serif headings, seal motifs), mobile-portrait layout, all CSS/JS inlined with no external dependencies, data access via the library SDK (a localStorage demo adapter also works).\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "creator": { "name": "高柒", "name_en": "Gao Qi", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "source": "community", "created_at": "2026-09-01T18:00:00+08:00", "updated_at": "2026-09-01T18:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "brand-content-ip-risk-analysis", "title": "品牌内容版权风险分析", "title_en": "Brand Content IP Risk Analysis", "subtitle": "把委托创作、素材授权与 AI 风险装进一张可执行的版权治理地图", "subtitle_en": "Turn commissioned creation, asset licensing, and AI risks into an actionable copyright governance map", "description": "品牌内容一旦上线到官网、社媒、电商和线下活动，最头疼的往往不是创意，而是海报、音乐、字体、图片和 AI 素材的权利链是否经得起追溯。\n\n本案例以品牌宣传委托创作为场景，演示一份可互动的法律风险分析卡片：识别权利对象，对比权利取得模式，拆解许可范围、合同条款、第三方素材核验、AI 专项风险、发布审核与证据清单。\n\n适合：品牌市场负责人、法务与知识产权团队、采购与供应商管理人员、内容创意负责人、负责广告投放和电商运营的团队。", "description_en": "When brand content goes live across websites, social media, e-commerce, and offline events, the hardest part is often not creativity but whether the rights chain for posters, music, fonts, images, and AI assets can stand up to review.\n\nThis case uses a commissioned brand-content project to demonstrate an interactive legal risk analysis card: it identifies rights objects, compares acquisition routes, and covers licence scope, contract clauses, third-party asset checks, AI-specific risks, launch review, and evidence retention.\n\nSuitable for: brand and marketing leaders, legal and IP teams, procurement and supplier managers, creative-content owners, and teams responsible for advertising and e-commerce operations.", "categories": [ "business" ], "tags": [ "知识产权", "著作权授权", "委托创作", "AI 生成内容", "品牌治理" ], "tags_en": [ "Intellectual Property", "Copyright Licensing", "Commissioned Creation", "AI Generated Content", "Brand Governance" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "preview": "brand-content-ip-risk-analysis-card.html", "preview_en": "", "prompt": "公司计划委托外部设计机构制作宣传海报、短视频和品牌音乐，并在官网、社交媒体、电商平台及线下活动中长期使用。请结合现行有效知识产权法律法规，分析这些作品的著作权应如何约定归属或取得授权；合同中应明确哪些使用范围、权利类型、交付要求和违约责任；如何核查素材是否涉及第三方版权、商标、肖像、字体、音乐或 AI 生成内容风险；以及公司应如何留存创作、授权和发布证据，建立可持续的版权治理机制。请形成一份包含法律依据、合同条款要点、风险清单、审核流程、证据清单和治理方案的法律分析报告。", "prompt_en": "The company plans to commission an external design agency to create promotional posters, short videos, and brand music for long-term use on its website, social media, e-commerce platforms, and offline events. Based on currently effective intellectual property laws and regulations, analyze how copyright ownership of these works should be agreed or how adequate licences should be obtained; which scope of use, rights types, delivery requirements, and liabilities should be specified in the contract; how to check whether assets involve third-party copyright, trademarks, portraits, fonts, music, or AI-generated-content risks; and how the company should retain evidence of creation, licensing, and publication to build a sustainable copyright governance mechanism. Produce a legal analysis report containing legal grounds, contract clause points, a risk list, review workflows, an evidence checklist, and a governance plan.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、对比与法规检索的资深合同法律顾问", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d1e4a1d28e92", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "artifacts": [ { "file": "brand-content-ip-risk-analysis-card.html", "label": "互动式知识产权法律风险分析卡片" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T06:31:04Z", "updated_at": "2026-08-13T06:31:04Z", "is_recommended": false, "recommended_order": 0 }, { "id": "brand-design-system-tokens", "title": "品牌设计 Token 系统展示页", "title_en": "Brand Design System Token Showcase", "subtitle": "完整的设计令牌系统可视化与交互式文档", "subtitle_en": "Complete design token system visualization and interactive documentation", "description": "如何规范化管理品牌设计资产？本案例展示一个完整的设计 Token 系统展示页面，涵盖颜色面板、字体排版、间距系统、圆角、阴影等所有设计令牌，每个 Token 配有视觉预览、CSS 变量名和点击复制功能。适合设计系统维护者、前端开发者和 UI 设计师。", "description_en": "How to systematically manage brand design assets? This case showcases a complete design token system with color palettes, typography scales, spacing systems, border radius, and shadow tokens — each with visual preview, CSS variable names, and click-to-copy functionality.", "categories": [ "business", "creation" ], "tags": [ "设计系统", "Design Token", "品牌规范", "UI组件" ], "tags_en": [ "Design System", "Design Token", "Brand Guidelines", "UI Components" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "设计 Token 展示页面" } ], "prompt": "请创建一个品牌设计 Token 系统展示页面（单文件 HTML，内联无外部依赖），品牌「NovaTech」，专业美观、可交互。\n\nToken 分类：色彩（主色/辅助色/中性色 + 色值）、字体排版（h1-h6 + 正文 + caption）、间距系统（基准网格）、圆角、阴影、断点。\n每个 Token 以卡片展示（名称 + CSS 变量名 + 视觉预览 + 数值），点击卡片复制 CSS 变量（toast 提示）。颜色面板展示对比度评级（AA/AAA）。布局：侧边栏导航 + 内容区。", "prompt_en": "Please create a brand design token system showcase page (a single-file HTML page, inline with no external dependencies) for the brand \"NovaTech,\" professional, polished, and interactive.\n\nToken categories: colors (primary / secondary / neutral + color values), typography (h1-h6 + body + caption), spacing system (base grid), border radius, shadows, breakpoints.\nDisplay each token as a card (name + CSS variable name + visual preview + value); clicking a card copies the CSS variable (with a toast notification). The color panel shows contrast ratings (AA/AAA). Layout: sidebar navigation + content area.", "skills": [ { "id": "brand-guidelines", "name": "品牌设计规范", "description": "应用品牌配色和排版规范", "skill_id": "skill_2061276041038688256" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "brand-event-poster", "title": "品牌活动宣传海报", "subtitle": "用设计哲学生成博物馆级视觉海报", "description": "做活动海报要么找设计师排期等三天，要么自己用 Canva 拼出来总觉得差点意思。\n\n这个案例演示了如何用 AI 从一个活动主题出发，先创建独特的设计哲学（配色、构图、空间语言），再将其表达为一张精美的视觉海报——不是套模板，而是真正的原创设计作品。\n\n适合：市场运营、品牌设计师、活动策划、需要快速出视觉稿的创意团队。", "categories": [ "creation", "business" ], "tags": [ "海报设计", "品牌视觉", "创意设计" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "活动宣传海报（PNG）" }, { "file": "design-philosophy.md", "label": "Luminous Convergence 设计哲学文档" } ], "prompt": "请为「2026 全球开发者大会」设计一张活动宣传海报。\n\n活动信息：\n- 活动名称：DevConnect 2026 全球开发者大会\n- 主题：「Code the Future — 连接创造者」\n- 时间：2026 年 6 月 15-17 日\n- 地点：深圳国际会展中心\n- 核心关键词：开源、AI、云原生、协作\n\n设计要求：\n1. 先创建一套设计哲学（风格参考：科技感 + 人文温度，不要冰冷的纯科技风）\n2. 基于设计哲学生成一张 PNG 海报\n3. 海报中包含活动名称、时间、地点等关键信息（文字作为视觉元素融入设计）\n4. 尺寸：适合社交媒体传播的竖版（1080x1920 或类似比例）\n5. 输出设计哲学文档（.md）+ 海报文件（.png）", "prompt_en": "Please design a promotional poster for the \"2026 Global Developer Conference.\"\n\nEvent details:\n- Event name: DevConnect 2026 Global Developer Conference\n- Theme: \"Code the Future — Connecting Creators\"\n- Date: June 15-17, 2026\n- Venue: Shenzhen International Convention and Exhibition Center\n- Core keywords: open source, AI, cloud native, collaboration\n\nDesign requirements:\n1. First, create a design philosophy (style reference: a sense of technology + human warmth; avoid a cold, purely tech aesthetic)\n2. Based on the design philosophy, generate a PNG poster\n3. The poster should include key information such as the event name, date, and venue (with the text integrated into the design as a visual element)\n4. Size: a portrait orientation suitable for social media sharing (1080x1920 or a similar ratio)\n5. Output a design philosophy document (.md) + the poster file (.png)", "skills": [ { "id": "canvas-design", "name": "canvas-design（视觉设计）", "description": "基于设计哲学创作精美视觉艺术（PNG/PDF）", "description_en": "Create beautiful visual art in PNG and PDF using design philosophy", "skill_id": "skill_2053081626394972160" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 10, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-11T15:09:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "title_en": "Brand Event Promotional Poster", "subtitle_en": "Generate museum-quality visual posters using design philosophy", "description_en": "Making event posters either means waiting three days for a designer or cobbling something together in Canva that never feels quite right.\n\nThis case shows how AI starts from an event theme, first creates a unique design philosophy (color palette, composition, spatial language), then expresses it as a stunning visual poster — not a template fill, but a genuinely original design piece.\n\nBest for: Marketing teams, brand designers, event planners, anyone who needs beautiful posters fast.", "tags_en": [ "Poster Design", "Brand Visual", "Creative Design" ], "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 19 }, { "id": "brand-visual-identity-manual", "title": "品牌视觉识别手册", "title_en": "Brand Visual Identity Manual", "subtitle": "在线交互式品牌规范，从 Logo 到应用全覆盖", "subtitle_en": "Interactive online brand guidelines from logo to applications", "description": "品牌手册还在用 PDF 传来传去，设计师找色值要翻半天、Logo 安全区永远记不住？这个交互式品牌视觉识别手册把传统 PDF 品牌规范变成在线体验——点击色块复制色值、Logo 安全区 SVG 标注、字体排版 Specimen 实时展示、间距刻度可视化、12 套应用图标、名片/邮件签名等应用展示。一键切换明暗模式查看所有元素的适配效果。适合 UI 设计师、品牌经理、市场团队建立统一的品牌设计语言。", "description_en": "Still passing brand guidelines around as PDFs where designers dig through pages for color values and never remember logo safe zones? This interactive brand identity manual transforms traditional PDF guides into a living online experience — click to copy color hex values, SVG-annotated logo safe zones, live typography specimens, spacing scale visualizations, 12 custom icons, and application mockups for business cards and email signatures. Toggle dark mode to preview all elements. Perfect for UI designers, brand managers, and marketing teams building a unified design language.", "categories": [ "creation", "business" ], "tags": [ "品牌手册", "视觉识别", "设计规范", "色彩系统", "Logo 规范" ], "tags_en": [ "Brand Manual", "Visual Identity", "Design Guidelines", "Color System", "Logo Standards" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "品牌视觉识别手册" } ], "cover_image": "cover.png", "cover_mode": "landscape", "skills": [ { "id": "impeccable", "name": "高品质前端设计", "description": "高品质 UI/UX 设计工具集，避免泛 AI 审美", "skill_id": "skill_2053082862415904768" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "prompt": "请为 AI 产品「CloudMind 云智」生成一个交互式品牌视觉识别手册单页（单文件 HTML，内联无外部依赖），编辑级排版、精致克制、避免泛 AI 审美。\n\n9 个章节：\n1. 封面（Logo + 品牌名 + 手册标题）\n2. 品牌理念（使命 + 智能/可信/温暖三价值观）\n3. 标识系统（全彩/反白/单色、最小尺寸、安全区、错误用法）\n4. 色彩系统（主色色阶 + 语义色 + 渐变，点击色块复制色值）\n5. 字体排版（H1-H4 + 正文 + 小字 Specimen）\n6. 间距与栅格（基准间距刻度 + 12 列栅格）\n7. 图标语言（一组自定义图标，多尺寸）\n8. 应用展示（名片/顶栏/邮件签名/社交头像 Mockup）\n9. 页脚\n\n交互：色值一键复制、栅格切换、暗色预览、章节导航、滚动渐入。", "prompt_en": "Please create an interactive brand visual identity manual single-page (single-file HTML, inline with no external dependencies) for the AI product \"CloudMind 云智,\" with editorial-grade typography, refined and restrained, avoiding generic \"AI aesthetic\" looks.\n\n9 chapters:\n1. Cover (logo + brand name + manual title)\n2. Brand Philosophy (mission + the three values of Intelligent / Trustworthy / Warm)\n3. Logo System (full color / reversed / monochrome, minimum size, clear space, incorrect usage)\n4. Color System (primary color scale + semantic colors + gradients, click a color swatch to copy its value)\n5. Typography (H1-H4 + body text + small text specimen)\n6. Spacing & Grid (base spacing scale + 12-column grid)\n7. Iconography (a set of custom icons in multiple sizes)\n8. Applications (business card / top bar / email signature / social avatar mockups)\n9. Footer\n\nInteraction: one-click copy of color values, grid toggle, dark mode preview, chapter navigation, scroll fade-in.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "badge": "official", "avatar": "" }, "source": "official", "quality_score": 94, "created_at": "2026-05-03T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 100, "featured_order": 0, "usage_count": 0, "is_featured": false, "mcps": [], "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "breakout-watchlist-ruleset", "title": "次日候选池初筛", "title_en": "Next-Day Watchlist Screen", "subtitle": "用 R01 与 R07 把收盘后候选股从热闹名单整理成可跟踪池", "subtitle_en": "Turn a noisy post-close stock list into a trackable pool with R01 and R07", "description": "收盘后最烦的不是没有股票可看，而是榜单太多、强弱混在一起。这个案例先用官方排行拿到 2026-07-27 的高分股，再结合真实行情、板块与资金数据，用 R01 和 R07 把真正值得跟踪的标的筛出来。\n\n页面里把重点关注、次级关注、排除说明和市场整体判断分层展示，还会明确说明当天为什么没有硬凑 R05。最终交付是一份单文件双语 HTML，右上角可切换中英，适合直接当次日观察清单使用。\n\n适合：想做收盘后候选池整理的短线交易者；需要把选股逻辑标准化的人；希望把“热股榜”转成执行名单的 A 股参与者。", "description_en": "The annoying part after the close is not a lack of stocks to watch, but the opposite: too many lists with strength and weakness mixed together. This case starts from official ranking results for 2026-07-27, then combines real quote, sector and flow data with R01 and R07 to isolate the names that actually deserve next-day attention.\n\nThe page separates priority names, secondary names, exclusions and the overall market stance, and it explicitly explains why R05 was not forced into the result. The final deliverable is a single-file bilingual HTML page with a top-right language switch, suitable for direct use as a next-day watchlist.\n\nSuitable for: short-term traders who build post-close candidate pools; anyone trying to standardize stock-screening logic; A-share participants who want to turn hot-stock rankings into execution-ready lists.", "categories": [ "finance", "data" ], "tags": [ "选股初筛", "候选池", "收盘后计划", "规则筛选" ], "tags_en": [ "stock screening", "watchlist", "post-close planning", "rule-based scan" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请按 A股短线交易 的标的初筛流程，用 2026-07-27 收盘后的真实 A 股数据做一份次日候选池。候选起点请固定为官方综合评分排行 `westock-tool ranking CompScore --limit 12` 的前 12 只股票，再结合 WeStock 官方行情、板块和资金数据，组合使用 R01 和 R07 规则做分类；如果 R05 没有出现硬触发，请明确说明，不要用假信号补齐。输出一份单文件、资源全内联的 HTML 页面。内容至少包含重点关注、次级关注、排除说明和市场整体判断。", "prompt_en": "Please follow the stock-screening workflow of A股短线交易 and build a next-day A-share watchlist with real post-close data from 2026-07-27. Fix the starting universe to the top 12 names from the official composite-score ranking `westock-tool ranking CompScore --limit 12`, then combine official WeStock quote, sector and flow data to classify the pool with R01 and R07. If R05 does not produce a hard trigger, say so explicitly instead of inventing a signal. Output a single-file HTML page with all resources embedded. The page should at least include priority names, secondary names, exclusions and the overall market stance.", "skills": [ { "id": "ashare-short-term-trading", "name": "A股短线交易", "description": "A股短线交易体系，支持交割单复盘分析、策略筛选选股、盘中关键时间点推送三大场景。", "skill_id": "skill_2076627819394322432" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "artifacts": [ { "file": "output.html", "label": "双语候选池页面" } ], "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:10:00+08:00", "updated_at": "2026-07-27T20:16:07+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "bronze-galloping-horse-phone-case-series", "title": "铜奔马 · 手机壳四风格系列", "title_en": "Bronze Galloping Horse · Four Stylistic Phone Case Series", "subtitle": "同一姿态 · 四种视觉风格", "subtitle_en": "同一姿态 · 四种视觉风格", "description": "以甘肃省博物馆藏东汉铜奔马（俗称马踏飞燕）为主题，只做手机壳一个品类的四个风格系列：Q 版卡通、儿童涂鸦、毕加索立体主义、中国水墨写意。案例先核对文物造型特征形成六项硬约束清单，再以该清单作为生图提示词的强制条件；同时约束排版，要求图案完整落在壳体内、不越出边缘。", "description_en": "以甘肃省博物馆藏东汉铜奔马（俗称马踏飞燕）为主题，只做手机壳一个品类的四个风格系列：Q 版卡通、儿童涂鸦、毕加索立体主义、中国水墨写意。案例先核对文物造型特征形成六项硬约束清单，再以该清单作为生图提示词的强制条件；同时约束排版，要求图案完整落在壳体内、不越出边缘。", "categories": [ "creation", "business" ], "tags": [ "文创提案", "铜奔马", "马踏飞燕", "手机壳", "风格化设计", "文物特征约束" ], "tags_en": [ "文创提案", "铜奔马", "马踏飞燕", "手机壳", "风格化设计", "文物特征约束" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "cover_image": "cover.jpg", "cover_image_en": "cover.jpg", "cover_mode": "landscape", "prompt": "以甘肃省博物馆藏东汉铜奔马为主题，做一份手机壳单品类文创提案，输出单文件 HTML。要求：一、先检索权威来源核对文物年代、出土、尺寸重量与造型特征，并整理成一份\"文物特征清单\"；二、参考国内已成功出圈的同源文创（如甘博绿马）的风格化思路，不要自行随意定义视觉；三、只做手机壳一个品类，给出 Q 版卡通、赛博朋克、毕加索立体主义、中国水墨写意四个风格系列；四、每款商品效果图的生成提示词必须把文物特征清单逐条写成强制约束，风格可变、文物特征不可变；五、每款注明保留了哪些文物特征、工艺与市场定位；六、声明效果图为 AI 设计示意，需文物与法务复核。", "prompt_en": "以甘肃省博物馆藏东汉铜奔马为主题，做一份手机壳单品类文创提案，输出单文件 HTML。要求：一、先检索权威来源核对文物年代、出土、尺寸重量与造型特征，并整理成一份\"文物特征清单\"；二、参考国内已成功出圈的同源文创（如甘博绿马）的风格化思路，不要自行随意定义视觉；三、只做手机壳一个品类，给出 Q 版卡通、赛博朋克、毕加索立体主义、中国水墨写意四个风格系列；四、每款商品效果图的生成提示词必须把文物特征清单逐条写成强制约束，风格可变、文物特征不可变；五、每款注明保留了哪些文物特征、工艺与市场定位；六、声明效果图为 AI 设计示意，需文物与法务复核。", "skills": [], "experts": [ { "id": "tanyuan-cultural-creative", "name": "文创提案专家", "description": "把一件文物做成文创产品线设计提案,含核心记忆点、产品线与商品示意图,事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cf052b46c196", "expert_id": "ex_HXd4FrKZutHg" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "设计创意", "usage_count": 0, "created_at": "2026-08-14T18:00:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "铜奔马 · 手机壳四风格系列" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "business-daily", "title": "经营日报与复盘：昨日经营日报，今日可执行", "title_en": "Yesterday's report, today's actions", "subtitle": "昨日经营日报，今日可执行", "subtitle_en": "Yesterday's report, today's actions", "description": "面向区域/店长场景：昨日经营日报，今日可执行。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向区域/店长场景：昨日经营日报，今日可执行。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "经营日报", "晨会", "行动清单" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请为 {门店范围} 生成 {营业日期} 经营日报。\n必须包含：核心指标、异常门店、重点问题、今日行动建议、明早复核项。\n要求：面向店长晨会可直接念；先结论后依据；一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请为 {门店范围} 生成 {营业日期} 经营日报。\n必须包含：核心指标、异常门店、重点问题、今日行动建议、明早复核项。\n要求：面向店长晨会可直接念；先结论后依据；一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435449149067264", "skill_id": "skill_2087435449149067264", "name": "经营日报与复盘", "description": "生成经营日报与复盘分析" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-detail", "title": "经营明细追踪：点开门店，追到账单明细", "title_en": "Drill into store bills and details", "subtitle": "点开门店，追到账单明细", "subtitle_en": "Drill into store bills and details", "description": "面向区域/店长场景：点开门店，追到账单明细。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向区域/店长场景：点开门店，追到账单明细。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "账单明细", "结算", "外卖", "下钻" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {门店名称} 在 {营业日期} 的账单、菜品、结算和外卖明细，定位营业异常。\n1) 汇总异常订单类型与金额影响\n2) 拆堂食/外卖结构变化\n3) 定位退菜、优惠、结算取消等明细证据\n4) 输出今日核查清单（岗位+时点）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {门店名称} 在 {营业日期} 的账单、菜品、结算和外卖明细，定位营业异常。\n1) 汇总异常订单类型与金额影响\n2) 拆堂食/外卖结构变化\n3) 定位退菜、优惠、结算取消等明细证据\n4) 输出今日核查清单（岗位+时点）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435070034563072", "skill_id": "skill_2087435070034563072", "name": "经营明细追踪", "description": "门店账单明细追踪" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-dish", "title": "菜品与结算诊断：退菜优惠支付，哪里不对劲", "title_en": "Spot dish return, discount and payment issues", "subtitle": "退菜优惠支付，哪里不对劲", "subtitle_en": "Spot dish return, discount and payment issues", "description": "面向区域/店长场景：退菜优惠支付，哪里不对劲。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向区域/店长场景：退菜优惠支付，哪里不对劲。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "菜品", "退菜", "优惠", "支付" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请对 {门店范围} 在 {统计区间} 做菜品与结算诊断。\n1) 找出销售下滑菜品及影响金额\n2) 列出退菜率偏高菜品\n3) 检查优惠过高与支付结构异常\n4) 给出核查建议与今日动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请对 {门店范围} 在 {统计区间} 做菜品与结算诊断。\n1) 找出销售下滑菜品及影响金额\n2) 列出退菜率偏高菜品\n3) 检查优惠过高与支付结构异常\n4) 给出核查建议与今日动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435070034563072", "skill_id": "skill_2087435070034563072", "name": "经营明细追踪", "description": "门店账单明细追踪" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-overview", "title": "综合经营诊断：昨日异常门店一键速诊", "title_en": "Traffic up, ticket down? One-tap diagnosis", "subtitle": "客流上来了却客单掉了？结论、证据、今日三件事一次给齐", "subtitle_en": "Traffic up, ticket down? One-tap diagnosis", "description": "面向区域/店长晨会：先标出异常门店，再追到客单下滑归因，输出可执行动作与明早复核项。本文为演示数据样例。", "description_en": "面向区域/店长晨会：先标出异常门店，再追到客单下滑归因，输出可执行动作与明早复核项。本文为演示数据样例。", "categories": [ "office" ], "tags": [ "经营诊断", "异常门店", "客单", "晨会", "超级日报" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请对 {门店范围} 在 {营业日期} 做综合经营诊断。\n1) 汇总营收、客流、客单、翻台率、折扣率，对比近四周同星期与目标达成\n2) 标出异常门店（红/黄灯），说明主要异常指标\n3) 对最需关注的门店做归因：客流/客单/结构哪条在驱动，给出证据链\n4) 输出「今天只做三件事」：动作、责任岗位、完成时点\n5) 列出明早复核指标与未恢复时的升级动作\n要求：先结论后依据；区分事实、判断、待验证；一次输出静态 HTML fragment，不要多轮追问；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请对 {门店范围} 在 {营业日期} 做综合经营诊断。\n1) 汇总营收、客流、客单、翻台率、折扣率，对比近四周同星期与目标达成\n2) 标出异常门店（红/黄灯），说明主要异常指标\n3) 对最需关注的门店做归因：客流/客单/结构哪条在驱动，给出证据链\n4) 输出「今天只做三件事」：动作、责任岗位、完成时点\n5) 列出明早复核指标与未恢复时的升级动作\n要求：先结论后依据；区分事实、判断、待验证；一次输出静态 HTML fragment，不要多轮追问；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435352724176896", "skill_id": "skill_2087435352724176896", "name": "综合经营诊断", "description": "综合经营诊断与趋势分析" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-rank", "title": "门店与品牌排行：门店增长榜：谁领先谁掉队", "title_en": "Growth board: leaders and laggards", "subtitle": "门店增长榜：谁领先谁掉队", "subtitle_en": "Growth board: leaders and laggards", "description": "面向区域/店长场景：门店增长榜：谁领先谁掉队。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向区域/店长场景：门店增长榜：谁领先谁掉队。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "排行", "增长", "门店对比" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请对 {门店范围} 在 {统计区间} 做门店排行诊断。\n1) 综合营收、增长率、客单价给出 Top3 与 Bottom3\n2) 解释领先/掉队门店的分化原因\n3) 标出分化最大的区域或品牌\n4) 给出本周关注重点与核查动作\n要求：一次输出静态 HTML fragment，含语义化排行与结论；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请对 {门店范围} 在 {统计区间} 做门店排行诊断。\n1) 综合营收、增长率、客单价给出 Top3 与 Bottom3\n2) 解释领先/掉队门店的分化原因\n3) 标出分化最大的区域或品牌\n4) 给出本周关注重点与核查动作\n要求：一次输出静态 HTML fragment，含语义化排行与结论；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435352724176896", "skill_id": "skill_2087435352724176896", "name": "综合经营诊断", "description": "综合经营诊断与趋势分析" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-trend", "title": "趋势与周期对比：本周 vs 上周经营趋势解读", "title_en": "This week vs last week trend insights", "subtitle": "本周 vs 上周经营趋势解读", "subtitle_en": "This week vs last week trend insights", "description": "面向区域/店长场景：本周 vs 上周经营趋势解读。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向区域/店长场景：本周 vs 上周经营趋势解读。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "趋势", "周环比", "客流", "复盘" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请对比 {门店范围} 在 {本周区间} 与 {上周区间} 的经营趋势。\n1) 对比营收、客流、客单、翻台的周环比\n2) 按品牌/区域拆解主要驱动因素\n3) 标出客流或营收下滑最明显的门店并解释\n4) 输出本周复盘要点与下周关注动作（含岗位）\n要求：先结论后依据；区分事实、判断、待验证；一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请对比 {门店范围} 在 {本周区间} 与 {上周区间} 的经营趋势。\n1) 对比营收、客流、客单、翻台的周环比\n2) 按品牌/区域拆解主要驱动因素\n3) 标出客流或营收下滑最明显的门店并解释\n4) 输出本周复盘要点与下周关注动作（含岗位）\n要求：先结论后依据；区分事实、判断、待验证；一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435352724176896", "skill_id": "skill_2087435352724176896", "name": "综合经营诊断", "description": "综合经营诊断与趋势分析" } ], "experts": [ { "id": "weimob-business-expert", "expert_id": "ex_wFmu6tMe5SV5", "name": "微盟经营分析专家", "description": "微盟餐饮经营数据分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "buy-sale-contract-pre-filing-checklist", "title": "买卖合同纠纷立案前核对清单", "title_en": "Pre-filing Checklist for Sale Contract Disputes", "subtitle": "把供货方的交易材料整理成可核验的立案准备工作台", "subtitle_en": "Turn a supplier's transaction records into a verifiable pre-filing preparation workspace", "description": "货款没收齐时，合同、签收、对账和催收材料常常分散在不同人和不同系统里，最容易遗漏的是交易时间线、金额底稿、管辖依据和电子数据原始载体。\n\n本案例以供货方主张 50 万元货款为示例，展示一份可交互的《立案前核对报告》：从事实时间线、法律与法院要求、证据矩阵到立案材料包，逐项标识风险、待补信息与优先行动，并保留材料文件名映射和报告下载能力。\n\n适合：供货企业法务、销售回款负责人、商事争议律师助理、拟作为原告的中小企业负责人。", "description_en": "When payment is overdue, contracts, delivery receipts, reconciliations, and collection records are often scattered across people and systems. The most common gaps are the transaction timeline, amount workbook, jurisdiction basis, and original electronic records.\n\nUsing a supplier's RMB 500,000 receivable as an example, this case demonstrates an offline interactive Pre-filing Checklist Report. It covers facts, legal and court requirements, an evidence matrix, and a filing package while flagging risks, missing information, and prioritized actions. It also supports filename-based material mapping and report download.\n\nSuitable for: in-house counsel at suppliers, receivables owners, commercial litigation paralegals, and SME owners preparing to sue as plaintiffs.", "categories": [ "business" ], "tags": [ "买卖合同", "诉讼准备", "证据清单", "立案核对" ], "tags_en": [ "Sale contract", "Litigation preparation", "Evidence checklist", "Filing review" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "立案前核对报告工作台.html", "preview_en": "", "prompt": "我是供货方，买方尚欠货款 50 万元，已持有合同、签收单、对账单和催款记录。请以买卖合同纠纷立案前准备为范围，按事实时间线、诉讼请求及计算、争议解决条款与管辖、证据目录、电子数据、立案材料包五个维度核对。请先核验现行法律、司法解释及拟受理法院公开要求；未核验处标注“待官方来源核验”。输出高/中/低风险、证据缺口、待补信息和按优先级排序的下一步行动。不要承诺立案、保全或胜诉结果。", "prompt_en": "I am a supplier. The buyer still owes RMB 500,000. I have the contract, delivery receipt, reconciliation statement, and collection records. Please limit the work to pre-filing preparation for a sale contract dispute, and review it across five dimensions: fact timeline; claims and calculation; dispute-resolution clause and jurisdiction; evidence catalogue and electronic data; and filing-package materials. First verify the current laws, judicial interpretations, and the public requirements of the intended receiving court; mark anything not verified as \"Pending verification against official sources\". Output high/medium/low risks, evidence gaps, information to be supplemented, and next actions in priority order. Do not promise filing acceptance, preservation, or a successful outcome.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "提供合同起草、签署、审查、对比和法规检索的资深合同法务能力", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "artifacts": [ { "file": "立案前核对报告工作台.html", "label": "可互动立案前核对工作台" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T14:13:57+08:00", "updated_at": "2026-08-13T15:43:26+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "campus-course-workbench", "title": "今天 · 校园课程工作台", "title_en": "Today · Campus Course Workbench", "subtitle": "课表 / DDL 看板 / 笔记库 / 课件库 / AI 陪读问答，一个页面管住学期", "subtitle_en": "Timetable / DDL board / notes / courseware / AI study Q&A — one page for the whole term", "description": "开学第一个月总在追赶：课在哪个教室、作业哪天截止、上周的课件还没翻、没听懂的问题转头就忘？本案例把大学生的学期日常收进一个页面：①「今天」视图给出现在的时间轴（进行中/已结束）、最近 DDL（红 ≤3 天/琥珀 4-7 天/绿 >7 天）和顶部信息 chips，还留了一句「说一句话，交给 WorkBuddy」——录入新课、新 DDL、归档课件都变成一句话对话；②「课表」是周一到周日的周网格 + 按天列表，添加/编辑/删除课程走弹窗表单，保存即写入资料库；③「DDL 看板」按紧急/临近/稍后三列分诊，一键顺延（截止 +1 天并记一笔延期次数）或完成；④「笔记库」自动生成本周复盘（课程数/待办/延期/笔记/问答 + 一句话小结），全部笔记按课程分组、每条可导出 Word；⑤点进任意课程进入「课程资料库」：课件按周归档可预览、所见即所得笔记编辑器（自动保存、可拖拽调高、「❓没听懂」一键标记）、AI 陪读问答带课件上下文，笔记里标过「❓没听懂」的问题会自动浮现在「待问清单」。全部数据存在资料库多维表（6 张表：课程/DDL/笔记/问答/课件/设置），换设备数据都在。适合在校大学生、研究生课程班、想用一句话管理学期事务的人。", "description_en": "The first month of term always feels like catch-up: which classroom, which deadline, which deck you never opened, which question you meant to ask and forgot. This case fits a student's whole term into one page: (1) \"Today\" shows a live timeline of today's classes (ongoing/ended), the nearest DDLs (red ≤3 days / amber 4-7 / green >7) and header chips — plus a \"just say it\" bar that turns adding a course, a deadline or filing courseware into one sentence of chat; (2) \"Schedule\" is a Mon–Sun grid plus a day-grouped list, with add/edit/delete via a modal form that writes straight to the library; (3) the \"DDL board\" triages cards into Urgent / Soon / Later, with one-tap snooze (due +1 day, postponement logged) or done; (4) the \"Note library\" auto-builds a weekly review (courses / to-dos / postponements / notes / Q&A plus a one-line recap) and groups all notes by course, each exportable to Word; (5) click any course to open its library: week-by-week courseware with inline preview, a WYSIWYG note editor (autosave, resizable, one-tap \"❓Unclear\" marking) and an AI study Q&A with courseware context — questions marked \"❓Unclear\" in notes surface automatically on the To-Ask list. All data lives in six library tables (courses / DDL / notes / Q&A / courseware / settings), synced across devices. Best for: university students, grad-school classes, and anyone who wants to run a term by saying one sentence.", "categories": [ "workspace" ], "tags": [ "校园课程", "课程表", "DDL 管理", "课堂笔记", "学生工作台" ], "tags_en": [ "Campus Courses", "Timetable", "DDL Tracking", "Class Notes", "Student Workbench" ], "artifact_type": "html", "created_at": "2026-09-02T16:20:00+08:00", "updated_at": "2026-09-06T15:45:00+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "今天 · 校园课程工作台（单体双语 HTML）" } ], "creator": { "name": "戴宇涵", "name_en": "Dai Yuhan", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/aNCyGL8LosHIpHJn9w1gX4，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「今天 · 校园课程工作台」 工作台，面向大学生：把每天要上的课、要交的作业、上课记的笔记、课程课件、没听懂的问题放进一个页面，数据全部存资料库数据表，页面运行时读写。五大模块：\n① 今天——问候语 + 日期周次 + 顶部信息 chips（今天 X 节课 / 最近 DDL 还剩 X 天 / 第 X 周 / 待办 X 项）；「今天的课」时间轴（进行中/已结束状态，点卡片进课程资料库）；「最近的 DDL」前 5 条（红 ≤3 天/琥珀 4-7 天/绿 >7 天档位，点卡片进课程）；底部「说一句话，交给 WorkBuddy」——一键复制录入模板按钮，把「添加课程 / DDL / 课件」变成一句话对话；\n② 周课表——周一到周日 × 5 个时段（1-2/3-4/5-6/7-8/9-10 节）的周网格，每格课程色块（悬停出现 ✎ 编辑）+ 按天分组的列表视图；「添加课程」弹窗表单（课程名称/星期/类别 必修·选修·实验/开始结束时间/教室/教师/周次范围/备注），保存即写入资料库，「刷新线上」重新拉取；\n③ DDL 看板——顶部四个统计（待办/本周新增/已完成/累计延期）；紧急（≤3 天）/临近（4-7 天）/稍后（>7 天）三列卡片，每张卡可「顺延」（截止 +1 天、顺延次数 +1）或「完成」；下方「已完成」列表，点卡片可进所属课程；\n④ 笔记库——本周复盘（本周课程数/待办 DDL/累计延期/笔记条数/问答沉淀 + 一句话周小结）；全部笔记按课程分组（组头有课程色点 + 进课程入口），每条可「导出 Word」；\n⑤ 课程资料库（点课程进入）——课程头（教师/教室/周次/类别/备注 + 课件·笔记·问答统计 + 主题色条）；课件库按周分组，可内联预览；课程笔记编辑器（所见即所得：H1-H3/加粗/行内代码/无序有序列表/待办/引用/代码块，「❓没听懂」一键标记，输入自动保存，可拖拽调高，「存入资料库」写回，可导出 Word）；AI 陪读问答——三个快捷问题一键复制提问（提示词带课件上下文），「待问清单」自动收集笔记里标记「❓没听懂」的问题并可「现在问」，问答历史沉淀；本课已存笔记列表。\n\n数据表：课程（课程ID/课程名/星期/开始结束时间/教室/教师/周次/类别/颜色/备注）、DDL（编号/任务/课程/截止/类型 小组·作业·实验·其他/状态 todo·done/顺延次数/预计分钟/创建日期）、笔记（日期/课程/周/内容 Markdown）、问答（课程/提问时间/问题/答复/关联课件）、课件（课程/文件名/周/归档日期/路径/大小）、设置（semester_start 学期起始日，用于计算第几周）。\n\n视觉走浅色低饱和雾蓝风（主色 #4C82EF，白卡片 + 大圆角 + 轻阴影 + 语义红/琥珀/绿），顶部品牌栏 + 四个主导航（今天/课表/DDL/笔记），PC/移动端自适应。演示数据：7 门课（高等数学A/大学英语/大学英语听说/数据结构/操作系统/操作系统实验/机器学习导论）、6 条 DDL、3 条笔记、2 条问答、4 份课件。\n\n交付：资料库在线工作台——在线 HTML 页面 + 5 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to one-click replicate https://www.workbuddy.cn/space/d/aNCyGL8LosHIpHJn9w1gX4 with online storage and two-way sync. If that is unavailable, build it as follows: Build a \"Today · Campus Course Workbench\" workbench for university students: today's classes, upcoming deadlines, class notes, courseware and unclear questions in one page, all data stored in library tables and read/written live by the page. Five modules: (1) Today — greeting, date & week, header chips (classes today / nearest DDL / week N / to-dos), a live timeline of today's classes, the nearest DDLs with color tiers, and a \"just say it\" bar that copies an entry template; (2) Schedule — Mon–Sun × 5 period grid with editable course chips plus a day-grouped list, add/edit/delete via modal form saved to the library; (3) DDL board — four stats, three columns (urgent/soon/later), one-tap snooze or done, completed list; (4) Note library — weekly review stats and a one-line recap, all notes grouped by course, each exportable to Word; (5) Course library — course header with stats, week-grouped courseware with inline preview, a WYSIWYG note editor (H1–H3, bold, inline code, lists, todo, quote, code block, autosave, resizable, \"❓Unclear\" one-tap marking, save-to-library, Word export), AI study Q&A with one-tap copyable questions, a To-Ask list auto-fed by \"❓Unclear\" marks, and Q&A history. Visual style: light, low-saturation misty blue (#4C82EF primary, white cards, large radii, soft shadows, semantic red/amber/green), top brand bar + four-section nav (Today/Schedule/DDL/Notes), responsive. Demo data: 7 courses, 6 DDLs, 3 notes, 2 Q&A entries, 4 courseware files.\n\nDeliverable: an online Library workbench — an online HTML page plus 5 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "is_recommended": false, "recommended_order": 0 }, { "id": "candy-university-map", "title": "糖果果大学成长地图", "title_en": "Candy's University Growth Map", "subtitle": "档案、经历、专业探索、机会库、方向与行动卡，一张地图管住大学四年", "subtitle_en": "Profile, experiences, major exploration, opportunities, directions and action cards — one map for all four years of college", "description": "大学的迷茫往往不是没有事情可做，而是经历、机会和想法散落在备忘录、聊天记录和脑子里，凑不成一张完整的地图。\n\n这个案例展示了一个大学成长工作台：总览页把能力画像、方向星图、本周聚焦和考试月历放在一屏；「过去」用时间线沉淀每一段经历、比赛和实习，能力标签自动统计占比；「现在」是机会库，竞赛、实习、考试按类型与截止时间管理，支持批量操作和倒计时提醒；「未来」把想探索的方向拆成一张张体验行动卡，每张卡带具体步骤和进度勾选。\n\n内置 90 条示例数据（含完成一次课程作业、参加大广赛、找实习等真实场景），打开即可对照填写；考试日历预置了数模竞赛、教资笔试等公开日程，可一键加入手机日历。\n\n适合：想系统复盘大学经历、规划探索方向的大学生，以及需要一套「记录—复盘—规划」闭环的成长模板使用者。", "description_en": "Feeling lost in college is rarely about having nothing to do — it's that experiences, opportunities and ideas end up scattered across notes, chats and memory, never forming one complete map.\n\nThis case demonstrates a university growth workbench: an overview screen combining a capability profile, a direction constellation, this week's focus and an exam calendar; a \"Past\" view that archives every experience, contest and internship on a timeline with auto-aggregated skill tags; a \"Now\" view acting as an opportunity hub where contests, internships and exams are managed by type and deadline with batch actions and countdown reminders; and a \"Future\" view that breaks exploration directions into action cards, each with concrete steps and checkable progress.\n\nIt ships with 90 example records (course projects, advertising competitions, internship hunts and more) so you can start by replacing them, plus a built-in exam calendar with public schedules such as math-modeling contests and teacher-certification exams that can be exported to your phone.\n\nBest for: college students who want to systematically review their experiences and plan exploration directions, and anyone looking for a \"record — review — plan\" growth template.", "categories": [ "workspace" ], "tags": [ "大学规划", "经历复盘", "机会管理", "方向探索", "能力画像" ], "tags_en": [ "College Planning", "Experience Review", "Opportunity Tracking", "Direction Exploration", "Skill Profile" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "糖果果大学成长地图（中英双语单文件 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/T5lqFuWr6vPIwLa2GgbqO8 ，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「大学成长地图」工作台，面向在校大学生，用 6 张数据表管理大学四年的记录与规划：① 个人档案（年级、专业、兴趣方向、当前困惑、每周可投入时间、正在考虑的方向、填写日期、地图阶段）；② 大学经历（名称、类型、角色、状态、开始/结束时间、职责与成果、能力标签、作品链接）；③ 专业探索（专业名称、了解程度、学习内容、相关职业、核心能力、相邻领域、体验建议）；④ 个人机会（名称、类型、重要程度、状态、来源、活动日期、活动结束日期、截止时间、安排选择、来源标识、活动信息、链接）；⑤ 发展方向（方向名称、一句话简介、所需能力、已有基础、关键阶段、准备事项、意向程度）；⑥ 方向体验卡（体验目标、关联方向、具体步骤、预计时间、完成状态、体验反馈、步骤进度）。\n\n页面分四个视图：总览——能力画像（能力标签占比条）、方向星图（按意向程度排布候选方向）、本周聚焦、未来 7 天截止提醒与月历；过去——经历时间线，按年份分组展示经历卡片，附带能力标签统计与「成长证据」表；现在——机会库表格（类型/状态筛选、关键词搜索、批量选择与批量操作）、月历视图（按日聚合事项）、报名截止倒计时列表；未来——方向卡片墙（相邻领域与体验建议）与方向体验卡（步骤清单勾选、进度自动汇总、体验反馈记录）。\n\n内置考试竞赛日历：预置数学建模竞赛、教师资格考试等公开日程（名称、日期、来源、官方链接、报名说明），支持一键下载 ICS 加入手机日历，机会可从日历「加入个人清单」。全部表格预填 90 条示例数据，示例记录带「示例（请替换）」标记并在统计中自动识别剔除；经历支持批量删除与撤销恢复；档案顶部按学位阶段显示「本科生/硕士生/博士生」。\n\n视觉：清新校园插画风，绿色主色调配大面积手绘背景插画，卡片式信息分区，顶部导航切换四个视图，PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/T5lqFuWr6vPIwLa2GgbqO8 with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows:\n\nBuild a \"University Growth Map\" workbench for college students, managing four years of records and planning with 6 tables: ① Profile (grade, major, interests, current confusions, weekly hours available, directions being considered, filled date, map stage); ② Experiences (name, type, role, status, start/end time, responsibilities & outcomes, skill tags, portfolio link); ③ Major Exploration (major name, familiarity, what you learn, related careers, core skills, adjacent fields, low-cost experience suggestions); ④ Personal Opportunities (name, type, importance, status, source, event date, event end date, deadline, schedule choice, source tag, event info, link); ⑤ Development Directions (direction name, one-line intro, required skills, existing foundation, key milestones, preparation items, interest level); ⑥ Direction Action Cards (experience goal, related direction, concrete steps, estimated time, completion status, experience feedback, step progress).\n\nFour views: Overview — skill profile (tag share bars), direction constellation (candidates arranged by interest level), this week's focus, next-7-days deadline reminders and a month calendar; Past — an experience timeline grouped by year with cards, plus skill-tag statistics and a \"growth evidence\" table; Now — an opportunities table (type/status filters, keyword search, batch selection and batch actions), a calendar view aggregating items by day, and a registration-deadline countdown list; Future — a direction card wall (adjacent fields and experience suggestions) and direction action cards (step checklists, auto-summed progress, feedback notes).\n\nBuilt-in exam & contest calendar: preset public schedules such as math-modeling contests and teacher-certification exams (name, date, source, official link, registration notes), one-click ICS download to add to your phone calendar, and opportunities can be added from the calendar into the personal list. All tables come prefilled with 90 example records marked \"Example (replace me)\" that are auto-excluded from statistics; experiences support batch delete with undo; the profile header shows the degree stage (undergraduate/master/PhD).\n\nVisual: fresh campus illustration style, green palette with a large hand-drawn background illustration, card-based sections, top navigation across the four views, responsive on PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "糖果果的未来要发光", "name_en": "Candy's Shining Future", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "community", "usage_count": 0, "created_at": "2026-09-10T12:00:00+08:00", "updated_at": "2026-09-10T08:28:27.170Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cangjie-great-work-distill", "title": "《如何做出伟大的工作》蒸馏精华", "title_en": "How to Do Great Work — Distilled", "subtitle": "把 Paul Graham 的长文蒸馏成一页可执行方法论", "subtitle_en": "Paul Graham's essay distilled into one page of executable methodology", "description": "一篇一万多字的经典长文，读完却记不住、更用不起来？本案用「仓颉」蒸馏 skill 把 Paul Graham 的《How to Do Great Work》真实蒸馏成一页结构化精华：从一句话内核（好奇心既是发动机也是方向舵），到三大可执行论点——选对方向（找到你既擅长、又天然好奇、还能做出伟大工作的交集）、进入前沿（staying upwind，持续走向让选择更多的方向）、抵御摩擦（把挫折、拖延、士气波动写进流程而非靠意志硬扛），再到常见陷阱与反例、作者本人的局限、以及可以立刻带走的三句话。交付物是一份克制精致的单文件 HTML 精华页：白底、红色点缀、衬线标题、粘性目录与滚动高亮，可直接收藏复用。适合创业者、研究者、产品与创作者，以及任何想把'读过'变成'用得上'的人。", "description_en": "A 10,000-word classic essay you read but can't remember, let alone apply? Here the 'Cangjie' distillation skill turns Paul Graham's 'How to Do Great Work' into one page of structured essence: from the one-line core (curiosity is both engine and rudder), to three executable theses — choosing direction (find the intersection of what you're good at, naturally curious about, and could do great work in), staying upwind (keep moving toward directions that expand your options), and resisting friction (build setbacks, procrastination, and morale swings into your process rather than brute-forcing them) — plus common traps and counterexamples, the author's own limits, and three takeaways you can use immediately. The deliverable is a restrained, refined single-file HTML page: white background, red accents, serif headings, a sticky TOC with scroll highlighting — ready to keep and reuse. Great for founders, researchers, product people, and creators, and anyone who wants to turn 'I read it' into 'I use it'.", "categories": [ "education", "productivity" ], "tags": [ "知识蒸馏", "方法论提取", "长文精读", "个人成长" ], "tags_en": [ "Knowledge Distillation", "Methodology Extraction", "Deep Reading", "Personal Growth" ], "artifact_type": "html", "prompt": "用「仓颉」蒸馏 skill 把 Paul Graham 的长文《How to Do Great Work》（原文见 paulgraham.com/greatwork.html）蒸馏成一份可执行的方法论精华，并做成一页单文件 HTML 精华页。要克制精致的阅读风格：白底、红色点缀、衬线标题、粘性目录与滚动高亮、Lucide 线性图标、不要 emoji。内容包括：一句话内核、三大可执行论点（选对方向 / 进入前沿 staying upwind / 抵御摩擦），每个论点给出可操作的做法而非空话；常见陷阱与反例；作者本人的局限与适用边界；可以立刻带走的三句话。所有资源内联，不引用任何外部 CSS/JS/字体/图片，导航用按钮+scrollIntoView 而非 hash 锚点。要响应式，手机上不裁切。", "prompt_en": "Use the 'Cangjie' distillation skill to distill Paul Graham's essay 'How to Do Great Work' (original at paulgraham.com/greatwork.html) into an executable methodology, and render it as a single-file HTML essence page. Restrained, refined reading style: white background, red accents, serif headings, a sticky TOC with scroll highlighting, Lucide line icons, no emoji. Include: a one-line core, three executable theses (choosing direction / staying upwind / resisting friction), each with concrete actions rather than platitudes; common traps and counterexamples; the author's own limits and boundaries of applicability; and three takeaways you can use immediately. Inline all resources — no external CSS/JS/fonts/images — and use button + scrollIntoView for navigation instead of hash anchors. Make it responsive with no cropping on mobile.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "《如何做出伟大的工作》蒸馏精华页" } ], "skills": [ { "id": "cangjie-skill", "name": "仓颉·长内容蒸馏", "description": "把书籍、长视频、播客、课程、长文蒸馏成一组可执行的方法论 skills", "description_en": "Distill books, long videos, podcasts, courses, and essays into a set of executable methodology skills", "skill_id": "skill_2079858719333715968" } ], "experts": [], "mcps": [], "scenario": "", "scenario_id": "", "scenario_en": "", "task_mode": "", "creator": { "name": "kangarooking", "name_en": "kangarooking", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-07-24T18:30:00Z", "updated_at": "2026-07-24T18:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cangjie-sunzi-strategy-distill", "title": "《孙子兵法》决策与谋略蒸馏", "title_en": "The Art of War — Strategy Distilled", "subtitle": "把两千年前的兵书蒸馏成一套现代决策系统", "subtitle_en": "A 2,000-year-old classic distilled into a modern decision system", "description": "《孙子兵法》人人听过，真要用起来却只剩几句金句？本案用「仓颉」蒸馏 skill 把公版《孙子兵法》真实蒸馏成一套可迁移到商业与个人决策的谋略系统：表面讲打仗，内核是一套对抗性、信息不完全环境下的决策与风险管理方法。核心提炼包括——五事七计（开打前先量化胜算）、兵者诡道（信息与预期管理）、先为不可胜（先把自己立于不败之地，再等对手露破绽）、不战而屈人之兵（以最小代价达成目标）、致人而不致于人（掌握主动权）、以正合以奇胜（稳态承接＋非对称破局）、兵形象水（因敌变化取胜）。附术语速查表把文言概念对应到现代语境。交付物是一份沉稳的单文件 HTML 精华页：白底、墨绿点缀、'兵法'印章装饰、术语速查表、粘性目录与滚动高亮，引文用楷体。适合创业者、管理者、产品与投资者，以及所有需要在竞争与不确定中做决策的人。", "description_en": "Everyone has heard of 'The Art of War', but when it's time to apply it, only a few slogans remain? Here the 'Cangjie' distillation skill turns the public-domain 'Art of War' into a strategy system transferable to business and personal decisions: on the surface it's about war, but at its core it's a method for decision-making and risk management under adversarial, incomplete-information conditions. Key distillations include: the Five Factors & Seven Assessments (quantify your odds before engaging), war as deception (managing information and expectations), first make yourself invincible (secure yourself, then wait for the opponent's opening), winning without fighting (achieve goals at minimum cost), control the enemy without being controlled (hold the initiative), engage with the orthodox and win with the extraordinary (stable base + asymmetric breakthrough), and be formless like water (adapt to win). A glossary maps classical terms to modern contexts. The deliverable is a composed single-file HTML page: white background, deep-green accents, a 'strategy' seal motif, a term glossary, a sticky TOC with scroll highlighting, and quotes set in a brush-script serif. Great for founders, managers, product people, and investors — anyone deciding under competition and uncertainty.", "categories": [ "education", "business" ], "tags": [ "知识蒸馏", "决策谋略", "经典精读", "战略思维" ], "tags_en": [ "Knowledge Distillation", "Strategy & Decisions", "Classic Deep Reading", "Strategic Thinking" ], "artifact_type": "html", "prompt": "用「仓颉」蒸馏 skill 把公版《孙子兵法》（十三篇原文，可从 zh.wikisource.org 获取）蒸馏成一套可迁移到商业与个人决策的谋略系统，并做成一页单文件 HTML 精华页。要沉稳的东方风格：白底、墨绿点缀、'兵法'印章装饰、粘性目录与滚动高亮、Lucide 线性图标、引文用楷体、不要 emoji。内容包括：一句话内核（这是一套信息不完全下的对抗性决策系统）、核心谋略提炼（五事七计、兵者诡道、先为不可胜、不战而屈人之兵、致人而不致于人、以正合以奇胜、兵形象水），每条都要把文言映射到现代商业/个人决策场景；一张术语速查表把关键概念对应到现代语境；可以立刻带走的三条决策原则。所有资源内联，不引用任何外部 CSS/JS/字体/图片，导航用按钮+scrollIntoView 而非 hash 锚点。要响应式，手机上不裁切。", "prompt_en": "Use the 'Cangjie' distillation skill to distill the public-domain 'Art of War' (the thirteen chapters, available from zh.wikisource.org) into a strategy system transferable to business and personal decisions, and render it as a single-file HTML essence page. Composed Eastern style: white background, deep-green accents, a 'strategy' seal motif, a sticky TOC with scroll highlighting, Lucide line icons, quotes in a brush-script serif, no emoji. Include: a one-line core (this is an adversarial decision system under incomplete information), core strategy distillations (Five Factors & Seven Assessments, war as deception, first make yourself invincible, winning without fighting, control the enemy without being controlled, orthodox-and-extraordinary, formless like water), each mapping the classical text to modern business/personal decision scenarios; a glossary table mapping key terms to modern contexts; and three decision principles you can use immediately. Inline all resources — no external CSS/JS/fonts/images — and use button + scrollIntoView for navigation instead of hash anchors. Make it responsive with no cropping on mobile.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "《孙子兵法》决策与谋略蒸馏页" } ], "skills": [ { "id": "cangjie-skill", "name": "仓颉·长内容蒸馏", "description": "把书籍、长视频、播客、课程、长文蒸馏成一组可执行的方法论 skills", "description_en": "Distill books, long videos, podcasts, courses, and essays into a set of executable methodology skills", "skill_id": "skill_2079858719333715968" } ], "experts": [], "mcps": [], "scenario": "", "scenario_id": "", "scenario_en": "", "task_mode": "", "creator": { "name": "kangarooking", "name_en": "kangarooking", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-07-24T18:30:00Z", "updated_at": "2026-07-24T18:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "canvas-generative-art", "title": "Canvas 流场生成艺术", "title_en": "Canvas Flow Field Generative Art", "subtitle": "基于 Perlin 噪声的粒子流场艺术生成器", "subtitle_en": "Perlin noise-based particle flow field art generator", "description": "想要用代码创造独一无二的艺术作品？本案例展示如何使用 HTML5 Canvas 和 Perlin 噪声构建粒子流场生成艺术，2000+ 粒子在噪声向量场中流动留下彩色轨迹，每次运行都产生不同的作品。支持重新生成和保存为 PNG。适合创意编程爱好者、NFT 艺术家和视觉设计师。", "description_en": "Want to create unique art with code? This case demonstrates how to build particle flow field generative art using HTML5 Canvas and Perlin noise, where 2000+ particles flow through noise vector fields leaving colorful trails. Supports regeneration and PNG export.", "categories": [ "creation" ], "tags": [ "生成艺术", "Canvas", "Perlin噪声", "创意编程" ], "tags_en": [ "Generative Art", "Canvas", "Perlin Noise", "Creative Coding" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "流场生成艺术页面" } ], "prompt": "请创建一个基于 Perlin 噪声的粒子流场生成艺术页面（单文件 HTML，内联无外部依赖），深色背景、暗角效果、流畅 60fps 动画。\n\n要求：Canvas 实现，5000+ 粒子在噪声向量场中流动留下半透明彩色轨迹，日落色调（珊瑚/琥珀/品红/靛蓝）。提供「重新生成」（新随机种子）和「保存为 PNG」按钮，信息面板显示种子号/粒子数/FPS。", "prompt_en": "Please create a Perlin-noise-based particle flow-field generative art page (a single-file HTML page, inline with no external dependencies), with a dark background, vignette effect, and smooth 60fps animation.\n\nRequirements: a Canvas implementation where 5000+ particles flow through a noise vector field leaving semi-transparent colored trails, in sunset tones (coral / amber / magenta / indigo). Provide \"Regenerate\" (new random seed) and \"Save as PNG\" buttons, and an info panel showing the seed number / particle count / FPS.", "skills": [ { "id": "canvas-design", "name": "canvas-design（视觉设计）", "description": "基于设计哲学创作精美视觉艺术", "skill_id": "skill_2053081626394972160" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "carmaker-research-workbench", "title": "上市车企研判工作台", "title_en": "Listed Carmaker Research Workbench", "subtitle": "AI 养数据、人做研判的车企跟踪工作台", "subtitle_en": "AI maintains the data, you make the calls", "description": "看车企研报永远停在生成那天？这个工作台把\"数据\"和\"判断\"分开养：AI 自动化任务每天盘后更新 15 家 A股港股车企的行情底表，每周盯着中报披露刷新分析；你的部分是真正的工作——星标自选、卡片上一秒记一笔（观察/事件/观点/复盘四类标签），笔记按时间线沉淀成你的跟踪档案，AI 更新数据永远碰不到你的记录。适合做个股跟踪、财报季研判的人，也适合任何想体验\"AI 养数据 + 人做判断\"新工作流的人。", "description_en": "Research reports that freeze on the day they were generated? This workbench raises \"data\" and \"judgment\" separately: AI automations refresh a 15-company A-share/HK carmaker quotes table daily after close and refresh analysis weekly on earnings disclosures; your part is the actual work — star your watchlist, jot a quick note in one tap (Observation/Event/View/Review tags), and let notes pile into a timeline archive. AI data updates never touch your records. For anyone tracking stocks through earnings season — and for anyone curious about the \"AI feeds data, human makes calls\" workflow.", "tags": [ "车企跟踪", "AI自动化任务", "笔记时间线", "数据底表" ], "tags_en": [ "Carmaker tracking", "AI automations", "Notes timeline", "Data base table" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "上市车企研判工作台（双语自包含演示）" } ], "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/4i6azRJ0g3u9D5LyavjbPh，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「上市车企研判工作台」 工作台，覆盖 A股+港股 15 家上市乘用车公司，数据全部存资料库数据表，页面运行时读写。五大模块：① 我的跟踪面板——首页顶部汇总自选公司（关注状态、每家最新笔记、累计笔记数），点击进详情；卡片星标为二态开关（未关注⇄自选，再点即取消），「重点关注」在详情页下拉设置；② 个股卡片——最新价/涨跌幅/总市值/TTM 财务/技术信号，卡片右上角「✎ 记一笔」快速笔记弹层（观察/事件/观点/复盘四类标签），卡片显示笔记计数；③ 详情页——公司概况、季度业绩图表、关键财务指标、资金面（A股显示主力净流入、港股显示南向持股占比）、同业对比、带「AI 初稿·更新于日期」标注的投资亮点/风险/综合评价、跟踪笔记时间线（一司多条、倒序、可追加可删除）；④ 行业分析——15 家业绩对比表 + 行业多维度解析，均由资料库表驱动；⑤ 使用手册——含数据更新机制说明。视觉走暗色金融工作台风，红涨绿跌（A股惯例），A股人民币/港股港币口径标注。\n\n数据结构（五张资料库表，创建后挂到页面节点下）：\n- 车企跟踪自选：公司代码/公司名称/关注状态（select：自选、重点关注）/最后操作——一司一条，页面读写\n- 车企跟踪笔记：公司代码/公司名称/笔记标签（select：观察、事件、观点、复盘）/笔记内容/笔记日期——一司多条追加式，页面读写\n- 车企行情底表：代码/名称/最新价/涨跌幅%/市值亿/PE_TTM/主力净流入元/南向持股占比/技术信号/财务季度JSON/投资亮点/投资风险/综合评价/更新日期——AI 维护、页面只读，与你的记录物理隔离\n- 车企业绩跟踪：公司名称/代码/市场/2026H1营收/2025H1营收/2026H1归母净利/H1营收同比/H1净利同比/H1毛利率/业绩摘要\n- 车企行业解析：条目类型（select：统计指标、分析段落）/标题/内容/排序/更新日期\n\n搭建完成后，再执行两步让工作台\"活\"起来：\n\n第一步：手动初始化数据。按「每日行情更新」的标准步骤执行一遍：用内置金融检索（agentic_search）查询 15 家公司当日行情（完整公司清单以你创建的行情底表为准），把最新价、涨跌幅%、市值亿、PE_TTM、技术信号写入你创建的行情底表（A股另写主力净流入元，港股另写南向持股占比），更新日期写当天实际日期；时点类字段查不到写空值、绝对值类字段查不到保持原值。\n\n第二步：创建两个自动化任务。**动手前先把两个任务的名称、触发周期和将写入的内容向用户说明，取得用户确认授权后再创建**（两家任务都用表驱动写法——公司清单从底表读；任务 prompt 中引用你刚创建的各表的实际 databaseId，不要使用来源页面里的表 ID）：\n\n任务一「车企研判工作台·每日盘后行情更新」：每个工作日 17:30 触发（RRULE：FREQ=WEEKLY;BYDAY=MO,TU,WE,TH,FR;BYHOUR=17;BYMINUTE=30），prompt 模板——\n「每日盘后更新资料库『车企行情底表』：1. query_database_record 读底表全部记录，公司清单以表内『代码』『名称』为准，记下 record_id；2. agentic_search 逐家查当日行情：最新价、涨跌幅、总市值、PE(TTM)、技术信号；A股另查主力资金净流入，港股另查南向持股占比；金融行情数据一律走 agentic_search，禁止用网络搜索凑数；3. batch_update_database_records 按 record_id 更新对应字段，『更新日期』写当天实际日期；4. 时点类字段当日查不到写空值、不保留旧值，绝对值类字段查不到不传保持原值；5. 输出摘要：更新家数、每家最新价/涨跌幅、缺失字段清单。」\n\n任务二「车企研判工作台·每周披露事件与分析更新」：每周一 9:30 触发（RRULE：FREQ=WEEKLY;BYDAY=MO;BYHOUR=9;BYMINUTE=30），prompt 模板——\n「每周检查披露事件并更新分析类内容（无事件即跳过，不强行改写）：1. 读行情底表公司清单与现有『财务季度JSON』『更新日期』；2. agentic_search 检索各公司最近一周的财报/业绩预告/盈利预警/重大公告；3. 有披露的公司：追加财务季度JSON新报告期、重写投资亮点/风险/综合评价（亮点3-4条、风险3-4条、评价200字内）、同步业绩跟踪表对应行；4. 行业数据（乘联会/中汽协）显著变化时更新行业解析表统计指标条目，分析段落仅格局显著变化才重写；5. 无事件输出『本周无披露事件』后结束，不修改任何内容。」\n\n交付：资料库在线工作台——在线 HTML 页面 + 5 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/4i6azRJ0g3u9D5LyavjbPh, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a 'Listed Carmaker Research Workbench' as a single HTML file covering 15 listed passenger-car makers across A-shares and HK, with all data in library databases, read and written at runtime. Five modules: 1) My Watchlist panel — top of home, aggregates followed companies (follow status, latest note per company, note counts), click to open details; card star is a two-state toggle (unfollowed ⇄ following, tap again to unfollow), 'Key focus' is set via the detail-page dropdown. 2) Stock cards — last price / change / market cap / TTM financials / technical signal, with a '✎ Quick note' button on each card (four tags: Observation/Event/View/Review) and a per-card note count. 3) Detail page — company profile, quarterly earnings charts, key metrics, money flow (main-force net inflow for A-shares, southbound holding % for HK), peer comparison, AI-drafted highlights/risks/overall view labeled 'AI draft · updated ', and a tracking-notes timeline (many per company, reverse-chronological, appendable and deletable). 4) Industry analysis — a 15-company results table plus a multi-dimension industry readout, both table-driven. 5) User guide — including the data-update mechanism. Visual: dark finance-workbench style, red-up/green-down (A-share convention), RMB for A-shares / HKD for HK stocks. Data structure (five library tables, mounted under the page node after creation):\n- Carmaker watchlist: Company code / Company name / Follow status (select: Following, Key focus) / Last action — one row per company, page read-write\n- Carmaker tracking notes: Company code / Company name / Note tag (select: Observation, Event, View, Review) / Note text / Note date — append-mode, many per company, page read-write\n- Carmaker quotes base table: Code / Name / Last price / Change % / Market cap (100M) / PE_TTM / Main-force net inflow (CNY) / Southbound holding % / Technical signal / Quarterly financials JSON / Highlights / Risks / Overall view / Update date — AI-maintained, page read-only, physically isolated from your records\n- Carmaker performance tracking: Company name / Code / Market / 2026H1 revenue / 2025H1 revenue / 2026H1 net profit / H1 revenue YoY / H1 profit YoY / H1 gross margin / Guidance summary\n- Carmaker industry analysis: Entry type (select: Statistic, Analysis paragraph) / Title / Content / Order / Update date After building, take two more steps to bring it alive: Step 1: Manual data initialization. Run the 'daily quotes update' standard procedure once: use the built-in financial search (agentic_search) to fetch the day's quotes for the 15 companies (the full list comes from the quotes base table you created), write last price, change %, market cap, PE_TTM and technical signal into that table (main-force net inflow for A-shares, southbound holding % for HK), with the actual date in 'Update date'; null out point-in-time fields when unavailable, keep level fields unchanged when unavailable. Step 2: Create two automations. **Before creating anything, explain both automations to the user — names, trigger schedules and what each will write — and get the user's explicit confirmation first** (both written table-driven — the company list is read from the base table; reference the actual databaseIds of the tables you just created in the automation prompts — never the table IDs found in the source page): Task 1 'Carmaker workbench · daily after-close quotes update': every weekday 17:30 (RRULE: FREQ=WEEKLY;BYDAY=MO,TU,WE,TH,FR;BYHOUR=17;BYMINUTE=30), prompt template —\n\"After close each day, update the library 'carmaker quotes base table': 1. query_database_record to read all rows; the company list comes from the table's own Code/Name columns; note the record_ids. 2. agentic_search per company for the day's last price, change, market cap, PE(TTM), technical signal; plus main-force flow for A-shares and southbound holding % for HK; market data must come from agentic_search — no web-search stand-ins. 3. batch_update_database_records by record_id for those fields; write the actual date into 'Update date'. 4. Null out point-in-time fields when unavailable — never stale; omit level fields to keep their values. 5. Output a summary: companies updated, each last price/change, missing-field list.\" Task 2 'Carmaker workbench · weekly disclosure-driven analysis update': every Monday 9:30 (RRULE: FREQ=WEEKLY;BYDAY=MO;BYHOUR=9;BYMINUTE=30), prompt template —\n\"Weekly, check disclosure events and update analysis content (skip when there are none; never force rewrites): 1. Read the company list plus existing 'Quarterly financials JSON' and 'Update date' from the base table. 2. agentic_search for each company's past-week earnings, guidance, profit alerts and major filings. 3. For disclosers: append new periods to the quarterly JSON, rewrite highlights/risks/overall view (3-4 bullets each, view within 200 words), sync the matching performance-table row. 4. When industry data (CPCA/CAAM) shifts materially, update the industry-analysis stat entries; rewrite paragraphs only on structural changes. 5. If no events, output 'No disclosure events this week' and change nothing.\"\n\nDeliverable: an online Library workbench — an online HTML page plus 5 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "趣解商业", "name_en": "趣解商业", "avatar": "", "badge": "community" }, "source": "community", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 86, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "created_at": "2026-08-24T04:56:36+00:00", "updated_at": "2026-08-24T04:56:36+00:00", "is_recommended": false, "recommended_order": 0 }, { "id": "cet-prep-workbench", "title": "四六级备考工作台", "title_en": "CET-4/6 Prep Workbench", "subtitle": "打卡 · 背词 · 听力 · 真题 · 错题闭环", "subtitle_en": "Check-ins · Word drills · Listening · Mock exams · Mistake loop", "description": "面向 CET-4/CET-6 备考的一体化学习工作台：今日打卡（连续/累计天数、近 18 周热力图、四张目标卡）、单词本（288 词内置词库覆盖四级 155 + 六级 133，看词选义/听音辨义/看义选词/听音拼写四题型学习引擎，答错的词自动排队重现，支持自定义生词与斩词）、听力训练（10 套仿真听力朗读播放可调速，在线作答自动判分，精听模式逐句播放/隐藏原文，支持导入本地音频）、真题演练（3 套仿真模拟卷自动判分折合 249 分制、710 分趋势图带 425 及格线、历年真题库 3 套 30 题覆盖写作/翻译/听力理解/仔细阅读/选词填空/段落匹配六题型并附答案解析）、长难句拆分器（光标断句 + 成分标注 + 句子库）、错题本（错误原因归档 + 状态流转）、数据统计与 11 项成就墙。内置演示数据开箱即用，所有学习记录保存在浏览器本地，刷新不丢，支持中英双语界面。适合备考四六级的大学生、英语教师布置自学任务、想系统管理英语学习进度的人。", "description_en": "An all-in-one study workbench for CET-4/CET-6 takers: daily check-ins (streak/total days, 18-week heatmap, four goal cards), vocabulary (a built-in 288-word bank — 155 CET-4 + 133 CET-6 — driven by a four-question-type engine: word→meaning, sound→meaning, meaning→word and spelling, with missed words automatically resurfacing, plus custom words and a 'cut' mechanic), listening training (10 mock sets read aloud at adjustable speed, online answering with auto-grading, an intensive line-by-line mode, local-audio import), mock exams (3 auto-graded papers scaled to the 249-point sections, a 710-point trend chart with the 425 pass line, and a past-exam bank of 3 sets / 30 questions across writing, translation, listening comprehension, careful reading, banked cloze and paragraph matching — with answers and explanations), a sentence splitter (cursor-based cuts, clause labels, sentence library), a mistake log (reason tagging and status flow) and stats with 11 achievements. Ships with demo data out of the box; all records persist in the browser; bilingual UI. Best for: college students preparing for CET-4/6, teachers assigning self-study, and anyone who wants structured English exam prep.", "categories": [ "workspace" ], "tags": [ "四六级", "备考", "英语学习", "学习打卡", "工作台" ], "tags_en": [ "CET-4/6", "Exam Prep", "English Learning", "Study Check-in", "Workbench" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "四六级备考工作台（双语自包含演示）" } ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/sYGavuhQEidofWLTLbpQeo，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「四六级备考工作台」 工作台，面向 CET-4/CET-6 备考学生，左侧深色侧栏导航 + 右侧内容区，共 7 个页面：\n\n1. 今日打卡——英雄区（问候语、连续/累计打卡、今日目标完成度、打卡按钮）、四张目标卡（背单词/听力/长难句/错题复习，进度条）、近 18 周打卡热力图（126 格 + 本周列表 + 悬浮明细）、今日学习建议 6 条轮换；\n2. 单词本——学习计划设置（目标考试四级/六级/全都要 × 每日新词量 20/40/60/80）、四题型学习引擎（看词选义/听音辨义/看义选词/听音拼写，组末战报结算，答错的词自动排队重现）、自定义生词添加、内置词库列表（四级 155 + 六级 133 共 288 词，支持斩词/恢复/搜索）；\n3. 听力训练——10 套仿真听力材料（speechSynthesis 朗读播放，语速 0.7×-1.2× 可调）、在线作答自动判分并计入记录、精听模式（逐句播放/点句显示隐藏原文/跟读）、导入本地音频配合原文精听、手动记录练习（材料类型/时长/正确率/心得）、听力提分三阶段建议；\n4. 真题演练——3 套仿真模拟卷（听力/阅读自动判分折合 249 分制、翻译写作自评 0-106.5、一键存入成绩）、710 分趋势折线图（425 及格虚线）、手动记录历年真题成绩、历年真题库（内置 3 套真题 30 题：写作/翻译/听力理解/仔细阅读/选词填空/段落匹配六题型，按套次折叠浏览，含选项、答案与解析，支持级别筛选）；\n5. 长难句——句子拆分器（光标处断句/五种成分标签：主干/定语/状语/从句/插入同位/整句翻译/结构笔记）、我的句子库；\n6. 错题本——错题录入（题型/出处/错误原因 7 选 1/题目/我的答案/正确答案/正确思路）、状态流转（未复习→已复习→已掌握→再错一次回滚）、按题型与状态筛选；\n7. 数据统计——六项汇总指标（已掌握单词/累计听力/真题套数/已拆句子/错题总数/连续打卡）、错题原因分布、听力材料分布、每日目标设置；11 个成就（初次打卡/七日之约/月度坚持/小试牛刀/百词过关/词霸降临/磨耳朵/听力狂人/身经百战/拆句学徒/错题清零）成就墙。\n\n数据全部存 localStorage（打卡/学习记录/自定义词/句子/错题/成绩），首次打开内置演示数据（打卡 11 天、听力 3 条、真题 3 套、错题 3 条），页面刷新后数据保留。视觉走清爽学习风（靛蓝主色 + 渐变英雄区 + 卡片阴影大圆角），所有 CSS/JS 无外部依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to one-click replicate https://www.workbuddy.cn/space/d/sYGavuhQEidofWLTLbpQeo with online data storage and two-way sync. Prefer the one-click replication path; if unavailable, build it as follows: Build a 'CET-4/6 Prep Workbench' as a workbench for CET-4/CET-6 students, with a dark left sidebar and a content area on the right, covering 7 pages: 1. Today — hero zone (greeting, streak/total days, daily goal completion, check-in button), four goal cards (words/listening/sentences/mistakes with progress bars), an 18-week check-in heatmap (126 cells + this-week list + hover details), and 6 rotating study tips;\n2. Vocabulary — study plan setup (target CET-4/6/both × 20/40/60/80 new words per day), a four-question-type learning engine (word→meaning, sound→meaning, meaning→word, spelling, with a round report and missed words auto-resurfacing), custom word entry, and a built-in word list (155 CET-4 + 133 CET-6 = 288 words, with cut/restore/search);\n3. Listening — 10 mock listening sets (speechSynthesis read-aloud, 0.7×-1.2× speed), online answering with auto-grading logged to records, an intensive mode (line-by-line playback, per-line script toggle, shadowing), local-audio import for script-aligned practice, manual practice logging (type/duration/accuracy/notes), and a three-stage score plan;\n4. Mock Exams — 3 mock papers (listening/reading auto-graded and scaled to the 249-point sections, self-scored translation/writing 0-106.5, one-click save), a 710-point trend chart with a 425 pass line, manual past-exam score logging, and a past-exam bank (3 built-in sets / 30 questions across writing, translation, listening comprehension, careful reading, banked cloze and paragraph matching, grouped by set with options, answers and explanations, filterable by level);\n5. Sentence Lab — a sentence splitter (cut at cursor, five clause labels: main/attributive/adverbial/subordinate/parenthetical, full translation, structure notes) and a sentence library;\n6. Mistakes — mistake entry (type/source/seven reasons/question/my answer/correct answer/reasoning), status flow (not reviewed → reviewed → mastered, with 'wrong again' rollback), filtering by type and status;\n7. Stats — six summary metrics (words mastered/total listening/exam sets/sentences split/total mistakes/streak), mistake-reason distribution, listening-material distribution, daily goal settings; an achievement wall with 11 achievements. All data lives in localStorage (check-ins/learning records/custom words/sentences/mistakes/scores); first open shows built-in demo data (11 check-in days, 3 listening records, 3 exam sets, 3 mistakes); data survives refresh. Visual: clean study style (indigo primary, gradient hero, card shadows, large radii). All CSS/JS inlined, no external dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "锦照", "name_en": "锦照", "avatar": "", "badge": "community" }, "source": "community", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 0, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "created_at": "2026-09-02T18:10:00+08:00", "updated_at": "2026-09-02T18:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "cet4-sprint-workbench", "title": "牟牟的四级冲刺台", "title_en": "CET-4 Sprint Workbench", "subtitle": "100 天从 300 到 450：动态派活、做题批改、错题歼灭、进度追踪", "subtitle_en": "100 days from 300 to 450: adaptive daily tasks, auto-graded practice, error-book review and progress tracking", "description": "四级备考最痛的不是缺资料，而是不知道今天该练什么：单词背了忘、真题刷不动、错题越堆越多，到考前才发现节奏全乱。\n\n本案例由 WorkspaceBuilder 工作台搭建师专家生成：一个 100 天冲刺工作台，目标是 300 → 450 分。「今日任务」是动态派活引擎——按剩余天数所处阶段（基础/专项/真题/定型）、各模块与目标分的差距、昨日完成率（低于 50% 自动减量、全完自动加量）、错题堆积类型，每天自动排出 4-6 项任务，昨天没做完的自动标红逾期、一键顺延；「训练场」内置 49 道词汇搭配、18 道听力、24 道仔细阅读、6 篇写作真题、6 段翻译，客观题提交即逐题批改给解析，错题自动进错题本，主观题做硬伤检测 + 采分点自评 + 范文对照；「进度追踪」有分数曲线（计划线/模考分/训练推算分/425 及格线）、四模块柱状图、打卡热力图和近 20 天明细。数据存资料库在线表，多设备同步。\n\n适合：备考四级的大学生（100 天冲刺节奏全托管）、拖延症患者（逾期任务不消失、自动减量防摆烂）、需要错题闭环的刷题党（错题自动归集、掌握后移出），以及想参考「动态任务引擎 + 在线表存储」实现的前端学习者。", "description_en": "The hardest part of CET-4 prep isn't finding materials — it's knowing what to practice today: vocab gets forgotten, past papers feel endless, wrong answers pile up, and the pace collapses right before the exam.\n\nBuilt by the WorkspaceBuilder expert: a 100-day sprint workbench targeting 300 → 450. \"Today\" is an adaptive task engine that schedules 4-6 tasks daily based on your phase (foundation / skills / past-paper sprint / final week), each module's gap to its target score, yesterday's completion rate (auto -25% if under 50%, +20% after a perfect day) and which question type is piling up in your error book; unfinished tasks turn red as overdue and can be carried over in one tap. The Training Ground ships 49 vocab collocation questions, 18 listening items, 24 reading items, 6 real-exam essay prompts and 6 translation passages — objective questions are graded instantly with per-item explanations and mistakes flow into the error book automatically, while subjective ones get automated flaw checks, a scoring-point self-review and model answers. Progress tracking offers a score curve (plan line / mock scores / training-estimated score / the 425 pass line), per-module bar charts, a check-in heatmap and a 20-day detail table. Data lives in cloud database tables and syncs across devices.\n\nFor: college students preparing for CET-4 (a fully paced 100-day plan), procrastinators (overdue tasks never vanish and workloads auto-adjust), drill-focused learners who want an error-closing loop, and front-end developers curious about the adaptive task engine + online table storage design.", "categories": [ "workspace" ], "tags": [ "四级备考", "英语学习", "备考计划", "错题本", "个人工作台" ], "tags_en": [ "CET-4 prep", "English learning", "study planner", "error book", "personal workspace" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "四级冲刺台 单文件 HTML（内置 i18n 中英文 + 中英双版示例数据与题库，localStorage 自包含）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/vCvKs5r65jDD1VDG59Fafs，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n我想做一个「四级冲刺」备考工作台，100 天从 300 分冲到 450 分，包含：\n1. 今日任务：动态派活引擎。按剩余天数落在哪个备考阶段（基础夯实/专项突破/真题冲刺/考前定型，各阶段有不同任务池）、各模块自评分与目标分的差距（差距越大权重越高）、昨日完成率（<50% 自动减量 25%、100% 自动加量 20%）、错题堆积类型（某类错题 ≥2 道该模块权重 ×1.5）四个输入，每天自动生成 4-6 项任务；听力+阅读每天必排，写作与翻译按权重轮换；过去 7 天未完成项标红显示逾期天数，可一键顺延到今天。\n2. 训练场：做题 + 自动批改。词汇搭配 10 题限时、听力 1 组、仔细阅读 1 篇按套餐组合；客观题提交即逐题标红绿、给正确答案和考点解析，错题自动进错题本；写作和翻译做自动硬伤检测（although/but 连用、词数不足、连接词偏少、缺高级句型）+ 采分点清单自评折算得分 + 范文/参考译文对照；错题本支持标记「已掌握」移出。\n3. 进度追踪：分数曲线（100 天匀速计划虚线 + 手动记录的模考分 + 按训练数据推算的实时水平 + 425 及格线，提示领先/落后计划 N 分）、听读写译四模块柱状图（起点/当前推算/阶段目标）、按完成率 5 级配色的打卡热力图、近 20 天任务完成数/训练时长/正确率/复盘摘要明细表。\n4. 设置：考试日期、起点分、目标分、每日训练时长、四模块当前水平自评；保存后今日任务按新参数重算。\n\n数据结构（资料库数据表）：备考设置（考试日期/开始日期/基础分/目标分/每日预算/四模块分）、每日任务（日期/任务名/模块/时长/已完成/是否训练场）、每日复盘（日期/心情/笔记）、模考成绩（日期/总分/听力/阅读/写作/翻译/来源）、错题本（题型/题目/我的答案/正确答案/解析/日期/已掌握）、训练记录（日期/类型/答对/总题数/用时）。\n\n视觉与图表为手写内联 SVG；首屏头部显示距考试天数、当前预估分、还需提升、连续打卡天数和目标进度条；PC 顶部横向 Tab、移动端底部 Tab，按钮 ≥44px；蓝色渐变头部 + 大圆角卡片风格。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the cloud database (资料库) capability to one-click clone this workbench: https://www.workbuddy.cn/space/d/vCvKs5r65jDD1VDG59Fafs — data must be stored online with bidirectional sync. Prefer the one-click clone path; if unavailable, build it this way: I want a \"CET-4 Sprint\" prep workbench: 100 days from 300 to 450 points, including:\n1. Today: an adaptive task engine. Every day it generates 4-6 tasks from four inputs — which prep phase the remaining days fall into (foundation / skill drills / past-paper sprint / final week, each with its own task pool), the gap between each module's self-rated score and its target (bigger gap = higher weight), yesterday's completion rate (-25% load if under 50%, +20% after a perfect day), and which question type is piling up in the error book (weight ×1.5 at ≥2 items). Listening + reading are scheduled daily; writing and translation rotate by weight. Unfinished items from the past 7 days show in red with overdue days and can be carried over to today in one tap.\n2. Training Ground: practice with instant grading. Timed sets combining 10 vocab collocations, 1 listening group and 1 reading passage; objective questions are graded on submit with per-item red/green marking, correct answers and exam-point explanations, and mistakes flow into the error book automatically; essays and translations get automated flaw checks (although/but together, word count below 120, too few connectives, no advanced clauses), a scoring-point checklist that converts to an estimated score, and model-answer comparison; mastered errors can be removed from the book.\n3. Progress: a score curve (the 100-day linear plan line, manually logged mock scores, a live training-estimated level, and the 425 pass line, showing how far ahead/behind plan you are), per-module bar charts (start / current estimate / phase target), a 5-level check-in heatmap by completion rate, and a 20-day table of tasks done, training minutes, accuracy and review notes.\n4. Settings: exam date, base score, target score, daily minutes and per-module self ratings; saving recomputes today's tasks. Data structure (cloud database tables): settings (exam date / start date / base / target / daily minutes / four module scores), daily tasks (date / name / module / minutes / done / is-training), daily reviews (date / mood / note), mock scores (date / total / listening / reading / writing / translation / source), error book (type / question / my answer / correct answer / explanation / date / mastered), training logs (date / type / correct / total / minutes). Visual & delivery: all CSS/JS/charts inline with zero external links; hand-written inline SVG charts; header shows days to exam, current estimated score, gap to target, check-in streak and a progress bar; horizontal tabs on desktop, bottom tab bar on mobile, buttons ≥44px; blue gradient header with large rounded cards.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "牟牟", "name_en": "Moumou", "avatar": "", "badge": "community" }, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "created_at": "2026-08-17T14:35:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "changan-changxin-palace-lamp-proposal", "title": "长信宫灯 · 文创产品线提案", "title_en": "Changxin Palace Lamp · Cultural-Creative Product Line Proposal", "subtitle": "3 个核心记忆点 + 4 条产品线 + AI 商品示意图", "subtitle_en": "3 memory anchors + 4 product lines + AI product mockups", "description": "想把一件深宫旧物做成年轻人愿意买单的周边，却摸不准从哪下手？看完一份又一份 PPT 提案，最后做出来还是「大红大绿旅游纪念品」。\n\n这个案例用探元文创提案专家，把「长信宫灯」这件西汉一级文物（河北博物院藏）当成主题。先用 `mcp__tanyuan-wb-mcp-server__search_relics` / `search_movable_relics` 拉出权威事实，提炼出 3 个核心记忆点（跽坐宫女造型 / 导烟环保系统 / 模块化拆装结构），再围绕每条产品线（鎏金铜灯香薰炉、鎏金纹样真丝披肩、六部分拆装木质拼装盲盒）给出定位人群、材质工艺与 ImageGen 商品示意图，最后沉淀出主色鎏金 + 汉漆朱红 + 竹青的色彩体系。一份单文件 HTML 提案，可直接发给设计 / 法务 / 供应商协同。\n\n适合：博物馆文创负责人、文化 IP 运营、文创品牌主理人、文旅项目策划、文物衍生品买手。", "description_en": "Want to turn a deep-palace antiquity into merch that younger buyers actually want, but unsure where to start? Hours of pitch decks later, the result still ends up looking like a generic tourist trinket.\n\nThis case uses the TanYuan Cultural-Creative Expert to take the Changxin Palace Lamp — a Western Han Grade-1 relic housed in the Hebei Museum — and turn it into a real product-line proposal. The agent first calls `mcp__tanyuan-wb-mcp-server__search_relics` and `search_movable_relics` to pull authoritative facts, distills 3 memory anchors (kneeling palace-lady silhouette, soot-absorbing eco flue, modular six-part assembly), then designs 4 product lines (gilded brass incense lamp, gilded silk shawl, six-part wooden assembly blind box, + one core rendition). Each line carries target users, materials, and an ImageGen product mockup, ending with a unified palette of gilded gold + lacquer red + bamboo green. Output is a single-file HTML deliverable ready to forward to design / legal / suppliers.\n\nBest for: Museum cultural-creative leads, cultural IP operators, cultural-creative brand owners, cultural-tourism project planners, relic-derivative product buyers.", "categories": [ "creation", "business" ], "tags": [ "文创提案", "长信宫灯", "汉代文物", "产品线设计", "AI 商品图" ], "tags_en": [ "Cultural Creative", "Changxin Palace Lamp", "Han Relic", "Product Line Design", "AI Product Mockup" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "cover_image": "cover.png", "cover_image_en": "cover.png", "cover_mode": "landscape", "prompt": "以长信宫灯为主题做一份文创产品线设计提案。\n\n背景：你是一位「腾讯探元文创提案专家」，把一件文物/一个馆藏主题转化为一份「文创产品线设计提案」，输出单文件 HTML（CSS/字体/图片全内联，可离线打开）。\n\n## 工作流程\n\n### Step 1 检索权威事实（必做，禁编造）\n调用 `mcp__tanyuan-wb-mcp-server__search_relics`，query=\"长信宫灯\"，datasourceType=0；再调用 `mcp__tanyuan-wb-mcp-server__search_movable_relics`，query=\"长信宫灯 工艺 寓意\"，topK=3。读 result 字段，把权威信息提炼为：年代 / 材质 / 等级 / 馆藏 / 尺寸 / 6 大可拆部件 / 导烟原理 / 铭文 9 处 65 字 / 流转历史（阳信侯→窦太后→窦绾）。所有事实标注 [探元文物库] / [探元知识库]，缺失留白不编。\n\n### Step 2 提炼 3 个核心记忆点\n从检索结果提炼：① 跽坐执灯宫女造型（汉代写实主义艺术代表）② 鎏金导烟环保系统（宫女体内中空 + 右臂虹管）③ 6 大部件拆装结构（灯盘 360° 转 + 灯罩双屏板可调）。这 3 个点同时驱动产品线设计与色彩体系。\n\n### Step 3 规划约 4 条产品线\n每条含：定位人群/场景 + 材质工艺 + 与记忆点的关联 + ImageGen 商品示意图。示例骨架：\n- 鎏金铜灯香薰炉（案头复制款）\n- 鎏金提灯纹真丝披肩（穿戴周边）\n- 6 部分拆装木质拼装盲盒（益智收藏玩具）\n- 长信宫灯造型铜灯茶器（茶饮店衍生）\n主色：鎏金 + 汉漆朱红 + 竹青；避免大红大绿旅游纪念品感。\n\n### Step 4 ImageGen 商品示意图\n参考 tanyuan-wb MCP/SKILL.md 与 generate_me 图片生成规范：每条产品线一张图，prompt 必须写明 ① 材质 ② 形态 ③ 提炼自文物的具象符号（如跽坐宫女、虹管导烟、可拆装灯罩）④ 使用场景 ⑤ 主色，末尾追加 \"absolutely no text, no letters, no words, no labels, no numbers, no watermark, no distortion, no messy background, no cheap plastic look\"。优先用 search_relics 返回的 cover 做 image-to-image 复刻款。\n\n### Step 5 合成单文件 HTML\n按预设版式（Hero / Fact strip / 3 记忆点 / 4 产品线 / 色彩体系 / AI 声明）输出 output.html。所有图片 base64 内嵌，所有字体子集化内嵌，无 https:// 外链、无 hash 锚点、无 emoji。", "prompt_en": "Create a cultural-creative product-line proposal for the Changxin Palace Lamp.\n\nBackground: You are the TanYuan Cultural-Creative Expert. Turn a relic / museum-curated theme into a single-file HTML proposal with all CSS / fonts / images inlined (fully offline).\n\n## Workflow\n\n### Step 1 — Retrieve authoritative facts (mandatory, no fabrication)\nCall `mcp__tanyuan-wb-mcp-server__search_relics` with query='长信宫灯' (Changxin Palace Lamp) and datasourceType=0. Then call `mcp__tanyuan-wb-mcp-server__search_movable_relics` with query='长信宫灯 工艺 寓意' and topK=3. Read `result`, distil: dynasty / material / grade / keeper / dimensions / 6 modular parts / soot-absorbing flue principle / 9 inscription marks / provenance (Yangxin Hou → Empress Dowager Dou → Dou Wan). Tag every fact `[探元文物库]` / `[探元知识库]`; leave blank rather than fabricate.\n\n### Step 2 — Distil 3 memory anchors\nFrom the search results, lift: ① kneeling palace-lady silhouette (Han realism) ② gilded soot-absorbing eco flue (hollow body + right-arm duct) ③ 6-part modular structure (360° dish + 2-panel shade). These three drive both product-line design and palette.\n\n### Step 3 — Plan ~4 product lines\nEach carries: target user / scenario + material & craft + link to a memory anchor + ImageGen mockup. Skeleton:\n- Gilded brass incense lamp (desktop reproduction)\n- Gilded silk shawl with kneeling-lady motif (wearable)\n- Six-part wooden assembly blind box (collectible puzzle)\n- Bronze-lamp-shaped tea ware (tea-house derivative)\nPalette: gilded gold + Han lacquer red + bamboo green. Avoid the loud red-green tourist-trinket feel.\n\n### Step 4 — ImageGen product mockups\nReference the tanyuan-wb MCP SKILL.md and standard image-gen specs: 1 image per product line, prompt must specify ① material ② form ③ relic-specific concrete symbol (kneeling lady, soot duct, modular shade) ④ use scenario ⑤ palette; append \"absolutely no text, no letters, no words, no labels, no numbers, no watermark, no distortion, no messy background, no cheap plastic look\". Prefer `search_relics` cover for the reproduction line's image-to-image.\n\n### Step 5 — Compose single-file HTML\nFollow the existing template (Hero / Fact strip / 3 memory anchors / 4 product lines / palette / AI disclaimer) as output.html. Base64-inline every image, subset-inline every font. No https:// external resources, no hash anchors, no emoji.", "skills": [], "experts": [ { "id": "tanyuan-cultural-creative", "name": "文创提案专家", "description": "把一件文物做成文创产品线设计提案,含核心记忆点、产品线与 AI 商品示意图,事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cf052b46c196", "expert_id": "ex_HXd4FrKZutHg" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "设计创意", "usage_count": 0, "created_at": "2026-08-14T18:00:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "长信宫灯 · 文创产品线提案" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "changxin-lamp-cultural-creative-brief", "title": "长信宫灯文创设计Brief", "title_en": "Changxin Palace Lamp: Cultural-Creative Brief", "subtitle": "指定一件文物，自动生成文创产品线设计 Brief", "subtitle_en": "Name any relic, auto-generate a cultural-creative product-line brief", "description": "博物馆想开发文创，最烦的是从文物到产品线之间那道坎：文物要点谁来提炼、做哪些 SKU、视觉调性怎么统一，一份像样的设计 Brief 常常要来回拉扯好几轮。\n\n这个案例把「长信宫灯」交给腾讯探元文博专家，一句话产出了一份单文件 HTML 文创设计 Brief：先梳理文物本体（西汉鎏金青铜灯、导烟净气的环保智慧、跪坐执灯造型等要点，信息源自探元资料库），提炼出「跪坐姿态 + 鎏金质感 + 环保智慧」三大记忆点，再落到复刻摆件、帆布袋、冰箱贴、笔记本四条产品线的商品示意，最后给出色彩体系与设计要点的视觉规范，供文创团队扩展 SKU 时保持一致。\n\n适合：博物馆文创与商品开发团队、品牌与视觉设计师、IP 授权与运营人员、文化空间策划者。", "description_en": "When a museum wants to develop cultural-creative products, the hardest gap is the leap from artifact to product line: who distills the relic's key points, which SKUs to make, how to unify the visual tone — a decent design brief often takes several painful rounds.\n\nThis case hands the 'Changxin Palace Lamp' to the TanYuan Cultural Heritage Expert and produces, from a single prompt, a single-file HTML cultural-creative brief: it first lays out the artifact itself (a Western-Han gilt-bronze lamp, its smoke-guiding eco-wisdom, the kneeling lamp-bearer form — facts sourced from the TanYuan knowledge base), distills three memory points ('kneeling posture + gilt texture + eco-wisdom'), then lands on product mock-ups across four lines (replica figurine, canvas bag, fridge magnet, notebook), and finally gives a visual guide of color system and design principles so the team stays consistent when expanding SKUs.\n\nIdeal for: museum cultural-creative and merchandise teams, brand and visual designers, IP licensing and operations staff, and cultural-space planners.", "categories": [ "creation", "business" ], "tags": [ "文创设计", "长信宫灯", "设计Brief", "博物馆IP", "单文件HTML" ], "tags_en": [ "Cultural Creative", "Changxin Palace Lamp", "Design Brief", "Museum IP", "Single-file HTML" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "长信宫灯文创设计 Brief（单文件 HTML）" } ], "prompt": "请以「长信宫灯」为主题，为博物馆文创团队生成一份文创周边产品线设计 Brief，输出为单文件 HTML。\n\n内容要求：\n1. 文物本体：梳理年代、出土、材质尺寸、结构与铭文、导烟净气的环保设计等关键信息（信息须来自探元文博知识库并标注来源）；\n2. 设计记忆点：从文物中提炼 3 个核心记忆点，作为整条产品线的叙事主线；\n3. 商品示意：规划 4 条产品线（如复刻摆件、帆布袋、冰箱贴、笔记本），每件给出定位、材质工艺与适配场景，并配设计示意图；\n4. 视觉规范：给出统一的色彩体系与设计要点（做旧质感、暖光打底、印章点睛、造型辨识等），供后续扩展 SKU 时保持系列视觉一致。\n\n请明确标注：文物文字信息来自权威资料，商品示意图为 AI 生成的设计示意（非真实产品、非最终打样），正式开发前需由文物与法务专业人员复核形象使用与版权边界。\n\n视觉风格：暖褐做旧背景、鎏金质感、朱红印章，古朴典藏气质。技术约束：单文件 HTML，所有 CSS/JS/图片资源全部内联，可离线打开，不引用任何外部链接。", "prompt_en": "Using the 'Changxin Palace Lamp' as the theme, generate a cultural-creative product-line design brief for a museum's cultural-creative team, output as a single-file HTML.\n\nContent requirements:\n1. The artifact itself: lay out key facts — era, excavation, material and dimensions, structure and inscriptions, and the smoke-guiding eco-friendly design (facts must come from the TanYuan cultural-heritage knowledge base with sources noted);\n2. Design memory points: distill 3 core memory points from the artifact as the narrative spine of the whole product line;\n3. Product mock-ups: plan 4 product lines (e.g., replica figurine, canvas bag, fridge magnet, notebook), each with its positioning, material/craft, and fitting scenario, accompanied by design mock-up images;\n4. Visual guide: provide a unified color system and design principles (aged texture, warm-light base, seal accents, silhouette recognizability) so the series stays visually consistent when expanding SKUs.\n\nExplicitly note: the relic's textual facts come from authoritative sources; the product mock-up images are AI-generated design references (not real products, not final samples), and before formal development, cultural-heritage and legal professionals must review image usage and copyright boundaries.\n\nVisual style: warm-brown aged background, gilt texture, vermilion seals, an archaic and archival feel. Technical constraints: single-file HTML with all CSS/JS/image resources fully inlined, openable offline, referencing no external links.", "skills": [], "experts": [ { "id": "TanyuanCulturalHeritageExpert", "name": "文博领域AI顾问", "name_en": "Cultural Heritage AI Consultant", "description": "基于腾讯探元文博知识库，提供文物与世界遗产查询、知识问答、文物对比、文博攻略及研学方案等专业服务。", "description_en": "Powered by TanYuan, providing cultural relic and world heritage queries, knowledge Q&A, relic comparison, museum guides and study programs.", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "fbca4e25911e", "expert_id": "ex_84afKR4m4iUc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-28T12:13:24+08:00", "updated_at": "2026-07-28T12:13:24+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "channel-customer-value-profile", "title": "渠道客户价值画像", "title_en": "Channel Customer Value Profile", "subtitle": "不同成交渠道的客户消费价值与商品偏好洞察", "subtitle_en": "A practical Weimob playbook for channel customer value profile", "description": "做渠道客户价值画像时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“渠道客户价值画像”展开：分析不同成交渠道带来的客户在消费金额、购买频率、偏好商品上的差异，找到值得长期经营的渠道和客群。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、渠道流量归因运营人员、经营分析团队。", "description_en": "The hardest part of channel customer value profile is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates channel customer value profile through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Channel Attribution specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "渠道客户", "客户价值", "购买频率", "商品偏好", "渠道经营" ], "tags_en": [ "Channel Customer Value Profile", "Business Diagnosis", "Channel Attribution", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我看看不同成交渠道各自擅长卖什么、客户有什么不同，以及我应该把哪些商品和活动放到哪个渠道重点经营。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a channel customer value profile covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "channel-payment-order-mix-analysis", "title": "渠道支付订单结构分析", "title_en": "Channel Payment Order Mix Analysis", "subtitle": "近30天可识别渠道的支付订单与商品结构洞察", "subtitle_en": "A practical Weimob playbook for channel payment order mix analysis", "description": "做渠道支付订单结构分析时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“渠道支付订单结构分析”展开：汇总近30天各可识别渠道的支付订单量、GMV和成交商品数，识别主要成交渠道及其商品结构差异。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、渠道流量归因运营人员、经营分析团队。", "description_en": "The hardest part of channel payment order mix analysis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates channel payment order mix analysis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Channel Attribution specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "渠道分析", "支付订单", "GMV", "成交商品", "商品结构" ], "tags_en": [ "Channel Payment Order Mix Analysis", "Business Diagnosis", "Channel Attribution", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "汇总近30天各可识别渠道的支付订单量、GMV和成交商品数，识别主要成交渠道及其商品结构差异；仅基于订单成交数据分析，不要求流量或投放成本。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a channel payment order mix analysis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "channel-product-strategy-ideas", "title": "渠道商品打法灵感", "title_en": "Channel Product Strategy Ideas", "subtitle": "基于渠道热销商品的组合、内容与活动策略", "subtitle_en": "A practical Weimob playbook for channel product strategy ideas", "description": "做渠道商品打法灵感时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“渠道商品打法灵感”展开：结合各渠道近期卖得好的商品和订单表现，为不同渠道设计更适合的商品组合、内容卖点和促销玩法。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、渠道流量归因运营人员、经营分析团队。", "description_en": "The hardest part of channel product strategy ideas is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates channel product strategy ideas through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Channel Attribution specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "渠道商品", "商品组合", "内容卖点", "促销玩法", "经营策略" ], "tags_en": [ "Channel Product Strategy Ideas", "Business Diagnosis", "Channel Attribution", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "结合各渠道近期卖得好的商品和订单表现，为不同渠道设计更适合的商品组合、内容卖点和促销玩法。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a channel product strategy ideas covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "content-creation", "name": "内容创作与分发", "description": "创作营销文案、公众号图文、小红书和朋友圈内容；可按渠道生成素材并发布或保存入库。", "skill_id": "skill_2087078523417223168" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "character-doodle-study-clock", "title": "汤匙不是钥匙的角色主题涂鸦学习时钟", "title_en": "Character-themed Doodle Study Clock", "subtitle": "上传角色素材，让它陪你专注与学习", "subtitle_en": "Upload your character and let it focus and study with you", "description": "番茄钟和学习工具千篇一律，用久了就腻；想要一个\"有陪伴感\"的、能跟着自己节奏变换表情的专属学习工作台。\n\n本案例由 WorkspaceBuilder 工作台搭建师专家生成：你上传的 Q 版角色图（默认/专注/欢呼/发呆四种表情）会被内嵌到一份单文件 HTML 里，Canvas 自动提取角色主色调生成整站配色方案，字体、边框、按钮、装饰、图表线条全部融入角色色系——再也不是千篇一律的黑白 UI。计时进行中角色切换专注表情并呼吸缩放，完成任务时切换欢呼表情、弹跳动画 + 星星涂鸦粒子 + 激励气泡；内置原生 Canvas 手绘风格的当日分类柱状图（学习/运动/日常/娱乐/阅读）和七日趋势折线图，点击柱状图查看当日分类详情；所有任务、分类耗时、每日历史 7 天数据、形象 Base64 都存进 localStorage，支持一键导出/导入备份。\n\n适合：学生与备考党（番茄钟陪伴学习）、番茄钟重度用户（想要专属感）、喜欢个性化定制的人（不满足于通用 App）、独立开发者与设计师（学习单文件 Canvas 应用开发的完整参考）。", "description_en": "Pomodoro timers and study tools all look the same — you get bored after a while. What if you had a companion-style study workspace where the character reacts with you?\n\nThis case is built by the WorkspaceBuilder expert: upload four Q-version character images (idle / focus / celebrate / sad expressions) and they get embedded into a single-file HTML. Canvas auto-extracts the character's dominant color and generates a full design system — colors, fonts, borders, buttons, decorations and chart lines all flow from your character, not generic black-and-white. While the timer runs, the character switches to focus mode with breathing scale; on task completion it bounces to celebrate mode with doodle particles and an encouragement bubble. Includes hand-drawn native Canvas charts: today's category bar chart (study / exercise / daily / entertainment / reading) and a 7-day trend line chart — click a bar to see the day's tasks for that category. All tasks, category durations, 7-day history and image base64 data persist in localStorage with one-click export / import backup.\n\nBest for: students and exam-prep users (Pomodoro study companion), heavy Pomodoro users who want personalization, people who love custom experiences, indie devs and designers learning single-file Canvas application development.", "categories": [ "workspace" ], "tags": [ "学习时钟", "番茄钟", "角色陪伴", "个人工作台", "Canvas" ], "tags_en": [ "study clock", "pomodoro", "character companion", "personal workspace", "canvas" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "角色主题涂鸦学习时钟 单文件 HTML（内置 i18n 中英文，4 张角色表情图已内嵌）" } ], "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "prompt": "请使用资料库的能力帮我一键做同款（https://www.workbuddy.cn/space/d/j0wDA5AnL0J1cmUop05fbC），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n你是一个资深前端工程师兼 UI 设计师。请为我生成一个\"角色主题涂鸦学习时钟\" HTML 单文件。开始编码前，你必须先做一件事：\n\n第一步：向我索要素材\n• 请回复：\"请上传 2~4 张你的 Q 版角色素材图（透明底 PNG 最佳，不同表情/动作）。我会提取角色的主色调，让整个工作台的配色、字体和装饰都长成它的样子。\"\n• 收到图片后，你把图片转成 Base64 内嵌，继续生成代码。\n\n==============================\n生成代码时，严格遵守以下要求：\n==============================\n\n【视觉美术规范（强制落地）】\n整体画风：复古低饱和度暖米色扁平卡通画风，简约细棕色轮廓描边 2D 插画风格，柔和莫兰迪配色，整体带极轻微纸张噪点肌理，无写实光影、无 3D 渲染、无高光反光。所有卡片、按钮、弹窗全部使用大圆角矩形设计，组件分层依靠柔和低饱和软阴影区分层级；配色基调参考样例：茶绿渐变背景、奶米白面板、暖棕系文字，低对比度视觉，不刺眼。字体选用圆润无衬线游戏黑体，排版宽松舒适，休闲小游戏治愈质感。禁止高饱和亮色、霓虹、厚重粗描边、金属光泽效果。\n\n【核心能力：角色深度融入整套 UI，而不是单纯贴图】\n1. 图片上传后，用读取第一张角色图，提取图像的主色调（取出现频率最高的颜色，忽略纯黑/纯白/透明背景）。\n2. 基于主色调自动生成一套配色方案：\n ◦ 主色（primary）：直接使用提取到的颜色。\n ◦ 辅色（secondary）：主色明亮化或柔和变体（HSL 调整）。\n ◦ 背景色：适配 UI 低饱和风格，使用极浅对应色系渐变底色（参考茶青低饱和渐变）。\n ◦ 强调色：用于按钮、边框、高亮，暖棕调描边基底。\n3. 使用 CSS 变量动态应用这套配色，整个页面的边框、文字、按钮、装饰、图表线条、阴影颜色全部使用这些角色色系，不再是纯黑白，边框统一为细描边风格，贴合样例 UI 轮廓质感。\n4. 字体风格倾向角色：\n ◦ 如果主色调偏暖、明亮或粉色系 → 使用圆体手写圆润风格 'Comic Sans MS', 'Chalkboard SE', 'KaiTi', cursive，匹配休闲游戏卡通感。\n ◦ 如果主色调偏冷、暗或蓝紫系 → 使用清爽的 'Segoe UI', 'PingFang SC', sans-serif 并增加字距。\n5. 角色形象全域融入细节：\n ◦ 页面顶部：角色以轻松姿态悬浮，较大尺寸，作为\"陪伴者\"，带有轻微浮动动画，外框采用参考风格细圆角描边。\n ◦ 统计板块、两张图表标题边角：放置迷你角色头像小贴片，作为装饰标识。\n ◦ 每个任务卡片左侧边缘：放置一张小尺寸角色头像，作为任务守护标记，带低饱和度细描边，贴合整体 UI 轮廓风格。\n ◦ 添加任务按钮旁边：一张蹲坐或招手的小角色图，按钮控件大圆角，复刻样例按钮样式。\n ◦ 计时进行中：角色切换为\"专注\"表情的图，并带有呼吸缩放动效。\n ◦ 完成任务时：角色切换为\"欢呼\"表情，弹跳动画 + 周围爆出小星星/火花涂鸦（用 ✦ ⚡ ✎ 等符号），同时对话气泡弹出；气泡采用大圆角卡片样式，匹配整套 UI。\n ◦ 柱状图、折线图无数据空白状态：展示角色发呆插画，提示\"快去完成今日任务吧！\"，不再使用枯燥文字。\n6. 所有角色图片的处理：\n ◦ 给图片加一层极细实线圆角边框（风格对标 UI 控件描边），颜色使用角色辅色。\n ◦ 图片周围偶尔出现手绘小圈、波浪线装饰（用 CSS 伪元素 + 角色色系）。\n\n【新增｜任务分类识别 + 每日时长柱状图 + 7 天历史折线图】\n1. 任务属性分类：新建任务时下拉选择任务类型：学习、运动、日常琐事、娱乐、阅读；允许用户自定义新增分类。每一类绑定一套基于角色主色调衍生出来的区分色（低饱和，贴合整体画风，不突兀）。\n2. 智能统计：自动归集每一类任务当日累计耗时。\n3. 【当日分类柱状图】内置手绘风柱状图（不引入 ECharts 等外部图表库，原生 Canvas 绘制，匹配整体细描边卡通 UI）：\n ◦ X 轴：任务分类（学习 / 运动 / 日常 / 娱乐 …）\n ◦ Y 轴：当日累计耗时（分钟）\n ◦ 柱子样式：大圆角柱体、细描边，低饱和填充，完全匹配整体 UI 风格\n ◦ 点击柱子：弹出该分类下今日完成任务清单小弹窗\n ◦ 当日暂无数据：显示角色发呆占位插画 + 提示文案\n4. 【7 日趋势折线图】同样原生 Canvas 手写绘制，独立圆角卡片：\n ◦ X 轴：最近 7 天日期\n ◦ Y 轴：每日全部任务总专注时长\n ◦ 线条：圆润手绘折线、圆点节点、细描边，低饱和配色，和整体画风统一\n ◦ 无数据时同样放置角色放空占位画面\n5. 两张图表放置顺序：今日统计卡片下方 → 当日分类时长柱状图卡片 → 七日趋势折线图卡片，之后再到任务列表。\n6. 所有分类耗时、每日历史数据存入 localStorage 持久保存，每天 0 点自动切分当日统计，保留最近 7 天历史数据，超过 7 天自动丢弃旧数据。\n\n【基础功能要求】\n• 任务管理：添加学习任务，默认 25 分钟番茄钟，可下拉选 15/25/45 分钟或自定义输入。计时使用 Date.now() 防后台错乱，支持暂停/继续。\n• 手动完成：直接标记完成，不计时长，但照样标记分类、计入时长统计、触发完成动画和角色切换。\n• 完成激励：随机话术气泡（≥10 条，与角色气质匹配）+ 涂鸦粒子 + 形象弹跳。\n• 今日总览统计：累计专注总时长 + 完成任务数 + 各类别耗时概览。\n• localStorage 持久化所有数据（任务、任务分类耗时、形象 Base64 数组、每日历史 7 天数据、今日统计、日期）。\n\n【UI：角色主题·卡通风格，手机界面完美适配】\n• 基础元素：大圆角控件、细统一轮廓描边，页面底层添加极低透明度纸张噪点，背景可柔和渐变。全部颜色都来自角色配色方案。\n• 卡片、按钮使用主色/辅色细描边边框，背景半透低饱和底色，hover 时填入极淡的主色；按钮样式复刻参考 UI：大圆角，内边距宽松，游戏休闲控件质感。\n• 最大宽度 480px，居中，手机满宽。严格禁止横向溢出：全部输入框、文本、图表 max-width:100%; box-sizing:border-box; word-break:break-word;。\n• 页面从上到下布局：顶部悬浮角色区 → 今日统计卡片 → 每日分类时长柱状图卡片 → 7 日趋势折线图卡片 → 任务列表（进行中/已完成） → 底部吸底添加任务栏。\n• 对话气泡：使用角色色系细描边大圆角卡片，白色背景，三角箭头，弹出弹性动画，2.5 秒消失。\n\n【技术约束】\n• 纯 HTML/CSS/JS，无外部 CDN、无第三方图表库。所有角色图 Base64 内嵌。\n• 主色提取必须用 Canvas 实现，并处理透明像素。柱状图、折线图全部使用原生 Canvas 手写绘制，保持整套画风统一。\n• 计时器基于 Date.now() 差值，setInterval 每秒刷新。\n• localStorage 保存近 7 天每日记录。\n• 代码注释清晰，结构干净，模块拆分：计时模块、分类统计模块、双图表绘图模块、存储模块、动画模块。\n\n【文案参考】\n• 标题：角色名占位（可由用户修改，或根据主色冷暖显示不同 emoji）+ \"的学习时钟\"，标题区域可增加复古标签样式，对标样图左上角标签设计。\n• 输入框占位：\"和 [角色] 一起，画下新任务…\"\n• 统计：\"今日专注 X 小时 X 分钟 | 打卡 X 次\"\n• 柱状图卡片标题：今日时间分配\n• 折线图卡片标题：近 7 天专注趋势\n• 激励语库示例：\"[角色] 为你撒花！\"、\"专注力满分，像 [角色] 一样闪闪发光！\"、\"叮！又成长了一点~\"等，内置不少于 10 条。\n\n【响应流程】\n1. 先向我要角色素材。\n2. 收到图片后，输出完整可运行的 HTML 文件。\n3. 输出不联网、无外部资源。\n\n【交付要求】\n请在我的个人资料库中复刻做同款并发布：将最终 HTML 保存到我的个人资料库，并发布为公开链接后返回给我。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/j0wDA5AnL0J1cmUop05fbC, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: You are a senior frontend engineer and UI designer. Generate a \"Character-themed Doodle Study Clock\" as a . Before you start coding, you MUST do this first: Step 1: Ask for the assets\n• Reply with: \"Please upload 2–4 Q-version character images (transparent PNG preferred, different expressions / poses). I will extract the character's dominant color so the whole workspace — colors, fonts and decorations — matches it.\"\n• Once you receive the images, convert them to Base64 and embed them, then continue generating the code. ==============================\nWhen generating the code, strictly follow these requirements:\n============================== [Visual art specifications — must land]\nOverall style: retro low-saturation warm beige flat cartoon, simple thin brown outline 2D illustration, soft Morandi palette, slight paper-noise texture. No realistic lighting, no 3D, no specular highlights. All cards, buttons, modals use big rounded rectangles; layering uses soft low-saturation shadows. Palette reference: tea-green gradient background, cream-white panels, warm brown text — low contrast, easy on the eyes. Use a rounded sans-serif \"game black\" font, generous spacing, casual game healing feel. Forbidden: high-saturation bright colors, neon, thick outlines, metallic shine. [Core ability: the character fully merges into the whole UI, not just a sticker]\n1. After upload, read the first character image and extract its dominant color (highest-frequency color, ignoring pure black / pure white / transparent background).\n2. Based on the dominant color, auto-generate a palette: ◦ Primary: use the extracted color directly. ◦ Secondary: a brighter or softer variant of the primary (HSL tweak). ◦ Background: very light tint of the corresponding hue, gradient (e.g. tea-cyan low-saturation gradient). ◦ Accent: warm-brown stroke baseline for buttons / borders / highlights.\n3. Apply the palette via CSS variables — borders, text, buttons, decorations, chart lines, shadows all use the character palette, never plain black/white; borders unified as thin outlines to match the sample UI.\n4. Font flavor follows the character: ◦ Warm / bright / pink dominant → use a rounded handwriting-style stack: 'Comic Sans MS', 'Chalkboard SE', 'KaiTi', cursive — matches casual cartoon game vibe. ◦ Cool / dark / blue-violet dominant → use a clean stack: 'Segoe UI', 'PingFang SC', sans-serif with extra letter-spacing.\n5. Character fully integrated into details: ◦ Page top: character floats in a relaxed pose at large size as a \"companion\" with subtle floating animation, framed with a thin rounded outline. ◦ Stats panel, two chart titles' corners: mini character avatar stickers as decorative markers. ◦ Each task card's left edge: a small character avatar as a task guardian badge, with low-saturation thin outline. ◦ Next to the add-task button: a sitting/waving small character; the button is a big rounded game-style control. ◦ While the timer is running: the character switches to the \"focus\" image and breathes (scale animation). ◦ On task completion: the character switches to \"celebrate\" — bouncing animation + doodle particles (use ✦ ⚡ ✎ etc.) + a speech bubble pops up; bubble uses big rounded card style matching the UI. ◦ Bar chart / line chart empty state: show the character's \"spacing out\" art with the prompt \"Go finish today's tasks!\" — no dry text.\n6. Image handling: ◦ Wrap each image in a thin solid rounded outline (matching UI controls), color uses the character secondary. ◦ Around the image, occasional hand-drawn circles / wavy decorations via CSS pseudo-elements using the character palette. [New — task category classification + daily duration bar chart + 7-day history line chart]\n1. Task categories: when creating a task, a dropdown picks category — study, exercise, daily chores, entertainment, reading; user can add custom categories. Each category gets a low-saturation distinguishing color derived from the character palette.\n2. Smart stats: auto-aggregate each category's total duration for the day.\n3. [Daily category bar chart] built-in hand-drawn bar chart (no ECharts or external chart libs — pure native Canvas, matching the thin-outline cartoon UI): ◦ X-axis: task category (study / exercise / daily / entertainment / …) ◦ Y-axis: total duration for the day (minutes) ◦ Bar style: big rounded bars, thin outline, low-saturation fill, fully consistent with the overall UI style ◦ Click a bar: popup listing the day's completed tasks in that category ◦ No data today: show the spacing-out character placeholder + prompt\n4. [7-day trend line chart] also native Canvas hand-drawn, in its own rounded card: ◦ X-axis: last 7 days ◦ Y-axis: total daily focus duration ◦ Line: rounded hand-drawn polyline, dot nodes, thin outline, low-saturation color matching the overall style ◦ No data: same character placeholder\n5. Chart placement order: today's stats card → daily category bar chart card → 7-day trend line chart card → then task list.\n6. All category durations and daily history persist in localStorage; auto-split at 0:00 each day, keep last 7 days, drop older data. [Core features]\n• Task management: add study tasks, default 25-minute Pomodoro, dropdown 15/25/45/custom minutes. Timer uses Date.now() to survive background throttling, supports pause / resume.\n• Manual complete: mark complete without timing, but still records category, counts toward stats, triggers the completion animation and character switch.\n• Completion reward: random encouragement bubble (≥10 lines, matching the character vibe) + doodle particles + character bounce.\n• Today overview stats: cumulative focus duration + completed task count + per-category breakdown.\n• localStorage persists everything (tasks, category durations, image Base64 array, 7-day history, today's stats, date). [UI — character-themed cartoon, mobile-perfect]\n• Basics: big rounded controls, unified thin outlines, page-level very-low-opacity paper-noise texture, soft gradient background. All colors come from the character palette.\n• Cards / buttons use primary/secondary thin outlines, semi-transparent low-saturation fill, hover fills with very faint primary. Buttons copy the reference UI: big rounded, generous padding, casual game control feel.\n• Max-width 480px, centered, full-width on mobile. Strictly no horizontal overflow: all inputs/text/charts use max-width:100%; box-sizing:border-box; word-break:break-word;.\n• Page layout top-to-bottom: top floating character → today's stats card → daily category bar chart card → 7-day trend line chart card → task list (active / completed) → bottom sticky add-task bar.\n• Speech bubble: character-palette thin-outline big rounded card, white background, triangle pointer, pop-in elastic animation, disappears after 2.5s. [Technical constraints]\n• Pure HTML/CSS/JS, no external CDN, no third-party chart libs. All character images embedded as Base64.\n• Dominant color extraction MUST use Canvas and handle transparent pixels. Bar chart and line chart MUST be hand-drawn in native Canvas to keep the style unified.\n• Timer based on Date.now() delta; setInterval refreshes every second.\n• localStorage holds the last 7 days.\n• Clear code comments, clean structure — modules: timer, category stats, dual-chart drawing, storage, animation. [Copy references]\n• Title: character name placeholder (user-editable, or an emoji chosen by warm/cool dominant color) + \"'s Study Clock\"; title area may include a retro tag style matching the sample top-left label.\n• Input placeholder: \"Draw a new task with [Character]…\"\n• Stats: \"Today's focus X hr X min | X check-ins\"\n• Bar chart card title: Today's Time Distribution\n• Line chart card title: 7-Day Focus Trend\n• Reward lines (≥10): e.g. \"[Character] sprinkles flowers for you!\", \"Focus maxed — shine like [Character]!\", \"Ding! You grew again~\" etc. [Response flow]\n1. Ask for character assets first.\n2. Once received, output the complete runnable HTML file.\n3. Deliver a with no network and no external resources. [Delivery requirement]\nPlease replicate the same workbench in my personal Library and publish it: save the final HTML to my personal WorkBuddy Library, publish it as a public link, and return the link to me.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "汤匙不是钥匙", "name_en": "汤匙不是钥匙", "avatar": "", "badge": "community" }, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-08-13T19:51:08+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "charity-moments-fundraising", "title": "朋友圈劝募文案一键生成", "title_en": "WeChat Moments Fundraising Copy", "subtitle": "300 字折叠预览里，同时塞下钩子、事实和行动引导", "subtitle_en": "Hook, verified facts, and a call-to-action — all inside the 300-character Moments preview", "description": "公益传播人写劝募稿常陷两难：手动查后台数据费时，AI 起草又容易带模板化口吻和未核验的数字；朋友圈 300 字折叠预览内，还要同时塞下钩子、受益对象、可信事实和行动引导。\n\n把「为 XXX 项目写一段朋友圈文案」这句话交给「公益推文专家」，它先按朋友圈折叠预览规范写一版方向稿，再自动读取机构后台的项目数据（累计筹款、覆盖省份与学校数、受益儿童数、机构资质、公开故事等），逐项核对后把可信事实写进正文。\n\n适合：公益机构内容运营、传播负责人，以及需要在朋友圈做高频劝募 / 项目传播的公益团队。", "description_en": "Writing fundraising copy for WeChat Moments is a squeeze: manually pulling backend data is slow, while AI drafts tend to sound generic or cite unverified numbers — and it all has to fit inside a 300-character folded preview along with a hook, beneficiary, credible facts, and a call-to-action.\n\nHand \"write a Moments post for the XXX project\" to the Public Welfare Article Expert: it first drafts against Moments' preview constraints, then pulls verified backend data (cumulative fundraising, province and school coverage, beneficiary count, organizational credentials, public stories) and folds those verified facts into the copy.\n\nFor: nonprofit content and communications leads, and any team running frequent Moments-based fundraising or project promotion.", "categories": [ "creation", "business" ], "tags": [ "公益", "朋友圈", "劝募文案", "短文案" ], "tags_en": [ "Public Welfare", "WeChat Moments", "Fundraising Copy", "Short-Form" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请你为本机构在腾讯公益的项目{项目名称}写一段劝募文案，我要发在朋友圈。", "prompt_en": "Please write a fundraising post for our organization's project {project name} on Tencent Public Welfare — I'll be sharing it on WeChat Moments.", "skills": [], "experts": [ { "id": "ex_ogQdtCq5YVhj", "name": "公益活动推文撰写专家", "description": "读取机构后台真实项目数据，一键生成公众号 / 小红书 / 朋友圈的公益传播内容", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ba7dbef889a1", "expert_id": "ex_ogQdtCq5YVhj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "净水计划朋友圈劝募文案" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "charity-wechat-article", "title": "公益公众号推文一键生成", "title_en": "One-Click WeChat Article for Nonprofits", "subtitle": "从后台真实数据拉出温暖可信的公众号叙事", "subtitle_en": "Warm, trustworthy WeChat storytelling drawn straight from backend data", "description": "公益机构在产出基于真实数据的公众号推文时，常常被散落在后台各处的筹款、进展、故事数据困扰——人工整理常常要花大半天，还总怕把数字写错、把故事写偏。\n\n让「公益推文专家」接住这个活：它直接拉取机构后台最近半年的项目进展、累计筹款与捐款人次、项目背景和受益人公开素材，按温暖可信的公众号叙事重写一版数据稿，关键事实逐条留痕可核验。\n\n适合：公益机构内容运营、传播负责人，以及基金会、高校公益团队的日常推文场景。", "description_en": "Writing a WeChat article that's grounded in real numbers means digging through scattered fundraising totals, project progress, and beneficiary stories in the backend — hours of copy-paste, plus the constant risk of misquoting a figure or missing the emotional beat.\n\nLet the Public Welfare Article Expert take it: pulls the last six months of project progress, cumulative fundraising and donor counts, project context, and publicly available beneficiary photos from your backend, then rewrites a warm, trustworthy WeChat narrative. Every key fact is traceable back to the source.\n\nFor: content and communications leads at nonprofits, and everyday storytelling teams at foundations or campus public-welfare groups.", "categories": [ "creation", "business" ], "tags": [ "公益", "公众号", "推文写作", "内容运营" ], "tags_en": [ "Public Welfare", "WeChat", "Article Writing", "Content Operations" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请你参考本机构在腾讯公益的项目{项目名称}最近半年的执行进展，撰写一篇公众号推文，文风温暖，需要突出受益人的真实变化。", "prompt_en": "Using our organization's last six months of execution progress for the project {project name} on Tencent Public Welfare, please write a WeChat public account article — warm in tone, and highlighting real changes for the beneficiaries.", "skills": [], "experts": [ { "id": "ex_ogQdtCq5YVhj", "name": "公益活动推文撰写专家", "description": "读取机构后台真实项目数据，一键生成公众号 / 小红书 / 朋友圈的公益传播内容", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ba7dbef889a1", "expert_id": "ex_ogQdtCq5YVhj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "壹基金净水计划公众号推文" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "charity-xiaohongshu-post", "title": "公益小红书运营帖生成", "title_en": "Xiaohongshu Post for Public Welfare Projects", "subtitle": "有料不会发？把项目进展写成小红书轻分享文风", "subtitle_en": "Sitting on great material? Turn project progress into shareable Xiaohongshu-style notes", "description": "公益机构想把项目近况搬上小红书时，常常卡在「有料不会发」——进展、筹款、故事素材散落在后台各处，拼成一篇短句轻快的种草文要反复搬运大半天，还总怕数字写错、语气写成硬广。\n\n让「公益推文专家」接住这个活：它直接拉取机构后台最近的项目进展、累计筹款与捐款人次、项目背景和公开故事图片，按小红书「场景钩子＋短句换行＋emoji 锚点＋话题标签」的轻分享风格生成一版可直接编辑的图文稿，项目图片自动本地化替换，关键数字逐条留痕可核验。\n\n适合：公益机构内容运营、传播负责人，以及基金会、高校公益团队的高频种草与项目推广场景。", "description_en": "Nonprofits want to post project updates on Xiaohongshu but get stuck: material is scattered across the backend, and turning it into short, breezy notes with the right vibe eats half a day — with the constant risk of misquoting numbers or slipping into hard-sell tone.\n\nLet the Public Welfare Article Expert handle it: pulls recent project progress, cumulative fundraising, donor counts, project context, and public story photos from your backend, then generates an editable Xiaohongshu-style post with scene hooks, short-line rhythm, emoji anchors, and hashtags. Project images are localized automatically; every key figure remains traceable.\n\nFor: content and communications leads at nonprofits, and high-frequency promotion teams at foundations or campus public-welfare groups.", "categories": [ "creation", "business" ], "tags": [ "公益", "小红书", "种草文案", "内容运营" ], "tags_en": [ "Public Welfare", "Xiaohongshu", "Lifestyle Copy", "Content Operations" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请你为本机构本在腾讯公益的项目{项目名称}，写一篇小红书推文，用于项目推广。", "prompt_en": "Please write a Xiaohongshu-style post promoting our organization's project {project name} on Tencent Public Welfare.", "skills": [], "experts": [ { "id": "ex_ogQdtCq5YVhj", "name": "公益活动推文撰写专家", "description": "读取机构后台真实项目数据，一键生成公众号 / 小红书 / 朋友圈的公益传播内容", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ba7dbef889a1", "expert_id": "ex_ogQdtCq5YVhj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "净水计划小红书运营帖" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "chilema-nin-workbench", "artifact_type": "html", "title": "吃了吗您 · 干饭盲盒工作台", "title_en": "Chi Le Ma Nin · Meal Blind-Box Workbench", "subtitle": "一日三餐抽盲盒，命运帮你决定吃什么，饭票额度管住干饭花销", "subtitle_en": "Turn every meal into a blind-box draw — let fate pick your food while a meal budget keeps spending in check", "description": "每天最难的不是挣钱，是决定\"今天吃啥\"——外卖软件翻半小时还定不下来。这个工作台把一日三餐变成\"抽盲盒\"：设定好饭票额度（默认 ¥50）后，点早餐/中餐/晚餐/宵夜卡片随机抽菜，抽到心仪的直接\"吃吃吃\"记录，不满意可以重抽；抽到超出饭票的贵菜还会弹窗问\"透支额度也要吃吗\"。菜品池页点\"嘴又馋了？加道菜\"就能添加自己的菜并选稀有度（普通 70%/稀有 25%/限定 5%），自己加的菜同样会被抽到；记录页可按日回看吃了什么、连续打卡与花销情况，还能生成可保存的 PNG 分享卡片，吃饭数据按日期归档，落在资料库数据表，支持在线存储与双向同步。适合：选择困难症的干饭人、想给三餐加点乐子的上班族和学生党、记录饮食习惯的生活记录爱好者。", "description_en": "The hardest part of every day is not earning money but deciding what to eat — scrolling takeout apps for half an hour and still undecided. This workbench turns three meals a day into blind-box draws: set your meal budget (default ¥50), tap the breakfast/lunch/dinner/late-night card to draw a random dish, tap \"eat it\" to record one you love, or redraw if you don't like it; if you draw a pricey dish over budget, a popup asks whether to overdraft the budget anyway. In the Dish Pool page, tap the \"craving something? add a dish\" bar to add your own dishes with a rarity tier (common 70% / rare 25% / limited 5%) — your own dishes join the draw too; the Records page shows per-day history, check-in streaks and spending, and can generate a saveable PNG share card. All eating records are archived by date, stored in Library data tables with online storage and two-way sync. Great for indecisive foodies, office workers and students who want more fun at mealtimes, and anyone who loves tracking daily-life habits.", "tags": [ "干饭盲盒", "选择困难", "一日三餐", "饭票额度", "连续打卡" ], "tags_en": [ "meal blind box", "decision fatigue", "three meals a day", "meal budget", "check-in streak" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/KfgpS3ldDwiAc9BkbAbEvf，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「吃了吗您」干饭盲盒工作台，专治\"今天吃啥\"选择困难，暖色卡通风、移动端优先、PC 自适应。用两张表管住数据：「菜品池」表（菜名、适合餐次（选择：早餐/中餐/晚餐/宵夜）、口味（选择：甜/咸/辣/酸/苦/鲜/香/淡/臭）、大卡、预估价格（元）、稀有度（选择：普通/稀有/限定）、一句话简介、来源（选择：默认池/自己加））；「吃饭记录」表（日期（日期）、餐次（选择：早餐/中餐/晚餐/宵夜）、菜名、口味、大卡、价格、稀有度、状态（选择：已吃/丢弃）、是否透支）。页面包含：主页（饭票余额与额度展示，点早餐/中餐/晚餐/宵夜卡片随机抽取该餐菜品，未抽显示\"未抽\"标签；抽到心仪的记录\"吃吃吃\"，不满意可重抽；抽到超出剩余饭票的贵菜时弹窗询问\"透支额度也要吃吗\"，确认后记录本次透支并允许余额为负）；菜品池页（展示全部菜品与抽中概率说明，点击\"嘴又馋了？加道菜\"引导条添加自定义菜品并自选稀有度，自定义菜同样进入抽取池）；记录页（按日查看吃了什么与当日合计大卡，连续打卡天数与累计抽卡次数统计，花销一目了然，可生成可保存的 PNG 分享卡片）。每次抽卡与吃下自动写入吃饭记录并同步扣减饭票余额，余额可在主页直接充值或重设额度。内置 29 道演示菜品（家常菜+外卖常见菜+宵夜饮品，含稀有度分级与手绘风 SVG 图标）。\n\n交付：资料库在线工作台——在线 HTML 页面 + 2 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/KfgpS3ldDwiAc9BkbAbEvf with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Chi Le Ma Nin\" meal blind-box workbench that cures the daily \"what to eat\" decision fatigue, in a warm cartoon style, mobile-first and PC responsive. Manage data with two tables: a \"Dish Pool\" table (dish name, meal type (select: breakfast / lunch / dinner / late-night), taste (select: sweet / salty / spicy / sour / bitter / umami / fragrant / mild / stinky), kcal, estimated price (CNY), rarity (select: common / rare / limited), one-line intro, origin (select: default pool / custom)); and an \"Eating Records\" table (date (date), meal type (select: breakfast / lunch / dinner / late-night), dish name, taste, kcal, price, rarity, status (select: eaten / tossed), overdraft flag). The page includes: Home (meal-ticket balance and budget display; tap the breakfast/lunch/dinner/late-night card to draw a random dish for that meal, with a \"not drawn\" tag when undecided; tap \"eat it\" to record a dish you love, or redraw if you don't like it; when you draw a pricey dish over the remaining balance, a popup asks \"overdraft the budget to eat it anyway?\" — once confirmed, the overdraft is recorded and a negative balance is allowed); a Dish Pool page (browse all dishes with draw-odds explanation, tap the \"craving something? add a dish\" bar to add custom dishes with a self-chosen rarity — custom dishes join the draw pool too); and a Records page (per-day view of what was eaten with the day's total kcal, check-in streak and total draw counts, spending at a glance, and a saveable PNG share card). Every draw and eat writes to the Eating Records table and deducts the meal-ticket balance accordingly; the balance can be topped up or reset right on the home page. Ships with 29 demo dishes (home-style dishes + common takeout + late-night drinks, with rarity tiers and hand-drawn SVG icons).\n\nDeliverable: an online Library workbench — an online HTML page plus 2 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "吃了吗您 · 干饭盲盒工作台（自包含单文件）" } ], "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "skill_id": null, "creator": { "name": "Yt", "name_en": "Yt", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "created_at": "2026-09-10T11:00:00+08:00", "updated_at": "2026-09-10T18:12:52+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "chili-oil-promo-page", "title": "蒋姐油辣子宣传页", "title_en": "Chili Oil Promo Page", "subtitle": "三个毛笔大字压在红油罐上，选辣度时罐里的油色跟着变深", "subtitle_en": "Three brush characters sit on a jar of red oil that darkens as you pick the heat", "description": "油辣子最怕被写成“香辣下饭、传统工艺”，买家无从判断自家吃得了多辣、锅里到底放了什么。\n\n这个案例给虚构的遵义家庭作坊「蒋姐油辣子」做了一张手机优先的宣传页：首屏是一罐红油，“油辣子”三个毛笔大字压在上面，滚动时被推散；接着让人选微辣、中辣、特辣，罐里的油色变深、辣椒图标点亮、碎辣椒落下；再用一张配料小票写清一锅放了什么、没放什么，最后是店主自述和 11.11 两罐的价格。页面同时有中文和英文两个版本。\n\n适合做食品、调味品的小微商家，也适合想参考“用一个可操作的小互动说明商品差别”的设计者。", "description_en": "Chili oil pages are easily reduced to spicy, great with rice and traditional craft, leaving buyers unable to judge how hot they can take or what is really in the pot.\n\nThis case builds a mobile-first promo page for a fictional home workshop in Zunyi. The first screen is a jar of red oil with three big brush characters sitting on it, pushed apart as the visitor scrolls. Visitors then choose mild, medium or extra hot, and the oil in the jar darkens, chili icons light up and crushed chili falls. An ingredient receipt shows what goes in a batch and what is left out, followed by the owner's note and the 11.11 two-jar price. The page comes in Chinese and English.\n\nUseful for small food and condiment merchants, and for designers exploring a small interaction that explains how a product differs.", "categories": [ "creation", "business" ], "tags": [ "宣传页", "调味品", "食品", "交互设计" ], "tags_en": [ "Promo Page", "Condiment", "Food", "Interaction Design" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "单页 HTML 宣传页" } ], "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "prompt": "为贵州遵义的家庭作坊「蒋姐油辣子」做一张手机优先的商品宣传页，用于 11.11 前把人引到店铺。卖点是每周三只熬一锅、一锅限 120 罐、熬完装罐就寄；11.11 两罐（每罐 260 克）¥68，平时一罐 ¥39、两罐 ¥78。\n\n首屏用整幅深红底的一罐红油照片，“油辣子”三个毛笔大字压在罐子上方，滚动时三个字被推向不同方向，罐子放大；左下一张价签“两罐 260 克 ¥68”，右侧一句“每周三熬一锅，熬完装罐就寄”。\n\n往下依次是：“你能吃多辣？”——微辣、中辣、特辣三个按钮，点选后罐里的油色变深、辣椒图标点亮对应个数、碎辣椒落下，并显示每档适合怎么吃；“一锅里放了什么”——一张微微倾斜的配料小票（遵义朝天椒 3.0 斤、花溪辣椒 1.5 斤、压榨菜籽油 4.0 斤、花生米 1.0 斤、白芝麻和花椒 0.6 斤，没放味精、防腐剂、香精）；“一周只做一锅”——周三熬、周四放凉装罐、周五顺丰发出；店主蒋丽的几句自述和一个点击就复制微信号的按钮；最后是 11.11 两罐的大号价格和下单时间说明，页尾写商家信息。\n\n风格：深红、奶油白、辣椒金黄，毛笔字做标题，数字用窄体。商品图为示意，正式发布前用商家实拍替换；页面只做展示和引流，下单在店铺完成。商家、数字均为演示，不编造真实销量、评价或功效。\n\n交付：一个开箱即用的单页 HTML 宣传页，手机和电脑都能打开，页面里没有需要保存的数据。", "prompt_en": "Create a mobile-first product promo page for Sister Jiang's Chili Oil, a home workshop in Zunyi, Guizhou, to bring people to the shop before 11.11. The selling points: one batch a week on Wednesday, 120 jars a batch, shipped as soon as it is jarred. For 11.11, two 260 g jars cost ¥68; usually ¥39 a jar or ¥78 for two.\n\nThe first screen is a full-bleed deep-red photo of a jar of red oil with three big brush characters for “chili oil” sitting above the jar, pushed in different directions and the jar enlarging as the visitor scrolls. A price tag at the lower left reads “Two 260 g jars ¥68”, and a line on the right says one batch every Wednesday, shipped as soon as it is jarred.\n\nNext, in order: “How hot can you take it?” with mild, medium and extra-hot buttons, where choosing one darkens the oil in the jar, lights the matching number of chili icons, drops crushed chili and shows how to eat that level; “What goes into a batch” as a slightly tilted ingredient receipt (Zunyi bird's-eye chilies 3.0 jin, Huaxi chilies 1.5 jin, pressed rapeseed oil 4.0 jin, peanuts 1.0 jin, white sesame and Sichuan pepper 0.6 jin, with MSG, preservatives and flavourings left out); “One batch a week” with Wednesday simmering, Thursday cooling and jarring, Friday shipping by SF Express; a few words from the owner, Jiang Li, with a button that copies her WeChat ID when tapped; and finally the large 11.11 two-jar price with ordering dates, and shop details in the footer.\n\nStyle: deep red, cream white and chili gold, brush lettering for headings and a condensed face for numbers. Product photos are illustrative and should be replaced with the merchant's own before publishing. The page only showcases and drives traffic; orders are placed in the shop. Shop and figures are demos; never invent real sales, reviews or efficacy claims.\n\nDeliverable: a ready-to-use single-page HTML promo page that opens on phones and desktops; the page keeps no data of its own.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 89, "created_at": "2026-10-15T15:30:00+08:00", "updated_at": "2026-10-15T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "china-internet-etf-sentiment-watch", "title": "中概互联网ETF持仓舆情观察", "title_en": "China Internet ETF Holding Sentiment Watch", "subtitle": "基于真实持仓和新闻，观察 513050 的重仓叙事如何向 ETF 传导", "subtitle_en": "Use real holdings and news to watch how 513050’s heavyweight narratives transmit into the ETF", "description": "看中概互联网ETF，最麻烦的往往不是基金名字，而是它背后那一篮子重仓股每天都在讲不同故事：AI、办公、汽车、供应链、本地生活，信息很多，却很难快速看清哪些新闻更可能传导到ETF层面。\n\n这个案例用真实持仓权重和近 7 天重点新闻，把 513050 做成一页式持仓舆情观察：先看一句话结论、前五大重仓集中度和编辑性温度条，再看时间线、重点公司卡片和风险提示，只保留对ETF可能有传导意义的线索与后续观察点。\n\n适合：关注中概互联网ETF波动来源的投资者；需要快速过一遍重仓新闻的人；想把“持仓 + 新闻 + 权重”整合成监控页的基金研究与内容用户。", "description_en": "What makes a China internet ETF difficult to track is not the fund name itself, but the fact that its heavyweight holdings are constantly telling different stories around AI, office software, autos, supply chains, and local services. There is plenty of information, but very little signal on which developments are most likely to transmit into the ETF.\n\nThis case turns 513050 into a one-page holding sentiment watch built from real holding weights and key news from the last seven days: it opens with a one-line verdict, top-five concentration, and an editorial temperature bar, then moves into a timeline, key company cards, and risk notes that focus only on transmission paths and watchpoints that matter to the ETF.\n\nBest for: investors who want to understand the drivers behind China internet ETF volatility; users who need a fast pass through heavyweight-holding news; and fund researchers or content creators who want to merge holdings, news, and weights into a monitoring page.", "tags": [ "基金分析", "ETF", "中概互联网", "持仓舆情", "新闻观察" ], "tags_en": [ "fund analysis", "ETF", "China internet", "holding sentiment", "news watch" ], "categories": [ "finance" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "中概互联网ETF持仓舆情观察" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "基于真实持仓和新闻做中概互联网ETF舆情观察，围绕 513050 输出一句话结论、前五大重仓集中度、编辑性舆情温度、近 7 天时间线、重点公司观察卡片与风险提示，只写可能影响路径和后续观察点。", "prompt_en": "Create a China internet ETF sentiment watch based on real holdings and news for 513050, centered on a one-line conclusion, top-five concentration, editorial sentiment temperature, a last-7-day timeline, key company watch cards, and risk notes that stay focused on possible paths and watchpoints.", "skills": [ { "id": "fund-analysis", "name": "基金分析", "description": "公募基金/ETF全面分析，支持持仓组合诊断和单只基金深度分析。", "skill_id": "skill_2076627911660621824" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T16:12:00+08:00", "updated_at": "2026-07-28T16:12:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "chinese-ceramics-5000y-history", "title": "中国五千年陶瓷发展史从新石器时代到现代瓷器", "title_en": "5000-Year History of Chinese Ceramics from Neolithic to Modern Porcelain", "subtitle": "彩陶→青瓷→青花→珐琅彩:中国陶瓷技术与艺术五千年演进的脉络与里程碑", "subtitle_en": "Painted pottery → celadon → underglaze blue → famille rose: the technical and artistic evolution of Chinese ceramics across five millennia", "description": "想一口气理清中国陶瓷从新石器时代到现代瓷器的整条脉络?朝代多、窑口多、釉色工艺多,做综述又怕漏掉关键转折(如东汉青瓷、唐三彩、北宋汝窑、元青花、清珐琅彩),也怕停留在「某某朝代某窑某色」的罗列上,看不出技术与艺术演进的内在逻辑。\n\n这个案例演示了如何借助腾讯探元文物数据库与知识库的并行检索,主代理亲自做 3-4 路网络调研(最新进展/典型案例/影响成效/历史脉络),覆盖原始瓷与彩陶、秦汉铅釉陶、唐三彩、宋五大名窑(汝/官/哥/钧/定)、元青花与釉里红、明清彩瓷与现代瓷器复兴等维度,最终交付包含 3-6 张配图与 1-2 张趋势图表的 Markdown / HTML / DOCX / PDF 四格式趋势综述报告。\n\n适合:陶瓷考古与艺术史研究者、博物馆策展与收藏鉴定人员、文物与博物馆学专业研究生、非物质文化遗产保护与陶瓷技艺传承研究者、文化创意与陶瓷文创产品开发从业者、艺术院校陶瓷与设计专业师生。", "description_en": "Want a single read that maps China's ceramics from the Neolithic through modern porcelain? With so many dynasties, kilns, glazes and techniques, surveys risk either missing critical inflections (Eastern Han celadon, Tang sancai, Song Five Great Wares, Yuan underglaze blue, Qing famille rose) or staying trapped in 'which dynasty, which kiln, which glaze' listings without surfacing the underlying technical-artistic logic.\n\nThis case shows how to leverage Tencent Tanyuan relic and knowledge bases for parallel retrieval, plus 3-4 concurrent main-agent WebSearch passes (latest progress / benchmark cases / impact / historical lines), covering proto-porcelain and painted pottery, Qin-Han lead-glazed ware, Tang sancai, Song Five Great Wares (Ru/Guan/Ge/Jun/Ding), Yuan underglaze blue and copper-red, Ming-Qing polychrome porcelain, and modern porcelain revival—delivering a 3-6 image + 1-2 trend-chart Markdown/HTML/DOCX/PDF four-format trend review report.\n\nBest for: ceramic archaeology and art-history researchers, curators and authentication professionals, museum and heritage graduate students, intangible heritage and ceramic-craft inheritors, ceramic cultural-product developers, art-and-design faculty and students.", "categories": [ "business" ], "tags": [ "中国陶瓷史", "新石器时代彩陶", "宋五大名窑", "青花瓷", "非遗传承" ], "tags_en": [ "Chinese Ceramics History", "Neolithic Painted Pottery", "Song Five Great Wares", "Underglaze Blue Porcelain", "Intangible Heritage" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含趋势图表)" } ], "prompt": "请按「腾讯探元学术研究-趋势综述」专家的工作流,综述「中国五千年陶瓷发展史,从新石器时代到现代瓷器」。\n\n研究范围(覆盖六段历史脉络):\n- 原始瓷与彩陶时期(约公元前 10000—前 2000):仰韶文化彩陶、马家窑文化彩陶、磁山文化早期陶\n- 商周至秦汉:原始青瓷、白陶、印纹硬陶、铅釉陶、东汉青瓷\n- 魏晋南北朝至隋唐:南北方青瓷、唐代邢窑/越窑/长沙窑、唐三彩、匣钵装烧技术\n- 宋代:汝/官/哥/钧/定五大名窑、磁州窑、耀州窑、建窑黑釉、景德镇青白瓷(影青)\n- 元代与明代:元青花、釉里红、景德镇崛起、明永乐/宣德青花、成化斗彩、嘉靖万历五彩\n- 清代与近代:康雍乾珐琅彩/粉彩/斗彩、清末浅绛彩、民国珠山八友、新中国国瓷复兴、改革开放后现代瓷器\n\n交付要求(强制):\n1. 探元两库 MCP(search_movable_relics + search_relics)并行检索 + 主代理 3-4 路 WebSearch 多源核验,优先权威博物馆官网、考古简报、学术期刊信源\n2. 配图 ≥ 3 张(建议 4-6 张,不同时代典型器物各一,优先 Wikimedia 公共领域 / 大馆官网 / 故宫博物院/上海博物馆公开图源),必须实图、与时代对应\n3. 趋势图表 ≥ 1 张(建议 2-3 张:中国陶瓷发展时间线 / 釉色与装饰技术演进桑基图 / 各时期代表器物与窑口矩阵图),matplotlib 渲染为 SVG,严格避免标签重叠\n4. 四格式报告:Markdown 母版 → HTML(嵌入 base64 图片) → DOCX → PDF,综述结构(领域背景→发展现状→典型案例→影响成效→趋势展望→结论→证据与局限→参考来源)\n5. 报告日期用真实当前日期;参考文献按时间倒序;只写结论不含建议;主要趋势判断同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Trend Review' expert workflow to survey the '5000-Year History of Chinese Ceramics from Neolithic to Modern Porcelain'.\n\nResearch scope (six historical periods):\n- Proto-porcelain and painted pottery period (c. 10,000—2,000 BCE): Yangshao painted pottery, Majiayao painted pottery, Cishan early pottery\n- Shang-Zhou to Qin-Han: proto-celadon, white pottery, stamped hard pottery, lead-glazed ware, Eastern Han celadon\n- Wei-Jin Northern-Southern Dynasties to Sui-Tang: north-south celadon, Tang Xing/Yue/Changsha kilns, Tang sancai, sagger firing\n- Song dynasty: Five Great Wares (Ru/Guan/Ge/Jun/Ding), Cizhou/Yao/Jian (black glaze), Jingdezhen Qingbai (shadowy blue)\n- Yuan and Ming: Yuan underglaze blue and copper red, Jingdezhen's rise, Ming Yongle/Xuande blue, Chenghua doucai, Jiajing/Wanli wucai\n- Qing and modern: Kangxi/Yongzheng/Qianlong enamel / famille rose / doucai, late-Qing qianjiangcai, Republican Zhushan Eight Friends, modern national porcelain revival and post-1978 contemporary porcelain\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 3-4 main-agent WebSearch passes for cross-source verification, prioritizing authoritative museum official sites, archaeological bulletins, and academic journal sources\n2. ≥ 3 images (4-6 recommended, one representative artefact per period, prioritizing Wikimedia public domain / major-museum / Palace Museum / Shanghai Museum public sources), each verified real and period-relevant\n3. ≥ 1 trend chart (2-3 recommended: ceramic history timeline / glaze-and-decoration evolution sankey / period-kiln-matrix), rendered to SVG via matplotlib with strict label overlap avoidance\n4. Four formats: Markdown master → HTML (base64-inlined images) → DOCX → PDF, trend-review structure (background → status → cases → impact → outlook → conclusion → limits → references)\n5. Report date is the true current date; references sorted reverse-chronologically; conclusions only, no recommendations; key trend judgments sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-trend", "name": "文博趋势综述专家", "description": "宽泛主题广谱调研,3-4 路网络调研强调时效性,案例卡片化呈现,交付四格式趋势综述报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "2e5f2aab7036", "expert_id": "ex_M24IASAeVHLO" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:20:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "civil-risk-goods-sale", "title": "买卖货款纠纷·诉讼风险提示", "title_en": "Civil Litigation Risk Report for a Goods-Sale Dispute", "subtitle": "一套自洽的货款纠纷案情，跑通「程序/实体/执行」三类风险识别与分级", "subtitle_en": "A self-consistent goods-sale dispute, showcasing procedural, substantive and enforcement risk identification.", "description": "接了一个货款纠纷，最怕的是不知道「哪里会翻车」——时效过了没有、证据够不够、赢了钱拿不拿得回来，心里没底。\n\n这个案例用一套自洽的买卖合同货款纠纷案情（货款 50 万 + 利息，被告以「质量不合格」抗辩），完整跑一遍「民事诉讼风险提示」方法论：识别出 11 项风险，按程序、实体、执行三类归集，逐项给出风险等级、评分、触发条件、法律依据和缓释动作，最后还附上给当事人讲清楚的专业话术。\n\n【产物形态说明】本案例默认产出排版专业、可交互的 HTML 风险提示报告（风险总览 + 逐项风险 + 缓释动作 + 当事人话术），文件保存为「风险提示报告.html」。\n\n适合：处理货款/合同纠纷的诉讼律师、需要向当事人做风险告知的律师、以及想系统梳理案件风险点的法律工作者。", "description_en": "Taking on a payment dispute, the biggest worry is not knowing where it could go wrong — has the limitation period lapsed, is the evidence sufficient, and can you actually collect after winning?\n\nThis case runs the full civil litigation risk method over a self-consistent goods-sale dispute (a 500k receivable plus interest, with the defendant pleading quality defects): it identifies 11 risks across procedural, substantive and enforcement categories, each with a rating, score, trigger condition, legal basis and mitigation actions, plus plain-language talking points to explain them to clients.\n\n[Note] Output format note: this case by default produces a professionally formatted, interactive HTML risk report (overview + itemized risks + mitigation actions + client talking points), saved as \"风险提示报告.html\".\n\nFor: litigation lawyers handling sales/contract disputes, lawyers who need to deliver risk disclosures to clients, and legal professionals who want a systematic view of a case's risk points.", "categories": [ "business" ], "tags": [ "民事诉讼", "风险提示", "买卖合同", "货款纠纷", "诉讼策略" ], "tags_en": [ "Civil Litigation", "Risk Disclosure", "Sales Contract", "Payment Dispute", "Litigation Strategy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "风险提示报告.html", "prompt": "请以「民事诉讼风险提示」的方法，帮我系统梳理一起买卖合同货款纠纷的风险，并以排版专业、可交互的 HTML 报告形式输出（保存为「风险提示报告.html」）。\n\n案件概况（脱敏示例）：甲公司（卖方/原告）与乙公司（买方/被告）于 2024-01-08 签订书面买卖合同，约定付款期为交货后 30 日；2024-03-15 交货、付款期于 2024-04-14 届满；乙公司未付款，甲公司 2024-06-01 发出书面催款函（依《民法典》第195条发生时效中断）；2025-05-15 乙公司首次以「货物质量不合格」为由明确拒付（注意：债务人单纯拒付不构成时效中断）；2027-05-10 甲公司再次书面催款（再次中断）。拟起诉时点 2027-06-15，距最近一次中断起算的 3 年时效届满不足 30 天。诉求：货款 50 万元及逾期利息。核心争议焦点：被告的质量异议抗辩。\n\n请输出：1）风险总览（总数、高/中/低风险计数）；2）按「程序风险 / 实体风险 / 执行风险」三类逐项列出风险，每项含风险等级、评分、触发条件、法律依据和缓释动作（标注优先级、截止时间、责任人）；3）针对高风险项目，生成给当事人做风险说明的通俗话术（通俗解释 + 对当事人的影响 + 应对措施）。\n\n要求：风险评分按「可能性 × 影响程度 × 可控性」综合，等级用 高/中/低 并配对应色块标签（红/黄/绿）；法律依据逐条标注到具体法条；最终结论须提示由执业律师复核。", "prompt_en": "Using the civil litigation risk method, systematically map the risks of a goods-sale payment dispute and output a professionally formatted, interactive HTML report (saved as \"风险提示报告.html\").\n\nCase overview (anonymized sample): Company A (seller/plaintiff) and Company B (buyer/defendant) signed a written sales contract on 2024-01-08, payment due within 30 days of delivery; goods delivered on 2024-03-15, payment due on 2024-04-14; Company B did not pay, and Company A sent a written demand letter on 2024-06-01 (causing limitation-period interruption under Art. 195 of the Civil Code); on 2025-05-15 Company B first refused payment citing \"non-conforming quality\" (note: a debtor's mere refusal does NOT interrupt the limitation period); on 2027-05-10 Company A sent another written demand (a second interruption). The intended filing date is 2027-06-15, less than 30 days before the 3-year limitation expires from the most recent interruption. Claims: RMB 500,000 plus overdue interest. Core issue: the defendant's quality-defect defense.\n\nOutput: 1) a risk overview (total count, high/medium/low counts); 2) itemized risks grouped under procedural / substantive / enforcement, each with risk level, score, trigger condition, legal basis and mitigation actions (marked with priority, deadline, responsible party); 3) for high-risk items, plain-language talking points for client disclosure (plain explanation + impact on the client + response measures).\n\nRequirements: score risks by probability × impact × controllability, use high/medium/low levels with corresponding colored block labels (red/yellow/green); cite each legal basis to the specific article; conclude by noting that final conclusions must be reviewed by a licensed lawyer.", "skills": [], "experts": [ { "id": "civil-litigation-agent", "name": "民商事诉讼专家", "description": "【法大大·睿契】面向律师与法务，覆盖民商事诉讼起诉应诉、保全、庭审、上诉、再审与执行全流程，内嵌执业安全红线", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ce9b5f0ae0bc", "expert_id": "ex_W5v0UgaMeBDz" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "community", "usage_count": 0, "created_at": "2026-09-07T11:22:57+08:00", "updated_at": "2026-09-07T11:22:57+08:00", "artifacts": [ { "file": "风险提示报告.html", "label": "交互式风险提示报告（主产物）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "college-100-things-workbench", "title": "大学四年，想做的 100 件小事", "title_en": "100 Little Things to Do in Four Years of College", "subtitle": "10 大分类 × 100 件小事打卡手账 ｜ 章节式进度 · 心情与照片记录", "subtitle_en": "10 categories × 100 little things | Chapter progress with moods & photos", "description": "「大学四年，想做的 100 件小事」打卡手账工作台。把大学愿望拆成 10 大分类（校园时光/青春日常/朋友之间/一个人的浪漫/小小远方/关于热爱/勇敢一点/感受世界/留下痕迹/好好告别），每类 10 件、共 100 张可打卡的卡片。每件小事都可以记录状态（想去做/等待发生/已经珍藏/留给以后）、完成日期、当下心情（平静/期待/惊喜/感动/放松/疲惫/迷茫/想念/自由/其他）、一句话记录和现场照片；顶部进度条实时显示已珍藏件数、当前章节等级和「距离下一章还差几件」；第二页「我的大学章节」用 Lv.1→Lv.6 六个章节（0/20/40/60/80/100 件）呈现大学进度线，配最近心情 Top3、大学关键词云、写给现在的自己的一句话和最近珍藏时间线。支持关键词搜索和分类筛选，中英双语一键切换，PC/手机自适应。打开即有 16 条演示记录，点底部「清空演示记录」即可从零开始写自己的 100 件。适合刚入学的准大学生、想认真记录大学生活的同学，以及想给朋友/孩子送一份「大学愿望清单」模板的人。", "description_en": "A check-in journal workbench for '100 little things to do in four years of college'. Your college wishes are organized into 10 categories (campus time, daily youth, between friends, solo romance, little faraways, about passion, be braver, feel the world, leave a trace, say goodbye) — 10 items each, 100 cards in total. Every item can record a status (want to do / waiting to happen / treasured / saved for later), completion date, mood (calm / expectant / surprise / moved / relaxed / tired / lost / missing / free / other), a one-line note and on-site photos. The sticky progress bar shows how many you have treasured, your current chapter level and how many more to the next chapter; the second tab 'My College Chapters' draws a Lv.1→Lv.6 roadmap (0/20/40/60/80/100 items) with recent moods Top 3, a college keyword cloud, a letter to your present self and a recent-treasures timeline. Keyword search, category filter, bilingual EN/中文 toggle, mobile-friendly. Ships with 16 demo records — click 'Clear demo records' at the bottom to start your own 100 from zero. Perfect for incoming freshmen, anyone journaling college life, or as a gift template for friends and kids.", "categories": [ "workspace" ], "tags": [ "大学手账", "清单打卡", "生活记录", "章节进度", "双语工作台" ], "tags_en": [ "College Journal", "Bucket List", "Life Logging", "Chapter Progress", "Bilingual" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "大学四年，想做的 100 件小事（单体双语 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/x8W4npYirwp1PSuLpBMYS9，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：做一个「大学四年，想做的 100 件小事」打卡手账工作台单页 HTML。\n\n1. 数据：一张「大学小事清单」资料库表，字段——编号（数字）、小事（文本）、分类（单选：校园时光/青春日常/朋友之间/一个人的浪漫/小小远方/关于热爱/勇敢一点/感受世界/留下痕迹/好好告别）、状态（单选：想去做/等待发生/已经珍藏/留给以后）、完成日期（日期）、心情（单选：平静/期待/惊喜/感动/放松/疲惫/迷茫/想念/自由/其他）、一句话记录（文本）、照片（图片）。预置 100 条小事（每类 10 条），编号 001-100；\n2. 页面两个页签：「100 件小事」卡片墙——按分类分组渲染 100 张卡片（编号/小事名/状态胶囊/完成日期/照片标记），支持关键词搜索 + 分类下拉筛选；「我的大学章节」统计页——Lv.1→Lv.6 六个章节（0/20/40/60/80/100 件）进度线，已珍藏件数、最近心情 Top3、大学关键词云、写给现在的自己、最近珍藏时间线；\n3. 交互：点卡片打开记录面板，可改状态/日期/心情、写一句话、添加照片（前端压缩后上传资料库），保存走资料库增量更新；页面读取资料库 schema 动态渲染分类与心情选项，订阅数据变更自动刷新；\n4. 视觉：奶油纸感手账风（米白底、衬线标题、圆角卡片、细线描边），顶部粘性导航 + 实时进度条（已珍藏 n/100、距下一章提示、当前章节等级），所有 CSS/JS 内联、无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to make a same-copy of https://www.workbuddy.cn/space/d/x8W4npYirwp1PSuLpBMYS9 with online data storage and two-way sync. Prefer the one-click same-copy flow; if unavailable, build it as follows: a single-page HTML journal workbench for '100 little things to do in four years of college'. 1. Data: a 'College Little Things' database table with fields — No. (number), Thing (text), Category (select: Campus Time / Daily Youth / Between Friends / Solo Romance / Little Faraway / About Passion / Be Braver / Feel the World / Leave a Trace / Say Goodbye), Status (select: Want to Do / Waiting to Happen / Treasured / Saved for Later), Date (date), Mood (select: Calm / Expectant / Surprise / Moved / Relaxed / Tired / Lost / Missing / Free / Other), One-line Note (text), Photos (image). Preload 100 items, 10 per category, numbered 001-100;\n2. Two tabs: '100 Little Things' card wall — cards grouped by category (number / name / status pill / date / photo mark) with keyword search + category filter; 'My College Chapters' stats page — Lv.1→Lv.6 roadmap (0/20/40/60/80/100 items) with treasured count, recent moods Top 3, college keyword cloud, a letter to your present self, and a recent-treasures timeline;\n3. Interactions: click a card to open the record panel — change status/date/mood, write a note, add photos (compressed client-side then uploaded to the database); saves go through incremental database updates; the page reads the database schema to render category/mood options dynamically and subscribes to data changes for auto-refresh;\n4. Visual: cream-paper journal style (off-white background, serif headings, rounded cards, hairline borders), sticky top nav with a live progress bar (treasured n/100, next-chapter hint, current level), all CSS/JS inlined, no external CDN.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "岳定谣", "name_en": "Yue Dingyao", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 100, "source": "community", "created_at": "2026-09-01T18:30:00+08:00", "updated_at": "2026-09-02T13:57:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "college-life-workbench", "title": "拾光集 · 大学生活手记", "title_en": "TimeTrace · College Life Journal", "subtitle": "把课业、习惯、重要日子和生活费收进一个会呼吸的大学生活工作台", "subtitle_en": "Coursework, habits, big days and allowance — one living college workbench", "description": "大学生活散落在备忘录、群通知和各种小本本里：待办追着截止日期跑，习惯三天打鱼两天晒网，生日和纪念日总要靠运气想起来，生活费月底永远对不上账。\n\n「拾光集」把这些收进一张工作台：今日总览一屏看清今天要处理什么、习惯是否打卡、最近的纪念日还有几天；待办按今日/未来 7 天/已逾期/已完成分栏，还能按学习、自媒体、杂事分类筛选；习惯打卡自动累计连续天数，断签从 1 天重新开始；重要日期带倒计时和准备进度，一键 +10%；生活费账本记一笔自动归类（写「食堂午餐」自动归餐饮），月预算剩余天数和日均可用实时可算；「人生第一次」收藏你的第一次独自旅行、第一次拿奖学金。全部数据存在资料库的结构化表里，手机电脑同步。\n\n适合：想把大学生活过出节奏感的人、习惯养成总是半途而废的人、总记不住重要日子的人、月底想搞清楚钱花哪了的人。", "description_en": "College life scatters across notes, group chats and random scraps: deadlines chase you, habits fade after three days, birthdays depend on luck, and the allowance never adds up by month end.\n\nTimeTrace gathers it all into one workbench: an overview showing today's load, habit streaks and the nearest big day; to-dos split into Today / Next 7 days / Overdue / Done with filters for study, content creation and chores; habit check-ins that auto-count streaks and restart from day 1 after a miss; countdowns with a one-tap +10% prep progress for big dates; an allowance ledger that auto-categorizes (type \"canteen lunch\" and it lands under Food) with monthly budget, days left and daily allowance computed live; and a \"First Times\" collection for your first solo trip, first scholarship. All data lives in structured WorkBuddy Library tables, synced across devices.\n\nBest for: students who want rhythm in their college life, serial habit-quitters, anyone who forgets big days, and anyone who wants to know where the money went.", "categories": [ "workspace" ], "tags": [ "大学生活", "待办管理", "习惯打卡", "生活费账本", "纪念日提醒" ], "tags_en": [ "College Life", "To-do Management", "Habit Tracking", "Allowance Ledger", "Countdown Reminder" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "拾光集 · 大学生活手记（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/1cLC6wMdvjvBp0KSudMZyJ?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n帮我做一个「拾光集 · 大学生活手记」生活工作台，单个完整 HTML 文件，六个模块：① 今日总览（五张卡：待办完成度、习惯打卡、最近纪念日倒计时、本月预算使用、人生最新一次）；② 待办事项（新增/编辑/删除，字段含任务名称、所属类型[学习/自媒体/杂事]、截止日期、优先级[P0-P3]、状态[未开始/进行中/已完成/已延期]、预计耗时、是否重复、备注；按今日/未来 7 天/已逾期/已完成分栏，可按类型筛选，完成后关联纪念日进度 +10%）；③ 习惯打卡（新增/暂停/打卡，自动累计连续天数与总次数，断签从 1 天重新开始）；④ 重要日期（生日/纪念日/节日，倒计时 hero 卡显示「距离 X 还有 N 天」与准备进度，按钮每次 +10%）；⑤ 生活费账本（记一笔：日期、收支类型[收入/支出]、消费分类[餐饮/交通/学习/娱乐/购物/住宿/其他]、金额、支付方式[微信/支付宝/现金/银行卡]、消费说明写「食堂午餐」自动归类；月预算设置，显示已使用百分比、剩余天数、日均可用）；⑥ 人生第一次（记录第一次事件、分类、日期、一句话描述）。\n\n视觉走温暖治愈的大学生手帐风：奶油底色 + 墨绿主色 + 珊瑚强调，卡片圆角软阴影，模块间留白呼吸感。制作完成后：将 HTML 存入我的个人资料库，创建 6 张资料库子数据表（待办/习惯/日期/账本/预算设置/第一次）绑定数据并预置示例数据，发布为公开链接返回给我。CSS/JS 无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/1cLC6wMdvjvBp0KSudMZyJ?source=2, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a \"TimeTrace · College Life Journal\" workbench as one complete HTML file, with six modules: ① Today overview (five cards: to-do completion, habit streaks, nearest big-day countdown, monthly budget usage, latest first-time); ② To-dos (add/edit/delete with name, category [Study / Content / Chores], deadline, priority [P0-P3], status [Not started / In progress / Done / Postponed], estimated hours, repeat, note; split into Today / Next 7 days / Overdue / Done with category filters; completing a linked to-do bumps its big-day progress +10%); ③ Habits (add/pause/check in, auto streak & total counts, streak restarts from 1 after a miss); ④ Big days (birthdays/anniversaries/holidays with a countdown hero showing \"N days until X\" and prep progress, +10% per tap); ⑤ Allowance ledger (add entry: date, income/expense, category [Food/Transport/Study/Fun/Shopping/Housing/Other], amount, payment method [WeChat/Alipay/Cash/Bank card], note auto-categorization — \"canteen lunch\" lands under Food; monthly budget with used %, days left, daily allowance); ⑥ First times (event, category, date, one-line description). Visual: warm healing college-journal style — cream background, ink-green primary, coral accents, rounded cards with soft shadows and breathing whitespace. When done: save the HTML to my personal WorkBuddy Library, create 6 bound child tables (todos / habits / big days / ledger / budget settings / first times), seed sample data, and publish a public link back to me. Inline all CSS/JS with no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "圈宁CeoxNim", "name_en": "圈宁CeoxNim", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "community", "usage_count": 0, "created_at": "2026-09-02T15:00:00+08:00", "updated_at": "2026-09-02T15:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "color-theory-basics", "title": "色轮、原色与互补色互动教学", "title_en": "色轮、原色与互补色互动教学", "subtitle": "色彩学入门 · 色轮、原色与互补色互动教学", "subtitle_en": "色彩学入门 · 色轮、原色与互补色互动教学", "description": "色彩理论入门互动课件，用可交互色轮讲清原色、二次色与互补色，并给出六组配色实战案例。", "description_en": "色彩理论入门互动课件，用可交互色轮讲清原色、二次色与互补色，并给出六组配色实战案例。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "color-theory-basics.html", "artifacts": [ { "file": "color-theory-basics.html", "label": "色轮、原色与互补色互动教学｜互动教学网页" } ], "prompt": "帮我做一个色彩理论基础的互动教学网页，要用可交互的色轮讲解原色、二次色和互补色。", "prompt_en": "帮我做一个色彩理论基础的互动教学网页，要用可交互的色轮讲解原色、二次色和互补色。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "color-theory-fundamentals", "title": "12 色相环与互补色交互探索", "title_en": "12 色相环与互补色交互探索", "subtitle": "设计基础 · 12 色相环与互补色交互探索", "subtitle_en": "设计基础 · 12 色相环与互补色交互探索", "description": "以 12 钟点色轮为核心的互动课件，支持悬停查看色相档案、锁定 HEX/RGB 值，讲清互补色与冷暖色。", "description_en": "以 12 钟点色轮为核心的互动课件，支持悬停查看色相档案、锁定 HEX/RGB 值，讲清互补色与冷暖色。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "color-theory-fundamentals.html", "artifacts": [ { "file": "color-theory-fundamentals.html", "label": "12 色相环与互补色交互探索｜互动教学网页" } ], "prompt": "帮我做一个设计基础色彩学的互动教学网页，核心是一个 12 色相的可交互色轮，能悬停查看 HEX 和 RGB。", "prompt_en": "帮我做一个设计基础色彩学的互动教学网页，核心是一个 12 色相的可交互色轮，能悬停查看 HEX 和 RGB。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "community-operation-scripts", "title": "民事律师社群运营话术全案", "title_en": "30-Day Operating Scripts for a Matrimonial & Family Law Community", "subtitle": "一句话输入领域和频率，产出 30 天社群运营话术 + SOP + 群规 + 转化话术全套", "subtitle_en": "One-line domain input produces a 30-day community script package: daily content, SOP, group rules, and conversion scripts", "description": "律师做私域最头疼的就是——社群建起来了，内容却不知道每天发什么，硬发广告又怕违规掉粉，咨询来了接不住、转化不了。\n\n这个案例演示了如何用「法律社群运营话术生成」技能，为一位专注民事律师 8 年的执业律师，一键生成一套可落地执行的 30 天社群运营话术全案：从每日运营 SOP（早/午/晚三段互动安排、周/月节奏），到入群欢迎话术、30 天逐日法律小常识与脱敏案例，再到问题解答、活动通知、转化引导和群规模板，全部内容均符合《律师执业行为规范》，标注「仅供参考」、规避「包赢」「保证胜诉」等违规表述，律师拿来稍作审核即可直接发布。\n\n适合：运营民事律师咨询群的执业律师、刚起步做私域的法律人、律所市场运营人员，以及想系统化社群内容输出的团队。", "description_en": "The biggest headache in private-domain marketing for lawyers is this: you build a community, but don't know what to post every day; hard-selling gets flagged as spam and loses members; and when inquiries come in, you can't convert them.\n\nThis case shows how the Legal Community Operation Script Generation skill produces a ready-to-use 30-day operating script package for a matrimonial & family lawyer with 8 years of experience: from the daily operating SOP (morning/noon/evening interaction rhythm, weekly and monthly cadence), to welcome scripts, 30 days of daily legal tips and de-identified cases, plus Q&A scripts, event announcements, conversion guidance, and group rules — all compliant with the Lawyers' Professional Conduct Rules, marked \"for reference only\" and avoiding prohibited wording like \"guaranteed win\".\n\nFor: practicing lawyers running matrimonial & family law communities, legal professionals starting out in private-domain marketing, law-firm marketing staff, and teams that want a systematic content output.", "categories": [ "business", "productivity" ], "tags": [ "社群运营", "律师私域", "民事律师", "话术模板", "客户转化" ], "tags_en": [ "community operation", "lawyer private domain", "matrimonial & family", "script templates", "client conversion" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "民事律师法律社群运营话术全案.html", "artifacts": [ { "file": "民事律师法律社群运营话术全案.html", "label": "30 天社群运营话术全案" } ], "prompt": "请为一位专注民事律师领域 8 年的执业律师，生成一份可落地执行的「民事律师法律咨询公益群」30 天运营话术全案。\n\n要求：\n1. 群定位：离婚 / 抚养权 / 财产分割 / 家暴 咨询人群；运营目标：专业展示 + 潜在客户培育；互动频率：3 次/日。\n2. 输出七大板块：①社群运营 SOP（每日早/午/晚安排 + 每周/每月安排）；②入群欢迎话术；③每日互动内容（30 天逐日内容日历，含早安法律小常识、午间问答、晚间热点/脱敏案例）；④问题解答话术；⑤活动通知话术（免费咨询日/专题分享/福利活动）；⑥转化引导话术（自然过渡、不生硬）；⑦群规模板（含违规处理话术）。\n3. 律师姓名、律所等信息用占位符（如 [王律师]、[XX律师事务所]）。\n4. 合规要求：全部话术符合《律师执业行为规范》，标注「仅供参考，具体案件请咨询律师」，禁止出现「包赢」「100%胜诉」等绝对化表述；案例须脱敏处理。", "prompt_en": "Please generate a ready-to-use 30-day operating script package for a matrimonial & family law public-welfare consultation group, for a practicing lawyer with 8 years of experience in matrimonial & family law.\n\nRequirements:\n1. Group positioning: divorce / child custody / property division / domestic violence audience; goal: professional showcase + lead nurturing; interaction frequency: 3 times per day.\n2. Output seven sections: 1) community operating SOP (daily morning/noon/evening schedule + weekly/monthly plan); 2) welcome script; 3) daily interaction content (30-day calendar, incl. morning legal tips, noon Q&A, evening hot-topic / de-identified cases); 4) Q&A reply scripts; 5) event announcement scripts (free consultation day / thematic sharing / welfare activities); 6) conversion guidance scripts (natural, not pushy); 7) group rules template (incl. violation handling).\n3. Use placeholders for lawyer name and firm (e.g. [Wang Lawyer], [XX Law Firm]).\n4. Compliance: all scripts comply with the Lawyers' Professional Conduct Rules, marked \"for reference only, please consult a lawyer for specific cases\", and must avoid absolute wording like \"guaranteed win\" or \"100% success\"; cases must be de-identified.", "skills": [], "experts": [ { "id": "brand-marketing", "name": "品牌营销专员", "description": "品牌营销专家，专注处理律师个人IP运营和律所品牌营销相关事务，涵盖律师个人IP定位、律所品牌营销方案、社交媒体内容运营、营销数据分析、获客转化方案、法律科普内容创作、短视频脚本创作等。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3d642417d5d0", "expert_id": "ex_AYR4CRp3puyw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "community", "usage_count": 0, "created_at": "2026-09-07T14:34:00+08:00", "updated_at": "2026-09-16T16:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "commuter-coffee-cup-sticker-article", "title": "通勤咖啡杯贴图号爆文", "title_en": "Commuter Coffee Cup Sticker Article", "subtitle": "封面、配图与 800 字正文一次生成", "subtitle_en": "A cover, inline images, and an 800-word article in one run", "description": "便携咖啡杯看起来都差不多，真正让通勤人崩溃的却是漏水、太重和杯盖洗不干净；只罗列容量和颜值，很难让用户产生购买理由。\n\n这个案例围绕“轻行锁温杯”生成一篇贴图号爆款图文，完整交付 1 张封面图、2 张内文配图、约 800 字正文和话题标签。文章从通勤包里的咖啡酸味切入，用杯盖可拆洗、物理锁扣、230 克轻量杯身三个判断标准承接产品卖点，并以互动问题自然收尾。\n\n适合：杯壶品牌、通勤好物商家、微信内容运营、生活方式类图文创作者。", "description_en": "Portable coffee cups can look nearly identical, but commuters actually struggle with leaks, excess weight, and lids that cannot be cleaned thoroughly. Listing capacity and appearance alone rarely creates a convincing purchase reason.\n\nThis case generates a viral sticker-style article for the Qingxing insulated travel cup, delivering one cover image, two inline images, an approximately 800-word article, and topic tags. It opens with the sour coffee smell inside a commuter bag, then connects three buying criteria—a removable lid, a physical lock, and a 230 g lightweight body—to the product’s selling points before ending with a natural discussion prompt.\n\nIdeal for: drinkware brands, commuter-product merchants, WeChat content teams, and lifestyle article creators.", "categories": [ "creation" ], "tags": [ "贴图号爆文", "通勤好物", "咖啡杯", "图文种草" ], "tags_en": [ "sticker-style viral article", "commuter products", "coffee cup", "visual product seeding" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "通勤咖啡杯贴图号爆款图文" } ], "prompt": "请生成一篇可直接发布的微信贴图号爆款图文，选题是“通勤人群如何选一只不漏又好洗的便携咖啡杯”，正文字数约 800 字，内容调性接地气。产品名为“轻行锁温杯”，价格 129 元，卖点包括 350ml 轻量杯身约 230g、防漏锁扣与单手开盖、宽口可拆洗杯盖。目标受众为 25—35 岁每天乘地铁或公交通勤的职场人；他们的痛点是杯子渗漏弄湿电脑和文件、大容量杯太重、杯盖缝隙难清洗且容易残留咖啡味。采用贴图号形式，封面与正文配图使用 3:4 比例，生成 1 张封面图、2 张正文配图和 1 个主标题。正文要先用具体通勤场景抓住注意力，再给出可执行的选购判断标准，让产品卖点自然承接这些标准，避免无法验证的绝对承诺，最后用真实问题引导评论。请将标题、封面、内文配图、正文和标签完整渲染为 Markdown。", "prompt_en": "Create a ready-to-publish viral WeChat sticker-style article on the topic “How commuters can choose a portable coffee cup that does not leak and is easy to clean.” Target approximately 800 Chinese characters and use a grounded tone. The product is the Qingxing insulated travel cup priced at RMB 129. Its selling points are a 350 ml lightweight body weighing about 230 g, a leak-resistant lock with one-handed opening, and a wide, removable lid for cleaning. The audience is working professionals aged 25–35 who commute by subway or bus every day. Their pain points are leaks that damage laptops and documents, bulky high-capacity cups, and lid crevices that are hard to clean and retain coffee odors. Use the sticker-content format with 3:4 cover and inline images, generating one cover, two inline images, and one main headline. Open with a concrete commuting scene, provide actionable selection criteria, connect the product benefits naturally to those criteria, avoid unverifiable absolute promises, and finish with a genuine question that encourages comments. Render the title, cover, inline images, body, and tags as complete Markdown.", "skills": [ { "id": "lingyi-article-sticker-viral-generator", "name": "公众号贴图号爆款生成", "description": "【零一数科·出品】输入选题、产品卖点和受众，一键生成封面、配图与可发布爆款图文。", "slug": "lingyi-article-sticker-viral-generator" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-28T10:08:20Z", "updated_at": "2026-08-31T16:31:11+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "company-fundamental-analysis", "title": "上市公司基本面分析报告", "title_en": "Listed-Company Fundamental Analysis Report", "subtitle": "财报 + 行情 + 资金 + 同业，一份看懂一家上市公司", "subtitle_en": "Financials + quotes + capital flow + peers — understand a listed company in one report", "description": "想搞懂一家上市公司值不值得投，要在财报、行情、资金流、同业之间来回查，普通投资者很难拼出全貌。本案例组合「NeoData 金融搜索服务」与「WeStock Data」，以宁德时代为例生成一份上市公司基本面分析报告——四季度财务趋势、关键指标、行情与主力资金、同业对比、投资亮点与风险一页呈现。适合做企业研究、想系统评估一家公司的投资者与从业者。", "description_en": "To judge whether a listed company is worth investing in, you have to dig across financial statements, quotes, capital flow, and peers — hard for an ordinary investor to assemble into a full picture. Combining NeoData Financial Search with WeStock Data, this case uses CATL as an example to generate a listed-company fundamental analysis report: four-quarter financial trends, key metrics, quotes and main-capital flow, peer comparison, and investment highlights vs risks — all on one page. Ideal for investors and professionals doing company research.", "categories": [ "finance" ], "tags": [ "公司分析", "基本面", "财报", "同业对比", "估值" ], "tags_en": [ "Company Analysis", "Fundamentals", "Financial Statements", "Peer Comparison", "Valuation" ], "artifact_type": "html", "prompt": "帮我生成一份「上市公司基本面分析报告」单页 HTML，分析宁德时代(300750)。\n\n用 NeoData 金融搜索服务查财务与宏观，用 WeStock Data 查行情、财报、资金流和技术指标。要直观看到：公司概况与估值、近四季营收/净利润/毛利率趋势、ROE 等关键财务指标、行情与主力资金（价格区间/主力净流入/北向持股/技术信号）、动力电池同业对比，以及投资亮点与风险，最后给一段综合评价（非投资建议）。涨跌按 A 股习惯（涨红跌绿）。", "prompt_en": "Generate a single-page HTML 'Listed-Company Fundamental Analysis Report' analyzing CATL (300750).\n\nUse NeoData Financial Search for financials and macro, and WeStock Data for quotes, statements, capital flow and technicals. Show at a glance: company overview and valuation; four-quarter revenue/net-profit/gross-margin trends; key financial metrics such as ROE; quotes and main-capital flow (price range / net main inflow / northbound holdings / technical signals); a power-battery peer comparison; and investment highlights vs risks, ending with an overall assessment (not investment advice). Use the China-market color convention (up=red, down=green).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "基本面分析报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-06-16T21:20:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "company-public-service-center", "title": "知危的公司公共服务中心", "title_en": "Company Public Service Center", "subtitle": "用资料库搭建可交互、数据在线存储的员工公共生活服务中心", "subtitle_en": "An interactive employee service hub backed by online database tables", "description": "公司里借书、报活动、早起打卡、积分兑换这些公共服务，往往散落在不同表格和群聊里——员工想借本书要看共享文档，报名活动要爬楼接龙，打卡和麻瓜币又各管各的，来回切换很累。\n\n这个案例用资料库搭建一个公司公共服务中心 Demo：员工端覆盖书架、影音推荐、早起打卡、公共物品、活动广场、同事互助、麻瓜币，运营端提供数据台和资料库数据表管理。所有核心操作（借还、打卡、报名、兑换）都真实写回在线数据表，换设备打开数据都在。\n\n适合：想快速搭建内部员工服务门户的行政/HR 团队；需要统一管理公司公共资源与活动的运营同学；以及想用资料库做数据驱动单页应用的产品与开发。", "description_en": "Company public services like book borrowing, event sign-ups, morning check-ins and point redemption are usually scattered across spreadsheets and group chats — employees have to open shared documents to borrow a book, scroll through chat history to join an event, and manage check-ins and coins separately, which gets exhausting.\n\nThis case builds a company public service center demo on top of the WorkBuddy database: the employee side covers a book shelf, media recommendations, morning check-ins, shared items, an event plaza, peer help, and magic coins, while the ops side provides a data dashboard and database table management. Every core action (borrow/return, check-in, sign-up, redeem) writes back to the online tables in real time, so the data is still there on another device.\n\nGreat for: HR/admin teams who want to spin up an internal employee service portal fast; ops folks who need to manage company resources and activities in one place; and product/engineering people who want to build data-driven single-page apps on the database.", "categories": [ "workspace" ], "tags": [ "HTML", "员工服务", "轻应用", "资料库", "数据管理" ], "tags_en": [ "HTML", "employee service", "lightweight app", "database", "data management" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "公司公共服务中心（资料库数据驱动）" } ], "prompt": "请使用资料库的能力帮我一键做同款（发布态url），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n设计并实现一个「公司公共服务中心」Demo，定位为员工内部公共生活入口，而非传统 OA。\n\n### 1. 核心模块\n员工端包括：\n- 首页：个人状态、近期动态、本周数据\n- 公司书架：搜索/筛选/借还/收藏\n- 影音推荐：推荐、评分、短评\n- 早起计划：目标、打卡、连续记录\n- 公共物品：查询、借用、归还\n- 活动广场：浏览、报名、取消、发布想法\n- 同事互助：发布、回复、解决\n- 我的：汇总个人借阅、活动、打卡、收藏等记录\n- 麻瓜币：积分流水、奖励兑换\n\n运营端包括数据台 + 资料库数据表管理，展示活跃、参与、借用等统计，并支持数据查看、导入和导出。\n\n### 2. 数据与交互\n使用资料库（WorkBuddy 资料库）的关联数据表存储全部数据，请按模块建好对应的资料库数据表。借还、打卡、报名、回复、兑换等操作必须真实写回在线数据，并同步影响个人记录、动态、通知和麻瓜币。数据要真正存在线上、不能只存在浏览器本地，换设备打开数据都在。\n\n提供全局搜索、通知、员工/运营视角切换和本地持久化。不要做静态展示页，主要按钮都必须可用。\n\n### 3. 视觉\n采用暖色编辑杂志风：奶油纸色背景、陶土橙点缀、深橄榄辅助色、衬线标题 + 无衬线正文。整体克制、轻量、有生活感，避免传统后台管理系统风格。\n\n### 4. 输出\n直接生成可运行的 HTML Demo，数据走资料库在线数据表，自带完整示例数据，无需用户继续补充信息即可体验完整流程。", "prompt_en": "Use the WorkBuddy database capability to build me the same thing in one click (published URL), with data stored online and synced both ways. Prefer the one-click same-as-this flow; if that capability is unavailable, build it as follows:\n\nDesign and implement a \"Company Public Service Center\" demo, positioned as an internal employee life service hub rather than a traditional OA system.\n\n### 1. Core modules\nEmployee side:\n- Home: personal status, recent activity, weekly stats\n- Company bookshelf: search / filter / borrow-return / favorite\n- Media recommendations: recommendations, ratings, short reviews\n- Morning plan: goals, check-ins, streaks\n- Shared items: search, borrow, return\n- Event plaza: browse, sign up, cancel, post ideas\n- Peer help: post, reply, resolve\n- My page: aggregated borrowings, events, check-ins, favorites\n- Magic coins: point ledger, reward redemption\n\nOps side includes a data dashboard + database table management, showing activity, participation and borrowing stats, with data view, import and export.\n\n### 2. Data & interactions\nStore all data in the related tables of the WorkBuddy database, and create the matching database tables per module. Borrow/return, check-in, sign-up, reply and redeem actions must write back to the online data in real time and sync to personal records, feeds, notifications and magic coins. The data must genuinely live online rather than only in browser storage, and remain available across devices.\n\nProvide global search, notifications, employee/ops view switching and local persistence. Don't make a static showcase page — the main buttons must all work.\n\n### 3. Visual\nUse a warm editorial-magazine style: cream paper background, terracotta-orange accents, deep olive secondary color, serif headings + sans-serif body. Keep it restrained, lightweight and lived-in, avoiding the traditional admin-dashboard look.\n\n### 4. Output\nGenerate a runnable HTML demo directly, with data backed by online database tables and complete sample data included, so users can experience the full flow without providing any extra information.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "description_en": "Builds custom digital workspaces for students, professionals, creators, parents, and lifestyle users, with responsive HTML, local data persistence, and one-click deployment.", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "知危", "name_en": "Zhiwei", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T12:00:00+08:00", "updated_at": "2026-08-13T12:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "construction-dispute-evidence-matrix", "title": "建设工程价款争议证据矩阵", "title_en": "Construction Price Dispute Evidence Matrix", "subtitle": "把装修工程价款的散乱材料，梳理成律师与业务可协作核对的证据矩阵", "subtitle_en": "Turn scattered construction payment materials into an evidence matrix lawyers and business teams can verify together", "description": "追装修尾款最头疼的，是合同、签证、聊天记录、竣工照片散落在微信、相册和文件夹各处，律师和业务对不上一张账，真到庭上才发现关键证据缺原件、没签字。\n\n这个案例演示了如何用「争点—举证责任—证据—证明目的—风险—缺口—补强」的证据导向工作流，把散乱材料梳理成一张可协作核对的证据矩阵，单列工程量、变更签证、验收、结算、质量异议和欠款金额，每份材料标注原始载体、来源、待补文件与责任人，并对照片、聊天记录等电子材料提示保留原始文件与导出路径。\n\n适合：装修/建设工程承包人、企业法务、诉讼律师、项目经理、合同履约与应收账款团队。", "description_en": "Chasing the final payment on a decoration project is painful — contracts, site instructions, chat records and completion photos are scattered across WeChat, photo albums and folders, so lawyers and the business team can never reconcile a single account, and by the time you reach court you find key evidence lacks originals or signatures.\n\nThis case shows how to apply the \"issue — burden of proof — evidence — proof purpose — risk — gap — reinforcement\" evidence-oriented workflow to turn scattered materials into a collaboratively verifiable evidence matrix, separately listing work quantity, change instructions, acceptance, settlement, quality objections and the outstanding amount, with each material tagged with its original carrier, source, missing documents and responsible person, and prompting you to keep original files and export paths for photos and chat records.\n\nFor: decoration/construction contractors, in-house counsel, litigation lawyers, project managers, and contract performance & accounts receivable teams.", "categories": [ "business" ], "tags": [ "建设工程", "价款争议", "证据矩阵", "诉讼", "装修工程" ], "tags_en": [ "Construction", "Price Dispute", "Evidence Matrix", "Litigation", "Decoration" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "建设工程价款争议证据矩阵互动卡片" } ], "prompt": "我方为装修工程承包人，项目已完工并交付使用，但发包方尚未支付尾款。现有施工合同、报价单、现场签证、进度款申请、微信沟通记录、竣工照片和部分付款流水。\n\n请按「争点/待证事实—举证责任—现有证据—证明目的—原始载体与来源—真实性、合法性、关联性风险—证据缺口—补强动作」建立证据矩阵，并单列工程量、变更签证、验收、结算、质量异议和欠款金额。\n\n请先核验现行规则和受理法院要求，未核验处标注「待官方来源核验」；输出待补信息及优先行动，不预测判决结果。输出为单文件、可互动、可离线打开的 HTML 页面，含提示词编辑框、复制按钮、案件角色切换（承包人/发包人）、材料完备度总览，并以预置案情展示案件摘要、金额关系图、时间线、证据矩阵、完备度热力图、风险说明与行动优先级清单。", "prompt_en": "We are the contractor of a decoration project. The project has been completed and delivered for use, but the owner has not yet paid the final payment. Existing materials include the construction contract, quotation, site instructions (visa), progress payment applications, WeChat chat records, completion photos, and partial payment records.\n\nPlease build an evidence matrix following \"issue / fact to be proved — burden of proof — existing evidence — proof purpose — original carrier and source — authenticity, legality, relevance risks — evidence gap — reinforcement actions\", and separately list work quantity, change instructions, acceptance, settlement, quality objections, and the outstanding amount.\n\nPlease first verify the current rules and the accepting court's requirements, marking unverified items as \"pending official source verification\"; output the information to be supplemented and priority actions, and do not predict the judgment result. Output as a single-file, interactive, offline-openable HTML page with a prompt editor, a copy button, a case role switcher (contractor / owner), and a material completeness overview, presenting the preset case with a case summary, an amount relationship diagram, a timeline, the evidence matrix, a completeness heatmap, risk notes, and a priority action list.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家，擅长合同起草、审查、对比、法规检索与在线签署", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d1e4a1d28e92", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-13T15:59:28+08:00", "updated_at": "2026-08-13T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "content-factory-newsletter", "title": "AI 周刊 Newsletter 模板", "title_en": "AI Weekly Newsletter Template", "subtitle": "精美的多栏邮件通讯模板设计", "subtitle_en": "Beautiful multi-column email newsletter template design", "description": "如何设计一封让人愿意打开的 Newsletter？本案例展示一个专业的 AI 科技周刊邮件模板，包含渐变 Hero 区域、精选文章卡片、数据统计板块、社区亮点和社交链接，兼容主流邮件客户端的 table 布局。适合内容运营者、自媒体创作者和市场营销团队。", "description_en": "How to design a newsletter people want to open? This case showcases a professional AI tech weekly email template with gradient hero, featured article cards, stats section, community highlights, and email-client-compatible table layout.", "categories": [ "business", "office" ], "tags": [ "Newsletter", "邮件模板", "内容运营", "AI周刊" ], "tags_en": [ "Newsletter", "Email Template", "Content Operations", "AI Weekly" ], "artifact_type": "html", "cover_image": "cover-portrait.png", "cover_mode": "portrait", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI 周刊 Newsletter 页面" } ], "prompt": "请创建一个名为「AI Weekly | 第 42 期」的 Newsletter 邮件模板（单文件 HTML，内联无外部依赖），现代精美但兼容 Gmail/Outlook（用邮件兼容的 table 布局）。\n\n包含：渐变 Hero 横幅、3 篇精选文章（图片+标题+2 行摘要+分类标签+Read more）、Quick Links 侧栏、「本周数字」统计、社区亮点、社交链接 Footer。文章内容用逼真的 AI/科技新闻（GPT-5 传闻、开源模型发布、法规更新等）。响应式流式 table。", "prompt_en": "Create a newsletter email template titled \"AI Weekly | Issue 42\" (single-file HTML, inlined with no external dependencies), modern and polished yet compatible with Gmail/Outlook (use an email-compatible table layout).\n\nInclude: a gradient hero banner, 3 featured articles (image + title + 2-line summary + category tag + Read more), a Quick Links sidebar, a \"By the Numbers This Week\" stats section, community highlights, and a social-links footer. Article content should use realistic AI/tech news (GPT-5 rumors, open-source model releases, regulatory updates, etc.). Responsive fluid tables.", "skills": [ { "id": "content-factory", "name": "内容工厂", "description": "多智能体内容生产系统", "skill_id": "skill_2053081490234478592" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "contentops-article-review", "title": "公众号文章质量评审仪表盘", "title_en": "Article Quality Review Dashboard", "subtitle": "专家面板给一篇公众号长文打分，递归迭代至 93 分", "subtitle_en": "An expert panel scores a long-form article, iterating to 93", "description": "写完一篇文章，最怕的不是没人看，而是自己觉得挺好、其实满篇 AI 味和空话。这个案例用 content-ops 专家面板给一篇「远程办公为何失败」的公众号长文做质量评审：8 位专家从声音契合度、具体性、去 AI 味、篇幅适配、互动潜力五个维度打分，递归迭代最多 3 轮，直到跨过 90 分阈值。\n\n报告采用深色科技风的数据仪表盘版式：环形总分、内联 SVG 五维雷达图、逐轮迭代进度轨迹、KPI 卡片，以及每一轮到底改了什么的修改动作清单。你能清楚看到分数如何从第一轮的 11.7 分（满篇 leverage/ecosystem/holistic 等禁用词）一路迭代到终稿的 93 分。\n\n所有分数均由 content-ops 官方脚本 content-quality-scorer.py 真实运行产出，非手写编造。适合：公众号作者、内容运营、newsletter 主理人、任何想在发布前给自己文章做一次严格质量闸门的人。", "description_en": "The scariest thing after finishing an article isn't no readers — it's thinking it's great when it's actually full of AI slop and empty phrases. This case uses the content-ops expert panel to review a long-form article on 'why remote work really fails': 8 experts score across five dimensions (voice match, specificity, anti-AI-slop, length fit, engagement) and iterate recursively for up to 3 rounds until it clears the 90 threshold.\n\nThe report uses a dark, tech-styled data dashboard: a ring total score, an inline-SVG five-dimension radar, a round-by-round iteration trail, KPI cards, and a checklist of exactly what changed each round. You can watch the score climb from a first-round 11.7 (packed with banned words like leverage/ecosystem/holistic) to a final 93.\n\nAll scores come from a real run of the official content-ops script content-quality-scorer.py, not hand-written. Best for: newsletter writers, content operators, and anyone who wants a strict quality gate on their article before publishing.", "categories": [ "creation" ], "tags": [ "内容评审", "质量评分", "专家面板", "文章优化", "去AI味" ], "tags_en": [ "content review", "quality scoring", "expert panel", "article optimization", "anti-AI-slop" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "文章质量评审仪表盘" } ], "prompt": "帮我用 content-ops 专家面板给这篇公众号长文（主题：远程办公为何失败）做质量评审。用 7-10 位专家从声音契合度、具体性、去 AI 味、篇幅适配、互动潜力五个维度打分，递归迭代到 90 分以上。产出一个深色科技风的数据仪表盘 HTML 报告：环形总分、内联 SVG 五维雷达图、逐轮迭代进度、KPI 卡片、每轮修改动作。分数必须真实运行评分脚本产出。", "prompt_en": "Use the content-ops expert panel to review my long-form article (topic: why remote work fails). Assemble 7-10 experts and score across five dimensions (voice match, specificity, anti-AI-slop, length fit, engagement), iterating recursively past 90. Produce a dark, tech-styled data-dashboard HTML report: ring total score, inline-SVG five-dimension radar, round-by-round iteration progress, KPI cards, and per-round edits. Scores must come from a real run of the scoring script.", "skills": [ { "id": "content-ops", "skill_id": "skill_2053081939666735104", "name": "Content Ops", "description": "内容质量评分与专家评审面板，递归迭代至 90+ 分", "description_en": "Content quality scoring with expert panel review, iterates until 90+" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 90, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-20T12:00:00Z", "updated_at": "2026-07-20T12:00:00Z", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "contentops-copy-review", "title": "产品文案三变体擂台赛", "title_en": "Product Copy Showdown — 3 Variants", "subtitle": "A/B/C 三版文案上擂台 PK，专家面板独立评分决出冠军", "subtitle_en": "Three copy variants battle it out; the panel scores and crowns a winner", "description": "写文案最纠结的时刻，就是三个版本都觉得还行、不知道该发哪个。这个案例用 content-ops 专家面板给同一款「会议削减 App」的 A/B/C 三版短文案独立评分，然后按分数排名，决出真正该上线的那一版。\n\n版式做成竞赛计分板／对战擂台：三条变体赛道横向 PK，每条挂金/银/铜奖牌与排名、分数大字与进度条、绿色优点标签与红色缺点标签，胜出者用金色冠军光效高亮，还原颁奖台的仪式感。你能一眼看出满篇 AI 套话的变体 A 只拿了 14.8 分，有真实故事但缺数据的变体 B 拿 58.6 分，而反直觉钩子+箭头数据的变体 C 夺冠。\n\n关键在于诚实：冠军并非天生达标——原始变体 C 只有 78.4 分，面板迭代 1 轮（命名产品、补数据、压字数）才升到 90.5 通过；变体 A、B 未迭代，如实保留 FAIL 分数。所有分数由 content-ops 官方脚本 content-quality-scorer.py 真实运行产出。适合：产品营销、广告文案、增长团队做上线前的文案 A/B 决策。", "description_en": "The most agonizing moment in copywriting is when all three versions feel 'fine' and you don't know which to ship. This case uses the content-ops expert panel to score three short copies (A/B/C) for the same 'meeting-cutting App' independently, then ranks them by score to crown the one that should actually go live.\n\nThe layout is a competition scoreboard / battle arena: three variant lanes PK horizontally, each with a gold/silver/bronze medal and rank, a big score with a progress bar, green pro-tags and red con-tags, and the winner highlighted with a golden champion glow for full podium ceremony. At a glance you see Variant A (a wall of AI slop) score just 14.8, Variant B (real story but no data) score 58.6, and Variant C (contrarian hook + arrow data) take the crown.\n\nThe key is honesty: the champion wasn't born passing — raw Variant C scored only 78.4, and the panel iterated once (named the product, added data, trimmed length) to reach 90.5 PASS; Variants A and B were not iterated and honestly keep their FAIL scores. All scores come from a real run of the official content-ops script content-quality-scorer.py. Best for: product marketers, ad copywriters, and growth teams making pre-launch A/B copy decisions.", "categories": [ "creation" ], "tags": [ "文案评审", "AB测试", "变体对比", "专家评分", "文案优化" ], "tags_en": [ "copy review", "A/B testing", "variant comparison", "expert scoring", "copy optimization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "产品文案三变体擂台赛" } ], "prompt": "帮我用 content-ops 专家面板给同一款会议削减 App 的 A/B/C 三版产品文案独立评分并排名，对最优版本迭代到 90 分以上。产出一个竞赛计分板／对战擂台风格的 HTML 报告：三变体做成横向 PK 赛道，每张卡带金银铜奖牌、分数条、绿色优点标签与红色缺点标签，胜出者用金色冠军光效高亮夺冠。未达标的变体如实标注 FAIL。不要用仪表盘或批注版式。分数必须真实运行评分脚本产出。", "prompt_en": "Use the content-ops expert panel to independently score and rank three product copies (A/B/C) for the same meeting-cutting App, iterating the best one past 90. Produce a competition-scoreboard / battle-arena HTML report: three variant lanes PK horizontally, each card with a gold/silver/bronze medal, a score bar, green pro-tags and red con-tags, and the winner highlighted with a golden champion glow. Non-passing variants are honestly marked FAIL. No dashboard or annotation layout. Scores must come from a real run of the scoring script.", "skills": [ { "id": "content-ops", "skill_id": "skill_2053081939666735104", "name": "Content Ops", "description": "内容质量评分与专家评审面板，递归迭代至 90+ 分", "description_en": "Content quality scoring with expert panel review, iterates until 90+" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 90, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-20T12:00:00Z", "updated_at": "2026-07-20T12:00:00Z", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "contentops-email-review", "title": "营销邮件评审 · 建议模式", "title_en": "Marketing Email Review · Suggesting Mode", "subtitle": "像 Google Docs 建议模式一样批注并修订一封营销邮件", "subtitle_en": "Annotate and redline a marketing email, Google-Docs style", "description": "一封营销邮件发出去只有一次机会，主题行踩雷、满篇企业腔、没有一个具体数字，打开率和转化率就全废了。这个案例用 content-ops 专家面板评审一封电商营销邮件，并把结果呈现成你最熟悉的「建议模式」——就像在 Google Docs 里开了修订建议。\n\n版式完全不同于仪表盘：左侧是邮件原文信笺，问题处用彩色下划线高亮并挂上编号图钉；右侧是专家评语气泡，一条条对应到正文；主题行和收尾用 diff 呈现修订建议——划线删除旧文、绿色标注新文。你能直观看到「我们很高兴分享变革性平台」被改成「你每月正在流失 $4,200」，去 AI 味从 0 分救回 100 分。\n\n所有分数由 content-ops 官方脚本 content-quality-scorer.py 真实运行产出，迭代轨迹 18.2 → 88.0 → 92.3 如实记录。适合：增长/邮件营销、SCRM 运营、独立开发者、任何要在点击发送前给邮件做最后一道质量闸门的人。", "description_en": "A marketing email gets one shot. A landmine subject line, wall-to-wall corporate-speak, and not a single concrete number will tank your open and conversion rates. This case uses the content-ops expert panel to review an e-commerce marketing email and presents the result in the format you already know — Suggesting mode, exactly like turning on suggested edits in Google Docs.\n\nThe layout is deliberately nothing like a dashboard: the left is the original email letter, with problem spots highlighted by colored underlines and numbered pins; the right is a rail of expert comment bubbles, each mapped to a line in the body; the subject line and closing show redline diffs — struck-through old text, green new text. You can literally watch 'We're excited to share our transformative platform' become 'You're losing $4,200 a month', rescuing anti-AI-slop from 0 back to 100.\n\nAll scores come from a real run of the official content-ops script content-quality-scorer.py, with the iteration trail 18.2 → 88.0 → 92.3 recorded honestly. Best for: growth/email marketers, CRM operators, indie makers, and anyone wanting one last quality gate before hitting send.", "categories": [ "creation" ], "tags": [ "邮件营销", "内容评审", "文案修订", "专家批注", "转化优化" ], "tags_en": [ "email marketing", "content review", "copy editing", "expert annotation", "conversion optimization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "营销邮件评审建议模式" } ], "prompt": "帮我用 content-ops 专家面板评审这封电商营销邮件，递归迭代到 90 分以上。产出一个「建议模式」风格的 HTML 报告（像 Google Docs 修订建议）：左侧显示邮件原文，问题处用彩色下划线高亮并加编号，右侧是对应的专家评语气泡，主题行和收尾用 diff（划线旧文+绿色新文）展示修订建议。不要用雷达图或仪表盘。分数必须真实运行评分脚本产出。", "prompt_en": "Use the content-ops expert panel to review this e-commerce marketing email, iterating past 90. Produce a Suggesting-mode HTML report (like Google Docs suggested edits): the original email on the left with problem spots highlighted by colored underlines and numbered pins, a rail of matching expert comment bubbles on the right, and redline diffs (struck-through old text + green new text) for the subject line and closing. No radar chart or dashboard. Scores must come from a real run of the scoring script.", "skills": [ { "id": "content-ops", "skill_id": "skill_2053081939666735104", "name": "Content Ops", "description": "内容质量评分与专家评审面板，递归迭代至 90+ 分", "description_en": "Content quality scoring with expert panel review, iterates until 90+" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 90, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-20T12:00:00Z", "updated_at": "2026-07-20T12:00:00Z", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-review-opinion-verification-1", "title": "EPC合同评审意见逐条核查", "title_en": "EPC Contract Review Opinion Fact-Check", "subtitle": "抓出评审意见里的虚构条款号、串年版编号与张冠李戴", "subtitle_en": "Catch fabricated clause numbers, cross-edition citations and misattributions in review opinions", "description": "最怕外部法律顾问的评审意见里夹带「幻觉」——条款号编得一本正经，实际在合同原文里根本不存在，照单全收就会谈错方向。这个案例把一份国际 EPC 交钥匙项目（FIDIC 银皮书 2017）的合同评审意见逐条拉回合同原文与权威条号结构核验，揪出 7 项事实错误、10 项覆盖缺口与 5 项措辞不严：保险被写成 1999 版的 18.1（2017 版实为第 19 条）、不可抗力被安到第 19 条（实为第 18 条例外事件）、「工程师」被张冠李戴成雇主代表、雇主风险明明已含战争却建议补充。最终产出按「事实错误 / 覆盖缺口 / 措辞不严」分级、自包含可离线打开的单文件 HTML 核对报告。\n\n适合：合同评审意见的复核人、法务与外部顾问对接人、EPC 项目商务经理、需要交叉核验 FIDIC 1999/2017 条号差异的律师，以及任何不想被「看似专业的条款号」带偏的合同审阅者。", "description_en": "Nothing is more frustrating than a law firm review opinion packed with \"hallucinations\" — clause numbers that sound authoritative but simply do not exist in the contract, which can send negotiations in the wrong direction if taken at face value. This case takes a contract review opinion for an international EPC turnkey project (FIDIC Silver Book 2017) and verifies every clause number and factual statement against the contract text and the authoritative clause structure, uncovering 7 factual errors, 10 coverage gaps and 5 loose-wordings: insurance cited as Cl. 18.1 from the 1999 edition (it sits in Clause 19 in 2017), force majeure pinned to Clause 19 (actually Clause 18 Exceptional Events), \"the Engineer\" misattributed to what is actually the Employer's Representative, and a suggestion to add \"war and hostilities\" to Employer's Risks even though they are already listed. The output is a self-contained single-file HTML report graded into factual errors / coverage gaps / loose wordings that opens offline.\\n\\nFor: reviewers of contract review opinions, legal teams liaising with outside counsel, EPC commercial managers, lawyers cross-checking FIDIC 1999/2017 clause renumbering, and anyone who does not want to be led astray by superficially professional-looking clause numbers.", "categories": [ "business" ], "tags": [ "合同审查", "FIDIC", "幻觉核查", "EPC", "法律合规" ], "tags_en": [ "contract review", "FIDIC", "hallucination check", "EPC", "legal compliance" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "分级核对报告（HTML，自包含可离线打开）" } ], "prompt": "请对一份国际 EPC 交钥匙项目的合同评审意见做逐条事实核查，找出其中的「幻觉」，输出一份分级核对报告（单文件 HTML，样式自包含、资源全内联，可离线打开）。\n\n材料（同一目录下两份文件，建议命名如「合同评审意见_待核.md」「合同原文_条款节选.md」）：①合同评审意见全文（待核对象）；②合同原文条款节选（核对基准，含通用条件与专用条件）。本合同采用 FIDIC 银皮书（Silver Book）2017 第二版。\n\n核查要求：\n1. 把评审意见当「待核对象」而非结论：意见中引用的每个条款号、每句事实与定性，都回到合同原文逐条比对，并用 FIDIC 2017 银皮书的条号结构交叉验证；\n2. 重点盯 FIDIC 1999 与 2017 两版的条号位移（保险 18↔19、不可抗力/例外事件 19↔18、索赔与争端 20/21），以及银皮书「无工程师、只有雇主代表」的特征；\n3. 核查维度：虚构/串年版条款号、把修改建议写成现行事实、张冠李戴、覆盖遗漏、措辞不严；拿不准的条目标「待核」，禁止凭记忆下条款号；\n4. 报告按「事实错误（必改）/ 覆盖缺口 / 措辞不严（打磨项）」分级，每条含位置、现状原文、问题所在、建议改法，另列覆盖缺口清单与检索溯源；\n5. 文件名命名为「<评审意见名>-错误与幻觉核对报告.html」，不改动评审意见原文。", "prompt_en": "Perform an item-by-item factual verification of a contract review opinion for an international EPC turnkey project, find the \"hallucinations\" in it, and produce a graded verification report (single-file HTML, self-contained styles, all resources inlined, openable offline).\n\nMaterials (two files in the same directory, e.g. named contract-review-opinion-to-verify.md and contract-excerpts.md): 1) the full text of the contract review opinion (the object to be verified); 2) excerpts of the contract (the verification baseline, including general conditions and particular conditions). The contract is based on the FIDIC Silver Book 2017 Second Edition.\n\nVerification requirements:\n1. Treat the review opinion as the \"object to be verified\", not as a conclusion: every clause number, factual statement and characterization must be checked item by item against the contract text, and cross-validated against the FIDIC 2017 Silver Book clause structure;\n2. Pay special attention to the renumbering between FIDIC 1999 and 2017 (Insurance 18↔19, Force Majeure/Exceptional Events 19↔18, Claims and Disputes 20/21), and the Silver Book feature of \"no Engineer, only the Employer's Representative\";\n3. Check dimensions: fabricated or cross-edition clause numbers, suggestions written as current facts, misattributions, coverage omissions, and loose wording; mark uncertain items as \"to be verified\" and never state clause numbers from memory;\n4. Grade the report into \"factual errors (must fix) / coverage gaps / loose wording (polish)\", each item containing location, current text, the problem, and a suggested fix, plus a coverage-gap list and a search provenance section;\n5. Name the file \"-错误与幻觉核对报告.html\" and do not modify the original review opinion.", "skills": [ { "id": "contract-review-opinion-verification", "name": "合同评审意见逐条核查", "description": "对合同评审意见、法律意见书或风险评审表做逐条事实核查：核条款号是否真实存在、事实与定性是否与合同原文及权威标准一致，找出虚构编号、张冠李戴、覆盖遗漏与措辞不严，输出分级核对报告", "skill_id": "skill_2090718994325864448" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "usage_count": 0, "created_at": "2026-08-21T14:55:00+08:00", "updated_at": "2026-08-21T14:55:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-review-opinion-verification-2", "title": "合同评审意见逐条核查", "title_en": "Clause-by-Clause Contract Review Verification", "subtitle": "逐条核验评审意见的条款号与事实，揪出虚构编号与张冠李戴", "subtitle_en": "Verify every cited clause and fact in review opinions, catch fabricated references", "description": "拿到同事手填的风险评审表，最怕的是条款号写错、内容张冠李戴，还拿着错误结论去谈合同。本案例演示了如何把一份半导体设备采购合同的风险评审表（10 行）逐条回到合同原文与民法典核验：7 处事实错误全部定位（含虚构条款号 14.5/16.7/11.4/9.3、违约金 0.05% 误写为千分之五、质保期 12 个月误写为 24 个月、把不存在的履约保证金写成现行约定等），并区分「合同现行约定」与「处理建议」，产出带严重度分级、可离线打开的自包含 HTML 核对报告。\n\n适合：合同与法务评审人员、采购与风控同事、需要复核他人评审意见的任何人。", "description_en": "When you receive a colleague's hand-filled risk review sheet, the worst part is wrong clause numbers, cross-wired content, and negotiating the contract based on wrong conclusions. This case shows how to verify a semiconductor equipment procurement risk review sheet (10 rows) line by line against the original contract text and the Civil Code: all 7 factual errors were located (including fabricated clause numbers 14.5/16.7/11.4/9.3, liquidated damages 0.05% miswritten as 5 per mille, warranty period 12 months miswritten as 24 months, and a performance bond written as an existing clause when it never existed), while distinguishing 'current contract terms' from 'recommendations'. The output is a severity-graded, self-contained HTML verification report that works offline.\n\nPerfect for: contract and legal reviewers, procurement and risk-control colleagues, and anyone who needs to double-check someone else's review opinions.", "categories": [ "business", "productivity" ], "tags": [ "合同评审", "条款核验", "幻觉核查", "风险评审", "法律合规" ], "tags_en": [ "contract review", "clause verification", "hallucination check", "risk assessment", "legal compliance" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "设备采购风险评审表-错误与幻觉核对报告.html", "prompt": "请对一份半导体设备采购合同的风险评审表做逐条事实核查：核每一行「所依据条款号」是否在合同原文里真实存在、引用内容是否与该条款原文一致，找出虚构编号、张冠李戴、把修改建议写成现行事实、覆盖遗漏与措辞不严，产出分级核对报告（HTML，自包含）。\n\n核对基准如下，两份本地材料都在工作区根目录：\n- 合同原文（节选）见《半导体设备采购合同_条款节选.md》（国内货物买卖合同，无单一强制示范文本，对照民法典合同编第 9 章及买卖合同司法解释）。\n- 待核查对象见《设备采购风险评审表_待核.xlsx》（一张表，列为：序号 / 风险事项 / 评审意见 / 所依据条款号 / 风险等级 / 处理建议）。\n\n这份风险评审表由买方的采购部同事用 Excel 手工填写，条款号极可能有写错、串行、虚构的情况，请你逐行回到合同原文去核实。请特别区分「合同现行约定」与评审表里的「处理建议」——二者混写是本类核对最易出错处。\n\n报告文件名用「设备采购风险评审表-错误与幻觉核对报告.html」，落盘到 ./output/，报告需自包含（内联 CSS，无外部依赖）。", "prompt_en": "Perform a line-by-line factual verification of a risk review sheet for a semiconductor equipment procurement contract: for each row, check whether the cited clause number actually exists in the original contract text, whether the quoted content matches the clause, and identify fabricated numbers, cross-wired citations, recommendations written as current facts, coverage gaps, and loose wording. Produce a severity-graded verification report (self-contained HTML).\n\nVerification baselines — both local materials are in the workspace root:\n- Contract excerpt: 半导体设备采购合同_条款节选.md (domestic goods sales contract; no single mandatory model text; cross-check against Book 9 of the Civil Code and the SPC Interpretation on Sales Contracts).\n- Object under review: 设备采购风险评审表_待核.xlsx (one sheet with columns: No. / Risk item / Review opinion / Cited clause number / Risk level / Recommendation).\n\nThe risk review sheet was filled in manually in Excel by the buyer's procurement team, so clause numbers may well be wrong, shifted, or fabricated. Verify each row against the original contract text. Pay special attention to distinguishing 'current contract terms' from 'recommendations' in the sheet — mixing the two is the most common error in this kind of review.\n\nName the report 设备采购风险评审表-错误与幻觉核对报告.html, save it to ./output/, and keep it self-contained (inline CSS, no external dependencies).", "skills": [ { "id": "contract-review-opinion-verification", "name": "合同评审意见逐条核查", "description": "对合同评审意见、法律意见书或风险评审表做逐条事实核查，核条款号真实性、事实与定性一致性，找出虚构编号、张冠李戴、覆盖遗漏与措辞不严，输出分级核对报告", "skill_id": "skill_2090718994325864448" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "official", "usage_count": 0, "created_at": "2026-08-21T15:00:00+08:00", "updated_at": "2026-08-21T15:00:00+08:00", "artifacts": [ { "file": "设备采购风险评审表-错误与幻觉核对报告.html", "label": "分级核对报告（自包含 HTML）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "contract-review-opinion-verification-3", "title": "合同评审意见逐条查错", "title_en": "Contract Review Opinion Fact-Check", "subtitle": "逐条核对条款号与法条时效，揪出评审意见里的张冠李戴与过期口径", "subtitle_en": "Line-by-line verification of clause numbers and legal currency to catch misattributed citations and outdated doctrine", "description": "拿到一份合同评审意见，最怕律师把条款号写串、把建议当成现行约定、引一条已经废止的法条——这份文件是要直接签合同的，错了就是真金白银的风险。本案例演示把一份施工总承包合同的律师评审意见全文交给 AI 逐条核查：对照 GF-2017-0201 合同原文与北大法宝现行法条，找出 9 处事实错误（示范文本编号串用、保险条号 17/18 张冠李戴、索赔期限 28/14 天对不上、优先受偿范围写反、把已废止的《合同法》第五十二条当现行法）、1 处覆盖缺口（含 9 个漏核风险块）和 3 处措辞不严，输出一份分级核对报告（自包含 HTML，按事实错误/覆盖缺口/措辞不严三档分级，附检索溯源）。特别提示 2026-06-30 施行的《建工解释二》已收紧「实际施工人直接诉发包人」路径，意见里仍在用《建工解释一》第 43 条的旧口径。适合：公司法务、外聘律师出具意见前的自查、合同审查 AI 应用的质检环节、法学院学生对司法解释时效的辨析训练。", "description_en": "Nothing is scarier than a contract review opinion that cites the wrong clause number, presents a suggestion as current fact, or quotes an abolished law — the document is about to be signed, and mistakes cost real money. This case hands an external lawyer's review opinion for a construction general-contracting contract to AI for line-by-line fact-checking: against the GF-2017-0201 contract text and current provisions retrieved from PKULaw, it surfaces 9 factual errors (wrong model-text GF number, insurance clause 17/18 mix-up, 28-day vs 14-day claim period mismatch, reversed priority-payment scope, and citing the abolished Article 52 of the Contract Law as current law), 1 coverage gap (covering 9 un-reviewed risk blocks), and 3 loose-wordings — delivered as a self-contained tiered HTML report (factual errors / coverage gaps / wording polish, with a retrieval trail). It highlights that the new Judicial Interpretation II on construction contracts, effective 2026-06-30, has tightened the 'actual constructor suing the employer directly' path, while the opinion still relies on the old Article 43 of Interpretation I. Suitable for: in-house counsel, pre-issuance self-check for external lawyers, QA of contract-review AI pipelines, and law students practicing legal-currency analysis.", "categories": [ "business", "productivity" ], "tags": [ "合同审查", "条款核查", "法条时效", "幻觉检测", "法律风控" ], "tags_en": [ "contract review", "clause verification", "legal currency", "hallucination detection", "legal risk control" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "scenario_cover": "", "preview": "科创园合同评审意见-错误与幻觉核对报告.html", "prompt": "请使用「合同评审意见逐条核查」skill，对一份已成稿的合同评审意见做逐条事实核查，找出其中的「幻觉」并输出分级核对报告（HTML，自包含，资源全内联）。\n\n任务要点：\n1. 输入材料：把你的合同原文文件（含通用条款与专用条款节选）与待核评审意见全文放到工作区并给出路径；如暂无自己的材料，可直接用案例目录内的示例（科创园施工总承包合同_条款节选.md、科创园合同评审意见_待核.md）演练，效果一致。\n2. 核查清单：①虚构或串套的条款号（意见引用的每条条号都要回合同原文对号）；②把修改建议写成现行事实、把 A 条款内容安到 B 条款；③引用已废止或被替代的法律口径当现行法——法条时效以检索当日为准，优先用北大法宝 MCP 核验，查不到就标「待核」，禁止凭记忆写条号；④覆盖遗漏（对照同类合同必审风险块，检查意见声称「已覆盖全部风险」是否成立）；⑤措辞不严。\n3. 特别留意：示范文本编号是否张冠李戴（如 GF-2017-0201 与 GF-2020-0216 串用）；通用条款号与专用条款已修改内容的关系（专用条款优先）；《建工解释一》（法释〔2020〕25 号）与 2026-06-30 施行的《建工解释二》（法释〔2026〕12 号）的口径差异，尤其「实际施工人直接诉发包人」路径已被解释二第 6 条收紧。\n4. 输出：按「事实错误（必改）→ 覆盖缺口 → 措辞不严（打磨项）」三级分级，每条含位置、现状原文、问题所在、建议改法；另附覆盖缺口清单、检索溯源（检索源、查询词、返回文号与时效）与下一步。不改动合同原文和意见原文。\n5. 报告文件名：<对象名>-错误与幻觉核对报告.html，落盘目录由用户指定。", "prompt_en": "Use the \"Contract Review Opinion Verification\" skill to run a line-by-line fact-check on a completed contract review opinion, identify its \"hallucinations\", and deliver a tiered verification report (HTML, self-contained, all resources inlined).\n\nTask points:\n1. Inputs: place your contract source file (with general and special terms excerpts) and the review opinion to be verified in the workspace and provide their paths; if you do not have your own materials yet, you may directly use the samples in the case directory (科创园施工总承包合同_条款节选.md, 科创园合同评审意见_待核.md) for a walkthrough with identical results.\n2. Checklist: ① fabricated or misplaced clause numbers (every cited number must be matched back to the contract source); ② suggestions stated as current fact, or content of clause A attributed to clause B; ③ abolished or superseded legal doctrine cited as current law — legal currency is determined as of the retrieval date, prefer the PKULaw MCP for verification, mark \"to be verified\" when not found, never write clause numbers from memory; ④ coverage gaps (compare against the mandatory risk blocks for that contract type and check whether a claim of \"full coverage\" holds); ⑤ loose wording.\n3. Particular attention: model-text number mix-ups (e.g. GF-2017-0201 vs GF-2020-0216); the relationship between general-term numbers and special-term modifications (special terms prevail); and the doctrinal differences between Construction Interpretation I (Fa Shi [2020] No. 25) and Construction Interpretation II (Fa Shi [2026] No. 12, effective 2026-06-30), especially the tightened \"actual constructor suing the employer directly\" path under Article 6 of Interpretation II.\n4. Output: tiered findings of \"factual error (must fix) → coverage gap → loose wording (polish)\", each with location, current text, problem, and suggested fix; plus a coverage-gap list, a retrieval trail (source, query, returned document number and currency), and next steps. Do not modify the contract source or the opinion itself.\n5. Report filename: -错误与幻觉核对报告.html, saved to the directory specified by the user.", "skills": [ { "id": "contract-review-opinion-verification", "name": "合同评审意见逐条核查", "description": "对合同评审意见、法律意见书或风险评审表做逐条事实核查：核条款号是否真实存在、事实与定性是否与合同原文及权威标准（FIDIC、示范文本、司法解释）一致，找出虚构编号、张冠李戴、覆盖遗漏与措辞不严，输出分级核对报告。", "skill_id": "skill_2090718994325864448" } ], "experts": [], "mcps": [ { "id": "pkulaw", "name": "北大法宝·法律智能检索", "description": "法条、案例与引注核验：法条时效、条号、案例原文" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-21T14:56:32+08:00", "updated_at": "2026-08-21T15:03:00+08:00", "artifacts": [ { "file": "科创园合同评审意见-错误与幻觉核对报告.html", "label": "分级核对报告（自包含 HTML）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "contract-term-review-revise-1", "title": "合同条款审查与修订（买方立场）", "title_en": "Contract Term Review & Revision (Buyer's Perspective)", "subtitle": "逐条改写不利条款，再查内部交叉矛盾，产出带风险批注的修订稿与冲突报告", "subtitle_en": "Rewrite unfavorable clauses one by one, then audit internal contradictions, delivering annotated revisions and a conflict report", "description": "签采购合同最怕的就是条款全是对方写的——验收标准对方说了算、违约金我方畸高对方却设上限、知识产权和源代码全归卖方，真出问题时才发现自己毫无还手之力。这个案例站在买方（甲方）立场，对一份软件采购及实施服务合同做了全流程处理：先逐条识别对买方不利的风险点并改写（改动标黑、重大风险批注），覆盖验收标准、违约金对等、责任限额、知识产权、源代码交付、维护费调整、单方解约权、数据安全责任等维度；再对修订稿做一轮内部交叉矛盾核查，专找条款互相打架、时间单位混用、责任主体写反、违约金上限被顶破等问题，最终产出修订文档、修改要点汇总表、核心条款修改对照表与独立的冲突审查报告（附统计）。\n\n适合：采购与商务负责人、法务与合规人员、中小企业合同经办人，以及任何需要站在己方立场审合同、把风险条款改回来的人。", "description_en": "The worst part of signing a procurement contract is that every clause is written by the other side — the acceptance standard is set unilaterally, your penalty is sky-high while theirs is capped, and all IP and source code belong to the seller, leaving you with zero leverage when things go wrong. This case takes the buyer's (Party A's) perspective and processes a software procurement & implementation services contract end to end: first it identifies and rewrites each buyer-unfavorable clause (changes bolded, major risks annotated), covering acceptance standards, penalty parity, liability caps, IP rights, source-code delivery, maintenance-fee adjustments, unilateral termination rights, and data-security responsibility; then it runs a cross-contradiction audit on the revised draft, hunting for conflicting clauses, mixed time units, reversed responsible parties, and penalty caps being blown through, finally producing a revision document, a summary table of changes, a clause-by-clause comparison table, and a standalone conflict review report (with statistics).\n\nPerfect for: procurement & business leads, legal and compliance staff, contract handlers at SMEs, and anyone who needs to review a contract from their own side and negotiate unfavorable terms back.", "categories": [ "business" ], "tags": [ "合同审查", "条款修订", "买方立场", "风险批注", "冲突核查" ], "tags_en": [ "contract review", "clause revision", "buyer perspective", "risk annotation", "conflict audit" ], "artifact_type": "html", "preview": "revision-document.html", "artifacts": [ { "file": "revision-document.html", "label": "合同修订文档（改动标黑 + 风险批注）" } ], "prompt": "我司「XX公司」是甲方（买方），准备采购一套软件并签实施服务合同。请站在买方立场，对 [粘贴合同全文或提供合同文件路径] 做全流程处理：\n\n1. 识别并逐条改写对我方（买方）不利的条款——包括但不限于：验收标准由卖方单方确定、我方违约金畸高而对方违约金过低且设上限、责任限制过低、知识产权与源代码全部归卖方、维护费单方调整、我方无解约权、数据安全责任全甩给买方等；\n2. 改写完成后，再对改好的稿子做一轮内部交叉矛盾核查，专找条款之间互相打架、逻辑不一致（如工作日与自然日混用、责任主体写反）、文字笔误（如主宾语倒置影响款项流向）的地方。\n\n产出要求：修改文档（HTML，改动标黑+风险批注）、修改要点汇总表、核心条款修改对照表，以及独立的冲突审查报告（HTML，末尾附矛盾/逻辑/笔误统计）。", "prompt_en": "My company [Company Name] is Party A (the buyer) and plans to procure a software product and sign an implementation services contract. Please process the contract [paste the full contract text or provide the contract file path] end to end from the buyer's perspective:\n\n1. Identify and rewrite, one by one, the clauses unfavorable to us (the buyer) — including but not limited to: acceptance standards set unilaterally by the seller, our penalty being disproportionately high while the seller's is low and capped, an overly low liability cap, IP and source code all belonging to the seller, unilateral maintenance-fee adjustments, no right of unilateral termination for us, and all data-security responsibility dumped on the buyer;\n2. After the rewrite, run a round of internal cross-contradiction audit on the revised draft, hunting for clauses that conflict with each other, logical inconsistencies (e.g., mixing working days with calendar days, reversed responsible parties), and typos (e.g., inverted subject/object that affects the direction of payment).\n\nDeliverables: a revision document (HTML, changes bolded + risk annotations), a summary table of revision points, a clause-by-clause comparison table, and a standalone conflict review report (HTML, ending with statistics of contradictions / logic issues / typos).", "skills": [ { "id": "contract-term-review-revise", "name": "合同条款审查与修订", "description": "代表合同中明确指定的一方，先逐条识别不利条款并做可追溯的修订，再对修订稿进行一轮内部交叉矛盾核查，输出带风险批注的修订文档与独立的冲突审查报告。", "skill_id": "skill_2090718439110635520" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-08-21T13:45:00+08:00", "updated_at": "2026-08-21T14:05:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-term-review-revise-2", "title": "买方立场合同条款审查与修订", "title_en": "Buyer-Side Contract Clause Review & Revision", "subtitle": "逐条识别 10 处不利条款并改写，再做交叉矛盾核查，一键产出修订稿与审查报告", "subtitle_en": "Identify and rewrite 10 unfavorable clauses, then run a cross-conflict audit — revised contract and review reports in one pass", "description": "收到对方起草的合同草稿最头疼的就是一堆「验收标准卖方单方定」「逾期一天付 1% 违约金」「买方不得解约」的条款，读着浑身难受却说不出哪里不对。这个案例以买方（澜石半导体）与钢材供应商的采购合同为样本，站在买方立场逐条识别出 10 处不利条款并全部改写：验收期从 3 日延到 15 日、甲方逾期付款违约金从日 1% 降到日万分之三并设 5% 封顶、乙方逾期交货违约金从日万分之一提到与买方对等、管辖改回买方所在地、解约权与赔偿上限全部双向对等化。改完后又对修订稿做了一轮独立的交叉矛盾核查，发现并处理了 2 处竞合/衔接矛盾，最终输出带风险批注的修订稿、修改要点汇总表、核心条款修改对照表和独立冲突审查报告四份文件。\n\n适合：采购与法务人员、中小企业经营者、合同经办人，以及任何需要站在己方立场审核对方起草合同、把风险压下来的人。", "description_en": "Nothing is more annoying than receiving a counterparty-drafted contract packed with clauses like “acceptance standards determined solely by the seller”, “1% penalty per day for late payment”, or “buyer may not terminate” — it feels wrong but you cannot pinpoint why. This case uses a steel purchase contract signed by a buyer (Lanshi Semiconductor) as the sample. Standing on the buyer's side, it identifies all 10 unfavorable clauses and rewrites every one: acceptance period extended from 3 to 15 days, buyer's late-payment penalty cut from 1%/day to 0.03%/day capped at 5%, seller's late-delivery penalty raised to match, jurisdiction moved back to the buyer's domicile, and termination rights and liability caps made fully reciprocal. After revision, an independent cross-conflict audit was run, surfacing and resolving 2 clashes, producing four deliverables: an annotated revised contract, a revision summary table, a clause-by-clause comparison table, and a standalone conflict audit report.\n\nFor: procurement and legal staff, small-business owners, contract handlers, and anyone who needs to review counterparty-drafted contracts from their own side and push risk down.", "categories": [ "business", "productivity" ], "tags": [ "合同审查", "合同修订", "风险条款", "买方立场", "冲突核查" ], "tags_en": [ "Contract Review", "Contract Revision", "Risk Clause", "Buyer Perspective", "Conflict Audit" ], "artifact_type": "html", "preview": "revision-report.html", "artifacts": [ { "file": "revision-report.html", "label": "修订版合同审查报告（改动标黑+风险批注）" } ], "prompt": "我是「澜石半导体有限公司」的采购负责人，我们作为买方要跟一家钢材供应商签采购合同，对方发来的草稿条款如下（纯文本，无附件）：\n\n货物采购合同（草案）\n甲方（买方）：澜石半导体有限公司\n乙方（卖方）：宏远金属材料有限公司\n\n第一条 标的：乙方向甲方供应螺纹钢、线材、钢板等货物。\n第二条 验收：货物到货后，甲方应在三日内完成验收，逾期未提出异议视为验收合格；验收标准由乙方单方确定。\n第三条 结算：甲方应在到货后六十日内付清全部货款。\n第四条 违约责任：甲方逾期付款的，每逾期一日按合同总额的百分之一支付违约金。\n第五条 违约责任：乙方逾期交货的，每逾期一日按合同总额的万分之一支付违约金，且累计不超过合同总额的百分之三。\n第六条 质量：货物质量标准由乙方按行业惯例自行掌握，甲方不得额外提出异议。\n第七条 管辖：因本合同发生争议，由乙方住所地人民法院管辖。\n第八条 解约：合同期内甲方不得单方解除合同；乙方可因甲方未按期付款单方解除合同。\n第九条 赔偿：因甲方原因造成的损失，甲方承担无限赔偿责任。\n第十条 税费：与本合同相关的全部税费由甲方承担。\n\n请站在我方（买方）立场，对上述合同做全流程处理：\n1. 逐条识别对买方不利的条款，说明风险点与修改措施，并做可追溯改写（改动处标黑加粗+风险批注）；\n2. 对改好的修订稿再做一轮内部交叉矛盾核查，重点检查验收标准单方化、双方违约金不对等、解约权单向、责任上限缺失等，并给出核查统计；\n3. 输出到 ./output/ 目录：修改文档（HTML，改动标黑+风险批注）、修改要点汇总表、核心条款修改对照表、独立的冲突审查报告（HTML，附统计）。\n\n要求：立场始终站在买方；每处改动可追溯（标黑+批注+对照）；合同类型与具体比例可按你的实际合同替换。", "prompt_en": "I am the procurement lead of \"Lanshi Semiconductor Co., Ltd.\" acting as the buyer signing a purchase contract with a steel supplier. The draft clauses the supplier sent are below (plain text, no attachment):\n\nGoods Purchase Contract (Draft)\nParty A (Buyer): Lanshi Semiconductor Co., Ltd.\nParty B (Seller): Hongyuan Metal Materials Co., Ltd.\n\nArticle 1 Subject: Party B supplies rebar, wire rod, steel plate and other goods to Party A.\nArticle 2 Acceptance: Party A shall complete acceptance within 3 days of arrival; failure to object within the period is deemed acceptance; acceptance standards are determined solely by Party B.\nArticle 3 Settlement: Party A shall pay all purchase prices within 60 days after arrival.\nArticle 4 Breach: If Party A pays late, Party A pays a penalty of 1% of the total contract amount per day.\nArticle 5 Breach: If Party B delivers late, Party B pays a penalty of 0.01% of the total contract amount per day, capped at 3% of the total contract amount.\nArticle 6 Quality: Quality standards are at Party B's own discretion per industry practice; Party A may not raise additional objections.\nArticle 7 Jurisdiction: Disputes arising from this contract are governed by the people's court at Party B's domicile.\nArticle 8 Termination: Party A may not unilaterally terminate during the contract term; Party B may unilaterally terminate if Party A fails to pay on time.\nArticle 9 Damages: Losses caused by Party A are borne by Party A with unlimited liability.\nArticle 10 Taxes: All taxes related to this contract are borne by Party A.\n\nPlease process the contract end-to-end from the buyer's (our) side:\n1. Identify every clause unfavorable to the buyer, explain the risk and the fix, and rewrite each change traceably (bold the changed text + add risk annotations);\n2. Run a second-pass internal cross-conflict audit on the revised draft, focusing on unilateral acceptance standards, asymmetric penalties, one-way termination rights, missing liability caps, etc., and provide audit statistics;\n3. Output to ./output/: the revised contract document (HTML, changed text bold + risk annotations), a revision summary table, a core-clause before/after comparison table, and a standalone conflict audit report (HTML, with statistics).\n\nRequirements: always take the buyer's side; every change must be traceable (bold + annotation + comparison); contract type and specific percentages can be replaced with your own actual contract.", "skills": [ { "id": "contract-term-review-revise", "name": "合同条款审查与修订", "description": "代表合同中明确指定的一方，先逐条识别不利条款并做可追溯的修订，再对修订稿进行一轮内部交叉矛盾核查，输出带风险批注的修订文档与独立的冲突审查报告", "skill_id": "skill_2090718439110635520" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-21T05:45:28Z", "updated_at": "2026-08-21T05:51:00Z", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-term-review-revise-3", "title": "Apache 2.0 许可审查与修订（被许可方）", "title_en": "Apache 2.0 License Review & Revision (Licensee)", "subtitle": "站在被许可方立场逐条识别不利条款并做可追溯改写，再对修订稿做一轮交叉矛盾核查，产出带风险批注的修订稿与独立冲突审查报告", "subtitle_en": "Identify unfavorable clauses from the licensee's standpoint, make traceable revisions, then run an internal cross-check for contradictions, delivering an annotated revised draft plus an independent conflict review report", "description": "引入开源组件最怕的就是许可文本里的隐藏风险——免责声明、责任限制、专利报复性终止条款，哪一条都可能让商业使用方吃了亏还没处说理。这个案例演示了把 Apache License 2.0 官方英文全文交给 AI 做全流程处理：先站在被许可方/使用方立场逐条识别 12 项不利与保护缺失的风险点（如第 7 条完全无担保、第 8 条完全免责、第 3 条专利报复性终止、缺失管辖与存续条款），逐条产出可追溯的改写建议并落实到修订稿（改动标黑 + 风险批注），再对修订稿做一轮独立的交叉矛盾核查，发现 5 处问题（2 处条款矛盾、2 处逻辑不一致、1 处笔误）并附修改建议与统计。全套交付 4 份文档：修订稿、修改要点汇总表、核心条款修改对照表、独立冲突审查报告，均自包含 HTML、可直接打开。\n\n适合：公司法务与外部律师做开源许可合规评估、软件研发团队在集成开源组件前做风险预判、合规与供应链安全负责人评估组件引入流程、以及任何需要站在己方立场审阅标准许可协议的人。", "description_en": "The scariest part of adopting an open-source component is the hidden risk buried in the license text — warranty disclaimers, liability limitations, patent retaliation clauses — any of which can quietly leave a commercial user without recourse. This case demonstrates a full-pipeline treatment of the official Apache License 2.0 English text: first, from the licensee/user standpoint, it identifies 12 unfavorable or missing-protection risk points (e.g., Section 7's total lack of warranty, Section 8's complete liability disclaimer, Section 3's patent retaliation termination, and missing governing-law and vesting clauses), produces traceable rewrite suggestions implemented in a revised draft (changes bolded, with risk annotations), then runs an independent cross-check on the revised draft that surfaces 5 issues (2 clause conflicts, 2 logic inconsistencies, 1 typo) with fixes and statistics. The full deliverable set includes 4 documents: the revised draft, a modification summary table, a core-clause comparison table, and an independent conflict review report — all self-contained HTML that opens directly.\n\nIdeal for: in-house counsel and external law firms assessing open-source license compliance, R&D teams pre-screening risk before integrating components, compliance and supply-chain security leads evaluating component intake processes, and anyone who needs to review standard license agreements from their own side's standpoint.", "categories": [ "business" ], "tags": [ "合同审查", "开源许可", "Apache 2.0", "风险条款", "法律合规" ], "tags_en": [ "Contract Review", "Open Source License", "Apache 2.0", "Risk Clause", "Legal Compliance" ], "artifact_type": "html", "preview": "revised-license-agreement.html", "prompt": "我是「北岸律师事务所」代理的一家企业客户（作为被许可方/使用方），客户准备在自研软件里引入 Apache License 2.0 许可的开源组件，想请律师先评估这份许可文本对它是否有利、有没有隐藏风险。请站在被许可方（使用方）立场，对 Apache License 2.0 许可协议文本做全流程处理：\n\n1. 逐条识别对被许可方不利、义务加重、保护缺失的条款，并做可追溯的改写建议；\n2. 改完后再对改好的稿子做一轮内部交叉矛盾核查（条款之间是否冲突、逻辑不一致、文字笔误）。\n\n请自行下载下述公开的 Apache License 2.0 英文原文后审阅：\n- 许可协议文本 URL（Apache 官网直链，txt 全文）：https://www.apache.org/licenses/LICENSE-2.0.txt\n\n直接产出全套终稿到 ./output/：修改文档（HTML，改动标黑+风险批注）、修改要点汇总表、核心条款修改对照表、独立的冲突审查报告（HTML，附统计），并在报告开头注明所引用的许可协议版本与来源链接。\n\n说明：输入为 Apache 官网公开的许可协议全文（txt，可 curl 直链下载），无登录墙、无版权风险；被许可方主体为虚构企业，律师为「北岸律师事务所」；一次性跑完，中途无需确认；如 URL 无法访问可改用 Apache 官网同页面镜像，并在报告中说明来源。", "prompt_en": "I am a corporate client represented by \"Beian Law Firm\" (as Licensee/User) that plans to integrate open-source components licensed under Apache License 2.0 into our self-developed software, and would like counsel to first assess whether this license text is favorable to us and whether it contains hidden risks. Please process the Apache License 2.0 license text end-to-end from the standpoint of the Licensee (User):\n\n1. Identify, clause by clause, the provisions that are unfavorable to the Licensee, that add obligations, or that lack protection, and provide traceable rewrite recommendations;\n2. After revising, run an internal cross-check for contradictions in the revised draft (conflicts between clauses, logic inconsistencies, and textual typos).\n\nPlease download the public Apache License 2.0 English original text below and review it:\n- License text URL (Apache official direct link, full txt): https://www.apache.org/licenses/LICENSE-2.0.txt\n\nProduce the complete final deliverables in ./output/: the revised document (HTML, with changes bolded and risk annotations), a modification summary table, a core-clause comparison table, and an independent conflict review report (HTML, with statistics), and note the referenced license version and source link at the beginning of the report.\n\nNotes: the input is the public license text published on the Apache website (txt, downloadable via curl direct link), no login wall, no copyright risk; the licensee is a fictional company and the law firm is \"Beian Law Firm\"; run it in one pass without mid-way confirmation; if the URL is unreachable, you may use the mirror of the same Apache page and state the source in the report.", "skills": [ { "id": "contract-term-review-revise", "name": "合同条款审查与修订", "description": "代表合同中明确指定的一方，先逐条识别不利条款并做可追溯的修订，再对修订稿进行一轮内部交叉矛盾核查，输出带风险批注的修订文档与独立的冲突审查报告", "skill_id": "skill_2090718439110635520" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-08-21T05:46:25Z", "updated_at": "2026-08-21T05:52:00Z", "artifacts": [ { "file": "revised-license-agreement.html", "label": "修订稿（改动标黑 + 风险批注）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "copyright-similarity-analysis", "title": "著作权侵权实质性相似分析", "title_en": "Copyright Substantial Similarity Analysis", "subtitle": "三层剥离 + 接触因果链，一键厘清「像不像」和「抄没抄」", "subtitle_en": "Three-layer stripping + access causal chain to untangle similarity and access", "description": "著作权侵权案子最头疼两件事：一是「像不像」说不清——相似点到底是公共领域、必然场景还是独创表达，混成一团；二是「抄没抄」证不了——实质性相似和接触可能性是两条独立的腿，缺一条都不成立。\n\n这个案例演示了把一篇品牌文案的抄袭争议交给 AI，自动完成思想表达二分法剥离、三层独创性剥离、逐层比对（含三性分类与错误复制识别）、接触可能性因果链评估、分权论证与被告抗辩预判，产出一份可直接作为律师工作底稿的 HTML 报告。\n\n适合：知识产权律师、版权代理人、企业法务，处理著作权侵权诉讼、发函前侵权评估、抗辩准备。", "description_en": "Copyright infringement cases hit two pain points: first, \"similar or not\" is hard to pin down — whether a similarity lies in public domain, scenes à faire, or original expression gets muddled; second, \"copied or not\" is hard to prove — substantial similarity and access are two independent legs, and both must hold.\n\nThis case shows how to hand a brand copywriting plagiarism dispute to AI, which automatically completes idea-expression dichotomy stripping, three-layer originality stripping, layer-by-layer comparison (with three-nature classification and error-copy detection), access causal-chain assessment, right-by-right argumentation, and defense preview, producing an HTML report ready to use as a lawyer's working draft.\n\nFor: IP lawyers, copyright agents, in-house counsel handling copyright litigation, pre-notice infringement assessment, and defense preparation.", "categories": [ "productivity" ], "tags": [ "著作权侵权", "实质性相似", "接触可能性", "思想表达二分法", "律师工具" ], "tags_en": [ "copyright infringement", "substantial similarity", "access", "idea-expression dichotomy", "legal tool" ], "artifact_type": "html", "preview": "similarity-analysis-report.html", "artifacts": [ { "file": "similarity-analysis-report.html", "label": "著作权侵权实质性相似分析报告" } ], "prompt": "请对以下文字作品进行著作权侵权实质性相似分析，输出自包含 HTML 分析报告。\n\n【作品原文提交要求（必要条件）】\n请提交原告作品与被控侵权作品的「原始全文」，而非对作品内容的文字描述或概括。分析必须严格以两份作品原文的逐字逐句比对为基础展开，确保结论可被独立复核。\n\n——原告作品原文（品牌软文《山谷里的咖啡馆》，2024年3月首发于品牌公众号）：\n清晨六点，山雾还没有散尽，我把一壶清水坐到炉火上。\n我先把咖啡豆倒进烘焙机，看它们在转笼里慢慢镀上一层琥珀色；再把豆子磨成粉，让焦糖与雪松的香气溢满整间木屋；最后把滚水一圈一圈注进粉床，看深褐色的液体顺着滤纸，一滴一滴落进壶里。\n这就是把清晨煮成一壶咖啡——山里的天光、雾气、还有第一声鸟鸣，都被我一点点收进了这只壶里。\n咖啡煮好了。我推开门，山谷正从夜色里醒过来。山风穿过松林，吹动屋檐下那串风铃。\n我想，山风记得每一粒豆子的名字。\n\n——被控侵权作品原文（竞品推广文案《山间咖啡时光》，2024年11月发布）：\n清晨六点，雾气还没散，我把一壶山泉放到炉子上。\n我先把生豆倒进烘豆机，看它们在滚筒里慢慢裹上一层焦糖色；再把豆子碾成粉，让焦糖与松木的香气漫过整间小屋；最后把沸水一圈一圈淋进粉层，看棕色的液流顺着滤杯，一滴一滴坠进壶中。\n这就是把清晨煮成一壶咖啡——山里的晨光、薄雾、还有第一声鸟鸣，都被我一点点收进了这只壶里。\n咖啡煮好了。我推开木门，山间正从黎明里醒过来。山风穿过松林，拨动屋檐下那串风铃。\n我想，山风记得每一颗豆子的名字。\n\n[注意] 如仅能提供作品名称、主题、结构或核心比喻等描述性信息而无法提供原文全文，分析结论将被明确标注为「条件性/参考性结论」，仅具方向提示价值，不具独立证据价值，不得直接作为诉讼或发函的事实依据。\n\n【接触证据提交要求】\n请提供接触可能性相关的具体证据材料（如联合推广合作协议/合同、业务往来记录、公开发表时间线及传播范围佐证、同城同业关系说明等）。\n\n——本案已掌握的接触线索：原告2024年3月先发表（公众号10万+粉丝、全网分发），被告为同城咖啡品牌竞品，双方曾有联合推广合作历史。\n\n[注意] 如无法提供上述接触证据材料，接触可能性评估将按「最低标准」进行（仅依据公开发表时间先后作基础推定），并在报告中明确标注该结论的证明力局限及需进一步补强的证据方向。\n\n涉诉权利：复制权、信息网络传播权、保护作品完整权。\n\n【分析任务要求】\n请严格按以下步骤完成分析：\n1. 按思想表达二分法完成三层独创性剥离（公共领域 → Scenes à faire → 独创性表达）；\n2. 对独创性表达层逐层比对，标注每个相似点的三性分类（共通表达 / 共同选择 / 共同错误）；\n3. 识别错误复制（本案提示：两份文本结尾句在量词上存在「同位置、近同、一字之差」的瑕疵——原告用「粒」、被告用「颗」，需在比对中单独标注该一致性瑕疵）；\n4. 评估接触可能性因果链四阶；\n5. 输出三阶交互判定（情形 A/B/C）与分权论证（逐权分析复制权、信息网络传播权、保护作品完整权）；\n6. 预判被告抗辩路径（含合法来源抗辩、委托创作抗辩、思想表达二分法抗辩、公有领域抗辩等），并给出应对要点。\n\n输出要求：自包含 HTML 报告（内联样式、无外部资源引用），可直接作为律师工作底稿使用。", "prompt_en": "Please perform a copyright substantial similarity analysis on the following literary work and output a self-contained HTML report.\n\n[Original-text submission requirement (mandatory)]\nPlease submit the ORIGINAL FULL TEXT of both the plaintiff's work and the accused infringing work, not a textual description or summary of the content. The analysis must be based strictly on a word-by-word, sentence-by-sentence comparison of the two original texts so that the conclusions can be independently verified.\n\n-- Plaintiff's original text (brand copy \"The Café in the Valley\", first published on the brand's official account in March 2024):\n\n\n-- Accused infringing original text (competitor's promotional copy \"Mountain Coffee Time\", published in November 2024):\n\n\n[注意] If only descriptive information (title, theme, structure, core metaphor, etc.) can be provided and the original full text is unavailable, the analysis conclusions will be explicitly marked as \"conditional / reference-only conclusions\", carrying only directional value and no independent evidentiary value, and must not be used directly as the factual basis for litigation or demand letters.\n\n[Access-evidence submission requirement]\nPlease provide specific evidence materials related to access (e.g., joint-promotion cooperation agreement/contract, business correspondence records, publication timeline with dissemination-scope support, same-city same-industry relationship, etc.).\n\n-- Access clues already available in this case: the plaintiff published first in March 2024 (official account with 100k+ followers, distributed across platforms); the defendant is a competing café brand in the same city, and the two parties have a history of joint promotion.\n\n[注意] If the above access evidence cannot be provided, the access assessment will be performed at the \"minimum standard\" (a basic presumption based only on the sequence of public publication), and the report will explicitly mark the probative-value limitations of that conclusion and the evidence directions that need to be further strengthened.\n\nClaimed rights: reproduction right, right of communication through information network, right of integrity.\n\nPlease strictly follow the idea-expression dichotomy to perform three-layer originality stripping (public domain → scenes à faire → original expression), compare the original expression layer by layer and label the three-nature classification of each similarity point (common expression / common choice / common error), detect error copying, assess the four-step access causal chain, output the three-tier interaction judgment (Case A/B/C) and right-by-right argumentation, and preview the defendant's defense paths.", "skills": [], "experts": [ { "id": "ip-team-lead", "name": "知识产权专家团", "description": "覆盖知识产权条款专项审查、权利归属与授权许可分析、侵权担保与责任评估、权利链风险识别与修改建议的知识产权法律专家团队", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "1bcf88d993f1", "expert_id": "ex_4IBuGIfQnBYT" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-09-07T11:56:00+08:00", "updated_at": "2026-09-07T11:56:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-inventory", "title": "库存风险预警：未来7天先处理谁", "title_en": "Which inventory risks to handle first in the next 7 days", "subtitle": "未来7天库存风险，先处理谁", "subtitle_en": "Which inventory risks to handle first in the next 7 days", "description": "面向成本/供应链场景：未来7天库存风险，先处理谁。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：未来7天库存风险，先处理谁。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "库存", "补货", "周转", "存货健康" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请预测 {组织范围} 未来 {天数} 天原料库存积压与短缺风险，给出处理优先级。\n1) 区分日配正常偏离、活动积压、停售残留、临期批次与短缺等场景\n2) 每条给出诊断、今天动作、责任人\n3) 标明已盘用实际/未盘用理论口径\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请预测 {组织范围} 未来 {天数} 天原料库存积压与短缺风险，给出处理优先级。\n1) 区分日配正常偏离、活动积压、停售残留、临期批次与短缺等场景\n2) 每条给出诊断、今天动作、责任人\n3) 标明已盘用实际/未盘用理论口径\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087437869308108800", "skill_id": "skill_2087437869308108800", "name": "库存风险预警", "description": "库存风险预警" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-margin", "title": "毛利与分差诊断：毛利掉了，分差出在哪一环", "title_en": "Where margin drops and variance comes from", "subtitle": "毛利掉了，分差出在哪一环", "subtitle_en": "Where margin drops and variance comes from", "description": "面向成本/供应链场景：毛利掉了，分差出在哪一环。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：毛利掉了，分差出在哪一环。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "毛利", "分差", "归因" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {门店范围} 在 {统计区间} 的毛利与原料分差异常。\n1) 定位高影响物料与责任环节\n2) 拆分差来源（采购/耗用/盘点/配方）\n3) 输出今日动作与复核项\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {门店范围} 在 {统计区间} 的毛利与原料分差异常。\n1) 定位高影响物料与责任环节\n2) 拆分差来源（采购/耗用/盘点/配方）\n3) 输出今日动作与复核项\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087436938187866112", "skill_id": "skill_2087436938187866112", "name": "毛利与分差诊断", "description": "毛利与原料分差诊断" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-price", "title": "食材价格波动：本周食材涨价影响哪些菜", "title_en": "Which dishes are hit by this week's price rises", "subtitle": "本周食材涨价影响哪些菜", "subtitle_en": "Which dishes are hit by this week's price rises", "description": "面向成本/供应链场景：本周食材涨价影响哪些菜。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：本周食材涨价影响哪些菜。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "食材涨价", "采购价", "成本冲击" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {组织范围} 在 {本周} 采购价涨跌最大的食材及其对菜品成本的影响。\n1) Top 涨价/跌价食材\n2) 受影响菜品与毛利冲击估算\n3) 给出调价/换料/限售建议\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {组织范围} 在 {本周} 采购价涨跌最大的食材及其对菜品成本的影响。\n1) Top 涨价/跌价食材\n2) 受影响菜品与毛利冲击估算\n3) 给出调价/换料/限售建议\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087437670490214400", "skill_id": "skill_2087437670490214400", "name": "食材价格波动", "description": "食材价格波动分析" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-purchase", "title": "采购调拨与盘点：采购调拨盘点，账实对得上吗", "title_en": "Do purchase, transfer and stocktake books match", "subtitle": "采购调拨盘点，账实对得上吗", "subtitle_en": "Do purchase, transfer and stocktake books match", "description": "面向成本/供应链场景：采购调拨盘点，账实对得上吗。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：采购调拨盘点，账实对得上吗。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "采购", "调拨", "盘点", "账实" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请检查 {组织范围} 在 {统计区间} 的采购、调拨、盘点是否存在异常波动或账实差异。\n1) 列出嫌疑单据类型与单号级线索\n2) 说明可能错误（单位/日期/档口/品项）\n3) 给出人工确认清单（不自动改账）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请检查 {组织范围} 在 {统计区间} 的采购、调拨、盘点是否存在异常波动或账实差异。\n1) 列出嫌疑单据类型与单号级线索\n2) 说明可能错误（单位/日期/档口/品项）\n3) 给出人工确认清单（不自动改账）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087436917139619840", "skill_id": "skill_2087436917139619840", "name": "采购调拨与盘点", "description": "采购调拨盘点检查" }, { "id": "skill_2087437273711214592", "skill_id": "skill_2087437273711214592", "name": "供应链报表", "description": "供应链综合报表查询" }, { "id": "skill_2091713208272846848", "skill_id": "skill_2091713208272846848", "name": "supply-chain-purchase-fulfillment", "description": "查询采购计划统计与外销订货履约" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-recipe", "title": "菜品成本卡异常：成本卡缺了、毛利负了怎么办", "title_en": "Fix missing cost cards and negative margins", "subtitle": "成本卡缺了、毛利负了怎么办", "subtitle_en": "Fix missing cost cards and negative margins", "description": "面向成本/供应链场景：成本卡缺了、毛利负了怎么办。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：成本卡缺了、毛利负了怎么办。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "成本卡", "理论成本", "负毛利" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请找出 {组织范围} 理论成本缺失、毛利为负或成本异常的菜品，并分析原料原因。\n1) 列出异常菜品清单与影响\n2) 归因到成本卡缺失/配方偏差/原料价\n3) 给出修复动作与责任岗位\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请找出 {组织范围} 理论成本缺失、毛利为负或成本异常的菜品，并分析原料原因。\n1) 列出异常菜品清单与影响\n2) 归因到成本卡缺失/配方偏差/原料价\n3) 给出修复动作与责任岗位\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087435637936848896", "skill_id": "skill_2087435637936848896", "name": "菜品成本卡异常", "description": "菜品成本卡异常诊断" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cost-report", "title": "供应链四类报表：库存盘点采购毛利，一句话问清", "title_en": "Ask inventory, stocktake, purchase and margin in one go", "subtitle": "库存盘点采购毛利，一句话问清", "subtitle_en": "Ask inventory, stocktake, purchase and margin in one go", "description": "面向成本/供应链场景：库存盘点采购毛利，一句话问清。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向成本/供应链场景：库存盘点采购毛利，一句话问清。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "供应链", "报表问数", "库存", "毛利" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请查询 {组织范围} 在 {统计月份} 的库存、盘点、调拨、采购、毛利和原料分差情况。\n1) 给出核心数字与异常亮点\n2) 标出高影响门店/品项\n3) 输出今日优先核查三件事\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请查询 {组织范围} 在 {统计月份} 的库存、盘点、调拨、采购、毛利和原料分差情况。\n1) 给出核心数字与异常亮点\n2) 标出高影响门店/品项\n3) 输出今日优先核查三件事\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087437273711214592", "skill_id": "skill_2087437273711214592", "name": "供应链报表", "description": "供应链综合报表查询" }, { "id": "skill_2087436938187866112", "skill_id": "skill_2087436938187866112", "name": "毛利与分差诊断", "description": "毛利与原料分差诊断" }, { "id": "skill_2087436917139619840", "skill_id": "skill_2087436917139619840", "name": "采购调拨与盘点", "description": "采购调拨盘点检查" }, { "id": "skill_2091713208272846848", "skill_id": "skill_2091713208272846848", "name": "supply-chain-purchase-fulfillment", "description": "查询采购计划统计与外销订货履约" } ], "experts": [ { "id": "weimob-cost-supply-chain-expert", "expert_id": "ex_yMGUEddylKGY", "name": "微盟成本供应链专家", "description": "微盟成本与供应链分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "coupon-redemption-review", "title": "优惠券整体核销复盘", "title_en": "Coupon Redemption Review", "subtitle": "优惠券领取、使用与核销表现复盘", "subtitle_en": "A practical Weimob playbook for coupon redemption review", "description": "做优惠券整体核销复盘时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“优惠券整体核销复盘”展开：复盘指定时间范围内优惠券的领取、使用和核销表现，识别高核销与低核销券，分析门槛、面额和有效期的影响，并提出下一轮优化建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、沉睡与流失挽回运营人员、经营分析团队。", "description_en": "The hardest part of coupon redemption review is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates coupon redemption review through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Dormancy and Churn Recovery specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "优惠券复盘", "领取情况", "使用情况", "核销表现", "券策略" ], "tags_en": [ "Coupon Redemption Review", "Customer Operations", "Dormancy and Churn Recovery", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我复盘指定时间范围内优惠券的领取、使用和核销表现，识别高核销与低核销券，分析门槛、面额和有效期可能带来的影响，并提出下一轮优化建议；如果我还没有指定时间范围，请先向我确认。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a coupon redemption review covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" }, { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "course-learning-compare", "title": "课程学习情况对比分析", "title_en": "Course Learning Performance Comparison", "subtitle": "两个课程的用户学习数据对比、统计与排名，生成可视化 HTML 报告", "subtitle_en": "Compare, aggregate and rank user learning data across two courses, rendered as a visual HTML report", "description": "同时开了多门课，却说不清哪门课学的人多、哪门课完课率高，逐个课程后台点开看数据又慢又乱。\n\n这个案例演示了如何把一个课程学习对比分析交给小鹅数据分析师：它自动查两个课程（test11、test22）的真实学习数据——累计用户数、学习人数、完课率、支付人数、订阅人数，做完横向对比、统计和排名后，生成一份可直接打开的可视化 HTML 报告，指标差异一目了然。\n\n适合：知识付费运营负责人、课程产品经理、在线教育机构运营、需要定期做课程效果复盘的内容操盘手。", "description_en": "Running multiple courses but can't tell at a glance which one attracts more learners or achieves a higher completion rate? Checking each course's dashboard one by one is slow and messy.\n\nThis case shows how to hand a course-learning comparison off to Xiaoe Data Analyst: it queries real learning data for two courses (test11, test22) — cumulative users, active learners, completion rate, payers, subscribers — then runs a side-by-side comparison, aggregation and ranking, and renders a shareable visual HTML report where metric differences pop out instantly.\n\nGreat for: knowledge-payment operations leads, course product managers, online education operators, and content operators who need routine course performance reviews.", "categories": [ "data" ], "tags": [ "课程分析", "学习数据", "数据对比", "数据可视化", "小鹅通" ], "tags_en": [ "Course Analysis", "Learning Data", "Data Comparison", "Data Visualization", "Xiaoe" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "对比分析一下课程「test11」和「test22」的用户学习情况，并做统计和排名，生成一个 HTML 页面。\n\n要求：\n1. 通过小鹅通数据分析能力分别查询这两个课程的真实学习数据（先按课程名搜索拿到 resource_id，再查学习概况指标：累计用户数、学习人数、完课数、完课率、支付人数、订阅人数）。\n2. 对两个课程做横向对比，输出关键指标的对比表和差异说明。\n3. 基于学习人数、支付人数等指标做统计与排名（说明排名口径）。\n4. 生成一个 HTML 可视化报告，用图表直观呈现对比结论，并附上数据来源与统计口径说明。", "prompt_en": "Compare and analyze the user learning data of courses \"test11\" and \"test22\", run statistics and ranking, and generate an HTML page.\n\nRequirements:\n1. Use the Xiaoe data-analysis capability to query the real learning data of these two courses (first search by course name to get resource_id, then query the learning overview metrics: cumulative users, active learners, completion count, completion rate, payers, subscribers).\n2. Compare the two courses side by side, output a comparison table of key metrics and difference notes.\n3. Rank and aggregate based on metrics such as learner count and payer count (state the ranking methodology).\n4. Generate an HTML visual report, present the comparison with charts, and include data source and scope notes.", "experts": [ { "id": "xiaoe-data-analyst", "name": "小鹅数据分析师", "description": "查询小鹅通平台直播、课程、交易等多维经营数据，组合分析并导出报表，用数据支撑商家决策", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_1M2LsZX1RbK9" } ], "skills": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:05:00+08:00", "updated_at": "2026-08-26T19:15:00+08:00", "artifacts": [ { "file": "output.html", "label": "课程学习对比分析报告（HTML 单页）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "criminal-defense-strategy-analysis", "title": "受贿案辩护策略分析", "title_en": "Bribery Case Defense Strategy Analysis", "subtitle": "审查起诉阶段三笔指控的分层攻防策略，一份报告看清无罪/罪轻/量刑三层辩护路径", "subtitle_en": "A layered defense strategy for three bribery charges at the prosecution stage — one report mapping out all three defense levels", "description": "接手刑事案件最头疼的就是拿不准辩护方向——认罪还是辩无罪？逐笔还是整体？一个选错就可能错失从宽机会。\n\n这个案例演示了如何把一起涉嫌受贿87万、共3笔指控的案件交给 AI，自动拆解出「分层攻防」策略：第1笔自首退赃做量刑之辩、第2笔有合同凭证做定性之辩、第3笔仅单一证言做事实之辩。产出一份带执行摘要、策略对比表、风险矩阵图、行动时间轴的完整辩护策略分析报告。\n\n适合：刑事辩护律师、法律顾问、律所合伙人，以及需要系统梳理案件辩护思路的法律从业者。", "description_en": "The hardest part of a criminal case is picking the right defense direction — plead guilty or claim innocence, per-charge or overall? One wrong call can cost you a lighter sentence.\n\nThis case shows how to hand a suspected bribery case (870,000 yuan across 3 charges) to AI, which automatically breaks it into a layered defense strategy: sentencing defense for the self-surrender charge, re-qualification defense for the contract-backed charge, and factual defense for the single-witness charge. It produces a complete defense strategy report with an executive summary, strategy comparison table, risk matrix, and action timeline.\n\nFor: criminal defense lawyers, legal counsel, law firm partners, and legal professionals who need to systematically map out a case defense approach.", "categories": [ "office" ], "tags": [ "刑事辩护", "策略分析", "受贿罪", "法律文书", "辩护策略" ], "tags_en": [ "Criminal Defense", "Strategy Analysis", "Bribery", "Legal Document", "Defense Strategy" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "受贿案辩护策略分析报告（HTML）" } ], "prompt": "请帮我分析这个刑事案件的辩护策略。\n\n王某涉嫌受贿罪，目前处于审查起诉阶段，涉嫌受贿87万元，共3笔：\n- 第1笔20万元：已自首并全额退赃\n- 第2笔35万元：辩称系正常业务往来，有合同和付款凭证\n- 第3笔32万元：仅有行贿人证言，证人与行贿方有利害关系\n\n证据情况：行贿人证言、银行流水、扣押笔记本。\n\n请制定本案辩护策略方案，重点分析第2笔的定性争议和第3笔的证据问题，输出一份完整的策略分析报告（HTML），包含执行摘要、案件事实评估、证据评估、法律适用评估、策略选项对比、推荐方案、阶段行动计划和风险提示。", "prompt_en": "Please analyze the defense strategy for this criminal case.\n\nWang is suspected of bribery and is currently at the prosecution review stage, accused of accepting 870,000 yuan across 3 charges:\n- Charge 1: 200,000 yuan — already surrendered and fully returned the money\n- Charge 2: 350,000 yuan — claims it was normal business dealings, backed by a contract and payment vouchers\n- Charge 3: 320,000 yuan — supported only by the briber's testimony, and the witness has a conflict of interest with the briber\n\nEvidence: the briber's testimony, bank transaction records, and a seized notebook.\n\nPlease draft a defense strategy for this case, focusing on the re-qualification dispute in Charge 2 and the evidentiary issues in Charge 3, and output a complete strategy analysis report (HTML) including an executive summary, case facts assessment, evidence assessment, legal application assessment, strategy comparison, recommended approach, staged action plan, and risk warnings.", "skills": [], "experts": [ { "id": "criminal-defense-team-lead", "name": "法大大·睿契提供的刑事辩护专家", "description": "面向刑事辩护律师，覆盖侦查、审查起诉与审判全流程的辩护策略制定、证据审查、定性争议与量刑辩护，内嵌执业安全红线", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "c2cecdf5e533", "expert_id": "ex_WCFaRFwJ3w2D" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "community", "usage_count": 0, "created_at": "2026-09-07T11:44:47+08:00", "updated_at": "2026-09-07T11:44:47+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "css-animation-loading", "title": "CSS 加载动画合集", "title_en": "CSS Loading Animation Collection", "subtitle": "8 种创意 Loading 动效，纯 CSS 即拿即用", "subtitle_en": "8 creative loading animations in pure CSS, ready to copy", "description": "每次做加载动画都要现写或找第三方库，要么效果太简陋要么引入太重，还得担心兼容性和性能问题，反复调参数浪费大量时间。\n\n这个案例用 web-animation-design 技能一次性生成了 8 种风格各异的纯 CSS 加载动画：旋转圆环（渐变边框）、脉冲圆点（三点缩放序列）、弹跳柱条（五柱交错弹跳）、波纹扩散（涟漪扩展圆环）、渐变旋转器（锥形渐变旋转）、打字指示器（聊天式输入动效）、骨架屏微光（内容占位闪烁）、DNA 双螺旋（双链旋转动画）。每个动画都封装在独立卡片中，深色背景突出动效，支持一键复制 CSS 代码，并尊重 prefers-reduced-motion 媒体查询。\n\n适合：前端开发者、UI 设计师、独立开发者、任何需要快速获取加载动画的人", "description_en": "Writing loading animations from scratch or hunting for the right library every time is a pain — results are either too plain or too heavy, with compatibility and performance concerns eating up your time.\n\nThis case uses the web-animation-design skill to generate 8 distinct pure CSS loading animations in one go: Spinning Ring (gradient border), Pulse Dots (3-dot scaling sequence), Bouncing Bars (5-bar staggered bounce), Wave Circle (ripple expanding rings), Gradient Spinner (conic-gradient rotation), Typing Dots (chat-style typing indicator), Skeleton Shimmer (content placeholder shimmer), and DNA Helix (double helix rotation). Each animation is presented in its own dark-background card with live preview, one-click CSS copy, and respects prefers-reduced-motion.\n\nIdeal for: frontend developers, UI designers, indie developers, anyone who needs ready-to-use loading animations", "categories": [ "creation", "dev" ], "tags": [ "CSS动画", "Loading", "前端开发", "动效设计", "Web动画" ], "tags_en": [ "CSS Animation", "Loading", "Frontend Dev", "Motion Design", "Web Animation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个纯 CSS 加载动画合集展示页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，深色卡片网格布局。\n\n包含 8 种加载动画：旋转圆环、脉冲圆点、弹跳柱条、波纹圆圈、渐变锥形旋转器、打字指示器、骨架屏微光、DNA 双螺旋。\n\n每个动画放独立卡片，显示实时预览、名称、简短描述和「复制 CSS」按钮。动画仅用 CSS keyframes + transform 实现，尊重 prefers-reduced-motion 降级。", "prompt_en": "Please generate a pure-CSS loading-animation collection showcase page (single-file HTML, inlined with no external dependencies). UI text should be in Simplified Chinese, with a dark card-grid layout.\n\nInclude 8 loading animations: spinning ring, pulsing dots, bouncing bars, ripple circles, gradient conic spinner, typing indicator, skeleton-screen shimmer, DNA double helix.\n\nPlace each animation in its own card, showing a live preview, the name, a short description, and a \"Copy CSS\" button. Animations must be implemented using only CSS keyframes + transform, and should gracefully degrade in accordance with prefers-reduced-motion.", "skills": [ { "id": "web-animation-design", "name": "Web 动画设计", "description": "设计和实现自然流畅的 Web 动画效果", "slug": "web-animation-design" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 11, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-22T09:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "CSS 加载动画" } ], "is_recommended": true, "recommended_order": 20 }, { "id": "customer-focus-decision", "title": "客户聚焦与拒绝决策", "title_en": "Customer Focus & Rejection Decisions", "subtitle": "在现金焦虑中守住第一个客户选择", "subtitle_en": "Protect the First Customer Choice Under Cash Pressure", "categories": [ "business", "education" ], "tags": [ "客户聚焦", "拒绝决策", "B2B SaaS", "GTM" ], "tags_en": [ "Customer Focus", "Rejection Decisions", "B2B SaaS", "GTM" ], "description": "痛点：8 人 B2B SaaS 团队在两个月内收到零售、制造、教育、物流四行业共 7 个定制需求，其中 3 个愿意付费但功能各不相同。创始人担心拒单会失去现金，团队已被定制拖慢，客户、产品与销售判断正在被拉散。\n\n演示：案例把讨论带回顺序——先把客户还原为具体的人、具体场景和具体痛点，再判断从哪个入口进入市场，最后用时间、现金、创始人精力与团队聚焦度约束哪些需求现在不能接。\n\n适用人群：适用于早期 B2B SaaS 创始人、产品负责人和小团队管理者，尤其是在付费机会分散、定制持续挤占核心产品时，需要做客户选择与拒绝决策的团队。", "description_en": "Pain point: An eight-person B2B SaaS team received seven custom requests across retail, manufacturing, education, and logistics in two months. Three are willing to pay but require different features. The founder fears losing cash by declining, while custom work has already slowed the team and scattered customer, product, and sales judgment.\n\nDemonstration: The case brings the discussion back to sequence—first restore the customer to a specific person, situation, and pain; then decide which entrance to use for the market; finally use time, cash, founder attention, and team focus to constrain which requests cannot be accepted now.\n\nWho it is for: Early-stage B2B SaaS founders, product leads, and small-team managers whose paid opportunities are fragmented and whose custom work is crowding out the core product, and who need to make customer-selection and rejection decisions.", "prompt": "你是一家 8 人 B2B SaaS 团队的创始人。过去两个月，团队收到来自零售、制造、教育、物流四个行业的 7 个定制需求，其中 3 个客户愿意付费，但所需功能各不相同。你担心拒绝会失去现金，团队却已被定制拖慢。请完成一次客户聚焦与拒绝决策：围绕“客户不是所有人”“拒绝是对有限资源的尊重”“先明确具体痛点再讨论产品和 GTM”展开。先用一句话给出核心判断；随后具体解释如何把客户从行业标签还原为具体的人、场景与痛点，为什么不应只按眼前收入比较需求，如何判断哪个入口能让客户、GTM、产品与商业结构更清楚，以及哪些定制现在不能接；最后以 1—2 个针对团队处境的具体反问结束。输出逐段内容，不编造任何外部数据或事实。", "prompt_en": "You are the founder of an eight-person B2B SaaS team. Over the past two months, the team has received seven custom requests from retail, manufacturing, education, and logistics. Three customers are willing to pay, but each needs different features. You fear that declining will lose cash, yet custom work has already slowed the team. Complete a customer-focus and rejection decision centered on these ideas: the customer is not everyone; rejection respects limited resources; and a specific pain must be clear before discussing product and GTM. First give the core judgment in one sentence; then explain specifically how to restore a customer from an industry label to a specific person, situation, and pain; why requests should not be compared only by immediate revenue; how to judge which entrance makes the customer, GTM, product, and business structure clearer; and which custom work cannot be accepted now. Finally end with one or two concrete questions for the team’s situation. Output paragraph-by-paragraph content without inventing any external data or facts.", "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "中英双语客户聚焦决策工作台" } ], "skills": [ { "id": "the-entrepreneurship-handbook", "name": "创业可以学", "description": "服务创业者和管理者，解答创业/商业/管理问题，引发深度思考", "skill_id": "skill_2053083248287678464" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-07-27T16:30:16+08:00", "updated_at": "2026-07-27T16:30:16+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "cyberfit-exercise-tracker", "title": "CyberFit 赛博朋克运动打卡工作台", "title_en": "CyberFit - Cyberpunk Exercise Tracker", "subtitle": "纯静态单体 HTML 运动打卡网站，赛博朋克霓虹风格，支持打卡/统计/图表", "subtitle_en": "Standalone static HTML exercise tracker with cyberpunk neon theme, check-in, analytics and Chart.js dashboards", "description": "厌倦了臃肿的运动 App？这个纯静态 HTML 工作台只需浏览器打开就能用，无需安装、无需后端。\n\n包含四个核心页面：首页显示连续打卡天数和本周概览，打卡页支持 8 种运动类型和感受记录，数据看板用 Chart.js 呈现周趋势和类型分布，历史记录支持筛选/编辑/删除。所有数据存于浏览器 localStorage，支持 JSON 导出导入。PC 端侧边栏导航，移动端底部 Tab 栏 + 滑出菜单。\n\n支持中英双语切换：访问时加 ?lang=en 即可切换为英文界面。适合想快速记录运动、不想要繁重 App 的个人用户；也适合前端开发者学习纯 HTML/CSS/JS SPA 架构、Chart.js 集成和前端 i18n 实践的实战参考。", "description_en": "Tired of bloated fitness apps? This standalone HTML workspace works with just a browser—no installation, no backend.\n\nFour pages: Home with streak tracking and weekly summary, Check-in with 8 exercise types and mood recording, Dashboard with Chart.js trends and distributions, History with filtering, editing and deletion. All data stored in browser localStorage, with JSON export/import support. PC sidebar navigation, mobile bottom tabs + slide-out menu.\n\nBuilt-in i18n: add ?lang=en to the URL to switch the entire interface to English. Perfect for individuals who want quick exercise logging without heavy apps, and for developers learning pure HTML/CSS/JS SPA architecture, Chart.js integration, and frontend i18n practices.", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/aggM27DD6gtosrTAoWybqC），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n生成一个纯静态运动打卡网站（HTML/CSS/JS），支持PC/移动端响应式。要求：1）多页面结构（首页、打卡页、数据看板、历史记录）；2）UI赛博朋克/霓虹风格，带粒子背景和CSS入场动画；3）使用localStorage存储数据，集成Chart.js展示运动图表；4）移动端底部Tab栏，PC端侧边栏；5）所有到单个HTML文件，离线可用。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/aggM27DD6gtosrTAoWybqC) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nCreate a pure static exercise tracker website (HTML/CSS/JS) with responsive PC/mobile support. Requirements: 1) Multi-page structure (Home, Check-in, Dashboard, History); 2) Cyberpunk neon UI style with particle background and CSS entry animations; 3) localStorage data storage with Chart.js charts; 4) Mobile bottom tab bar, PC sidebar;\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "artifact_type": "html", "preview": "cyberfit.html", "artifacts": [ { "file": "cyberfit.html", "label": "CyberFit 运动打卡工作台" } ], "cover_image": "cover_cn.png", "cover_image_en": "cover_en.png", "cover_mode": "landscape", "categories": [ "workspace", "productivity", "creation" ], "tags": [ "运动", "打卡", "赛博朋克", "Chart.js", "SPA", "响应式" ], "tags_en": [ "exercise", "check-in", "cyberpunk", "Chart.js", "SPA", "responsive" ], "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "is_recommended": false, "recommended_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-01T18:30:00+08:00", "updated_at": "2026-08-01T18:30:00+08:00", "preview_en": "cyberfit.html", "scenario_cover": "scenario-cover.png" }, { "id": "daily-life-manager", "title": "刘派派的每日生活管理工作台", "title_en": "Daily Life Manager Workbench", "subtitle": "待办、习惯、心情、购物、便签，一页管好每一天", "subtitle_en": "To-dos, habits, moods, shopping and notes — one page for your whole day", "description": "每天的事散落在各处：待办记在脑子里、习惯靠意志力、购物写在小票背面，到晚上总觉得忙了一天又说不出忙了什么；想要一个不用装 App、打开就能用的生活管理页。\n\n本案例由 WorkspaceBuilder 工作台搭建师专家生成：一份单文件 HTML 收进五个日常模块——「今天要处理」自动汇总当天到期/逾期的待办、未打卡的习惯和待买的物品，一键完成/打卡/已购；待办按日期排序、逾期标红、优先级三色徽章；习惯打卡、心情记录（六种心情各配代表色）、购物清单带已购进度、灵感便签可打标签；数据存浏览器 localStorage，内置一版示例数据打开即体验，支持备份/恢复 JSON 和一键清空；界面内置 i18n 中英双语，?lang=zh / ?lang=en 直接切换，示例数据也有中英两个版本。暖米色纸张质感 + 大圆角卡片的治愈手账风，手机和电脑都好用。\n\n适合：想轻装管理每天的上班族（待办+购物+便签一页收）、习惯养成期的自律党（打卡+心情记录）、手账/治愈系美学爱好者（暖米色纸张质感），以及想参考单文件 localStorage 应用与内联 i18n 实现的前端学习者。", "description_en": "Your day is scattered everywhere: to-dos live in your head, habits rely on willpower, shopping items sit on the back of receipts — by evening you feel busy but can't say what you actually did. You want a life-management page that needs no app install and works the moment you open it.\n\nThis case is built by the WorkspaceBuilder expert: a single-file HTML packs five daily modules — \"Today's To-Do\" auto-collects tasks due or overdue today, unchecked habits and items to buy, with one-tap Done / Check in / Bought; to-dos sort by date with overdue in red and tri-color priority badges; habit check-ins, a mood log (six moods, each with its own color), a shopping list with bought progress, and quick notes with tags. Data persists in browser localStorage, a sample dataset is built in for instant try-out, with JSON backup/restore and one-click clear. The UI ships with inline i18n in Chinese and English — switch with ?lang=zh / ?lang=en, and the sample data comes in both languages too. Warm beige paper texture and big rounded cards give it a healing journal feel, friendly on both phone and desktop.\n\nBest for: office workers who want to manage the day light-weight (to-dos + shopping + notes in one page), habit builders (check-ins + mood log), journal aesthetics lovers (warm paper texture), and frontend learners studying single-file localStorage apps with inline i18n.", "categories": [ "workspace" ], "tags": [ "生活管理", "待办清单", "习惯打卡", "心情记录", "个人工作台" ], "tags_en": [ "life management", "to-do list", "habit tracker", "mood log", "personal workspace" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "每日生活管理工作台 单文件 HTML（内置 i18n 中英文 + 中英双版示例数据，localStorage 自包含）" } ], "prompt": "请使用资料库的能力帮我一键做同款：https://www.workbuddy.cn/space/d/o6ziMSnpgANV3IqdLeaRyB，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n我想做一个「每日生活管理」个人工作台，把一天里杂七杂八的事收进一个页面，包含：\n1. 今天要处理：自动汇总今天到期或逾期的待办、今日还没打卡的习惯、还没买的购物项，每条右侧有快捷按钮（完成 / 打卡 / 已购）。\n2. 待办：可添加事项（内容、日期、优先级 高/中/低），列表按日期升序，逾期标红，完成后划线，可删除。\n3. 习惯打卡：记录喝水、阅读、早睡这类习惯，当天点一下就打勾，区分「已打卡 / 未打卡」。\n4. 心情记录：从开心 / 平静 / 一般 / 难过 / 焦虑 / 疲惫里选心情（各有代表色小圆点），可附一句话。\n5. 购物清单：物品 + 数量，买完打勾，顶部显示已购进度（已购 X / 共 Y 件）。\n6. 便签：随手记灵感，可打标签（如 工作 / 生活）。\n7. 数据存浏览器 localStorage，首次打开内置一版示例数据方便直接体验；支持备份 / 恢复 JSON、一键清空。\n\n视觉走治愈手账风：暖米色纸张质感背景、棕色系点缀、大圆角卡片、serif 标题字，整体低饱和柔和；响应式，手机和电脑都好用。要求CSS / JS 全部内联，不引用任何外部资源。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the library's capability to make me an exact replica of this workbench: https://www.workbuddy.cn/space/d/o6ziMSnpgANV3IqdLeaRyB — data must be stored online with two-way sync. Prefer the one-click replica path; if that's not available, build it as follows:/n/nI want a \"Daily Life Manager\" personal workbench that gathers all the bits and pieces of my day into one page, including:/n1. Today's To-Do: auto-collect to-dos due today or overdue, habits not yet checked in today, and shopping items not yet bought, each with a quick action button (Done / Check in / Bought).\n2. To-dos: add items (content, date, priority High/Medium/Low); list sorted by date ascending, overdue in red, completed struck through, deletable.\n3. Habit check-ins: track habits like drinking water, reading, sleeping early — one tap checks them off for the day, with clear checked / unchecked states.\n4. Mood log: pick from Happy / Calm / Okay / Sad / Anxious / Tired (each with its own colored dot), plus an optional one-line note.\n5. Shopping list: item + quantity, tick when bought, progress shown on top (Bought X of Y items).\n6. Notes: jot down ideas with optional tags (e.g. Work / Life).\n7. Data persists in browser localStorage; ship a built-in sample dataset so it works on first open; support JSON backup / restore and one-click clear. Visual style: a healing journal look — warm beige paper-texture background, brown accents, big rounded cards, serif headings, soft low-saturation palette; responsive for both phone and desktop. Deliver a with all CSS / JS inline and zero external resources.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "刘派派", "name_en": "Liu Paipai", "avatar": "", "badge": "community" }, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "created_at": "2026-08-17T10:21:49+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "data-county-development-profile", "title": "云岭县发展画像 · 县域多维发展与振兴专题", "title_en": "Yunling County Development Profile: A Multidimensional Rural Revitalization Story", "subtitle": "五维一体的县域发展画像，看清短板与振兴路径", "subtitle_en": "A five-dimension county development portrait that reveals gaps and revitalization paths", "description": "想把一个县的发展讲清楚，却困在零散指标里，既看不出全貌也找不到短板？本案由「探数数」数据探索专家，以一个虚构的西南山区县「云岭县」为样本，沿「数据探索→校验→多维下钻→洞察生成」的方法论，搭建收入、教育、医疗、基础设施、产业多元、生态宜居六维发展指数，并通过收入折线、产业结构演进、乡镇梯度对比等图表识别短板、给出振兴行动建议。交付物是一份纪实摄影风的单文件 HTML 专题作品，全幅大图、图文交错、数据图谱穿插，积极正向地呈现县域振兴成效。适合做县域/区域发展分析、乡村振兴汇报、政务数据可视化的人参考。", "description_en": "Struggling to tell a county's development story when scattered metrics hide both the big picture and the real gaps? In this case, the DataExplorationExpert builds a six-dimension development index (income, education, healthcare, infrastructure, industry diversity, livability) for a fictional southwestern mountain county, following an explore-validate-drill-insight workflow. Through income trend lines, industry-structure evolution, and township comparisons, it pinpoints weaknesses and proposes revitalization actions. The deliverable is a single-file HTML feature in a documentary-photography style—full-bleed imagery, magazine layout, and data charts woven together. Ideal for regional development analysis, rural revitalization reporting, and government data visualization.", "categories": [ "data" ], "tags": [ "数据可视化", "县域发展", "乡村振兴", "多维分析" ], "tags_en": [ "Data Visualization", "County Development", "Rural Revitalization", "Multidimensional Analysis" ], "artifact_type": "html", "prompt": "我要做一份「县域多维发展与振兴画像」的单文件 HTML 数据专题，用积极正向的视角呈现一个县的发展成效（不要写成负面贫困叙事）。可以用合成数据虚构一个山区县，各维度指标内部自洽。视觉走纪实摄影/国家地理式图文专题风格：全幅大图 hero 配渐变蒙版压标题，大地色系（土黄/橄榄绿/赭石/米色），图文交错的杂志排版。页面要包含：1) 全幅图片 Hero（标题+副标+关键结论）；2) 收入/教育/医疗/基础设施/产业多元/生态宜居六维发展指数雷达图与综合指数；3) 产业振兴、人口城镇化、公共服务覆盖等图文专题板块；4) 近5年人均可支配收入折线、产业结构堆叠条、各乡镇发展水平对比等 SVG 图表；5) 短板识别 + 振兴行动建议卡片；6) 请联网搜索该县或同类县域的真实公开统计数据（统计年鉴/政府公报等）并在页脚注明来源；确无公开数据的指标才标注为建模测算。所有 CSS/JS 内联，图标用固定尺寸内联 SVG，界面简体中文，货币用¥。", "prompt_en": "I want a single-file HTML data feature on a county's multidimensional development and revitalization profile, told from a positive perspective (not a negative poverty narrative). Use self-consistent real data for a fictional mountain county. Visual style: documentary / National Geographic feature—full-bleed hero image with gradient overlay and title, earth-tone palette (ochre, olive green, sienna, beige), magazine layout interleaving photos and data. Include: 1) full-bleed photo hero (title, subtitle, key takeaway); 2) a six-dimension development radar (income, education, healthcare, infrastructure, industry diversity, livability) plus a composite index; 3) photo-feature sections on industry revitalization, urbanization, and public-service coverage; 4) SVG charts: 5-year per-capita income line, industry-structure stacked bars, township comparison; 5) gap-identification and revitalization-action cards; 6) a footer noting the data is real (web-searched) modeling with a referenced framework. All CSS/JS inline, fixed-size inline SVG icons, simplified Chinese UI, currency in ¥.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "云岭县发展画像数据专题" } ], "skills": [], "experts": [ { "id": "DataExplorationExpert", "name": "探数数", "description": "数据探索专家，支持 SQL 查询、数据探索、可视化、仪表板构建、数据验证和洞察生成", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3323f4a73129", "name_en": "Lake", "description_en": "SQL querying, data exploration, visualization, dashboard building, data validation, and insight generation with complete data analysis workflows", "expert_id": "ex_sRRTBWMNdyed" } ], "mcps": [], "scenario": "数据分析及可视化", "scenario_id": 102, "scenario_en": "Visualization", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T10:00:00Z", "updated_at": "2026-06-22T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "data-ecommerce-rfm-value", "title": "电商用户价值 RFM 分析驾驶舱", "title_en": "E-commerce Customer Value RFM Dashboard", "subtitle": "RFM 模型 + 8 层用户分层运营策略", "subtitle_en": "RFM model with 8-tier customer segmentation & playbooks", "description": "拉新越来越贵，但多数店铺不知道钱到底是谁贡献的、该把预算押在谁身上。本案由「舒明析」数据分析专家以 RFM 模型为框架，对虚构跨境电商「星野严选」48,600 名活跃用户做价值分层：按最近购买、购买频次、累计金额三维打分，切出重要价值、重要挽留、流失预警等 8 个可运营分层，量化各层人数占比与 GMV 贡献，揭示「28.5% 重要用户贡献 70.9% GMV」的二八结构，并为每层配套运营动作与可衡量目标。交付一份深色 SaaS 驾驶舱风格的单文件 HTML 报告，含 RFM 矩阵、人数/GMV 对比图、策略卡与趋势分析。适合电商运营、用户增长、CRM 与数据分析岗位。", "description_en": "Acquisition keeps getting pricier, yet most shops can't say who actually drives revenue or where to bet their budget. Using the RFM framework, data expert Shu segments 48,600 active users of fictional cross-border shop 'Xingye Select' by recency, frequency, and monetary value into 8 actionable tiers—from high-value to at-risk and churn-warning—quantifying each tier's headcount and GMV share. The analysis surfaces the 80/20 reality (28.5% of users drive 70.9% of GMV) and pairs every tier with concrete actions and measurable targets. Delivered as a single-file, dark SaaS-cockpit HTML report with an RFM matrix, headcount/GMV comparison, strategy cards, and trend charts. Built for e-commerce ops, growth, CRM, and analytics roles.", "categories": [ "data" ], "tags": [ "数据分析", "RFM模型", "用户分层", "电商运营" ], "tags_en": [ "Data Analysis", "RFM Model", "Customer Segmentation", "E-commerce Ops" ], "artifact_type": "html", "prompt": "我是电商运营，想要一份用户价值分析报告（单文件 HTML，深色数据驾驶舱风格）。请用 RFM 模型把我们的活跃用户分成重要价值/重要保持/重要发展/重要挽留/一般价值/一般保持/一般发展/流失预警共 8 层。页面要有：店铺信息和分析周期的 Hero、活跃用户数/GMV/客单价/复购率四个关键指标卡、RFM 三维说明、用 SVG 画的分层矩阵热力格、各层人数占比 vs GMV 贡献占比的对比条形图（体现二八结构）、4-8 张分层运营策略卡（含人群特征、典型 RFM 分数、运营动作、预期目标）、近 6 个月趋势折线，以及方法口径脚注。请使用真实电商公开数据集或用户提供的真实数据；如仅演示可用合成数据但须在脚注明确标注为合成、且内部勾稽一致，纯 SVG/CSS 不要外链。", "prompt_en": "I run e-commerce ops and need a customer-value report as a single-file HTML in a dark data-cockpit style. Use the RFM model to split active users into 8 tiers (high-value, loyal, potential, at-risk, general-value, general-loyal, general-developing, churn-warning). Include: a hero with shop info and analysis period; four KPI cards (active users, GMV, AOV, repurchase rate); an RFM explanation; an SVG segmentation heat-matrix; a bar chart comparing each tier's headcount share vs GMV share (showing the 80/20 structure); 4-8 strategy cards per tier (traits, typical RFM scores, actions, targets); a 6-month trend line; and a methodology footnote. real real (web-searched) data is fine but must reconcile internally; pure SVG/CSS, no external resources.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "电商用户价值 RFM 分析驾驶舱" } ], "skills": [], "experts": [ { "id": "DataAnalyticsReporter", "name": "舒明析", "description": "将复杂数据转化为战略洞察，提供指标诊断、KPI框架设计、数据质量评估与决策报告", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "44c28ef1212d", "name_en": "Shu", "description_en": "Transforms complex data into strategic insights with metric diagnostics, KPI frameworks, quality assessment, and decision-ready reports", "expert_id": "ex_eYU4NfSZb2Bm" } ], "mcps": [], "scenario": "数据分析及可视化", "scenario_id": 102, "scenario_en": "Visualization", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T10:00:00Z", "updated_at": "2026-06-22T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "data-employment-climate-report", "title": "2025 中国就业景气年度报告", "title_en": "2025 China Employment Climate Annual Report", "subtitle": "《经济学人》式社论报刊风 · 总量稳、结构分", "subtitle_en": "Economist-style editorial broadsheet: stable totals, divergent structure", "description": "年度就业数据散落在公报、发布会与月度调查里，普通人很难一眼看清\"到底好不好\"。本案由「析数数」数据分析专家以官方权威数据为主干，整理出 2025 城镇新增就业、调查失业率、就业人员总量与城镇占比等核心指标，并补充月度走势、分年龄段对比与年度新增就业测算图示，最终交付一份《经济学人》/FT 社论报刊风的单文件 HTML 年度报告：含报头导语、核心指标条、两条主线社论正文、2–3 张 SVG 图表、结构性议题卡片与来源脚注。适合做行业研究、政策解读与汇报展示的人参考。", "description_en": "Employment data is scattered across bulletins, press conferences and monthly surveys, making it hard to see the full picture. Here the data-analysis expert 'Dash' uses authoritative official data as the backbone, organizing 2025 urban new jobs, surveyed unemployment, total employment and urban share, plus monthly trends, by-age comparisons and annual new-jobs estimates. The deliverable is a single-file HTML annual report in an Economist/FT editorial broadsheet style: masthead lede, headline metric bar, two-thread editorial body, 2-3 SVG charts, structural-issue cards and sourced footnotes. Ideal for industry research, policy reading and presentations.", "categories": [ "data" ], "tags": [ "数据分析", "数据可视化", "就业市场", "年度报告" ], "tags_en": [ "Data Analysis", "Data Visualization", "Labor Market", "Annual Report" ], "artifact_type": "html", "prompt": "帮我做一份《经济学人》/FT 财经社论报刊风格的「2025 中国就业景气年度报告」单文件 HTML。要求：米白浅暖底、衬线大标题、暗红主色点缀，多栏报刊排版、首字下沉。内容包含：刊名报头与导语摘要；核心指标条（城镇新增就业 1267 万、调查失业率均值 5.2%、就业人员 72504 万、城镇占比 65.6%，用国家统计局公报真实数据）；两条主线的社论正文（总量稳、结构分）；2-3 个内联 SVG 图表（月度失业率折线含 5% 参考线、分年龄段失业率柱状青年 vs 全龄、新增就业年度对比）；青年就业/招工难/慢就业的结构性议题卡片；页脚明确列出数据来源机构与报告名年份，数据须来自联网搜索的真实公开来源（统计局/人社部等），页脚注明来源机构与报告年份；仅推算指标标注为测算。全部 CSS/JS 内联、无外链、无 emoji、简体中文、响应式。", "prompt_en": "Create a single-file HTML '2025 China Employment Climate Annual Report' in an Economist/FT editorial broadsheet style. Requirements: cream warm background, serif headlines, dark-red accents, multi-column layout, drop caps. Include: masthead with lede summary; a headline metric bar (urban new jobs 12.67M, average surveyed unemployment 5.2%, total employment 725.04M, urban share 65.6%, using real National Bureau of Statistics data); a two-thread editorial body (stable totals, divergent structure); 2-3 inline SVG charts (monthly unemployment line with a 5% reference line, by-age unemployment bars youth vs all-age, annual new-jobs comparison); structural-issue cards on youth employment, labor shortage and slow employment; a footer listing data-source institutions, report names and years, marking which figures are real (web-searched) estimates. All CSS/JS inline, no external resources, no emoji, Simplified Chinese, responsive.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "2025 中国就业景气年度报告" } ], "skills": [], "experts": [ { "id": "DataAnalysisExpert", "name": "析数数", "description": "数据分析专家，支持 Excel 电子表格的创建、编辑、分析、公式计算、格式化和数据可视化", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1e22c5145cd0", "name_en": "Dash", "description_en": "Data analysis with Excel spreadsheet creation, editing, analysis, formulas, formatting, and visualization capabilities", "expert_id": "ex_fgKMhO3xPx9p" } ], "mcps": [], "scenario": "数据分析及可视化", "scenario_id": 102, "scenario_en": "Visualization", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T10:00:00Z", "updated_at": "2026-06-22T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "data-gdp-hdi-explorer", "title": "GDP 之外：用 HDI 重新丈量发展", "title_en": "Beyond GDP: Remeasuring Development with HDI", "subtitle": "经济总量 ≠ 人类发展水平的跨国数据星图", "subtitle_en": "A cross-country data star-map on why wealth isn't the same as well-being", "description": "想用一组数据讲清「有钱不等于发展好」，却常常陷在表格里讲不出故事？本案由「图说说」视觉叙事专家操刀，把联合国开发计划署 HDI 真实数据转化为一页深空数据星图：从太空夜灯 Hero 切入，拆解 HDI 健康/教育/生活水平三维度，再用「人均 GDP × HDI 散点图」与「HDI 横向排行榜」呈现美国财富顶级却 HDI 第 17、冰岛小国登顶、中国与中国香港高效转化等反差洞察。交付单文件 HTML，内联可直接打开。适合做经济社会议题科普、数据新闻、教学演示与汇报的人。", "description_en": "Want to show that 'rich isn't the same as well-developed' but keep getting stuck in spreadsheets? Created by visual-storytelling expert 'Eva', this case turns UNDP's real HDI data into a single deep-space data star-map: an Earth-at-night hero, a breakdown of HDI's three dimensions, plus a GDP-per-capita × HDI scatter plot and a horizontal HDI ranking that surface contrasts—the US ranks only 17th despite top wealth, tiny Iceland tops the chart, and China/Hong Kong convert income into development efficiently. Delivered as a single self-contained HTML file. Ideal for socioeconomic explainers, data journalism, teaching and presentations.", "categories": [ "data" ], "tags": [ "数据可视化", "数据叙事", "经济社会", "国际对比" ], "tags_en": [ "Data Visualization", "Data Storytelling", "Socioeconomics", "Cross-country Comparison" ], "artifact_type": "html", "prompt": "帮我做一个单文件 HTML 数据可视化页面，主题是『GDP 之外：用 HDI 重新丈量发展』，讲清楚经济总量不等于人类发展水平。要深空天文台/数据星图风格，深蓝近黑宇宙底色，用一张从太空看地球城市夜灯的大图做 Hero（叠深色渐变蒙版保证白色标题清晰）。内容包含：HDI 三维度（健康/教育/生活水平）拆解说明、核心的人均 GDP × HDI 散点图（标注各国、能看出高收入未必高 HDI）、HDI 横向排行榜条形图（中国和中国香港高亮）、几张反差洞察卡片，以及标注 UNDP 人类发展报告 2025 的页脚。用真实 HDI 数据，所有 CSS/JS 和图片都内联，界面中文。", "prompt_en": "Build me a single-file HTML data-visualization page titled 'Beyond GDP: Remeasuring Development with HDI', making clear that economic size isn't the same as human development. Use a deep-space observatory / data-star-map style with a near-black blue cosmic background, and an Earth-at-night city-lights photo as the hero (with a dark gradient overlay so the white title stays readable). Include: a breakdown of HDI's three dimensions (health/education/standard of living), a core GDP-per-capita × HDI scatter plot with country labels (showing high income doesn't guarantee high HDI), a horizontal HDI ranking bar chart (highlighting China and Hong Kong), several contrast-insight cards, and a footer citing UNDP Human Development Report 2025. Use real HDI data, inline all CSS/JS and images, Chinese UI.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "GDP 之外：HDI 发展星图" } ], "skills": [], "experts": [ { "id": "VisualStorytellingExpert", "name": "图说说", "description": "擅长将复杂信息转化为引人入胜的视觉故事", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "4aca8292f173", "name_en": "Eva", "description_en": "Expert at transforming complex information into compelling visual stories", "expert_id": "ex_id1QnlViWLJi" } ], "mcps": [], "scenario": "数据分析及可视化", "scenario_id": 102, "scenario_en": "Visualization", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T10:00:00Z", "updated_at": "2026-06-22T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "data-global-population-structure", "title": "八十二亿之后：全球人口结构与趋势图志", "title_en": "After 8.2 Billion: Atlas of Global Population Structure & Trends", "subtitle": "杂志信息图风 · 人口峰值、老龄化、生育率下降与区域分化", "subtitle_en": "Editorial infographic on peak, aging, fertility decline & regional divergence", "description": "做人口议题的科普或战略汇报，最怕数字堆砌却看不出趋势。本案由「舒明析」数据分析师以决策导向的方法论，把联合国《World Population Prospects 2024》的核心结论提炼成一张杂志信息图风的单文件 HTML：超大数字指标墙、人口达峰曲线（标注 2084 年约 103 亿峰值）、多区域生育率下降折线（含 2.1 更替线）、65+ 与 18− 老龄化交叉图、印度超越中国的大国易位曲线，以及区域增长分化与决策启示。主干数字均来自联合国官方，建模细分均标注示意。适合研究者、媒体编辑、战略与政策从业者快速理解人口长波。", "description_en": "Population briefings often pile up numbers while hiding the trend. Here data analyst 'Shu' applies a decision-first methodology to distill the UN World Population Prospects 2024 into a single-file, editorial-infographic HTML page: a big-number metric wall, the peak curve (10.3B around 2084), multi-region fertility decline with the 2.1 replacement line, the 65+ vs under-18 aging crossover, the India-overtakes-China shift, plus regional divergence and decision takeaways. Headline figures come from the UN; modeled breakdowns are clearly labeled. Built for researchers, editors, and strategy/policy professionals.", "categories": [ "data" ], "tags": [ "数据可视化", "人口趋势", "信息图设计", "战略洞察" ], "tags_en": [ "Data Visualization", "Population Trends", "Infographic Design", "Strategic Insight" ], "artifact_type": "html", "prompt": "帮我做一个单文件 HTML 的全球人口结构与趋势数据可视化页面，杂志信息图/编辑设计风（明亮大胆、强对比大色块、超大数字排版、图片裁成色块拼贴）。要用联合国 World Population Prospects 2024 的真实数据：2024 年世界人口约 82 亿，预计 2084 年前后达峰约 103 亿、2100 年约 102 亿，全球总和生育率约 2.3（更替水平 2.1），印度 2023 年超越中国，65 岁以上人口将在 2070 年代后期超过 18 岁以下儿童。页面要包含：带城市航拍配图的 Hero 和 82 亿→103 亿叙事、关键数字指标墙、人口达峰曲线（SVG，标注峰值点）、多区域生育率下降折线（世界/非洲/亚洲/欧洲含更替线）、老龄化 65+ vs 18− 交叉图、印度 vs 中国人口交叉、区域增长分化板块（穿插人群配图）、来源与建模说明页脚。所有 CSS/JS 和两张图片都内联，不要外链，界面用简体中文，图表用 SVG 绘制不要溢出坐标轴。", "prompt_en": "Build a single-file HTML data-visualization page on global population structure and trends in an editorial / magazine-infographic style (bright, bold, high-contrast color blocks, oversized numbers, photos cropped into blocks). Use real UN World Population Prospects 2024 data: ~8.2B in 2024, peak ~10.3B around 2084, ~10.2B by 2100, global TFR ~2.3 (replacement 2.1), India overtook China in 2023, and the 65+ population overtaking under-18 in the late 2070s. Include: a hero with an aerial city photo and the 8.2B→10.3B narrative, a big-number metric wall, the peak curve (SVG with a labeled peak point), multi-region fertility-decline lines (World/Africa/Asia/Europe with the replacement line), the 65+ vs under-18 aging crossover, an India-vs-China crossover, a regional divergence section (with a crowd photo), and a sources/modeling footnote. Inline all CSS/JS and both images, no external resources, Simplified-Chinese UI, SVG charts that never overflow the axes.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "全球人口结构与趋势图志" } ], "skills": [], "experts": [ { "id": "DataAnalyticsReporter", "name": "舒明析", "description": "将复杂数据转化为战略洞察，提供指标诊断、KPI框架设计、数据质量评估与决策报告", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "44c28ef1212d", "name_en": "Shu", "description_en": "Transforms complex data into strategic insights with metric diagnostics, KPI frameworks, quality assessment, and decision-ready reports", "expert_id": "ex_eYU4NfSZb2Bm" } ], "mcps": [], "scenario": "数据分析及可视化", "scenario_id": 102, "scenario_en": "Visualization", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T10:00:00Z", "updated_at": "2026-06-22T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "database-schema-design", "title": "数据库 Schema 设计与迁移脚本", "title_en": "Database Schema Design and Migration Scripts", "subtitle": "一句 Prompt 生成完整的 MySQL Schema 与版本化迁移脚本", "subtitle_en": "Generate complete MySQL schema and versioned migration scripts from a single prompt", "description": "手动设计数据库 Schema 需要考量表关系、命名规范、索引策略、字段类型选择、外键约束和存储引擎等诸多细节，任何疏漏都可能导致性能瓶颈或数据一致性问题。本案例演示如何通过 MySQL Skill，一次性为在线教育平台生成完整的数据库设计方案，涵盖用户管理、课程体系、订单系统和学习记录四大核心模块，包含规范的命名约定、完整的外键关系、索引设计、字段注释以及 3 个版本的渐进式迁移脚本。适合后端开发者、DBA、全栈工程师及需要快速落地数据库架构设计的技术团队。", "description_en": "Designing database schemas manually requires careful consideration of table relationships, naming conventions, indexing strategies, data type selection, foreign key constraints, and storage engine choices -- any oversight can lead to performance bottlenecks or data consistency issues. This case demonstrates how to use the MySQL Skill to generate a complete database design for an online education platform in one shot, covering four core modules: user management, course system, order system, and learning records, with standardized naming conventions, complete foreign key relationships, index design, column comments, and 3 versioned progressive migration scripts. Ideal for backend developers, DBAs, full-stack engineers, and teams looking to quickly implement database architecture designs.", "tags": [ "MySQL", "数据库设计", "迁移脚本", "Schema", "后端" ], "tags_en": [ "MySQL", "Database Design", "Migration", "Schema", "Backend" ], "artifact_type": "code", "categories": [ "dev", "data" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.sql", "prompt": "请为一个在线教育平台设计完整的 MySQL 数据库 Schema，包含以下模块：1) 用户管理（用户、角色、权限）；2) 课程体系（课程、章节、课时、课程分类）；3) 订单系统（订单、支付记录、优惠券）；4) 学习记录（学习进度、考试成绩、证书）。要求：使用规范的命名约定，包含完整的外键关系、索引设计、字段注释、存储引擎选择，以及 3 个版本的迁移脚本（V1 基础表结构 → V2 添加索引和分区 → V3 新增审计日志表）。", "prompt_en": "Design a complete MySQL database schema for an online education platform, covering the following modules: 1) user management (users, roles, permissions); 2) course system (courses, chapters, lessons, course categories); 3) order system (orders, payment records, coupons); 4) learning records (learning progress, exam scores, certificates). Requirements: use standardized naming conventions, include complete foreign key relationships, index design, column comments, storage engine selection, and 3 versioned migration scripts (V1 base table structure → V2 add indexes and partitioning → V3 add audit log table).", "skills": [ { "id": "mysql", "name": "MySQL", "description": "MySQL 数据库 Schema 设计与 SQL 脚本生成", "slug": "mysql" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 93, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "source": "official", "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.sql", "label": "数据库 Schema 设计" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "decision-career-pivot", "title": "高维决策：35岁程序员该不该 All-in AI 创业", "title_en": "High-Dimensional Decision: Should a 35-Year-Old Engineer Go All-In on an AI Startup", "subtitle": "用九大决策模块把一场人生豪赌拆成可推演、可校准、可执行的结构化判断", "subtitle_en": "Turning a life-defining gamble into a structured, simulatable, and executable decision with nine decision modules", "description": "「35 岁了，到底该不该辞掉大厂高薪去 AI 创业？」这种人生级选择，最难受的是它没有标准答案，凭感觉决定又怕赌错一生。脑子里全是「机会难得」和「万一失败呢」在打架，越想越乱。\n\n这个案例演示了如何把这道模糊的焦虑题，交给高维决策师用完整九步法拆解：先用仓颉编码把「要不要创业」重定义为「如何不赌上家庭安全垫地把技术资本变成长期复利」；再用周易辩证列清阴阳两面与转化条件、系统科学定位真正的短板节点、康波周期判断当前正处 AI 技术革命孕育期；接着用世界模型给出乐观/基准/悲观三种未来的概率推演，设计出 All-in、骑驴找马、曲线入场三套方案；最后用女娲范式给出「有条件推进」的推荐方案、分三阶段的行动路线，并用宇宙视角做终极价值校准。产出一份排版精致、可离线打开的 HTML 决策报告。\n\n适合：面临职业转型的技术人与职场人、纠结创业与否的从业者、想学系统决策方法的人、给团队/自己做重大选择复盘的管理者、对东西方决策智慧感兴趣的思考者。", "description_en": "\"I'm 35 — should I quit my well-paid big-tech job and go start an AI company?\" The hardest part of a life-level choice like this is that there's no right answer, yet deciding on gut feel risks betting your whole life wrong. \"This is a rare window\" and \"but what if I fail\" just keep fighting in your head.\n\nThis case shows how to hand that vague anxiety to the High-Dimensional Decision Maker and break it down with a full nine-step method: Cangjie encoding first redefines \"should I start up?\" into \"how to convert my technical capital into long-term compounding without betting the family safety net\"; then Yin-Yang dialectics map the pros/cons and turning conditions, systems science locates the real bottleneck node, and the Kondratieff cycle places us in the incubation phase of the AI tech revolution; a world model then simulates optimistic/base/pessimistic futures with probabilities and designs three options — all-in, moonlighting-first, and a curved entry; finally the Nuwa paradigm delivers a \"conditional go\" recommendation, a three-phase action roadmap, and a cosmic-view value calibration. The output is a polished, offline-ready HTML decision report.\n\nIdeal for: engineers and professionals facing a career pivot, anyone torn over whether to start a business, learners of systematic decision-making, managers reviewing major choices, and thinkers curious about East-meets-West decision wisdom.", "categories": [ "education", "business" ], "tags": [ "决策分析", "职业规划", "系统思维", "认知升级", "创业决策" ], "tags_en": [ "Decision Analysis", "Career Planning", "Systems Thinking", "Cognitive Upgrade", "Startup Decision" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我今年 35 岁，是一线大厂的资深后端工程师，年包约 80 万，有房贷和一个孩子，家庭有另一份稳定收入，可支配启动资金约 120 万，风险承受能力中等。我在纠结要不要辞职 all-in 去做 AI 应用创业。\n\n请用你的高维决策九步法，帮我做一份完整的决策分析：\n1）先用仓颉编码把我的模糊焦虑重定义成真正要决策的问题；\n2）用周易辩证列出阴阳两面（有利/风险）和阴阳转化的关键条件；\n3）用系统科学分析技术能力、产品商业化、现金流、心力等要素的关系与反馈回路，指出真正的短板节点；\n4）用康波周期判断当前 AI 赛道处在周期的哪个阶段，并给出对应的周期策略；\n5）用世界模型给出乐观/基准/悲观三种未来场景及各自概率；\n6）设计至少 3 套可选方案（含 all-in、骑驴找马、曲线入场），列出核心思路、预期收益、主要风险和适合条件；\n7）用女娲范式给出推荐方案（含核心理由、关键成功因素、应急止损预案）；\n8）给出分三阶段（0–30 天 / 1–3 月 / 3–12 月）的行动路线和反馈复盘机制；\n9）最后用宇宙视角做价值校准反思。\n\n请把结果做成一份排版精致、适合阅读的单文件 HTML 决策报告：浅色专业风格、信息层次清晰、含阴阳对比卡片、康波周期时间轴、场景概率条、方案对比表、行动时间线等可视化元素；要求单文件、所有 CSS/JS/图片资源全部内联、可离线打开、不引用任何外部链接。", "prompt_en": "I'm 35, a senior backend engineer at a top tech company with a total annual package of about 800K RMB. I have a mortgage and a child, my family has another stable income, my available startup capital is about 1.2M RMB, and my risk tolerance is medium. I'm torn over whether to quit and go all-in on an AI application startup.\n\nPlease use your High-Dimensional nine-step decision method to produce a complete decision analysis:\n1) First use Cangjie encoding to redefine my vague anxiety into the real problem to decide;\n2) Use Yin-Yang dialectics to list both sides (pros/risks) and the key conditions for Yin-Yang transformation;\n3) Use systems science to analyze the relationships and feedback loops among technical ability, product commercialization, cash flow, and mental energy, and point out the real bottleneck node;\n4) Use the Kondratieff cycle to judge which phase the AI track is currently in, and give the matching cycle strategy;\n5) Use a world model to give optimistic/base/pessimistic future scenarios with their probabilities;\n6) Design at least 3 options (including all-in, moonlighting-first, and a curved entry), listing core idea, expected return, main risks, and suitable conditions;\n7) Use the Nuwa paradigm to give a recommended option (with core reasoning, key success factors, and an emergency stop-loss plan);\n8) Provide a three-phase action roadmap (0–30 days / 1–3 months / 3–12 months) and a feedback/retrospective mechanism;\n9) Finally, do a cosmic-view value calibration reflection.\n\nPlease turn the result into a polished, readable single-file HTML decision report: light professional style, clear information hierarchy, with Yin-Yang comparison cards, a Kondratieff cycle timeline, scenario probability bars, an options comparison table, an action timeline and other visual elements; it must be a single file with all CSS/JS/image resources fully inlined, openable offline, and referencing no external links.", "skills": [ { "id": "gaoweijuece", "name": "高维决策师", "description": "整合东西方智慧的九大模块系统决策框架，帮你在不确定中做出结构化决策", "slug": "gaoweijuece" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T17:50:00+08:00", "updated_at": "2026-07-24T17:50:00+08:00", "artifacts": [ { "file": "output.html", "label": "职业跃迁高维决策报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "decision-kondratieff-strategy", "title": "高维决策：制造企业在康波切换点如何布局新赛道", "title_en": "High-Dimensional Decision: How a Manufacturer Should Position for New Industries at a Kondratieff Turning Point", "subtitle": "用康波周期、系统科学与女娲范式，给传统制造企业一份可执行的转型战略", "subtitle_en": "An executable transformation strategy for a traditional manufacturer, built with Kondratieff cycles, systems science, and the Nuwa paradigm", "description": "主营业务增速见顶、毛利率一年比一年薄——这是很多传统制造企业老板半夜睡不着的事。死守主业怕被时代淘汰，砸钱进新赛道又怕押错方向、把家底赔光。到底是精益提效守存量，还是切进不确定的新兴产业？\n\n这个案例演示了如何把这道企业级战略难题交给高维决策师系统拆解：先用仓颉编码把「要不要投新赛道」重定义为「如何在主业还能供血的窗口期完成向下一轮康波技术的能力迁移」；再用康波周期把「增速见顶」判断为周期性信号而非管理问题，配一张波形图定位当前正处旧动能衰退、新旧切换段；接着用系统科学梳理主业现金牛、新业务、人才组织、制度激励四要素的反馈回路，用周易辩证列清转型的阴阳两面与转化条件；用世界模型推演三种未来概率，设计死守、双轮驱动、激进转型三套方案；最后用女娲范式（创造造人 / 修复补天 / 延续立制）给出「双轮驱动」推荐方案与分三阶段行动路线，并用宇宙视角校准企业的长期价值。产出一份含康波波形图、系统结构图、方案对比表的精致 HTML 战略报告，可离线打开。\n\n适合：传统企业创始人与高管、战略/投资部门负责人、面临第二增长曲线焦虑的经营者、企业转型顾问、想学周期思维与系统决策的商业研究者。", "description_en": "Core-business growth peaking and margins thinning year after year — that's what keeps many traditional manufacturing owners up at night. Holding the line risks being left behind by the era; pouring money into a new track risks betting wrong and losing everything. So which is it: lean out the existing business, or cut into an uncertain emerging industry?\n\nThis case shows how to hand that enterprise-level strategy dilemma to the High-Dimensional Decision Maker for a systematic breakdown: Cangjie encoding first redefines \"should we invest in a new track?\" as \"how to complete the capability migration to the next Kondratieff technology while the core business can still fund it\"; the Kondratieff cycle then reframes \"peaking growth\" as a cyclical signal rather than a management failure, with a wave chart placing us at the decline-and-switchover segment; systems science maps the feedback loops among the cash-cow core business, the new business, talent/organization, and incentives, while Yin-Yang dialectics lay out the pros/cons and turning conditions of transformation; a world model simulates three future probabilities and designs three options — hold the line, dual-engine, and aggressive pivot; finally the Nuwa paradigm (create / repair / sustain) delivers a recommended \"dual-engine\" option with a three-phase roadmap, and a cosmic view calibrates the firm's long-term value. The output is a polished, offline-ready HTML strategy report with a Kondratieff wave chart, a system diagram, and an options table.\n\nIdeal for: founders and executives of traditional firms, heads of strategy/investment, operators anxious about a second growth curve, transformation consultants, and business researchers wanting to learn cyclical and systematic decision-making.", "categories": [ "business", "education" ], "tags": [ "战略决策", "康波周期", "企业转型", "系统思维", "第二增长曲线" ], "tags_en": [ "Strategic Decision", "Kondratieff Cycle", "Enterprise Transformation", "Systems Thinking", "Second Growth Curve" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我是一家成立 20 年的中型装备制造企业负责人，年营收约 15 亿，主营产品市占率高但行业增速已降到个位数，毛利率三年内从 28% 降到 19%，现金流健康但没有第二增长曲线。我在纠结：是继续死守主业做精益提效，还是投入资源切入新兴产业（如智能制造/新能源相关赛道）。\n\n请用你的高维决策框架帮我做一份完整的企业战略决策分析：\n1）先用仓颉编码把我的焦虑重定义成真正要决策的战略命题；\n2）用康波周期判断主营业务当前处于周期的哪个阶段（回升/繁荣/衰退/萧条/恢复），说明增速见顶是不是周期性信号，并给出对应的周期策略；\n3）用系统科学分析主业现金流、新业务、人才组织、制度激励等要素之间的关系和反馈回路，指出关键节点；\n4）用周易辩证列出转型的阴阳两面（有利/风险）与阴阳转化的关键条件；\n5）用世界模型给出乐观/基准/悲观三种未来场景及概率；\n6）设计至少 3 套战略方案（含死守主业、双轮驱动、激进转型），对比核心思路、预期收益、主要风险、适合条件；\n7）用女娲范式（创造/修复/延续三阶段）给出推荐方案，含核心理由、关键成功因素、止损应急预案；\n8）给出分三阶段（0–6 月 / 6–18 月 / 18–36 月）的落地行动路线与反馈复盘机制；\n9）最后用宇宙视角做企业长期价值校准。\n\n请把结果做成一份排版精致、适合决策层阅读的单文件 HTML 战略报告：浅色专业商务风格、信息层次清晰，包含康波周期波形图、系统结构图、阴阳对比卡片、场景概率条、方案对比表、女娲三阶段卡片、行动时间线等可视化元素；要求单文件、所有 CSS/JS/图片（含 SVG 图表）全部内联、可离线打开、不引用任何外部链接。", "prompt_en": "I run a mid-sized equipment manufacturer founded 20 years ago, with annual revenue around 1.5B RMB. Our main product has high market share but industry growth has dropped to single digits, gross margin fell from 28% to 19% over three years, cash flow is healthy but there's no second growth curve. I'm torn: keep holding the core business with lean efficiency, or invest resources to enter an emerging industry (e.g., smart-manufacturing / new-energy related tracks).\n\nPlease use your High-Dimensional decision framework to produce a complete enterprise strategy analysis:\n1) First use Cangjie encoding to redefine my anxiety into the real strategic proposition to decide;\n2) Use the Kondratieff cycle to judge which phase the core business is in (upturn/prosperity/recession/depression/recovery), explain whether peaking growth is a cyclical signal, and give the matching cycle strategy;\n3) Use systems science to analyze the relationships and feedback loops among core-business cash flow, the new business, talent/organization, and incentives, pointing out the key nodes;\n4) Use Yin-Yang dialectics to list both sides of transformation (pros/risks) and the key conditions for Yin-Yang transformation;\n5) Use a world model to give optimistic/base/pessimistic future scenarios with probabilities;\n6) Design at least 3 strategic options (including hold-the-line, dual-engine, aggressive pivot), comparing core idea, expected return, main risks, and suitable conditions;\n7) Use the Nuwa paradigm (create/repair/sustain three phases) to give a recommended option, including core reasoning, key success factors, and a stop-loss contingency plan;\n8) Provide a three-phase implementation roadmap (0–6 months / 6–18 months / 18–36 months) and a feedback/retrospective mechanism;\n9) Finally, do a cosmic-view calibration of the firm's long-term value.\n\nPlease turn the result into a polished single-file HTML strategy report suitable for decision-makers: light professional business style, clear information hierarchy, including a Kondratieff cycle wave chart, a system diagram, Yin-Yang comparison cards, scenario probability bars, an options comparison table, Nuwa three-phase cards, and an action timeline; it must be a single file with all CSS/JS/images (including SVG charts) fully inlined, openable offline, and referencing no external links.", "skills": [ { "id": "gaoweijuece", "name": "高维决策师", "description": "整合东西方智慧的九大模块系统决策框架，帮你在不确定中做出结构化决策", "slug": "gaoweijuece" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T17:50:00+08:00", "updated_at": "2026-07-24T17:50:00+08:00", "artifacts": [ { "file": "output.html", "label": "企业战略康波布局决策报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "deep-research-ai-agents", "title": "AI Agent 技术深度调研报告", "title_en": "AI Agent Technology Deep Research Report", "subtitle": "2025 年 AI Agent 技术全景深度调研与趋势分析", "subtitle_en": "2025 AI Agent Technology Landscape Deep Research and Trend Analysis", "description": "需要快速了解 AI Agent 技术全貌？本案例展示一份结构化的深度调研报告，包含市场规模、关键玩家、技术栈分层、应用场景矩阵、挑战与风险、未来展望等六大模块，配有 SVG 图表、数据卡片和可导航的目录结构。适合技术决策者、产品经理和投资分析师。", "description_en": "Need to quickly understand the AI Agent technology landscape? This case presents a structured deep research report covering market size, key players, technology stack layers, use case matrices, challenges, and future outlook with SVG charts and navigable structure.", "categories": [ "agent", "business" ], "tags": [ "深度调研", "AI Agent", "技术趋势", "行业报告" ], "tags_en": [ "Deep Research", "AI Agent", "Tech Trends", "Industry Report" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI Agent 深度调研报告页面" } ], "prompt": "请创建一份「2025 年 AI Agent 技术全景」深度调研报告页面（单文件 HTML，内联无外部依赖），专业海军蓝配色、清爽长页排版、支持打印。\n\n章节：Executive Summary、市场概览（含 SVG 柱状图）、关键玩家（卡片）、技术栈分层图、应用场景（图标网格）、挑战与风险、未来展望、研究方法论。\n左侧目录 scroll-spy 高亮当前章节；关键数据高亮卡片（市场规模 $12.8B、340+ 公司、5 大类别）。数据可模拟但需合理可信。", "prompt_en": "Please create a \"2025 AI Agent Technology Landscape\" in-depth research report page (a single-file HTML page, inline with no external dependencies), with a professional navy-blue color scheme, a clean long-page layout, and print support.\n\nSections: Executive Summary, Market Overview (with an SVG bar chart), Key Players (cards), a technology stack layered diagram, Application Scenarios (icon grid), Challenges and Risks, Future Outlook, Research Methodology.\nA left-side table of contents with scroll-spy highlighting of the current section; highlighted key-data cards (market size $12.8B, 340+ companies, 5 major categories). Data may be simulated but must be reasonable and credible.", "skills": [ { "id": "deep-research", "name": "Deep Research", "description": "结构化深度调研工作流", "skill_id": "skill_2053082035566706688" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "design-system-doc-site", "title": "AI 产品设计系统文档站", "title_en": "AI Product Design System Documentation Site", "subtitle": "Linear/Vercel 风格的专业设计系统文档", "subtitle_en": "Professional Design System Docs in Linear/Vercel Style", "description": "设计系统文档散落各处、缺乏统一规范，团队协作时频繁出现样式不一致的问题。本案例演示如何用 AI 一键生成包含 Design Tokens（颜色、字体、间距）、组件文档（按钮、卡片、表单）、暗色主题和代码片段的完整设计系统文档站。适合前端工程师、UI 设计师和产品团队快速建立统一的设计语言。", "description_en": "Design system documentation scattered across tools leads to inconsistent styles in team collaboration. This case demonstrates using AI to generate a complete design system doc site with Design Tokens (colors, typography, spacing), component docs (buttons, cards, forms), dark theme, and code snippets. Ideal for frontend engineers, UI designers, and product teams building a unified design language.", "categories": [ "dev", "creation" ], "tags": [ "设计系统", "文档站", "暗色主题", "组件库", "前端" ], "tags_en": [ "design-system", "documentation", "dark-theme", "component-library", "frontend" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "设计系统文档站" } ], "prompt": "请生成一个设计系统文档站单页（单文件 HTML，内联无外部依赖），Linear / Vercel 风格暗色主题，左侧固定导航 + 右侧滚动内容。\n\n章节：\n1. Overview 概览（理念与使用方式）\n2. Colors 颜色系统（主色/中性色/语义色，每个色块显示色值与名称）\n3. Typography 字体排版（H1-H6 + 正文 + 小字，标注字号行高）\n4. Spacing 间距系统（基准间距刻度可视化）\n5. Components 组件（按钮/输入框/卡片各状态，附使用代码片段）\n6. Icons 图标集（常用图标）\n\n导航点击平滑滚动，组件有 hover 过渡，代码片段简单语法高亮。", "prompt_en": "Please create a single-page design-system documentation site (single-file HTML, fully inline with no external dependencies), in a Linear / Vercel-style dark theme, with a fixed navigation panel on the left + scrolling content on the right.\n\nSections:\n1. Overview (philosophy and how to use it)\n2. Colors (primary / neutral / semantic colors, with each swatch showing its color value and name)\n3. Typography (H1-H6 + body text + small text, annotated with font size and line height)\n4. Spacing (a visualization of the base spacing scale)\n5. Components (buttons / input fields / cards in their various states, with accompanying usage code snippets)\n6. Icons (commonly used icons)\n\nClicking the navigation scrolls smoothly, components have hover transitions, and code snippets have simple syntax highlighting.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 7, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-05-03T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": true, "recommended_order": 16 }, { "id": "developer-api-docs-portal", "title": "开发者 API 文档门户", "title_en": "Developer API Documentation Portal", "subtitle": "Stripe 风格三栏布局，代码示例多语言切换", "subtitle_en": "Stripe-style three-column layout with multi-language code samples", "description": "手动编写 API 文档费时费力，格式不统一、代码示例过时、多语言维护成本高，开发者体验差。本案例一键生成 Stripe 风格的三栏 API 文档门户，支持深色/浅色主题切换、多语言代码示例（Python/JavaScript/cURL）、语法高亮、参数表格自动渲染，让你的 API 文档专业感拉满。适合后端工程师、技术写作者、API 产品经理，以及任何需要快速搭建高质量开发者文档的团队。", "description_en": "Manually writing API documentation is time-consuming, inconsistent in format, with outdated code samples and high multi-language maintenance costs, resulting in poor developer experience. This case generates a Stripe-style three-column API documentation portal with dark/light theme toggle, multi-language code samples (Python/JavaScript/cURL), syntax highlighting, and auto-rendered parameter tables — making your API docs look professionally polished. Perfect for backend engineers, technical writers, API product managers, and any team needing high-quality developer documentation fast.", "categories": [ "dev" ], "tags": [ "API文档", "开发者体验", "技术文档", "代码示例", "接口设计" ], "tags_en": [ "API Docs", "Developer Experience", "Technical Docs", "Code Samples", "API Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "API 文档门户" } ], "prompt": "请参考 Stripe 官方文档风格，生成一个 AI API 服务文档门户页面（单文件 HTML，内联无外部依赖），支持深色/浅色主题切换（默认深色），响应式。\n\n经典三栏布局：\n- 左栏：可折叠导航（Introduction、Authentication、Chat Completions、Models、Error Handling）\n- 中栏：每个 API 端点展示 HTTP Method、Path、描述、参数表（类型/required 标记）、响应说明\n- 右栏：代码示例（Python / JavaScript / cURL 三语言 Tab 切换 + 语法高亮 + 一键复制 + JSON 响应示例）\n\n左栏点击平滑滚动，小屏时右栏折叠到文档下方。整体简洁、信息密度高、专业。", "prompt_en": "Referencing the style of Stripe's official documentation, please generate an AI API service documentation portal page (single-file HTML, inline with no external dependencies), supporting dark/light theme toggling (dark by default), responsive.\n\nClassic three-column layout:\n- Left column: collapsible navigation (Introduction, Authentication, Chat Completions, Models, Error Handling)\n- Center column: each API endpoint shows HTTP Method, Path, description, parameter table (type / required markers), and response notes\n- Right column: code samples (Python / JavaScript / cURL three-language tab switching + syntax highlighting + one-click copy + JSON response examples)\n\nClicking left-column items smoothly scrolls; on small screens the right column collapses below the documentation. Overall clean, high information density, professional.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-05-03T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "diet-analysis", "title": "饮食分析", "title_en": "Diet Analysis", "subtitle": "每天吃得好不好，不再凭感觉：全餐热量拆解 + 结合身体状况提示风险 + 下一餐调整方案，一页讲清", "subtitle_en": "Stop guessing whether you're eating well: meal calorie breakdown, risk alerts based on your health, and next-meal adjustments in one page", "description": "一日三餐天天吃，但真正说得清自己吃得健不健康的人并不多——热量多少、搭配合不合理、跟自己的身体状况合不合拍，大多凭感觉。对有基础病的人来说，这种“凭感觉”其实藏着不小的风险。\n\n以一位 42 岁、有风湿性心脏病的男士为例：他把自己的一顿晚餐发给微小医，问「这顿饭健不健康、以后怎么吃」。微小医帮他算清了全餐热量、逐道菜拆解，结合他的病史指出了心脏方面的风险，还告诉他下一餐怎么换主食、改做法、控分量——最后生成一页清清爽爽的图文报告，存成图片发给家人，一起照着吃。", "description_en": "We eat three meals a day, yet few of us can say with confidence whether we're eating well — the calories, the balance, whether it suits our body, all left to guesswork. For people living with chronic conditions, that guesswork carries real risk.\n\nTake a 42-year-old man with rheumatic heart disease: he sent his dinner to Weixiaoyi and asked, \"Is this healthy, and how should I eat going forward?\" Weixiaoyi worked out the total calories, broke down each dish, flagged the cardiac risks tied to his condition, and showed him how to adjust the next meal — all wrapped in a clean one-page report he could save as an image and share with his family.", "categories": [ "business" ], "tags": [ "饮食分析", "热量计算", "健康饮食", "微医", "HTML生成" ], "tags_en": [ "Diet Analysis", "Calorie Calculation", "Healthy Eating", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "diet-analysis.html", "artifacts": [ { "file": "diet-analysis.html", "label": "晚餐健康分析报告 HTML 页面" } ], "prompt": "帮我分析这顿晚饭健不健康，卡路里分别是多少如何，未来可以怎么进行调整。我的个人信息是：男，42岁，176cm，74kg，患有风湿性心脏病和脱发。将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203），logo裁成圆形。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果；5. 重点内容用流程图、指标卡片等特殊样式进行说明。", "prompt_en": "Analyze whether this dinner is healthy, how many calories each dish contains, and how I can adjust my diet going forward. My personal info: male, 42 years old, 176cm, 74kg, with rheumatic heart disease and hair loss. Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for sharing; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203), with the logo cropped into a circle. HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation; 5. Key content should be explained with special styles such as flow charts and metric cards.", "skills": [], "experts": [ { "id": "we-health-expert", "name": "微小医", "description": "微医官方 AI 健康管理专家：报告解读、就医推荐、健康与用药咨询", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_iLQ9gn0VnEOh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-20T18:30:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "digital-life-curator-personal-h5", "title": "数字生活策展人 · 个人介绍互动网页", "title_en": "Digital Life Curator · Interactive Personal Profile", "subtitle": "把简历改造成「交朋友」的社交名片", "subtitle_en": "Turn your resume into a social calling card", "description": "找工作的简历写多了，可是想交朋友的时候总说不清自己是谁。\n\n这个案例用腾讯文档 Smart Page 跑一份『数字生活策展人』风格的个人主页——按七组件叙事（Hero / Stats / Timeline / Bento / Skills / Testimonial / Contact）把你的硬经历、软经历、爱好、趣味数据、好友吐槽、破冰话题全部编排进去，最终产出一个零外部依赖的单文件 H5，双击即开、手机可看、有滚动入场和点击微交互。\n\n适合：想留好第一印象的产品经理、设计师、内容创作者、自由职业者，以及想给自己的博客/数字名片配一个 About Me 的人。", "description_en": "Tired of writing job-hunting resumes when all you want is to make friends?\n\nThis case uses Tencent Docs Smart Page to generate a 'Digital Life Curator' style personal profile — arranging your life across seven components (Hero / Stats / Timeline / Bento / Skills / Testimonial / Contact) and producing a single-file H5 page with zero external dependencies: double-click to open, mobile-friendly, with scroll reveals and tiny interactions.\n\nTarget audience: PMs, designers, creators, and freelancers who want a warm, memorable first impression — or anyone looking for a polished About Me page for a blog or digital business card.", "categories": [ "creation" ], "tags": [ "个人主页", "互动H5", "数字名片", "About Me", "Smart Page" ], "tags_en": [ "personal page", "interactive H5", "digital card", "About Me", "Smart Page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "个人介绍 H5 互动网页" }, { "file": "persona.json", "label": "结构化数据（可复用）" } ], "prompt": "我想用腾讯文档 Smart Page 做一份『数字生活策展人』风格的个人介绍互动网页——不是简历，是交朋友用的社交名片。\n\n请按下面的七组件叙事骨架输出单文件 H5（CSS/JS 全内联、系统字体栈、适配移动端、尊重 prefers-reduced-motion）：\n\n1. HeroCard：身份混合体标题 + 拉近距离的大白话 slogan + Vibe 标签。\n2. StatsBoard：6 项有趣的生活化数据，而非 KPI。\n3. TimelinePro：至少 5 条双轨时间轴，混入生活节点。\n4. BentoGallery：6 格兴趣宫格，把爱好包装成『项目』，每格一句心得；配图用 CSS 渐变或 SVG。\n5. SkillRadar：8 项技能，混入 2–3 项软技能。\n6. Testimonial：6 条朋友的真实吐槽或暖心评价，标签墙风格，带关系。\n7. ContactSocial：每个渠道给出具体破冰话题和行动号召。\n\n主题：Creative_Light（画廊风，琥珀 + 墨绿）。\n\n【个人信息】\n[此处粘贴你的简历 + 爱好 + 碎片经历，或口述一段；信息缺失处请合理脑补，让内容像一个活生生的人。]", "prompt_en": "I want to use Tencent Docs Smart Page to create a \"digital life curator\"-style personal-introduction interactive web page — not a resume, but a social calling card for making friends.\n\nPlease output a single-file H5 following the seven-component narrative skeleton below (all CSS/JS inlined, system font stack, mobile-friendly, respecting prefers-reduced-motion):\n\n1. HeroCard: a hybrid-identity title + a down-to-earth, approachable slogan + Vibe tags.\n2. StatsBoard: 6 fun, life-flavored stats rather than KPIs.\n3. TimelinePro: at least 5 dual-track timeline entries, mixing in life milestones.\n4. BentoGallery: a 6-cell interest grid that packages hobbies as \"projects,\" each with a one-line takeaway; use CSS gradients or SVG for imagery.\n5. SkillRadar: 8 skills, mixing in 2-3 soft skills.\n6. Testimonial: 6 genuine roasts or heartwarming comments from friends, in a tag-wall style, each with the relationship.\n7. ContactSocial: a concrete ice-breaker and a call-to-action for each channel.\n\nTheme: Creative_Light (gallery style, amber + dark green).\n\n[Personal Info]\n[Paste your resume + hobbies + bits of experience here, or just describe yourself; fill any gaps with reasonable inferences so the content reads like a living, breathing person.]", "skills": [ { "id": "smart-page", "name": "腾讯文档智能页面 · Smart Page", "description": "下一代可交互在线汇报网页生成工具，内置叙事骨架与动效模板，生成即获永久在线链接，覆盖汇报/邀请函/官网/H5/个人主页等场景。", "skill_id": "skill_2053082927661453312" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 6, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-11T00:00:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": true, "recommended_order": 15 }, { "id": "dingtalk-group-digest", "title": "项目群消息智能日报", "title_en": "Project Group Chat Digest", "subtitle": "一晚 300 条群消息，浓缩成 3 分钟读完的日报", "subtitle_en": "300 overnight messages distilled into a 3-minute digest", "description": "项目群一晚上刷几百条消息，早上爬楼半小时还抓不住重点？这个案例演示用钉钉套件把指定项目群的消息做成「智能日报」：按讨论话题自动聚类，每个话题给出参与人、结论与关键原文摘录；待办自动提取并标注负责人和截止时间；未决分歧与风险单独标出；结尾附三句话管理者摘要。所有数据来自钉钉群消息真实查询，人名按需脱敏，产出一份可离线打开的单页日报。适合：项目经理、Scrum Master、多项目并行的团队负责人、不想爬楼的深度协作者、需要向上汇报群内进展的同学。", "description_en": "A project group can pile up hundreds of messages overnight — and scrolling through them costs half an hour before you even find the point. This case shows the DingTalk suite turning a project group's messages into a smart daily digest: discussions auto-clustered by topic with participants, conclusions and key quotes; action items extracted with owners and deadlines; unresolved disagreements and risks flagged separately; and a closing three-sentence manager summary. All data comes from real DingTalk group-message queries, names anonymized on demand, delivered as one offline-ready page. Ideal for: project managers, Scrum Masters, multi-project team leads, deep collaborators who hate scrolling, and anyone reporting group progress upward.", "categories": [ "office" ], "tags": [ "钉钉", "群聊摘要", "话题聚类", "项目管理", "信息降噪" ], "tags_en": [ "DingTalk", "chat digest", "topic clustering", "project management", "noise reduction" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "项目群消息智能日报（双语单页 HTML）" } ], "prompt": "读取钉钉项目群「[群名]」昨天的消息，生成一份群消息智能日报，单文件 HTML、可离线打开。内容：1）按话题聚类，每个话题给出参与人、结论和关键原文摘录；2）提取待办与负责人，标注截止时间；3）单独标出未决分歧与风险；4）结尾给 3 句话管理者摘要。数据来自钉钉群消息真实查询，人名按需脱敏，页面注明来源与时间。", "prompt_en": "Read yesterday's messages from the DingTalk project group [group name] and generate a smart group-chat digest as a single offline-ready HTML file. Include: 1) topics auto-clustered, each with participants, conclusions and key quotes; 2) extracted action items with owners and deadlines; 3) unresolved disagreements and risks flagged separately; 4) a closing three-sentence manager summary. Data must come from real DingTalk group-message queries, names anonymized where needed, with source and time noted on the page.", "skills": [ { "id": "dingtalk-unified", "name": "钉钉套件", "description": "钉钉 CLI 套件，覆盖消息、日历、待办、审批、考勤、日志、文档、AI 表格、钉盘、AI 听记、邮箱等 14 个产品能力", "skill_id": "skill_2054903319109931008" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-27T17:05:00+08:00", "updated_at": "2026-08-27T17:05:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "dingtalk-morning-briefing", "title": "钉钉晨间一站简报", "title_en": "DingTalk Morning Briefing Hub", "subtitle": "日程、待办、审批、邮件，一杯咖啡的时间全知道", "subtitle_en": "Calendar, todos, approvals and mail — all before your first coffee", "description": "每天早上打开钉钉，日程要看日历、待办翻列表、审批切 OA、邮件再进一遍邮箱，四五个入口切下来，晨会都开始了。这个案例演示用钉钉套件一句话拉起「今日晨间简报」：自动读取今天的日程并标出连开会议与空闲窗口、按优先级整理待办、汇总待我处理的审批及已等待时长、摘要重要未读邮件，最后在页首给出「今日最该优先做的三件事」及理由，生成一份可离线打开的单页简报。适合：每天日程密集的职场人、管理者与团队负责人、行政与项目助理、想把钉钉多个入口收敛成一屏看的效率党。", "description_en": "Every morning in DingTalk you hop between calendar, todo list, OA approvals and mail — by the time you are done, the standup has already started. This case shows the DingTalk suite pulling up a one-shot morning briefing: today's schedule with back-to-back meetings and free windows flagged, todos grouped by priority, approvals waiting on you with pending duration, key unread mails digested, and a top-three-things-to-do verdict at the very top — delivered as a single offline-ready page. Ideal for: schedule-heavy professionals, managers and team leads, admin and project assistants, and anyone who wants five DingTalk entrances collapsed into one screen.", "categories": [ "office", "productivity" ], "tags": [ "钉钉", "晨报", "日程管理", "待办清单", "办公自动化" ], "tags_en": [ "DingTalk", "morning briefing", "calendar", "todo list", "office automation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "今日晨间一站简报（双语单页 HTML）" } ], "prompt": "读取我的钉钉数据，生成今日晨间简报，单文件 HTML、可离线打开。内容：1）今天的日程，标出连开会议和空闲时段；2）待办按优先级分组，今天到期的突出显示；3）待我处理的审批列表及已等待时长；4）重要未读邮件摘要；5）页首「今日聚焦」：最该优先做的三件事及理由。数据来自钉钉真实查询（日历/待办/审批/邮箱），页面注明数据来源与生成时间。", "prompt_en": "Read my DingTalk data and generate today's morning briefing as a single offline-ready HTML file. Include: 1) today's schedule with back-to-back meetings and free windows flagged; 2) todos grouped by priority with today's due items highlighted; 3) approvals waiting on me with pending duration; 4) digests of important unread mail; 5) a 'Today's Focus' verdict at the top: the three things to prioritize and why. Data must come from real DingTalk queries (calendar / todo / approvals / mail), with source and generation time noted on the page.", "skills": [ { "id": "dingtalk-unified", "name": "钉钉套件", "description": "钉钉 CLI 套件，覆盖消息、日历、待办、审批、考勤、日志、文档、AI 表格、钉盘、AI 听记、邮箱等 14 个产品能力", "skill_id": "skill_2054903319109931008" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-27T17:05:00+08:00", "updated_at": "2026-08-27T17:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "dingtalk-team-insights", "title": "考勤审批月度洞察看板", "title_en": "Attendance & Approval Monthly Insights", "subtitle": "考勤、加班、请假一屏看清，异常自动浮出", "subtitle_en": "Attendance, overtime and leave on one screen — anomalies surface themselves", "description": "每月做人力盘点，考勤导一张表、审批再导一张表，Excel 里对半天还算不清加班和请假的重叠。这个案例演示用钉钉套件生成「部门月度人力洞察看板」：出勤率与迟到缺卡分布、加班时长分布与 TOP 名单（脱敏）、请假/出差审批通过率与平均审批时长、以及与上月的环比和异常提醒，全部来自钉钉考勤与 OA 审批的真实查询，产出一份带图表、可离线打开的单页看板。适合：HR 与人力分析师、部门管理者、行政考勤负责人、需要月度人力汇报的 HRBP。", "description_en": "Monthly workforce reviews mean exporting one sheet for attendance and another for approvals, then wrestling Excel for half a day to reconcile overlapping overtime and leave. This case shows the DingTalk suite generating a monthly team insights dashboard: attendance rate with late/missed-punch distribution, overtime hours and the anonymized top list, leave and business-trip approval pass rates with average turnaround, plus month-over-month deltas and anomaly alerts — all from real DingTalk attendance and OA queries, delivered as a chart-rich offline-ready page. Ideal for: HR and people analysts, department managers, admin owners of attendance, and HRBPs preparing monthly workforce reports.", "categories": [ "office", "data" ], "tags": [ "钉钉", "考勤分析", "审批流程", "数据看板", "人力洞察" ], "tags_en": [ "DingTalk", "attendance analysis", "approval workflow", "dashboard", "people insights" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "部门月度人力洞察看板（双语单页 HTML）" } ], "prompt": "读取钉钉考勤与审批数据，为我的部门生成本月人力洞察看板，单文件 HTML、可离线打开。内容：1）出勤率、迟到/缺卡分布；2）加班时长分布与 TOP 名单（脱敏）；3）请假/出差审批通过率与平均审批时长；4）与上月的环比和异常提醒。数据来自钉钉考勤与 OA 审批真实查询，页面注明数据来源与生成时间。", "prompt_en": "Read DingTalk attendance and approval data and generate this month's team insights dashboard for my department as a single offline-ready HTML file. Include: 1) attendance rate with late/missed-punch distribution; 2) overtime hour distribution and the top list (anonymized); 3) leave/business-trip approval pass rates and average turnaround; 4) month-over-month deltas and anomaly alerts. Data must come from real DingTalk attendance and OA queries, with source and generation time noted on the page.", "skills": [ { "id": "dingtalk-unified", "name": "钉钉套件", "description": "钉钉 CLI 套件，覆盖消息、日历、待办、审批、考勤、日志、文档、AI 表格、钉盘、AI 听记、邮箱等 14 个产品能力", "skill_id": "skill_2054903319109931008" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-27T17:05:00+08:00", "updated_at": "2026-08-27T17:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "disease-consultation", "title": "疾病咨询", "title_en": "Disease Consultation", "subtitle": "体检指标异常别慌：把病讲明白 + 管理路线 + 饮食运动 + 用药须知 + 真实专家推荐，一页管全", "subtitle_en": "Don't panic over an abnormal checkup result: the condition explained, a management roadmap, diet & exercise, medication notes, and real specialist picks in one page", "description": "体检季一过，不少人拿着报告单陷入同一种焦虑：指标箭头忽上忽下，说病不是病、说没事又不放心。网上说法五花八门，越查越不知道该信谁、要不要紧、下一步该怎么办。\n\n以「糖尿病前期」为例：把心里的疑问讲给微小医听，它会用大白话把这个病说明白，给你一条清清楚楚的控糖路线、一日三餐的饮食示例、运动方案和用药须知，还帮你推荐杭州真实的内分泌专家——汇成一页指南，存图发给家人，一起照着做，心里踏实多了。", "description_en": "After every checkup season, many people find themselves stuck in the same anxiety: the arrows on the report point up and down — not quite a disease, but not nothing either. The advice online is all over the place, and the more you read, the less you know whom to trust or what to do next.\n\nTake prediabetes as an example: share your worries with Weixiaoyi, and it explains the condition in plain language, lays out a clear glucose-control roadmap, sample meals, an exercise plan and medication notes, and even points you to real endocrinology specialists in Hangzhou — all in a one-page guide you can save and share with family, so you can face it together with peace of mind.", "categories": [ "business" ], "tags": [ "疾病咨询", "糖尿病前期", "健康管理", "微医", "HTML生成" ], "tags_en": [ "Disease Consultation", "Prediabetes", "Health Management", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "prediabetes-guide.html", "artifacts": [ { "file": "prediabetes-guide.html", "label": "糖尿病前期健康管理指南 HTML 页面" } ], "prompt": "什么是糖尿病前期？糖尿病前期控糖该怎么做？饮食上该怎么吃？运动怎么做？可以吃什么药？还有什么注意事项？有什么推荐医生或医院？将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203），logo裁成圆形。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果。5. 重点内容用流程图、指标卡片等特殊样式进行说明。", "prompt_en": "What is prediabetes? How should I control blood sugar in the prediabetes stage? What should I eat? How should I exercise? What medications can I take? What other precautions are there? Any recommended doctors or hospitals? Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for the sharing scenario; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203), with the logo cropped into a circle. HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation; 5. Key content should be illustrated with special styles such as flowcharts and metric cards.", "skills": [], "experts": [ { "id": "we-health-expert", "name": "微小医", "description": "微医官方 AI 健康管理专家：报告解读、就医推荐、健康与用药咨询", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_iLQ9gn0VnEOh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-20T17:15:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "dividend-lowvol-vs-bank-income-play", "title": "红利低波 vs 银行ETF", "title_en": "Dividend Low Vol vs Banks", "subtitle": "给“防御收益型底仓”一个真正能落地的条件化对比框架。", "subtitle_en": "A practical framework for comparing defensive income sleeves with real conditions attached.", "description": "做防御收益型底仓时，最怕一句“红利”和一句“银行”就把事情说完。真正影响持有体验的，是回撤、波动、集中度和流动性，而不是标签本身。\n\n这个案例真实对比了红利低波ETF华泰柏瑞和银行ETF华宝，把费率、收益、回撤、持仓差异和配置场景整理成一份像财富顾问手册的双语 HTML 页面，更适合直接拿去展示或复用。\n\n适合：关注防御收益 ETF 的个人投资者、做资产配置沟通的财富顾问、需要稳定风格案例的金融产品团队、想展示 Skill 交付质感的运营同学。", "description_en": "When you build a defensive income sleeve, the worst habit is to stop at the labels “dividend” and “banks”. What really shapes holding experience is drawdown, volatility, concentration and liquidity rather than the label itself.\n\nThis case compares Huatai Dividend Low Vol ETF and HuaBao Bank ETF with real data, then turns fees, return, drawdown, holdings divergence and allocation scenarios into a bilingual HTML page that feels closer to a wealth-advisor booklet than a plain fact sheet.\n\nBest for: individual investors looking at defensive-income ETFs, wealth advisors communicating allocation choices, finance product teams that need a stable-style case, and operators who want to show delivery quality of the skill.", "categories": [ "finance", "data" ], "tags": [ "红利低波", "银行ETF", "防御收益", "资产配置", "双语HTML" ], "tags_en": [ "dividend low vol", "bank ETF", "defensive income", "asset allocation", "bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于 ETF深度对比 skill 的官方能力，对红利低波ETF华泰柏瑞（sh512890）和银行ETF华宝（sh512800）做一次真实公开数据对比。使用 westock-data 官方 CLI 获取 ETF detail、holdings、quote、kline 等数据，重点比较规模、费率、过去 1 个月 / 3 个月 / 年初至今收益、波动、最大回撤、前十大重仓、流动性与适用场景。最终交付一个单文件 HTML 报告，资源全部内联，可离线打开。结论请写成“什么账户适合哪一只 + 主要风险提醒”，不要输出确定性买卖建议。", "prompt_en": "Please use the official capability of the ETF depth comparison skill to compare Huatai Dividend Low Vol ETF (sh512890) and HuaBao Bank ETF (sh512800) with real public data. Use the official westock-data CLI to fetch ETF detail, holdings, quote and kline data, then compare scale, fees, 1-month / 3-month / YTD return, volatility, max drawdown, top holdings, liquidity and scenario fit. Deliver a single-file HTML report, fully inlined and offline-ready. Present the conclusion as “which account setup fits which ETF” plus key risk notes instead of a deterministic buy or sell call.", "skills": [ { "id": "etf-depth-comparison", "name": "ETF深度对比", "description": "两只A股ETF多维度深度对比", "skill_id": "skill_2076627838067363840" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "ETF 深度对比 HTML 报告" }, { "file": "data.json", "label": "ETF 对比原始数据摘要" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:20:38+08:00", "updated_at": "2026-07-30T16:22:28+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-api-reference", "title": "Orders API 接口文档", "title_en": "Orders API Reference", "subtitle": "订单服务 REST API 开发者参考文档", "subtitle_en": "Developer reference for the Orders Service REST API", "description": "接入方常被零散、过时、缺少可运行示例的接口文档困扰，导致联调反复返工。本案由「文通通」专家按开发者文档方法论，为虚构的「订单服务 Orders API」编写了一份完整自洽的 REST 参考文档：覆盖 API 概览、Bearer Token 鉴权、基础 URL，以及 GET/POST/PATCH/DELETE 共 5 个端点的参数表、可运行 curl 请求与 JSON 响应示例、状态码，并附通用错误码、限流策略与版本变更。交付为单文件 HTML，左侧端点导航、右侧详情、深色代码卡呈现，适合需要快速产出规范接口文档的后端与平台团队。", "description_en": "Developers often struggle with scattered, outdated API docs that lack runnable examples, causing repeated integration rework. Here the 'Technical Documentation Engineer' expert applies developer-docs methodology to write a complete, self-consistent REST reference for a fictional 'Orders API': covering overview, Bearer Token auth, base URL, plus parameter tables, runnable curl requests, JSON response examples and status codes for 5 GET/POST/PATCH/DELETE endpoints, with common error codes, rate limiting and a changelog. Delivered as a single-file HTML with left endpoint nav, right detail pane and dark code cards — ideal for backend and platform teams who need standardized API docs fast.", "categories": [ "office" ], "tags": [ "API文档", "接口参考", "开发者文档", "REST" ], "tags_en": [ "API Docs", "API Reference", "Developer Docs", "REST" ], "artifact_type": "html", "prompt": "帮我写一份「订单服务 Orders API」的 REST 接口参考文档，做成单文件 HTML，开发者文档风（类似 Stripe/ReadMe）：左侧是端点导航、点击平滑滚动到右侧详情，深色代码卡呈现请求/响应。内容要自洽完整：API 概览、Bearer Token 鉴权与基础 URL，以及 GET /orders、POST /orders、GET /orders/{id}、PATCH、DELETE 共五个端点，每个含方法/路径/参数表/curl 请求示例/JSON 响应示例/状态码；再加通用错误码表、限流说明和版本变更。HTTP 方法用语义色标签，要响应式、不要外链。", "prompt_en": "Write a REST API reference for a fictional 'Orders Service Orders API' as a single-file HTML in a developer-docs style (like Stripe/ReadMe): a left endpoint nav that smooth-scrolls to the right detail pane, with dark code cards for request/response. Make the content complete and self-consistent: API overview, Bearer Token auth and base URL, plus five endpoints — GET /orders, POST /orders, GET /orders/{id}, PATCH, DELETE — each with method/path/parameter table/curl request/JSON response/status codes; add a common error-code table, rate limiting, and a changelog. Use semantic-colored HTTP method tags, make it responsive, and use no external resources.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "Orders API 接口参考文档" } ], "skills": [], "experts": [ { "id": "TechnicalDocumentationEngineer", "name": "文通通", "description": "将复杂技术概念转化为清晰准确的文档，让技术知识可传播", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f779afe385ae", "name_en": "Technical Documentation Engineer", "description_en": "Turns complex technical concepts into clear, accurate, shareable documentation", "expert_id": "ex_3zidva69KVGL" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-book-summary-notes", "title": "《思考，快与慢》精读笔记卡", "title_en": "Thinking, Fast and Slow — Reading Notes", "subtitle": "把认知心理学经典整理成可随时翻阅的知识卡片", "subtitle_en": "A classic on cognitive psychology, distilled into reference cards", "description": "读完一本好书，观点却很快忘光、难以复用？本案由「文通通」专家把丹尼尔·卡尼曼的《思考，快与慢》拆解整理为一份结构化精读笔记：从一句话总览、系统1与系统2双系统框架，到锚定、可得性、损失厌恶、框架效应等关键认知偏误卡片、全书脉络图、5条金句，以及对工作与决策的6条行动启发。交付物是一份优雅阅读笔记风的单文件HTML摘要卡，米色纸感配衬线标题与书签色带，可直接收藏复用。适合管理者、产品运营、投资者，以及想系统理解'人为何非理性'的读者。", "description_en": "Finished a great book but forgot the ideas a week later? Here the Technical Documentation Engineer expert distills Daniel Kahneman's 'Thinking, Fast and Slow' into structured reading notes: a one-line overview, the System 1 vs System 2 framework, cards for key cognitive biases (anchoring, availability, loss aversion, framing), a chapter map, five quotes, and six action takeaways for work and decisions. The deliverable is an elegant single-file HTML summary card with a paper-toned, serif, bookmark-ribbon style — ready to keep and reuse. Great for managers, product and ops folks, investors, and anyone wanting to understand why people are predictably irrational.", "categories": [ "office" ], "tags": [ "读书笔记", "知识卡片", "认知科学", "决策思维" ], "tags_en": [ "Reading Notes", "Knowledge Cards", "Cognitive Science", "Decision Making" ], "artifact_type": "html", "prompt": "帮我把《思考，快与慢》（丹尼尔·卡尼曼）整理成一份精美的读书笔记网页，单文件HTML。要优雅的阅读笔记风格：米色纸感背景、衬线标题、书签色带。内容包括书籍信息（作者/领域/核心命题）、一句话总览、系统1与系统2的双系统核心框架对比、关键认知偏误卡片（锚定、可得性、损失厌恶、框架效应等）、全书章节脉络图、3-5条金句引用块、对工作和决策的行动启发、以及适合谁读。用图书馆书架的大图做hero并叠渐变蒙版压标题。要响应式，手机上不裁切。", "prompt_en": "Help me turn 'Thinking, Fast and Slow' by Daniel Kahneman into a beautiful reading-notes web page as a single-file HTML. Elegant reading-notes style: paper-toned beige background, serif headings, a bookmark ribbon. Include: book info (author/field/core thesis), a one-line overview, the System 1 vs System 2 framework comparison, key cognitive bias cards (anchoring, availability, loss aversion, framing, etc.), a chapter map, 3-5 quote blocks, action takeaways for work and decisions, and who should read it. Use a library bookshelf hero image with a gradient overlay for the title. Make it responsive with no cropping on mobile.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "《思考，快与慢》精读笔记卡" } ], "skills": [], "experts": [ { "id": "TechnicalDocumentationEngineer", "name": "文通通", "description": "将复杂技术概念转化为清晰准确的文档，让技术知识可传播", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f779afe385ae", "name_en": "Technical Documentation Engineer", "description_en": "Turns complex technical concepts into clear, accurate, shareable documentation", "expert_id": "ex_3zidva69KVGL" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-competitive-analysis-brief", "title": "协作办公工具赛道竞品调研分析简报", "title_en": "Collaboration Tools Competitive Analysis Brief", "subtitle": "4 款竞品横向对比 · 功能矩阵 · SWOT · 定位结论", "subtitle_en": "4-product benchmark · feature matrix · SWOT · positioning", "description": "要进入已有头部玩家的协作办公赛道，团队往往缺一份口径统一、结论可追溯的横向对比。本案由「专业文档生成团队」专家以企业级长文档工作流（需求分析→检索→撰写→合规审核逐章闭环）产出一份竞品调研分析简报：涵盖调研方法、4 款竞品概览卡、13 项功能对比矩阵、六维能力评分条、价格×功能完整度定位象限图、逐家 SWOT 与差异化机会建议，最终给出清晰的进入定位结论。交付为单文件 HTML 长页，配真实签约场景配图，适合做竞品调研、产品立项与战略汇报的产品、战略与市场团队参考。", "description_en": "Entering a mature collaboration-tools market needs a consistent, traceable benchmark. Here the Professional Document Team expert applies an enterprise long-document workflow (analyze → research → write → audit) to produce a competitive analysis brief: research method, 4 competitor cards, a 14-item feature matrix, six-dimension score bars, a price-by-completeness positioning quadrant, per-product SWOT, gap-based differentiation ideas, and a clear positioning conclusion. Delivered as a single-file HTML long page with a real signing-scene photo, it suits product, strategy and marketing teams doing competitive research and go-to-market planning.", "categories": [ "office" ], "tags": [ "竞品分析", "功能矩阵", "市场定位", "调研简报" ], "tags_en": [ "Competitive Analysis", "Feature Matrix", "Market Positioning", "Research Brief" ], "artifact_type": "html", "prompt": "帮我做一份「协作办公工具赛道」的竞品调研分析简报，单文件 HTML 长页、咨询报告风、专业浅色配色。要包含：调研背景与方法；3-4 个竞品概览卡（定位/目标客户/定价/亮点）；功能对比矩阵（行=功能项，列=各竞品，用✓/部分/✗ 图标）；多维能力评分条对比；价格×功能完整度的市场定位象限图（SVG，坐标轴正确）；各竞品 SWOT 卡；机会缺口与差异化建议；以及定位结论。请联网搜索各竞品的真实公开信息（官网/财报/媒体报道）作为分析依据并注明来源；确无公开数据处才标注为估计，配一张签约/文书场景的图作 hero。", "prompt_en": "Create a competitive analysis brief for the collaboration-tools market as a single-file HTML long page in a consulting-report style with a clean light palette. Include: research background and method; 3-4 competitor overview cards (positioning/target/pricing/highlights); a feature comparison matrix (rows=features, columns=competitors, using check/partial/cross icons); multi-dimension capability score bars; a price-by-feature-completeness positioning quadrant (SVG with correct axes); per-competitor SWOT cards; gap and differentiation recommendations; and a positioning conclusion. real real data labeled as real (web-searched) is fine, with a signing/document scene image as the hero.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "竞品调研分析简报" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-contract-risk-radar", "title": "合同条款风险审查雷达", "title_en": "Contract Clause Risk Review Radar", "subtitle": "逐条标注风险等级，给出修改建议与处置清单", "subtitle_en": "Per-clause risk grading with redline suggestions and a triage checklist", "description": "业务和法务看一份几十页的合同，最怕漏掉藏在条款里的风险点。本案例由「专业文档生成团队」把一份企业 SaaS 采购合同做成可视化风险审查页——逐条标注高/中/低风险等级、风险说明、修改建议与参考依据，支持按等级筛选，并给出必改/可协商/可接受处置清单。适合需要快速评估合同风险的法务、采购与业务负责人。", "description_en": "When business and legal teams review a dozens-of-pages contract, the biggest fear is missing risks buried in the clauses. Built by the Professional Document Team, this case turns an enterprise SaaS procurement contract into a visual risk-review page — per-clause high/medium/low grading, risk notes, redline suggestions and references, with filter-by-level and a must-fix/negotiable/acceptable triage list.", "categories": [ "office" ], "tags": [ "合同审查", "风险管理", "法务", "文档处理" ], "tags_en": [ "Contract Review", "Risk Management", "Legal", "Document Processing" ], "artifact_type": "html", "prompt": "帮我审查一份企业 SaaS 采购合同，生成一个可视化风险审查页（单文件 HTML）。要直观看到：合同概要与总体风险评级、高/中/低风险条款统计与分布、逐条风险审查卡片（条款名/风险等级/风险说明/修改建议/参考依据）、按风险等级筛选，以及必改/可协商/可接受处置清单。法务严谨的暖红警示风格，界面中文，金额用 ¥。", "prompt_en": "Help me review an enterprise SaaS procurement contract and generate a visual risk-review page (single-file HTML). Show contract summary with overall risk rating; high/medium/low clause counts and distribution; per-clause review cards (clause/risk level/risk note/redline/reference); filter by risk level; and a must-fix/negotiable/acceptable triage list. Rigorous legal style with a warm-red alert palette; UI in Simplified Chinese; amounts in ¥.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "合同条款风险审查雷达" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-06-20T11:00:00Z", "updated_at": "2026-06-20T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-meeting-decision-digest", "title": "会议决议精炼看板", "title_en": "Meeting Decision Digest Board", "subtitle": "把冗长会议记录提炼成决议、待办与时间线", "subtitle_en": "Turn a long meeting record into decisions, action items and a timeline", "description": "开完一场长会，最有价值的是决议和待办，但它们常常淹没在大段记录里。本案例由「专业文档生成团队」把一场跨部门季度评审会提炼成决议看板——决议事项与待办清单双栏呈现、议题时间线、分歧与风险、下次会议安排。适合需要快速对齐会议结论、追踪落地的团队。", "description_en": "After a long meeting, the decisions and action items matter most — yet they're often buried in walls of notes. Built by the Professional Document Team, this case distills a cross-functional quarterly review into a decision board — decisions and action items side by side, an agenda timeline, disagreements and risks, and the next-meeting plan.", "categories": [ "office" ], "tags": [ "会议决议", "待办", "时间线", "文档处理" ], "tags_en": [ "Decisions", "Action Items", "Timeline", "Document Processing" ], "artifact_type": "html", "prompt": "帮我把一场跨部门季度评审会的记录提炼成会议决议看板（单文件 HTML）。要直观看到：会议主题/时间/参会人/一句话结论与会议健康度、决议事项与待办事项双栏（待办含负责人/截止/优先级）、议题时间线、分歧与风险、下次会议安排。浅灰极简办公风、时间线 + 双栏布局，界面中文。", "prompt_en": "Help me distill a cross-functional quarterly review meeting into a decision board (single-file HTML). Show topic/time/attendees/one-line conclusion and meeting-health tags; decisions and action items in two columns (with owner/deadline/priority); an agenda timeline; disagreements and risks; and the next-meeting plan. Light-gray minimalist office style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "scenario_cover_en": "scenario-cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "会议决议精炼看板" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-06-20T11:00:00Z", "updated_at": "2026-06-20T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-paper-visual-abstract", "title": "学术论文可视摘要", "title_en": "Paper Visual Abstract", "subtitle": "把一篇长论文浓缩成一页图解：问题→方法→结果→结论", "subtitle_en": "Condense a long paper into a one-page graphic: problem to conclusion", "description": "读一篇学术论文动辄十几页，想快速抓住贡献点很费力。本案例由「文通通」技术文档工程师把一篇论文提炼成一张可视摘要（Visual Abstract）——用图解串起研究问题、方法架构、关键结果与结论，配关键数据小图表和引用卡片。适合科研、技术选型与需要快速读懂论文的人。", "description_en": "Academic papers run a dozen-plus pages, making it hard to grab the contributions fast. Built by the Technical Documentation Engineer, this case distills a paper into a visual abstract — a graphic flow linking research question, method architecture, key results and conclusion, with small result charts and a citation card.", "categories": [ "office" ], "tags": [ "论文", "可视摘要", "信息图", "学术" ], "tags_en": [ "Paper", "Visual Abstract", "Infographic", "Academic" ], "artifact_type": "html", "prompt": "帮我把一篇学术论文提炼成一张可视摘要单页（单文件 HTML）。论文方向为基于检索增强的长文档问答。要直观看到：标题/作者/TL;DR、研究问题→方法→关键结果→结论的图解流程、方法架构框图、关键结果小图表、贡献点与局限、如何引用。浅蓝学术期刊风格、图解为主，界面中文（标题可中英）。", "prompt_en": "Help me distill an academic paper into a one-page visual abstract (single-file HTML). The paper is about retrieval-augmented long-document QA. Show title/authors/TL;DR; a graphic flow of research question to method to key results to conclusion; a method architecture diagram; small result charts; contributions and limitations; and how to cite. Light-blue academic-journal style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "学术论文可视摘要" } ], "skills": [], "experts": [ { "id": "TechnicalDocumentationEngineer", "name": "文通通", "description": "将复杂技术概念转化为清晰准确的文档，让技术知识可传播", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f779afe385ae", "name_en": "Technical Documentation Engineer", "description_en": "Turns complex technical concepts into clear, accurate, shareable documentation", "expert_id": "ex_3zidva69KVGL" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-06-20T11:00:00Z", "updated_at": "2026-06-20T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-policy-interpretation-card", "title": "政策文件解读卡", "title_en": "Policy Interpretation Card", "subtitle": "原文与白话解读对照，讲清政策对企业的影响", "subtitle_en": "Source-text vs plain-language interpretation of policy impact on businesses", "description": "政策文件读起来晦涩，企业最关心的是这条对我有什么影响、我该做什么。本案例由「专业文档生成团队」把一份企业数据合规政策做成解读卡——原文与白话解读左右对照、重点变化前后对比、合规自查清单与关键时间线。适合需要快速读懂政策并落地合规的企业法务与管理者。", "description_en": "Policy documents are dense, but what businesses care about is how a clause affects them and what to do. Built by the Professional Document Team, this case turns a corporate data-compliance policy into an interpretation card — source text vs plain-language reading side by side, before/after change comparisons, a self-check list and a key timeline.", "categories": [ "office" ], "tags": [ "政策解读", "合规", "对照", "文档处理" ], "tags_en": [ "Policy Interpretation", "Compliance", "Side-by-side", "Document Processing" ], "artifact_type": "html", "prompt": "帮我把一份企业数据合规政策做成政策解读卡（单文件 HTML）。要直观看到：政策名称/发文机关/生效日期/适用对象/核心要义、逐条原文与白话解读对照（含对企业影响与应对动作）、重点变化前后对比、合规自查清单、关键时间线。红金庄重政务风、左右对照布局，界面中文；请联网搜索并引用真实政策原文条款（注明政策名与发文机关），解读部分标注为参考建议。", "prompt_en": "Help me turn a corporate data-compliance policy into an interpretation card (single-file HTML). Show policy name/issuing body/effective date/applicable parties/core gist; clause-by-clause source-text vs plain-language interpretation (with business impact and actions); before/after change comparisons; a compliance self-check list; and a key timeline. Solemn red-gold government style; clauses real policy text via web search, marked for-reference-only; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "政策文件解读卡" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-06-20T11:00:00Z", "updated_at": "2026-06-20T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-prd-spec-sheet", "title": "智能日程助手 PRD 可视化一页", "title_en": "Smart Scheduling Assistant PRD on One Page", "subtitle": "把背景、目标、用户故事与功能规格整合为可滚动的单页需求文档", "subtitle_en": "Background, goals, user stories and specs unified into a scrollable one-page PRD", "description": "产品需求文档常散落在多处、字段不一致、验收标准模糊，开发与评审反复对齐成本高。本案由「文通通」专家运用文档工程方法论：先定结构再写正文、一节一概念、术语统一、给出可验收的明确标准，把虚构功能「智能日程助手」的文档元信息、背景与问题、目标与非目标、用户故事、功能规格表、交互流程、验收清单、非功能需求、里程碑与待决问题，整理为一份配章节锚点导航的精美单页 HTML PRD。交付物可直接用于需求评审与开发对齐，适合产品经理、技术文档工程师与研发团队。", "description_en": "PRDs are often scattered, inconsistent in fields, and vague on acceptance criteria, making review and dev alignment costly. Here the Technical Documentation Engineer applies a docs-engineering method—structure first, one concept per section, consistent terms, testable acceptance criteria—to turn a fictional Smart Scheduling Assistant into a polished single-page HTML PRD with anchor navigation, covering metadata, background, goals/non-goals, user stories, a spec table, interaction flow, acceptance checklist, non-functional requirements, milestones and open questions. Ready for review and dev alignment—great for PMs, tech writers and engineering teams.", "categories": [ "office" ], "tags": [ "产品需求文档", "PRD", "功能规格", "单页文档" ], "tags_en": [ "PRD", "Product Spec", "Functional Spec", "One-Page Doc" ], "artifact_type": "html", "prompt": "帮我做一个单文件 HTML 的产品需求文档（PRD）可视化页面，主题是虚构功能「智能日程助手」。要规范文档风、专业浅色，左侧有章节锚点导航（点击平滑滚动定位）。内容包含：文档元信息（版本/负责人/状态/目标发布）、背景与问题、目标与非目标、用户故事（As a..I want..so that..）、功能规格表（功能点/描述/优先级 P0-P2/状态标签）、交互流程说明、验收标准 checklist、非功能需求（性能/安全/兼容）、里程碑时间线、待决问题。要有规格表、用户故事卡、状态标签等清晰结构，配一张文档撰写的真实图片做 hero 大图叠渐变蒙版，响应式适配手机，全部内联无外链。", "prompt_en": "Build a single-file HTML product requirements document (PRD) page for a fictional feature \"Smart Scheduling Assistant\". Use a clean, professional spec-doc style with a left-side section anchor nav (smooth scroll on click). Include: document metadata (version/owner/status/target release), background & problem, goals & non-goals, user stories (As a..I want..so that..), a functional spec table (feature/description/priority P0-P2/status tags), interaction flow, an acceptance-criteria checklist, non-functional requirements (performance/security/compatibility), a milestone timeline, and open questions. Use a real photo of document writing as a hero with a gradient overlay, make it responsive for mobile, fully inlined with no external resources.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "智能日程助手 PRD 单页文档" } ], "skills": [], "experts": [ { "id": "TechnicalDocumentationEngineer", "name": "文通通", "description": "将复杂技术概念转化为清晰准确的文档，让技术知识可传播", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f779afe385ae", "name_en": "Technical Documentation Engineer", "description_en": "Turns complex technical concepts into clear, accurate, shareable documentation", "expert_id": "ex_3zidva69KVGL" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-rfp-proposal-onepager", "title": "投标技术方案速览", "title_en": "RFP Technical Proposal One-Pager", "subtitle": "把冗长投标方案浓缩成一页：架构、计划、资质、报价", "subtitle_en": "Condense a long bid into one page: architecture, plan, credentials, pricing", "description": "投标技术方案动辄上百页，评审和客户很难快速抓住要点。本案例由「专业文档生成团队」把一个政务云数据中台投标方案浓缩成一页速览——需求理解、方案架构图、实施里程碑、团队资质与报价一页呈现，并附评分维度。适合做投标、方案汇报、需要快速过会的售前与商务团队。", "description_en": "RFP technical proposals can run hundreds of pages. Built by the Professional Document Team, this case condenses a government-cloud data-platform bid into a one-pager — needs understanding, an architecture diagram, an implementation timeline, team credentials and pricing on one page, plus scoring dimensions.", "categories": [ "office" ], "tags": [ "投标方案", "技术方案", "一页纸", "文档处理" ], "tags_en": [ "RFP Proposal", "Tech Proposal", "One-Pager", "Document Processing" ], "artifact_type": "html", "prompt": "帮我把一个政务云数据中台投标技术方案浓缩成一页速览（单文件 HTML，评审/客户决策用）。要直观看到：项目名/投标方/方案价值主张/综合自评分、需求理解、方案架构分层框图、实施里程碑时间线、团队资质与案例徽章、评分维度（技术/价格/服务/案例）、报价摘要。深青墨绿专业商务风、信息密度高，界面中文，金额 ¥。", "prompt_en": "Help me condense a government-cloud data-platform RFP technical proposal into a one-pager (single-file HTML). Show project name/bidder/value proposition/overall self-score; needs understanding; a layered architecture diagram; an implementation milestone timeline; team credentials and case badges; scoring dimensions (technology/price/service/cases); and a pricing summary. Deep teal-green professional style; UI in Simplified Chinese; amounts in ¥.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "投标技术方案速览" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-06-20T11:00:00Z", "updated_at": "2026-06-20T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "doc-sop-workflow", "title": "客户投诉处理标准作业流程（SOP）", "title_en": "Customer Complaint Handling SOP", "subtitle": "六步主流程 + RACI 矩阵 + 分级时限的企业级流程文档", "subtitle_en": "Enterprise process doc: 6-step flow, RACI matrix, tiered SLA", "description": "企业内部投诉处理常因环节不清、责任不明、时限失控而导致客户流失与重复返工。本案由「专业文档生成团队」专家按 6 阶段工作流（需求分析→检索→生成→审核→汇编→交付）协作，产出一份结构完整、参数自洽的标准作业流程文档：覆盖目的与适用范围、6 步主流程图、各步骤输入/动作/输出/负责人/时限详述、投诉三级分级标准、RACI 角色责任矩阵、四级异常升级路径、KPI 质量指标与趋势图及附则。交付物为单文件 HTML，含真实配图与清晰流程可视化，可直接用于规范发布或培训。适合客服中心、质量管理与运营团队沉淀标准化作业规范。", "description_en": "Internal complaint handling often fails due to unclear steps, ownership and SLA control, causing churn and rework. Here the Professional Document Team applies a 6-phase workflow (analysis→research→generation→audit→assembly→delivery) to produce a complete, self-consistent SOP: purpose and scope, a 6-step flow diagram, per-step input/action/output/owner/SLA details, three-tier complaint grading, a RACI matrix, a four-tier escalation path, KPIs with a trend chart, and appendix. Delivered as a single-file HTML with real imagery and clear process visuals, ready for publishing or training. Ideal for customer service, QA and operations teams standardizing their procedures.", "categories": [ "office" ], "tags": [ "标准作业流程", "SOP文档", "RACI矩阵", "流程可视化" ], "tags_en": [ "SOP", "Process Doc", "RACI Matrix", "Flow Visualization" ], "artifact_type": "html", "prompt": "帮我做一份「客户投诉处理标准作业流程（SOP）」的单文件 HTML 文档，企业规范风、专业浅色配色。要有：SOP 元信息（编号/版本/生效日/责任部门）、适用范围与目的、6 步主流程图（受理→分级→处理→升级→回访→归档，带角色与时限的步骤条+箭头）、各步骤详述卡（输入/动作/输出/负责人/时限）、投诉轻中重三级分级与响应时限、RACI 角色责任矩阵表、四级异常升级路径、质量指标（首响时长/解决率/满意度）含趋势折线图、附则。要专业精美、配真实图片做 hero、响应式、简体中文，CSS/JS 全内联。", "prompt_en": "Create a single-file HTML 'Customer Complaint Handling SOP' document in a professional enterprise style with a clean light palette. Include: SOP metadata (number/version/effective date/owner department), scope and purpose, a 6-step main flow diagram (intake→grading→handling→escalation→follow-up→archiving, as a step bar with arrows showing roles and SLAs), per-step detail cards (input/action/output/owner/SLA), three-tier complaint grading with response SLAs, a RACI matrix table, a four-tier escalation path, KPIs (first-response time/resolution rate/satisfaction) with a trend line chart, and an appendix. Make it polished, with a real hero photo, responsive, in Simplified Chinese, all CSS/JS inlined.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "客户投诉处理 SOP 流程文档" } ], "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4 角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "4f057377c883", "name_en": "Professional Document Team", "description_en": "Enterprise long-document generation: a 4-role team covering research, outline, writing and compliance audit", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "文档处理", "scenario_id": 100, "scenario_en": "Documentation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "dockerfile-generation", "title": "Dockerfile 一键生成", "title_en": "One-Click Dockerfile Generation", "subtitle": "从零到生产级容器化配置，一句话搞定", "subtitle_en": "From zero to production-grade containerization in one prompt", "description": "手动编写 Dockerfile 和 docker-compose.yml 需要熟悉多阶段构建、安全加固、健康检查等大量细节，稍有不慎就会导致镜像臃肿或存在安全隐患。本案例演示如何通过一句 Prompt，为 Python FastAPI + PostgreSQL + Redis 微服务项目自动生成生产级 Dockerfile、docker-compose.yml 和 .dockerignore，涵盖多阶段构建、非 root 用户、健康检查、环境变量配置及开发/生产环境分离。适合后端开发者、DevOps 工程师及希望快速容器化项目的技术团队。", "description_en": "Writing Dockerfiles and docker-compose.yml manually requires deep knowledge of multi-stage builds, security hardening, health checks, and many other details -- small mistakes can lead to bloated images or security vulnerabilities. This case demonstrates how a single prompt can generate production-grade Dockerfile, docker-compose.yml, and .dockerignore for a Python FastAPI + PostgreSQL + Redis microservice project, covering multi-stage builds, non-root user execution, health checks, environment variable configuration, and dev/prod environment separation. Ideal for backend developers, DevOps engineers, and technical teams looking to containerize projects quickly.", "tags": [ "Docker", "容器化", "微服务", "DevOps", "FastAPI" ], "tags_en": [ "Docker", "Containerization", "Microservices", "DevOps", "FastAPI" ], "artifact_type": "code", "categories": [ "dev" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "Dockerfile", "artifacts": [ { "file": "Dockerfile", "label": "Dockerfile 一键生成" } ], "prompt": "请为一个 Python FastAPI + PostgreSQL + Redis 的微服务项目生成生产级 Dockerfile 和 docker-compose.yml。要求：多阶段构建（multi-stage build）减小镜像体积、非 root 用户运行、健康检查、环境变量配置、.dockerignore 文件、开发/生产环境分离。", "prompt_en": "Please generate a production-grade Dockerfile and docker-compose.yml for a Python FastAPI + PostgreSQL + Redis microservice project. Requirements: multi-stage build to reduce image size, running as a non-root user, health checks, environment variable configuration, a .dockerignore file, and separation of development/production environments.", "skills": [ { "id": "docker", "name": "Docker", "description": "Dockerfile 与 docker-compose 生成，容器化部署方案", "slug": "docker" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "source": "official", "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "document-digitization", "title": "拍一张照，单据秒变结构化数据", "title_en": "Document Digitization", "subtitle": "磅单送货单发票拍照即录，自动勾稽校验", "subtitle_en": "Snap a weighbridge slip, delivery note or invoice — structured data with cross-checks", "description": "拍照上传磅单/送货单/发票等单据，自动OCR识别为结构化数据表，并做勾稽校验，告别手工录入。", "description_en": "Photograph weighbridge slips, delivery notes, invoices and other documents; OCR turns them into structured data tables with automatic cross-checking, ending manual entry.", "categories": [ "productivity" ], "tags": [ "单据数字化", "OCR", "内勤", "B2B贸易" ], "tags_en": [ "Document Digitization", "OCR", "Back Office", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "单据数字化结构化表（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易内勤专家。我上传了一张单据照片（{单据类型，如 过磅单/送货单/入库单/发票}），请帮我把它数字化。\n\n【处理要求】\n1. 用【OCR识别】把单据上的关键字段识别出来（如单据号、日期、车牌、货物名称、规格、毛重/皮重/净重、供货/收货单位、经办人、金额等，按单据类型自动判断需要哪些字段）；\n2. 输出为一张整洁的结构化数据表，字段名 + 识别值一一对应；\n3. 做智能勾稽校验：如净重=毛重−皮重是否一致、车牌/单号是否与发货计划匹配，并标注校验结果；\n4. 给出归档建议（如与材质单、发货单三单核对）。\n\n若识别不清或字段缺失，请标注「待人工确认」，不要臆造数值。请生成识别结果表。", "skills": [ { "id": "document-digitization", "name": "单据数字化", "description": "TradeBuddy把内勤最累的“录单+归档”包圆了：合同、订单、收发货单、收付款凭证、发票……拍张照或导个聊天记录，就自动认字理成规范表格、自动判断是什么单归到对应档，还能按模板生成台账并导出。相当", "skill_id": "skill_2091155881432920064" } ], "experts": [ { "id": "neiqin-expert", "name": "B2B贸易内勤专家", "description": "懂B2B贸易的内勤专家＝AI版录单员+档案室：识单、发票管理、智能对账、出表出单、盯订单进度。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Hu9b9KuHQwa" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "dormant-customer-risk-segmentation-90d", "title": "90天沉睡客户风险分层", "title_en": "90-Day Dormant Customer Risk Segmentation", "subtitle": "90天未消费客户风险分层与挽回策略", "subtitle_en": "A practical Weimob playbook for 90-day dormant customer risk segmentation", "description": "做90天沉睡客户风险分层时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“90天沉睡客户风险分层”展开：识别90天未消费的沉睡客户，按历史消费金额分层，设计阶梯式唤醒权益和多轮触达话术，并预估回流效果。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、沉睡与流失挽回运营人员、经营分析团队。", "description_en": "The hardest part of 90-day dormant customer risk segmentation is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates 90-day dormant customer risk segmentation through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Dormancy and Churn Recovery specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "沉睡客户", "流失风险", "消费金额", "风险分层", "挽回策略" ], "tags_en": [ "90-Day Dormant Customer Risk Segmentation", "Customer Operations", "Dormancy and Churn Recovery", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我圈出90天未消费的沉睡会员，按历史消费金额分层，设计阶梯式唤醒权益和多轮触达话术，并预估回流效果。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a 90-day dormant customer risk segmentation covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "dormant-member-reactivation-90d", "title": "90天沉睡会员唤醒", "title_en": "90-Day Dormant Member Reactivation", "subtitle": "90天沉睡会员分层唤醒与回流预估", "subtitle_en": "A practical Weimob playbook for 90-day dormant member reactivation", "description": "做90天沉睡会员唤醒时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“90天沉睡会员唤醒”展开：识别90天未消费的沉睡会员，按历史消费金额分层，设计阶梯式唤醒权益和多轮触达话术，并预估回流效果。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、沉睡与流失挽回运营人员、经营分析团队。", "description_en": "The hardest part of 90-day dormant member reactivation is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates 90-day dormant member reactivation through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Dormancy and Churn Recovery specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "沉睡会员", "客户分层", "唤醒权益", "触达话术", "回流预估" ], "tags_en": [ "90-Day Dormant Member Reactivation", "Customer Operations", "Dormancy and Churn Recovery", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我圈出90天未消费的沉睡会员，按历史消费金额分层，设计阶梯式唤醒权益和多轮触达话术，并预估回流效果。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a 90-day dormant member reactivation covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "categories": [ "business" ], "skills": [], "experts": [ { "id": "tanyuan-heritage-article", "name": "科普推文专家", "description": "为博物馆公众号写文物科普图文，正文+封面+配图，排版复用 doocs/md，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "076b70b41214", "expert_id": "ex_aPGVu1473LOP" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:05:37+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "id": "dragon-boat-festival-relic-article", "title": "端午主题文物科普推文", "title_en": "Dragon Boat Festival Relic Science Article", "subtitle": "用一件战国船纹青铜缶，把端午竞渡的源头摆给读者看", "subtitle_en": "Trace the visual origin of Duanwu racing to a single Warring States bronze fou", "description": "端午推文常见两种写法：复述屈原故事、列端午习俗清单。这个案例用第三种——让一件战国船纹青铜缶（宁波中国港口博物馆镇馆之宝、国内仅存）做主角：它把竞渡画在器腹上，比文献早、比传说直观。\n\n演示内容：调用探元知识库确认器物细节（高 37.8 cm、腹径 50.5 cm；下腹四组羽人竞渡纹；学界两种解读——战船凯旋 / 龙舟祭祀；百越故地出土），配合三闾庙浮梁龙舟竞渡 2400 年史、宋代史浩《花心动·端午》词作，写出一篇「看完再去过节」的科普图文。\n\n适合：博物馆宣教编辑、节庆科普栏目、博物馆教育课程开发人员，以及任何想用「实物 + 文献」双支撑，写出有质感的端午推文的公众号主笔。", "description_en": "Dragon Boat Festival articles usually re-tell Qu Yuan's biography or list customs. This case picks a third path: let one artifact — the Warring States bronze fou with boat-race decoration (Ningbo China Port Museum, sole surviving example of its kind) — do the storytelling. The boats are right there on the vessel's belly, earlier than any text and more literal than any legend.\n\nDemo: pull the fou from the TanYuan KB (h. 37.8 cm, max dia. 50.5 cm; four feathered-man boat-race bands; two academic interpretations — victorious war-fleet vs. dragon-boat ritual; recovered from the ancient Hundred-Yue homeland) and pair it with the Fuliang dragon-boat tradition at Sanlu Temple (2,400+ years) and Shi Hao's Song-dynasty ci poem 《Hua Xindong · Duanwu》. Deliver a \"read this before you celebrate\" science article.\n\nBest for: museum WeChat editors, holiday science columns, museum-education curriculum designers, and any public-account writer who wants a Duanwu article grounded in both objects and texts.", "tags": [ "端午", "文物科普", "战国青铜", "竞渡", "公众号推文" ], "tags_en": [ "Dragon Boat Festival", "Relic Science", "Warring States Bronze", "Boat Racing", "WeChat Article" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "doocs/md 风格公众号图文 HTML" } ], "prompt": "请基于以下要点，写一篇端午主题的文物科普公众号图文。\n\n写作要点：\n1. 不写屈原故事开头：以「战国缶上的竞渡，比文献更早」做钩子，让读者理解端午看得见的源头在青铜器上。\n2. 真实文物：所有数据必须调用 mcp__tanyuan-wb-mcp-server__search_relics / search_movable_relics 检索。常见支撑：\n - 战国船纹青铜缶：战国，国家一级，宁波中国港口博物馆镇馆之宝；高 37.8 cm、腹径 50.5 cm；下腹饰四组羽人竞渡纹；每船 5 人左右、戴羽饰；广西北海合浦出土；具有明显楚国青铜器特征；学界两种解读（战船 vs 龙舟祭祀）；国内仅存。\n - 三闾庙（江西浮梁）：屈原曾到过浮梁；浮梁以龙舟竞渡方式庆祝端午，迄今 2400 多年；庙门口对古码头；庙中 3.6 m 屈原青铜像。\n3. 观点分层：\n - 第 2 章：器物本身（尺寸/纹饰/出土/孤品价值）。\n - 第 3 章：为什么说它画了端午——羽人船纹的鸟神崇拜 + 楚地巫祭传统 → 接近龙舟祭祀。\n - 第 4 章：战国竞渡接住屈原故事后，宋代史浩《花心动·端午》「菖歙碎琼，角黍堆金，龙舟竞夺，锦标方彻」——一条水脉延续到今天。\n - 第 5 章：端午文物小目录（缶、三闾庙、铜鉴缶）。\n4. 结尾：把「看完再去过节」的视角落到读者身上——当晚如遇博物馆开放夜场，欢迎去现场对照。\n5. 风格：doocs/md 风格——677px 居中、淡蓝/水纹底（端午色系）、章节小标题加蓝条、卡片块列出端午文物小目录；克制、美观。\n6. 配图：探元知识库已授权的战国缶图（multi image_url），download 后 base64 内嵌。\n7. 字数：1500–1800 字正文。\n8. 逐处标源：每个事实声明「来自探元知识库《...》词条」或宋代史浩《花心动·端午》。\n\n约束：输出单文件 HTML，CSS 全内联，图片 data:image/...;base64,... 内嵌，禁止任何外链。封面图单独导出 cover.png。", "prompt_en": "Write a Dragon Boat Festival (Duanwu) themed WeChat science article. The expert persona is the TanYuan Heritage Article Expert — a heritage science writer.\n\nWriting requirements:\n1. Do not start with the Qu Yuan biography. Open with \"the Warring States bronze fou shows Duanwu boat-racing earlier than any text does\" — let the reader see the visual origin of the festival on a vessel's belly.\n2. Real artifacts: every fact must come from mcp__tanyuan-wb-mcp-server__search_relics / search_movable_relics. Useful anchors:\n - Warring States boat-pattern bronze fou: Warring States, National Grade-1, Ningbo China Port Museum keeper-piece; h. 37.8 cm, max dia. 50.5 cm; four bands of feathered-man boat-race decoration; 5 figures per boat with feathered headdresses; recovered from Beihai, Guangxi; strong Chu-bronze style; two academic readings (war-fleet vs. dragon-boat ritual); sole surviving example in China.\n - Sanlu Temple (Fuliang, Jiangxi): Qu Yuan once visited Fuliang; the place has celebrated Duanwu with dragon-boat racing for 2,400+ years; temple faces the old dock; houses a 3.6 m bronze statue of Qu Yuan.\n3. Layered argument:\n - §2 the object itself (dimensions, decoration, provenance, singleton status).\n - §3 why the fou \"paints Duanwu\" — bird-god totem of feathered figures + Chu-region shamanic rite → dragon-boat-ritual reading.\n - §4 how the Warring States racing absorbed the Qu Yuan story and stayed alive through the Song dynasty poet Shi Hao's 《Hua Xindong · Duanwu》(\"菖歙碎琼，角黍堆金，龙舟竞夺，锦标方彻\") — one continuous water-line to today.\n - §5 a \"Duanwu relic directory\" (the fou, Sanlu Temple, bronze jian-fou).\n4. Close: pull the \"read this before you celebrate\" frame to the reader — if a local museum has a night opening tonight, invite them to verify in person.\n5. Visual style: doocs/md — 677px column, soft blue / water-palette background (Duanwu color), blue-bullet section headings, soft cards; restrained, WeChat-ready.\n6. Imagery: authorized TanYuan KB images of the fou (multiple image_url fields), downloaded and embedded as base64.\n7. Length: 1500–1800 Chinese characters of body copy.\n8. Cite every fact: each claim labels source.\n\nOutput: a single HTML file, all CSS inlined, all images as data:image/...;base64,..., no external links. Export a separate cover.png.", "is_recommended": false, "recommended_order": 0 }, { "id": "drug-box-scan", "title": "拍个药盒，它就认得这药", "title_en": "Snap the Box, It Knows the Drug", "subtitle": "手机拍药盒，识别药品并解读用法、副作用与禁忌", "subtitle_en": "Photograph a drug box — identify it and explain usage, side effects, contraindications", "description": "长辈寄来一盒没见过的药，或旅游带回的进口药，名字陌生、看不懂。打字查又拼不准。\n\n这个案例演示把一张手机拍的药盒照片交给智能问药：它用多模态识别药盒上的通用名、商品名、规格、厂家，对反光/模糊处如实标注「识别不清」而非瞎猜；接着解读这是什么药、每天吃几次、常见副作用、哪些人禁用，并提醒长期管理药不能自行停用。\n\n适合：代父母管理用药、整理家庭药箱、跨境/进口药识别、药店用药咨询服务。", "description_en": "An elder mails you an unfamiliar box of medicine, or you bring back an imported drug whose name you cannot read or spell.\n\nThis case shows handing a phone photo of the box to the drug-query skill: it identifies the generic name, brand, strength and manufacturer via multimodal recognition, marking glare/blur areas as \"unclear\" rather than guessing; then explains what the drug is, how often to take it, common side effects, who must not use it, and reminds that long-term drugs should not be stopped on your own.\n\nGood for: managing parents' medications, tidying the home medicine cabinet, identifying imported drugs, pharmacy consultation.", "categories": [ "education", "productivity" ], "tags": [ "拍药盒", "药品识别", "说明书", "多模态", "健康科普" ], "tags_en": [ "box scan", "drug ID", "leaflet", "multimodal", "health literacy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "老人药箱里有盒阿托伐他汀钙片，20mg×7 片，辉瑞的。这药是治什么的、每天什么时候吃、长期吃要注意什么、要定期查什么？", "prompt_en": "There is a box of atorvastatin calcium tablets (20 mg × 7, Pfizer) in my parent's medicine cabinet. What is it for, when should it be taken daily, what to watch for long term, and which tests need regular checking?", "skills": [ { "id": "tencent-ai-drug-consultant", "name": "腾讯健康-AI就医助手-智能问药", "description": "AI 智能问药：对接 ADP 问药 API，支持文字提问与药盒拍照识别，覆盖用法用量、副作用与药物相互作用", "slug": "tencent-ai-drug-consultant" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "拍个药盒，它就认得这药" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "drug-interaction-check", "title": "我同时吃这几种药，会冲突吗", "title_en": "Do These Medicines Conflict", "subtitle": "列出正在吃的药，自动排查相互作用与重复成分", "subtitle_en": "List your current drugs — auto-check interactions and duplicate ingredients", "description": "支架术后在吃抗血小板药，平时还有高血压药，牙痛想加片布洛芬——会不会出血？普通人手里好几种药，谁能一下说清冲突。\n\n这个案例演示把一份用药清单交给智能问药：它逐项梳理药品类别，重点排查高危相互作用（抗血小板药 + 布洛芬叠加消化道出血、NSAID 减弱降压），检查有无同类成分重复，再给给药间隔与出血信号的实用建议。\n\n适合：慢病多药患者、支架/抗凝人群、药店用药安全教育、互联网医院用药审核前置。", "description_en": "After a stent you take antiplatelets, plus hypertension drugs daily, and want ibuprofen for a toothache — will it bleed? Ordinary people hold several drugs and cannot tell conflicts at a glance.\n\nThis case shows handing a medication list to the drug-query skill: it sorts each drug's class, flags high-risk interactions (antiplatelet + ibuprofen raising bleeding, NSAID weakening antihypertensives), checks for duplicate active ingredients, then gives dosing-interval and bleeding-signal tips.\n\nGood for: polypharmacy chronic patients, stent/anticoagulation users, pharmacy safety education, pre-review at tele-health.", "categories": [ "education", "productivity" ], "tags": [ "相互作用", "多药同服", "抗血小板", "出血风险", "健康科普" ], "tags_en": [ "interaction", "polypharmacy", "antiplatelet", "bleeding risk", "health literacy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我爸每天吃这几种：阿司匹林肠溶片（预防血栓）、氯吡格雷（心脏支架后）、氨氯地平（降压），最近膝盖疼又想加布洛芬。这几个一起吃会不会有问题？", "prompt_en": "My father takes aspirin enteric tablets (thrombosis prevention), clopidogrel (after cardiac stent), and amlodipine (blood pressure) daily. His knee hurts and he wants to add ibuprofen. Is combining these a problem?", "skills": [ { "id": "tencent-ai-drug-consultant", "name": "腾讯健康-AI就医助手-智能问药", "description": "AI 智能问药：对接 ADP 问药 API，支持文字提问与药盒拍照识别，覆盖用法用量、副作用与药物相互作用", "slug": "tencent-ai-drug-consultant" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "我同时吃这几种药，会冲突吗" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "drug-query-usage", "title": "常用药怎么吃 · 一查就明白", "title_en": "Look Up How to Take Any Medicine", "subtitle": "文字问「布洛芬怎么吃」，用法用量、禁忌、副作用一次说清", "subtitle_en": "Ask \"how to take ibuprofen\" in text — dosage, contraindications, side effects in one go", "description": "家里药箱翻出一盒药，说明书字小又专业，用法用量、能不能和别的药一起吃，全看不明白。问药店远、问医生要挂号。\n\n这个案例演示把一句文字提问「布洛芬混悬液孩子发烧吃多少、成人牙痛怎么吃」交给智能问药：它从药品说明书知识库识别出药品，分人群给出成人与儿童的用法用量，把禁忌症（孕妇禁用、消化道溃疡禁用等）和常见副作用单独列出，再提示与其他药（抗凝药、降压药）的相互作用。\n\n适合：家庭常用药自查、慢病患者多药管理、药店/互联网医院用药咨询前置、企业员工健康问答。", "description_en": "You find a box of medicine at home, but the leaflet is tiny and jargon-heavy — dosage, whether it conflicts with other drugs, all unclear. Asking a pharmacist is far, a doctor needs an appointment.\n\nThis case shows handing a text question \"how much ibuprofen suspension for a feverish child, how for adult toothache\" to the drug-query skill: it identifies the drug from the package-insert knowledge base, gives adult and pediatric dosage by population, lists contraindications (pregnancy, active ulcer) and common side effects separately, then flags interactions with anticoagulants and antihypertensives.\n\nGood for: household medicine lookup, polypharmacy management for chronic patients, pre-consultation drug Q&A at pharmacies / tele-health, corporate employee health Q&A.", "categories": [ "education", "productivity" ], "tags": [ "用药查询", "用法用量", "禁忌症", "药物相互作用", "健康科普" ], "tags_en": [ "drug lookup", "dosage", "contraindications", "drug interaction", "health literacy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "家里翻出一盒布洛芬混悬液，孩子 5 岁发烧想吃，我自己牙疼也想吃。这药我们俩各自该吃多少、间隔多久，有什么不能一起吃的？", "prompt_en": "I found a bottle of ibuprofen suspension at home. My 5-year-old has a fever and I have a toothache. How much should each of us take, how often, and what shouldn't be taken with it?", "skills": [ { "id": "tencent-ai-drug-consultant", "name": "腾讯健康-AI就医助手-智能问药", "description": "AI 智能问药：对接 ADP 问药 API，支持文字提问与药盒拍照识别，覆盖用法用量、副作用与药物相互作用", "slug": "tencent-ai-drug-consultant" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "常用药怎么吃 · 一查就明白" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "drumming-storyteller-lifestyle-series", "title": "击鼓说唱俑 · 快乐生活文创系列", "title_en": "Drumming Storyteller Figurine · Joyful Lifestyle Series", "subtitle": "四品类 · 情绪先行", "subtitle_en": "四品类 · 情绪先行", "description": "以中国国家博物馆藏东汉击鼓说唱陶俑为主题，覆盖摇头娃娃摆件、等比复原冰箱贴、鼓形陶艺笔筒、氛围夜灯四个品类。除六项造型特征硬约束外，本案额外约束材质表现：所有商品须呈无釉灰陶的粗糙哑光肌理、五官简化模糊如手捏泥坯，不得做成写实人脸；冰箱贴采用等比复原路线，效果图由文物原始照片直接处理合成，未经生成模型重绘。", "description_en": "以中国国家博物馆藏东汉击鼓说唱陶俑为主题，覆盖摇头娃娃摆件、等比复原冰箱贴、鼓形陶艺笔筒、氛围夜灯四个品类。除六项造型特征硬约束外，本案额外约束材质表现：所有商品须呈无釉灰陶的粗糙哑光肌理、五官简化模糊如手捏泥坯，不得做成写实人脸；冰箱贴采用等比复原路线，效果图由文物原始照片直接处理合成，未经生成模型重绘。", "categories": [ "creation", "business" ], "tags": [ "文创提案", "击鼓说唱俑", "多品类", "等比复原", "陶质质感约束" ], "tags_en": [ "文创提案", "击鼓说唱俑", "多品类", "等比复原", "陶质质感约束" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "cover_image": "cover.jpg", "cover_image_en": "cover.jpg", "cover_mode": "landscape", "prompt": "以中国国家博物馆藏东汉击鼓说唱陶俑为主题，做一份多品类文创产品线提案，输出单文件 HTML。要求：一、以国博官网藏品页的造型描述为准，整理\"文物特征清单\"；二、参考国博说唱俑冰箱贴等已上市文创的品类思路，不要自行臆造视觉；三、覆盖摇头娃娃摆件、冰箱贴、创意花瓶、氛围夜灯四个品类，每款的功能点必须由文物特征派生（如花瓶借鼓腹、夜灯借鼓面）；四、每款商品效果图的生成提示词必须把文物特征清单逐条写成强制约束；五、每款注明保留了哪些文物特征、材质与市场定位；六、声明效果图为 AI 设计示意，需文物与法务复核。", "prompt_en": "以中国国家博物馆藏东汉击鼓说唱陶俑为主题，做一份多品类文创产品线提案，输出单文件 HTML。要求：一、以国博官网藏品页的造型描述为准，整理\"文物特征清单\"；二、参考国博说唱俑冰箱贴等已上市文创的品类思路，不要自行臆造视觉；三、覆盖摇头娃娃摆件、冰箱贴、创意花瓶、氛围夜灯四个品类，每款的功能点必须由文物特征派生（如花瓶借鼓腹、夜灯借鼓面）；四、每款商品效果图的生成提示词必须把文物特征清单逐条写成强制约束；五、每款注明保留了哪些文物特征、材质与市场定位；六、声明效果图为 AI 设计示意，需文物与法务复核。", "skills": [], "experts": [ { "id": "tanyuan-cultural-creative", "name": "文创提案专家", "description": "把一件文物做成文创产品线设计提案,含核心记忆点、产品线与商品示意图,事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cf052b46c196", "expert_id": "ex_HXd4FrKZutHg" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "设计创意", "usage_count": 0, "created_at": "2026-08-14T18:00:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "击鼓说唱俑 · 快乐生活文创系列" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "dunhuang-mural-digital-preservation", "title": "数字化技术对敦煌壁画保护的影响研究", "title_en": "Impact of Digital Technology on Dunhuang Mural Preservation", "subtitle": "从1980年代胶片摄影到数字敦煌全球访问:三十年技术演进与保护范式重塑", "subtitle_en": "From 1980s film photography to global Digital Dunhuang access: three decades of technology evolution and preservation paradigm shift", "description": "敦煌莫高窟壁画是脆弱到禁止拍照的国宝级文物,但我们今天却能在网上看到毫米级高清的数字化壁画,这中间到底经历了一场怎样的技术革命?\n\n这个案例演示了如何借助腾讯探元文物数据库与知识库的并行检索,主代理亲自做 3-4 路网络调研(最新进展/典型案例/影响成效/历史脉络),覆盖胶片数字化、高清壁画全球访问、数字洞窟、虚拟漫游、AI 辅助病害识别、永续保存与海外巡展等维度,最终交付包含 3-6 张配图与 1-2 张趋势图表的 Markdown / HTML / DOCX / PDF 四格式趋势综述报告。\n\n适合:文物保护工程与考古学研究者、博物馆数字化从业人员、丝绸之路与敦煌学研究者、文化遗产政策与法规研究者、数字人文与艺术史专业研究生、文物保护材料与装备研发从业者。", "description_en": "Dunhuang Mogao Cave murals are so fragile that photography is banned, yet today we can view millimeter-resolution digital murals online—what technological revolution took place in between?\n\nThis case shows how to leverage Tencent Tanyuan relic and knowledge bases for parallel retrieval, plus 3-4 concurrent main-agent WebSearch passes (latest progress / benchmark cases / impact / historical lines), covering film scanning, global high-resolution mural access, digital caves, virtual tours, AI-assisted disease identification, perpetual preservation, and overseas touring exhibitions—delivering a 3-6 image + 1-2 trend-chart Markdown/HTML/DOCX/PDF four-format trend review report.\n\nBest for: conservation engineering and archaeology researchers, museum digitization practitioners, Silk Road and Dunhuang studies researchers, cultural-heritage policy researchers, digital humanities and art-history graduate students, conservation-materials and equipment professionals.", "categories": [ "business" ], "tags": [ "敦煌莫高窟", "数字化保护", "数字敦煌", "壁画病害", "丝绸之路" ], "tags_en": [ "Dunhuang Mogao Caves", "Digital Preservation", "Digital Dunhuang", "Mural Disease", "Silk Road" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含趋势图表)" } ], "prompt": "请按「腾讯探元学术研究-趋势综述」专家的工作流,综述「数字化技术对敦煌壁画保护的影响研究」。\n\n研究范围(覆盖五维):\n- 数字化采集技术演进:1980 年代胶片摄影→冷光照明+大幅面胶片→高精度多频段采集(紫外/可见/红外)→毫米级结构光三维扫描→多视角图像建模\n- 全球访问与平台:「数字敦煌」国际网站、敦煌研究院数字资源库、海外数字洞窟展览(英国大英博物馆/法国吉美博物馆)、谷歌艺术与文化合作\n- 病害分析与预警:盐害、起甲、酥碱、霉变、烟熏等病害图像识别,AI 卷积神经网络模型,洞窟微环境监测与风险预警\n- 虚拟现实与沉浸体验:VR 漫游洞窟、增强现实洞窟复原、互动展项\n- 永久保存与活化利用:壁画「数字档案」、标准编制、海外巡展、文创 IP\n\n交付要求(强制):\n1. 探元两库 MCP(search_movable_relics + search_relics)并行检索 + 主代理 3-4 路 WebSearch 多源核验,优先近 1-2 年(2025/2026)信源,以及敦煌研究院、英国国家美术馆、学术期刊信源\n2. 配图 ≥ 3 张(Wikimedia 公共领域 / 敦煌研究院官网 / 公开新闻图源),必须实图、与主题相关\n3. 趋势图表 ≥ 1 张(建议 2-3 张:数字化采集技术演进时间线/病害识别 AI 模型对比/全球访问桑基图),matplotlib 渲染为 SVG,严格避免标签重叠\n4. 四格式报告:Markdown 母版 → HTML(嵌入 base64 图片) → DOCX → PDF,综述结构(领域背景→发展现状→典型案例→影响成效→趋势展望→结论→证据与局限→参考来源)\n5. 报告日期用真实当前日期;参考文献按时间倒序;只写结论不含建议;主要趋势判断同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Trend Review' expert workflow to survey the 'Impact of Digital Technology on Dunhuang Mural Preservation'.\n\nResearch scope (five dimensions):\n- Digitization acquisition technology evolution: 1980s film photography → cold-light + large-format film → high-precision multi-band (UV/visible/IR) → millimeter-grade structured-light 3D scanning → multi-view image modelling\n- Global access platforms: Digital Dunhuang international website, Dunhuang Academy digital archive, overseas digital cave exhibitions (British Museum / Guimet), Google Arts & Culture collaboration\n- Disease analysis and early warning: salt damage, flaking, efflorescence, mould, soot contamination identification; AI CNN models; micro-environment monitoring and risk warning\n- Virtual reality and immersive experience: VR cave tours, AR cave restoration, interactive exhibits\n- Permanent preservation and activation: digital mural archives, standards drafting, overseas touring, cultural-product IP\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 3-4 main-agent WebSearch passes for cross-source verification, prioritizing sources from the last 1-2 years (2025/2026) plus Dunhuang Academy, British Museum, Guimet, academic journal sources\n2. ≥ 3 images (Wikimedia public domain / Dunhuang Academy official / public news), each verified real and topic-relevant\n3. ≥ 1 trend chart (2-3 recommended: digitization technology evolution timeline / AI disease identification model comparison / global-access sankey), rendered to SVG via matplotlib with strict label overlap avoidance\n4. Four formats: Markdown master → HTML (base64-inlined images) → DOCX → PDF, trend-review structure (background → status → cases → impact → outlook → conclusion → limits → references)\n5. Report date is the true current date; references sorted reverse-chronologically; conclusions only, no recommendations; key trend judgments sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-trend", "name": "文博趋势综述专家", "description": "宽泛主题广谱调研,3-4 路网络调研强调时效性,案例卡片化呈现,交付四格式趋势综述报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "2e5f2aab7036", "expert_id": "ex_M24IASAeVHLO" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:20:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "early-spring-harbin-map-tour", "title": "跟着《早春晴朗》游哈尔滨", "title_en": "Tour Harbin with the Drama Early Spring", "subtitle": "剧集取景地打卡地图：4 大主题 11 处实景一键巡游", "subtitle_en": "Drama filming-location map: 11 real spots across 4 themed routes", "description": "想跟着热播剧打卡同款取景地，却总要在大杂烩攻略里一个个搜坐标？《早春晴朗》里男女主漫步的中央大街、定情的教堂长椅、重逢的老酒馆，到底都在哪、怎么走最顺路？\n\n这个案例用腾讯地图把剧中 4 大主题、11 处哈尔滨实景搬进一张可交互地图：一念难桃三大名场面、冰城底色城市地标、城外\"瑞士\"雪原山水和同款味道菜馆，每个点位都带剧情注解和打卡提示，支持按主题筛选、一键巡游和飞线动画预览路线，扫码即可存入腾讯地图小程序，手机随时导航、分享同好。\n\n适合：剧迷打卡党、哈尔滨旅行规划者，以及想复刻\"影视+地图\"玩法的内容创作者和文旅运营。", "description_en": "Want to visit your favorite drama's filming spots but stuck searching coordinates one by one in messy travel notes? Where exactly are the Central Boulevard promenade, the church bench and the reunion restaurant from Early Spring, and what's the best route?\n\nThis case brings 11 real Harbin locations from the drama into one interactive Tencent map, organized into 4 themes: the three iconic romance scenes, city landmarks, the \"Swiss\" snow-forest scenery outside town, and same-flavor restaurants. Every spot carries scene notes and check-in tips, with theme filtering, one-tap auto-tour and flight-line animation to preview routes. Scan the QR code to save the guide into the Tencent Maps mini program for on-the-go navigation and sharing.\n\nIdeal for: drama fans on location pilgrimages, Harbin trip planners, and content creators or tourism operators looking to replicate the \"drama + map\" play.", "categories": [ "travel" ], "tags": [ "腾讯地图", "取景地打卡", "哈尔滨旅行", "影视剧巡游", "地图指南" ], "tags_en": [ "Tencent Maps", "filming location tour", "Harbin travel", "drama location tour", "map guide" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "用腾讯地图的自定义地图能力，为热播剧《早春晴朗》制作一份「跟着电视剧游哈尔滨」取景地打卡指南页，输出一份可直接打开的单页 HTML。数据用剧中真实取景地：按主题分组——一念难桃·三大名场面（中央大街、圣·索菲亚教堂、老俄侨俄侨菜馆(西十三道街店)）、冰城底色·城市地标（松花江滨洲铁路桥、斯大林公园、中华巴洛克历史文化街区、哈尔滨大剧院、哈尔滨站）、城外山水·瑞士其实在哈尔滨（郡王山露营基地、横头山金龙山国家森林公园）、同款味道（拾楼壹号哈埠菜馆），共 4 个主题 11 处点位，每个点位含名称、坐标、剧情注解与打卡提示。页面布局：左侧剧名卡片与分组点位列表（序号徽标、简介、打卡提示，支持按分组筛选和查看全部），右侧全屏地图打点展示，提供巡游模式与飞线动画两种路线预览方式，并生成二维码引导用户存入腾讯地图小程序（手机查看、一键导航、分享同好）。", "prompt_en": "Use Tencent Maps' custom map capability to create a \"Tour Harbin with the drama Early Spring\" filming-location guide page, delivered as a single HTML file that opens directly in a browser. Use real filming locations from the drama, grouped by theme: three iconic romance scenes (Central Street, Saint Sophia Cathedral, Old Russian emigre restaurant), city landmarks (Songhua River Railway Bridge, Stalin Park, Chinese Baroque Historic District, Harbin Grand Theatre, Harbin Railway Station), snow-forest scenery outside town (Junwang Mountain campsite, Jintou Mountain National Forest Park), and same-flavor restaurants (Shilou No.1 Harbin Cuisine), 4 themes and 11 spots in total, each with a name, coordinates, scene notes and check-in tips. Layout: a drama title card and grouped spot list on the left (numbered badges, intros, check-in tips, with group filtering and an all-spots view), a full-screen map with pins on the right, two route-preview modes (auto-tour and flight-line animation), plus a QR code guiding users to save the guide into the Tencent Maps mini program (view on phone, one-tap navigation, share with fellow fans).", "skills": [ { "id": "tencentmap-map-assistant", "name": "腾讯地图·地图助手", "description": "腾讯地图·地图助手 Skill，一句自然语言调用腾讯地图全套能力，开箱即用。提供 AI 旅游攻略、地点搜索（含评分/人均/营业时间）、关键词提示、路线规划（驾车/步行/公交/骑行）、地址解析与逆解析、行政区划、IP 定位、距离计算、天气查询，并可生成自定义地图（网页 / 小程序指南）。涉及找地点、规划路线、旅游行程、查天气、坐标转换等出行场景时使用。", "slug": "tencentmap-map-assistant" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "腾讯地图", "name_en": "Tencent Maps", "avatar": "", "badge": "community" }, "source": "community", "quality_score": 94, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-24T08:05:16Z", "updated_at": "2026-09-24T08:05:16Z", "artifacts": [ { "file": "output.html", "label": "剧集取景地地图指南页" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "earth-online-university", "title": "地球Online（大学版）", "title_en": "Earth Online (College Edition)", "subtitle": "把大学生活过成一场角色扮演游戏", "subtitle_en": "Turn college life into a role-playing game", "description": "大学生活过得像一团浆糊，上过的课、做过的事转头就忘，更别说看清自己到底会什么、想学什么。\n\n这个案例展示了一个游戏化的「大学成长人物面板」：像玩游戏一样给自己建玩家档案、挑形象，随手记录真实经历自动整理成结构化复盘、提取技能，想学的技能拆成每天能打勾的小任务，经验值、等级、技能树、日历、每日一句全部联动，还能随时生成一张游戏风人物存档卡。演示数据是一位建筑系大二学生的真实感日常。所有数据落在资料库数据表，手机电脑双向同步。\n\n适合：想把大学生活过明白一点的人、需要一点游戏化动力去学习和记录的人、想给自己攒一份成长存档的人。", "description_en": "College life blurs by—classes attended and things done are quickly forgotten, let alone a clear picture of what you can do and what you want to learn.\n\nThis case shows a gamified \"college growth character panel\": create a player profile and pick a look like in a game, jot down real experiences that get auto-structured into reflections and skill suggestions, break skills you want to learn into daily checkable tasks, with EXP, levels, a skill tree, a calendar and a daily quote all interconnected—plus a game-style character archive card you can generate anytime. Demo data follows a believable day-to-day of a sophomore architecture student. All data lives in Library tables with two-way sync across devices.\n\nBest for: students who want to live college life a bit more deliberately, anyone needing gamified motivation to learn and record, and anyone who'd love to keep a growing archive of themselves.", "categories": [ "workspace" ], "tags": [ "游戏化成长", "大学生活", "技能树", "经历记录", "成长存档" ], "tags_en": [ "Gamified Growth", "College Life", "Skill Tree", "Experience Log", "Growth Archive" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "地球Online（大学版）工作台（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/dGEWUioQbbmcx82MUAP2Uz，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n帮我做一个「地球Online（大学版）」游戏化大学成长工作台，做成单个完整 HTML 文件。\n\n功能模块：\n1. 启动流程：启动页（品牌视觉 + 点击进入）→ 建立玩家档案表单（昵称必填，学校、专业、年级选择、一句话描述、兴趣关键词、职业关键词）→ 选择角色形象（5 张预设背景 + 14 个预设角色，点选实时合成预览）→ 进入首页。\n2. 首页人物面板：人物卡（形象图、昵称、学校/专业/年级、一句话气泡、兴趣标签）；等级经验条（LV.1~30，加经验弹 +EXP 提示、升级弹庆祝）；两个主操作（记录一段经历 / 帮我学习一个技能）；我的技能等级（从经历和学习任务自动识别 22 个分类并累计等级）；本月日历（有经历或完成任务的日子打点、周末与今天高亮）；今日清单（自建待办回车添加、勾选 +8 EXP，每个学习中的技能显示今天第一个未完成任务、勾选 +5 EXP，今天记过的经历也在列）；最近经历 5 条；正在学习 Top3（进度条 + 下一步任务）；每日答案之书（按日期稳定抽一句，可换一句）。底部导航：人物 / 经历 / 技能 / 存档。\n3. 经历库：记录经历（日期、标题、发生了什么），本地规则自动整理成「发生了什么 / 我做了什么 / 结果 / 我的感受」四段结构，并按关键词给出技能建议（只有我勾选确认才加入面板，可自己补充）；经历列表可展开原始输入、可删除；记录经历 +30 起 EXP、每确认一个技能 +20 EXP。\n4. 技能树：左边「我已经拥有的技能」（来自经历确认），右边「我想学习的新技能」；可新建成长方向并勾选技能归入、按方向 Tab 筛选。想学新技能时用一句话描述目标，内置领域知识库（雅思 / 视频剪辑 / 设计 / 做饭 / 健身 / 写作等）自动拆成 2-3 个能力单元和按天或按周的具体学习任务（每个任务含做什么、时长、达标标准），最后一个是验证任务；勾选任务推进进度与单元状态，全部完成且验证任务通过后技能正式点亮。\n5. 存档：把当前昵称 / 学校 / 年级 / 等级 / 技能 / 经历汇总成游戏风人物存档卡，可命名保存多版历史存档并点击回看，可生成存档长图保存或分享。\n\n数据落在资料库数据表，支持在线存储与双向同步：eo_玩家档案（昵称/学校/专业/年级/一句话描述/兴趣关键词/职业关键词/总经验）、eo_经历库（日期/标题/描述/结果/感受/确认技能）、eo_技能（技能名/状态[已点亮/学习中/想学]/目标/进度说明/确认日期）、eo_任务（类型[学习任务/今日待办]/所属技能/日期/标题/内容/时长/达标标准/是否完成）。页面加载时从数据表拉取渲染，页面内的新增与修改异步回写数据表。\n\n视觉方向：像素游戏风 + 清新手账感，奶油色卡片、像素描边、游戏化图标，mobile-first 单栏布局。无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/dGEWUioQbbmcx82MUAP2Uz — with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build me a \"Earth Online (College Edition)\" gamified college-growth workbench as a single complete HTML file. Feature modules:\n1. Onboarding flow: splash page (brand visual + tap to enter) → player profile form (nickname required; school, major, grade selector, one-line bio, interest keywords, career keywords) → character look picker (5 preset backgrounds + 14 preset characters with live composited preview) → home.\n2. Home character panel: character card (avatar, nickname, school/major/grade, one-line bubble, interest tags); level & EXP bar (LV.1–30, +EXP toast on gains, celebration toast on level-up); two main actions (Log an Experience / Help Me Learn a Skill); \"My Skill Levels\" (auto-detects 22 categories from experiences and learning tasks, accumulating levels); monthly calendar (dots on days with experiences or completed tasks, weekends and today highlighted); Today's list (add custom todos with Enter, +8 EXP per check; each learning skill shows today's first unfinished task, +5 EXP per check; today's logged experiences also listed); 5 recent experiences; top-3 learning goals (progress bar + next task); Daily Book of Answers (a stable quote per date, refreshable). Bottom nav: Character / Log / Skills / Archive.\n3. Experience log: log an experience (date, title, what happened); local rules structure it into \"what happened / what I did / result / how I felt\", and suggest skills by keywords (only my confirmed ticks join the panel; custom additions allowed); the list supports expanding raw input and deletion; logging an experience grants +30 EXP and up, each confirmed skill +20 EXP.\n4. Skill tree: left \"Skills I already have\" (confirmed from experiences), right \"New skills I want to learn\"; create growth tracks, assign skills to tracks, filter by track tabs. Describe a learning goal in one sentence; a built-in domain knowledge base (IELTS / video editing / design / cooking / fitness / writing, etc.) breaks it into 2–3 ability units and concrete daily or weekly tasks (each with action, duration, and pass standard), ending with a verification task; checking tasks advances progress and unit states, and once everything plus verification is done, the skill is officially lit.\n5. Archive: compile current nickname / school / grade / level / skills / experiences into a game-style character archive card; save multiple named historical archives and revisit them; generate a tall archive image to save or share. Data lives in Library tables with online storage and two-way sync: eo_玩家档案 (nickname/school/major/grade/bio/interests/career keywords/total EXP), eo_经历库 (date/title/description/result/feeling/confirmed skills), eo_技能 (skill name/status[已点亮/学习中/想学]/goal/progress note/confirmation date), eo_任务 (type[学习任务/今日待办]/skill/date/title/content/duration/standard/done). The page pulls from the tables on load; additions and edits in the page are asynchronously written back. Visual direction: pixel-game style meets fresh journal aesthetic—cream cards, pixel borders, game icons, mobile-first single-column layout. no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd9b7744a331", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "塑料叉FOKU", "name_en": "塑料叉FOKU", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 98, "created_at": "2026-09-07T00:05:03+08:00", "updated_at": "2026-09-07T01:16:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "edgeone-ai-trend-agent", "title": "定时 AI 资讯聚合 Agent 应用", "title_en": "Scheduled AI Trend Aggregation Agent", "subtitle": "多 Agent 流水线每天自动产出 AI 资讯日报，部署到 EdgeOne Makers", "subtitle_en": "A multi-agent pipeline that auto-generates a daily AI news digest, deployed on EdgeOne Makers", "description": "每天想快速掌握 AI 圈动态，却要在 Hacker News、Dev.to 之间反复刷、人工筛选摘要，既费时又容易漏。\n\n这个案例由「Makers 开发专家团」协作完成：主理人齐上线做技术选型与编排，前端裴知页搭 React+Vite 工作台，后端范云申用 LangGraph 写「采集 → 过滤分类 → 摘要 → 打分 → 生成报告」五阶段流水线，全程通过 SSE 实时把进度推到前端。支持 cron 每天 09:00 自动触发，也能点按钮手动跑；产出一份带分类、重要性评分和中文摘要的 Markdown 日报并落库，可随时回看历史。案例同时附带完整可部署的项目工程包。\n\n适合：① 想在 EdgeOne Makers 上做定时 Agent 应用的全栈开发者；② 需要搭建资讯/趋势聚合看板的产品团队；③ 想学习 LangGraph 多节点流水线 + SSE 流式推送 + 定时触发完整范式的工程师。", "description_en": "Want to stay on top of the AI world every day, but tired of jumping between Hacker News and Dev.to, manually filtering and summarizing — slow and easy to miss things?\n\nThis case is delivered by the \"Makers Dev Team\": the lead orchestrates the tech stack, the frontend engineer builds a React+Vite workbench, and the backend engineer implements a five-stage LangGraph pipeline (collect → filter/classify → summarize → score → report), streaming progress to the frontend over SSE in real time. It supports a cron trigger at 09:00 daily as well as a manual run button; it outputs a categorized, importance-scored Chinese Markdown digest that is persisted for history browsing. The case also ships the full deployable project source.\n\nFor: (1) full-stack developers building scheduled Agent apps on EdgeOne Makers; (2) product teams that need a news/trend aggregation dashboard; (3) engineers learning the complete pattern of LangGraph multi-node pipelines + SSE streaming + scheduled triggers.", "categories": [ "dev" ], "tags": [ "EdgeOne Makers", "AI Agent", "LangGraph", "定时任务", "全栈部署" ], "tags_en": [ "EdgeOne Makers", "AI Agent", "LangGraph", "Cron Job", "Full-stack Deploy" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "ai-daily-report.md", "artifacts": [ { "file": "ai-daily-report.md", "label": "AI 资讯日报（Agent 真实产出）" } ], "prompt": "我想在 EdgeOne Makers 上做一个定时运行的 AI 资讯/趋势聚合 Agent 应用，请按团队协作流程完成开发与部署。\n\n应用形态：\n- Agent 项目（根目录有 agents/，用 edgeone makers dev/deploy 驱动），TypeScript\n- 触发方式：① cron 定时（默认每天一次）+ ② 前端按钮手动触发\n- Agent 编排：多个节点协作完成「采集 → 过滤/分类 → 摘要 → 打分 → 生成 Markdown 报告」，过程通过 SSE 实时推送到前端\n- 前端：一个简洁的 React + Vite 面板，能看到流水线各阶段进度、最终 Markdown 报告，并提供「停止」按钮中断当前运行\n\n默认参数：\n1. 主题方向：AI 行业资讯\n2. 数据源：Hacker News、Dev.to 两个公开 API\n3. 定时频率：每天一次\n4. 单次条目数上限：30 条\n5. 模型：@makers/minimax-m2.7\n\n交付要求：主流程跑通，能稳定完成「定时触发 → Agent 流水线 → 出报告 → 前端可看报告+文章」；最后本地预览 + 部署，给我完整的访问地址。", "prompt_en": "I want to build a scheduled AI news/trend aggregation Agent app on EdgeOne Makers. Please complete development and deployment following the team collaboration workflow.\n\nApp shape:\n- Agent project (agents/ in the root, driven by edgeone makers dev/deploy), TypeScript\n- Triggers: (1) cron schedule (default once a day) + (2) a manual button in the frontend\n- Agent orchestration: multiple nodes collaborate on \"collect → filter/classify → summarize → score → generate Markdown report\", streaming progress to the frontend via SSE\n- Frontend: a clean React + Vite panel showing pipeline stage progress and the final Markdown report, with a \"Stop\" button to interrupt the current run\n\nDefaults:\n1. Topic: AI industry news\n2. Data sources: Hacker News and Dev.to public APIs\n3. Schedule: once a day\n4. Max items per run: 30\n5. Model: @makers/minimax-m2.7\n\nDelivery: get the main flow working so it reliably completes \"scheduled trigger → Agent pipeline → report → frontend shows report + articles\"; finally do a local preview + deploy and give me the full access URL.", "skills": [], "experts": [ { "id": "edgeone-makers-experts", "expert_id": "ex_EDEph9cW7RjS", "name": "Makers 开发专家团", "description": "在 EdgeOne Makers 上构建并部署 Web 应用，涵盖前端、Serverless 后端、AI Agent 开发与快速部署", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "2d3417bfb566" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "source": "official", "created_at": "2026-06-25T07:56:34Z", "updated_at": "2026-06-25T07:56:34Z", "is_recommended": false, "recommended_order": 0 }, { "id": "efund-chinext-computing-etf-listing", "title": "如何了解新上市的ETF：创业板算力ETF易方达", "title_en": "Getting to Know a Newly Listed ETF: E Fund ChiNext Computing Power ETF", "subtitle": "158050！创系列上新，这次轮到算力", "subtitle_en": "158050 is live — this time, computing power", "description": "新ETF上市，最麻烦的不是买不到，是看不懂——它跟踪的指数是什么、最近的行情走到了哪一步，信息散在公告、指数公司和七八篇公众号文章里。\n\n这份报告替你把这些收进一页：创业板算力指数近一年涨了多少、波动多大、最新点位和当日表现，一张表看完；再挑出 5 篇最相关的公众号原文，点标题直接跳原文，不用自己搜。\n\n你看完会知道：这只 ETF 到底装的是什么资产。", "description_en": "The hard part of a newly listed ETF is never access — it is understanding what is inside. What index does it track? Where does the market stand right now? The answers are scattered across filings, the index provider's site and half a dozen WeChat articles.\n\nThis report gathers them onto one page: how much the ChiNext computing-power index has returned over the past year, how volatile it has been, its latest level and daily move — all in a single table. Plus five of the most relevant E Fund WeChat articles, each a click away from the original.\n\nBy the end you will know what this ETF actually holds.", "categories": [ "finance" ], "tags": [ "算力", "创业板", "ETF上市", "指数基金", "一页看懂" ], "tags_en": [ "Computing Power", "ChiNext", "ETF Listing", "Index Fund", "One-pager" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "创业板算力ETF易方达 · 跟踪指数与相关资讯（单页 HTML 报告）" } ], "prompt": "围绕「易方达创业板算力ETF（158050）2026-09-18 在深交所上市」，用「指数查询」+ efunds 连接器出一份单页 HTML 简报，只讲两块：\n\n1）跟踪指数：创业板算力基础设施指数（970083）的编制属性、申万一级行业分布、前十大成分股权重、近一年收益与年化波动率、PE/PB 当前倍数、最新行情（点位、涨跌幅、昨收、时间）。\n2）相关资讯：挑 4–5 篇易方达系公众号相关文章，给出标题、公众号、发布时间与原文逐字摘录（不得改写为归纳），标题可点开原文。\n\n数据须真实查询并注明来源与日期，取不到就留空，不要编造或估算。单文件 HTML、样式内联、可离线打开；红涨绿跌；末尾附数据来源与风险提示。", "prompt_en": "For \"E Fund ChiNext Computing Power ETF (158050) listing on the Shenzhen Stock Exchange on 2026-09-18\", use \"index query\" plus the efunds connector to produce a single-page HTML briefing covering just two things:\n\n1) Tracked index: methodology attributes of the ChiNext Computing Power Infrastructure Index (970083), SW level-1 industry distribution, top-10 constituent weights, 1-year return and annualised volatility, current PE/PB multiples, and the latest quote (level, change rate, previous close, time).\n2) Related articles: pick 4–5 relevant E Fund WeChat articles and give the title, account, publication time and a verbatim excerpt (do not rewrite it as a summary), with the title clickable through to the original article.\n\nAll figures must be genuinely queried and labelled with source and date; leave blanks rather than inventing or estimating. Single HTML file, inlined styles, openable offline; gains in red, losses in green; end with data sources and risk disclosure.", "skills": [ { "id": "指数查询", "name": "指数查询", "description": "查询证券指数的基础信息、行情、成分股、行业分布、估值、收益及相关基金等数据。", "skill_id": "skill_2089730438115336192" } ], "experts": [], "mcps": [ { "id": "efunds", "name": "易方达基金MCP服务", "description": "按基金代码查询基金画像（净值、业绩、基金经理、持仓、费率等），并检索与读取易方达官方公众号文章。" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-09-17T18:03:39+08:00", "updated_at": "2026-09-18T12:12:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-cpo-theme-funds", "title": "CPO主题的易方达基金", "title_en": "Efunds Funds on the CPO Theme", "subtitle": "想了解易方达旗下哪些基金与CPO（共封装光学）相关？帮你从AI算力、信息技术、通信主题基金中挑出最相关的几只", "subtitle_en": "Want to know which Efunds funds are tied to CPO (co-packaged optics)? We surface the most relevant AI-computing, information-technology and telecom theme funds.", "description": "痛点：想了解哪些易方达基金聚焦 CPO（共封装光学）方向？跟踪什么指数、风险几何、官方最近有哪些行业解读？\n\n演示内容：这个案例通过易方达官方渠道检索名称与业绩基准含「CPO/光模块」的基金，通过语义理解将 CPO 对应至 AI 算力/信息技术/通信方向最相关的主题基金，按今年以来收益排序展示前 5 只，并附上官方最近发布的行业解读文章，一屏看清基金跟踪的指数、风险等级与最新官方资讯。案例演示了如何让 AI 通过易方达官方渠道检索行业主题基金，并同步整理官方最新的行业观点文章。\n\n适用人群：想了解易方达旗下行业主题基金、关注主题纯度与官方资讯的投资者", "description_en": "Pain point: Want to know which Efunds funds focus on CPO (co-packaged optics)? What index do they track, how risky are they, and what industry commentary has the fund company published recently?\n\nDemo content: This case searches Efunds' official channels for funds whose names and performance benchmarks mention \"CPO / optical modules\", uses semantic understanding to map CPO to the most relevant AI-computing / information-technology / telecom theme funds, sorts the top 5 by YTD return, and attaches the latest official industry commentary — one screen shows the tracked index, risk level and newest official news. The case demonstrates how to have AI search industry theme funds through official Efunds channels and collect the latest official industry views at the same time.\n\nTarget audience: Investors who want to explore Efunds' industry theme funds and care about theme purity and official information", "categories": [ "finance", "data" ], "tags": [ "基金选基", "CPO", "共封装光学", "易方达", "行业主题基金" ], "tags_en": [ "Fund Selection", "CPO", "Co-packaged Optics", "Efunds", "Industry Theme Funds" ], "artifact_type": "html", "external_url": "", "preview": "cpo-theme-funds.html", "artifacts": [ { "file": "cpo-theme-funds.html", "label": "CPO主题的易方达基金·可视化选基报告" } ], "prompt": "CPO行业的易方达基金有哪些？按今年以来收益率从高到低展示前 5 只，并给出每只基金的风险等级、基金经理、近1年收益率与业绩比较基准收益率；同时整理易方达官方最近发布的CPO行业观点文章。数据须全部来自易方达官方渠道（官网基金信息 + 官方公众号文章），官方未披露的字段如实标注，不得编造、不得引用第三方。", "prompt_en": "Which Efunds funds are in the CPO industry? Show the top 5 sorted by YTD return from high to low, and give each fund's risk level, fund manager, 1-year return, and the corresponding benchmark return; also compile the latest official Efunds articles on the CPO industry. All data must come from Efunds' official channels (official fund information + official WeChat account articles); fields not officially disclosed must be labeled as such. Do not fabricate, and do not cite third-party sources.", "skills": [ { "id": "efunds-fund-selector", "name": "易方达智能选基", "description": "易方达旗下基金智能筛选：按重仓股、行业、大类资产、量化指标四个分支选基，输出对话简表与可视化报告。", "skill_id": "skill_2092527229591240704" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-31T14:51:00+08:00", "updated_at": "2026-08-31T14:51:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-grid-theme-funds", "title": "电网主题的易方达基金及经理观点", "title_en": "Efunds Grid-Theme Funds & Manager Views", "subtitle": "哪些易方达基金与电网相关？基金经理最近怎么看？选基结果与经理观点一次看齐", "subtitle_en": "Which Efunds funds are tied to the power grid? What do fund managers think lately? Fund picks and manager views in one place.", "description": "痛点：想了解哪些易方达基金聚焦电网方向？跟踪什么指数、风险几何、官方最近有哪些行业解读与基金经理观点？\n\n演示内容：这个案例通过易方达官方渠道检索名称与业绩基准含「电网」的基金（名称直接命中的电网设备基金之外，通过语义理解将电网对应至电力产业链的发电、供电环节，纳入绿色电力与电力公用事业主题基金），按今年以来收益排序展示前 5 只，并附上官方最近发布的行业解读文章与基金经理近期观点，一屏看清基金跟踪的指数、风险等级与最新官方资讯。案例演示了如何让 AI 通过易方达官方渠道检索行业主题基金，并同步整理官方最新的行业观点文章与基金经理观点。\n\n适用人群：想了解易方达旗下行业主题基金、关注主题纯度与官方资讯的投资者", "description_en": "Pain point: Want to know which Efunds funds focus on the power grid? What index do they track, how risky are they, and what industry commentary and fund-manager views has the fund company published recently?\n\nDemo content: This case searches Efunds' official channels for funds whose names and performance benchmarks mention \"power grid\" (beyond funds directly named grid-equipment, semantic understanding maps the grid to the power-generation and supply links of the power industry chain, adding green-power and electricity-utility theme funds), ranks the top 5 by YTD return, and attaches the latest official industry commentary and recent fund-manager views — one screen shows the tracked index, risk level and newest official news. The case demonstrates how to have AI search industry theme funds through official Efunds channels and compile the latest official industry views and fund-manager opinions.\n\nTarget audience: Investors who want to explore Efunds' industry theme funds and care about theme purity and official information", "categories": [ "finance", "data" ], "tags": [ "基金选基", "电网", "电力", "易方达", "经理观点" ], "tags_en": [ "Fund Selection", "Power Grid", "Electricity", "Efunds", "Manager Views" ], "artifact_type": "html", "external_url": "", "preview": "grid-theme-funds.html", "artifacts": [ { "file": "grid-theme-funds.html", "label": "电网主题的易方达基金及经理观点·可视化选基报告" } ], "prompt": "电网行业的易方达基金有哪些？按今年以来收益率从高到低展示前 5 只，并给出每只基金的风险等级、基金经理、近1年收益率与业绩比较基准收益率；同时整理易方达官方最近发布的电网行业观点文章，以及这些基金基金经理的近期观点。数据须全部来自易方达官方渠道（官网基金信息 + 官方公众号文章），官方未披露的字段如实标注，不得编造、不得引用第三方。", "prompt_en": "Which Efunds funds are in the power-grid industry? Show the top 5 sorted by YTD return from high to low, and give each fund's risk level, fund manager, 1-year return, and the corresponding benchmark return; also compile the latest official Efunds articles on the grid industry and the recent views of these funds' managers. All data must come from Efunds' official channels (official fund information + official WeChat account articles); fields not officially disclosed must be labeled as such. Do not fabricate, and do not cite third-party sources.", "skills": [ { "id": "efunds-fund-selector", "name": "易方达智能选基", "description": "易方达旗下基金智能筛选：按重仓股、行业、大类资产、量化指标四个分支选基，输出对话简表与可视化报告。", "skill_id": "skill_2092527229591240704" }, { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-31T14:59:00+08:00", "updated_at": "2026-08-31T14:59:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-hotspot-recent", "title": "近期市场热点梳理", "title_en": "E Fund Hotspot Combing: Recent Investment Hotspots", "subtitle": "基于易方达官方公众号，梳理近一周投资热点", "subtitle_en": "Three recent investment hotspots combed from E Fund official WeChat articles, with real index quotes", "description": "每天财经新闻铺天盖地，却总觉得抓不住重点？这个案例帮你用一分钟看清最近一周市场到底在聊什么。\n\n它会自动读取易方达基金官方公众号（易方达基金 / 易方达指数通）近一周的公开文章，把被反复提及的热点方向挑出来，并为每个热点配上：来源文章的原文摘录、关联指数的真实涨跌数据（来自指数直通车 etf.com.cn）、对应的基金产品链接，最后一键生成一份清爽的 HTML 速览报告，还自带风险提示。\n\n适合这样的你：想快速跟上市场节奏的基金投资者；需要给客户做每周行情简报的理财顾问；关注板块轮动与产业趋势的研究员；想看看 AI 热点梳理到底靠不靠谱的产品同学。", "description_en": "Drowning in financial news but can't pin down what actually matters? This case helps you see what the market has been talking about over the past week — in about a minute.\n\nIt automatically reads the past week's public posts from E Fund's two official WeChat accounts (E Fund / E Fund Index), surfaces the topics that keep coming up, and for each one pairs a verbatim quote from the source article, the real return data of the related indices (from etf.com.cn), and links to the matching E Fund products — then turns it all into a clean, single-page HTML overview with built-in risk disclaimers.\n\nThis is for you if: you're a fund investor who wants to keep up with the market; a wealth advisor who needs a quick weekly briefing for clients; an analyst following sector rotation and industry trends; or a product person curious whether AI hotspot combing actually holds up.", "categories": [ "finance" ], "tags": [ "热点梳理", "投资理财", "市场速览", "基金", "易方达" ], "tags_en": [ "Hotspot Combing", "Investment", "Market Overview", "Fund", "E Fund" ], "artifact_type": "html", "preview": "efunds-hotspot-recent.html", "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "用「易方达热点梳理」技能出近一周热点速览 HTML：抓两官号文章归并 3 个热点，配原文 + 真实行情（涨红跌绿）+ 基金链接，过敏感词合规，数据真实可溯源。", "prompt_en": "Help me comb the investment hotspots of the past week and produce an HTML overview report titled \"近期市场热点梳理\" (Recent Market Hotspot Combing). Use the \"易方达热点梳理\" skill (skill_2092287909491437568):\n\n1. Search: with the efunds connector, pull the past 7 days of public posts from both \"易方达基金\" and \"易方达指数通\" accounts in parallel, then merge and deduplicate;\n2. Pick: cluster by topic (keep only the latest post as the primary source per topic), then choose 3 hotspots by timeliness and attention, with event-style titles that name the news;\n3. Quote verbatim: open the matched articles and read the full text; quote any market viewpoint word-for-word (100–200 chars, making one complete point), and cite the source (title + account + time + link). Do not summarize or rephrase;\n4. Fill in quotes: for each hotspot's related indices, fetch real data from etf.com.cn (latest level, plus 1D / 1W / 1M / 3M / 1Y % changes, red for up / green for down) and show it in a table; hide any index without real data — never fabricate;\n5. Match products: map each sector to the corresponding E Fund product (look up the product code table) and turn the product name into a link to its fund profile;\n6. Build the report: use the \"近期热点速览\" HTML template to generate a single-file report (all images and styles inlined, no external references), in E Fund blue, with the fixed risk disclaimer and index risk statement built in;\n7. Compliance: before finalizing, check against the sensitive-words list line by line — banned words must never appear, and warning words should be swapped for near-synonyms.\n\nAll data must be real, traceable, and attributed.", "skills": [ { "id": "efunds-hotspot-combing", "name": "易方达热点梳理", "description": "易方达基金「热点梳理」总入口：基于 efunds-mcp 检索「易方达基金」「易方达指数通」公众号公开文章，覆盖事件热点梳理/近期热点速览/热点周报三类成稿，AI 只做资讯与事实梳理，市场观点一律原文引用。", "skill_id": "skill_2092287909491437568" } ], "experts": [], "mcps": [ { "id": "efunds", "name": "易方达基金", "description": "易方达基金公众号文章检索与正文读取（search_wx_articles / get_wx_article_detail）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T09:34:36+08:00", "updated_at": "2026-08-27T09:34:36+08:00", "artifacts": [ { "file": "efunds-hotspot-recent.html", "label": "近期市场热点梳理报告（HTML 速览）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-ningde-times-holdings", "title": "重仓宁德时代的易方达基金", "title_en": "Efunds Funds Heavy on CATL (Ningde Times)", "subtitle": "想知道哪些易方达基金重仓了宁德时代？按持仓占比帮你挑出最相关的几只", "subtitle_en": "Want to know which Efunds funds hold CATL heavily? We surface the most relevant ones ranked by position weight.", "description": "痛点：想了解某只股票被哪些基金重仓？持仓占比多少、风险几何、官方最近有哪些披露？\n\n演示内容：这个案例以宁德时代为例，通过易方达官方渠道找出前十大重仓中包含该股票的基金，按持仓占比排序取前 5 只，一表看清持仓权重、风险等级与基金经理，并附可视化报告。案例演示了如何让 AI 通过易方达官方渠道查询重仓某只股票的基金，按持仓占比排序，逐只核对关键信息。\n\n适用人群：想通过基金了解个股持仓情况、关注持仓透明度的投资者", "description_en": "Pain point: Want to know which funds hold a given stock heavily? What's the position weight, how risky is it, and what has the fund company officially disclosed lately?\n\nDemo content: Using CATL (Ningde Times) as an example, this case finds Efunds funds whose top-10 holdings include the stock via official Efunds channels, sorts them by position weight and takes the top 5 — one table shows position weight, risk level and fund manager, plus a visual report. The case demonstrates how to have AI query funds heavy on a given stock through official Efunds channels, sort by position weight, and verify key facts fund by fund.\n\nTarget audience: Investors who want to understand a stock's fund ownership and care about holdings transparency", "categories": [ "finance", "data" ], "tags": [ "基金选基", "宁德时代", "易方达", "重仓股反查", "可视化报告" ], "tags_en": [ "Fund Selection", "CATL", "Efunds", "Heavy-held Reverse Lookup", "Visual Report" ], "artifact_type": "html", "external_url": "", "preview": "ningde-times-holdings.html", "artifacts": [ { "file": "ningde-times-holdings.html", "label": "重仓宁德时代的易方达基金·可视化选基报告" } ], "prompt": "重仓宁德时代的易方达基金有哪些？按宁德时代占基金净值的比例从高到低展示前 5 只，并给出每只基金的风险等级、基金经理、近1年收益率与对应的业绩比较基准收益率。数据须全部来自易方达官方渠道（官网定期报告披露的持仓数据），官方未披露的字段如实标注，不得编造、不得引用第三方。", "prompt_en": "Which Efunds funds are heavy on CATL (Ningde Times)? Show the top 5 sorted by CATL's weight in each fund's net asset value from high to low, and give each fund's risk level, fund manager, 1-year return, and the corresponding benchmark return. All data must come from Efunds' official channels (holdings disclosed in official periodic reports); fields not officially disclosed must be labeled as such. Do not fabricate, and do not cite third-party sources.", "skills": [ { "id": "efunds-fund-selector", "name": "易方达智能选基", "description": "易方达旗下基金智能筛选：按重仓股、行业、大类资产、量化指标四个分支选基，输出对话简表与可视化报告。", "skill_id": "skill_2092527229591240704" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T15:04:00+08:00", "updated_at": "2026-08-31T14:39:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-stable-bond-funds", "title": "稳健债券型的易方达基金", "title_en": "Stable Bond Funds from Efunds", "subtitle": "想要波动相对较小、收益相对稳健的债券基金？帮你从易方达偏债混合（中低风险）中挑出今年以来表现相对稳健的几只", "subtitle_en": "Looking for bond funds with relatively low volatility and steady returns? We pick out Efunds' partial-bond hybrid (medium-low risk) funds that have been relatively steady so far this year.", "description": "痛点：想找偏稳健的债券方向基金，但不知道易方达旗下哪些产品波动相对较小、历史回撤控制得相对较好。\n\n演示内容：这个案例通过易方达官方渠道，按「稳健」偏好从纯债（短债/同业存单）与偏债混合（中低风险）中筛选，按今年以来收益排序取前 5 只，一表看清基金类型、风险等级、基金经理与业绩基准，并附可视化报告。案例演示了如何让 AI 通过易方达官方渠道按风险偏好筛选债券方向基金。\n\n适用人群：希望筛选低波动债券方向基金、关注风险等级与回撤控制的投资者", "description_en": "Pain point: Looking for bond funds with a stable tilt, but unsure which Efunds products have relatively low volatility and a decent drawdown history.\n\nDemo content: This case uses official Efunds channels and a \"stable\" preference to screen pure-bond (short-term bond / certificates of deposit) and partial-bond hybrid (medium-low risk) funds, ranks the top 5 by YTD return, and shows fund type, risk level, fund manager and performance benchmark in one table, plus a visual report. The case demonstrates how to have AI screen bond funds by risk preference through official Efunds channels.\n\nTarget audience: Investors who want to screen low-volatility bond funds and care about risk ratings and drawdown control", "categories": [ "finance", "data" ], "tags": [ "基金选基", "债券基金", "稳健", "易方达", "低波动" ], "tags_en": [ "Fund Selection", "Bond Funds", "Stable", "Efunds", "Low Volatility" ], "artifact_type": "html", "external_url": "", "preview": "stable-bond-funds.html", "artifacts": [ { "file": "stable-bond-funds.html", "label": "稳健债券型的易方达基金·可视化选基报告" } ], "prompt": "帮我找较为稳健的易方达债券基金。按今年以来收益率从高到低展示前 5 只，并给出每只基金的风险等级、基金经理、近1年收益率与业绩比较基准收益率。数据须全部来自易方达官方渠道（官网基金信息），官方未披露的字段如实标注，不得编造、不得引用第三方。", "prompt_en": "Help me find relatively stable Efunds bond funds. Show the top 5 sorted by YTD return from high to low, and give each fund's risk level, fund manager, 1-year return, and the corresponding benchmark return. All data must come from Efunds' official channels (official fund information); fields not officially disclosed must be labeled as such. Do not fabricate, and do not cite third-party sources.", "skills": [ { "id": "efunds-fund-selector", "name": "易方达智能选基", "description": "易方达旗下基金智能筛选：按重仓股、行业、大类资产、量化指标四个分支选基，输出对话简表与可视化报告。", "skill_id": "skill_2092527229591240704" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-31T14:55:00+08:00", "updated_at": "2026-08-31T14:55:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-top-active-equity-funds", "title": "收益居前的易方达主动权益基金", "title_en": "Top-YTD Active Equity Funds from Efunds", "subtitle": "想了解主动权益基金里今年哪些收益居前？帮你从易方达主动股票与偏股混合中挑出今年以来表现居前的几只", "subtitle_en": "Curious which active equity funds lead this year? We pick out Efunds' active-stock and partial-equity hybrid funds with the best YTD performance.", "description": "痛点：想找收益居前的主动权益基金，但不知道易方达旗下主动股票/偏股混合类产品里哪些今年表现居前、风险等级与回撤情况如何。\n\n演示内容：这个案例通过易方达官方渠道，从主动股票与偏股混合（主动权益）基金中按今年以来收益排序取前 5 只，一表看清基金类型、风险等级、基金经理、近1年收益与业绩比较基准，并附可视化报告（含关键指标雷达图与风险-收益散点图）。案例演示了如何让 AI 通过易方达官方渠道按量化指标筛选主动权益基金。\n\n适用人群：想了解易方达主动权益基金表现、关注收益水平与风险等级的投资者", "description_en": "Pain point: Looking for top-return active equity funds, but unsure which Efunds active-stock / partial-equity hybrid products lead this year, and what their risk levels and drawdowns look like.\n\nDemo content: This case uses official Efunds channels to rank active-stock and partial-equity hybrid (active equity) funds by YTD return and takes the top 5 — one table shows fund type, risk level, fund manager, 1-year return and performance benchmark, plus a visual report (key-indicator radar chart and risk-return scatter chart). The case demonstrates how to have AI screen active equity funds by quantitative indicators through official Efunds channels.\n\nTarget audience: Investors who want to track Efunds' active equity fund performance and care about return levels and risk ratings", "categories": [ "finance", "data" ], "tags": [ "基金选基", "主动权益", "易方达", "收益排名", "量化指标" ], "tags_en": [ "Fund Selection", "Active Equity", "Efunds", "Return Ranking", "Quantitative Indicators" ], "artifact_type": "html", "external_url": "", "preview": "top-active-equity-funds.html", "artifacts": [ { "file": "top-active-equity-funds.html", "label": "收益居前的易方达主动权益基金·可视化选基报告" } ], "prompt": "帮我找今年收益居前的易方达主动权益基金。按今年以来收益率从高到低展示前 5 只，并给出每只基金的风险等级、基金经理、近1年收益率与业绩比较基准收益率。数据须全部来自易方达官方渠道（官网基金信息），官方未披露的字段如实标注，不得编造、不得引用第三方。", "prompt_en": "Help me find Efunds active equity funds with the best returns this year. Show the top 5 sorted by YTD return from high to low, and give each fund's risk level, fund manager, 1-year return, and the corresponding benchmark return. All data must come from Efunds' official channels (official fund information); fields not officially disclosed must be labeled as such. Do not fabricate, and do not cite third-party sources.", "skills": [ { "id": "efunds-fund-selector", "name": "易方达智能选基", "description": "易方达旗下基金智能筛选：按重仓股、行业、大类资产、量化指标四个分支选基，输出对话简表与可视化报告。", "skill_id": "skill_2092527229591240704" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-31T14:57:00+08:00", "updated_at": "2026-08-31T14:57:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "enterprise-fsr-energy-storage", "title": "企业投资可研第1-3章编写实操：新能源储能电站", "title_en": "Enterprise FSR Chapters 1-3: New Energy Storage Station", "subtitle": "基于2023版大纲，逐章演示企业投资可研与政府项目的关键差异", "subtitle_en": "A chapter-by-chapter demo of enterprise feasibility study with differences from government projects", "description": "企业可研报告和政府项目可研大纲相似但写法迥异——最头疼的是不知道哪些地方该写「赚钱逻辑」、哪些地方该写「公共服务」。\n\n本案例以某制造企业 50MW/100MWh 储能电站为样本，按2023版10章框架逐章编写第1-3章（项目概况、市场与商业模式、选址与要素保障），每章标注与政府项目写法的关键差异，让读者一眼看清企业项目可研的正确姿势。\n\n适合：企业可研编制人员、银行审贷经理、投资经理、咨询顾问、储能项目开发者。", "description_en": "Enterprise feasibility reports and government project reports share similar outlines but require fundamentally different approaches — the hardest part is knowing where to write 'profit logic' and where to write 'public service'.\n\nThis case uses a manufacturing enterprise's 50MW/100MWh energy storage station as a sample, demonstrating Chapters 1-3 (Project Overview, Market & Business Model, Site Selection & Factor Assurance) following the 2023 edition 10-chapter framework, with key differences from government project approaches clearly annotated for each chapter.\n\nFor: feasibility report writers, bank credit reviewers, investment managers, consultants, energy storage project developers.", "categories": [ "business", "finance", "productivity" ], "tags": [ "可研报告", "储能电站", "企业投资", "编写指引", "2023版大纲" ], "tags_en": [ "feasibility study", "energy storage", "enterprise investment", "writing guide", "2023 outline" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "第1-3章编写稿（含企业vs政府差异对比）" } ], "prompt": "请根据《项目任务书_新能源储能电站企业投资.md》中的项目参数（50MW/100MWh储能电站，总投资约2.1亿元，商业模式为峰谷套利+需求响应+备用容量），按《企业投资项目可行性研究报告编写参考大纲》（2023年版）10章框架，编写第1-3章（项目概况、市场与商业模式、选址与要素保障）的编写稿。\n\n要求：\n1. 严格按照企业投资可研的编写指引，每章包含指引要求的全部子节\n2. 每章末尾以「与政府项目的差异」框单独标注企业项目与政府项目写法的关键区别\n3. 项目概况需包含项目基本画像（含技术经济指标表）、企业综合画像（含能力匹配论证）、编制依据、结论与建议综述\n4. 市场与商业模式需包含规划政策符合性、企业战略必要性、市场维度分析（行业全景/目标市场/容量预测/竞争格局/价格分析）、建设内容与产出方案、商业模式设计（收入结构/成本结构/盈利评估/外部支持条件）\n5. 选址与要素保障需包含多方案比选、建设条件分析、要素保障评估\n6. 数据使用真实查询或基于行业公开数据的合理推算，标注数据来源和推算依据\n7. 输出为单文件HTML，资源全内联、可离线打开", "prompt_en": "Write Chapters 1-3 of a feasibility study report based on the project parameters (50MW/100MWh energy storage station, total investment ~210 million CNY, business model: peak-valley arbitrage + demand response + backup capacity), following the 2023 edition \"Reference Outline for Enterprise Investment Project Feasibility Study Reports\" 10-chapter framework.\n\nRequirements:\n1. Follow enterprise investment FSR writing guidelines strictly, include all subsections required by the outline\n2. Mark key differences from government project approaches in a separate \"Differences from Government Projects\" box at the end of each chapter\n3. Project Overview must include project profile (with technical-economic indicators table), enterprise profile (with capability matching assessment), compilation basis, and conclusions & recommendations summary\n4. Market & Business Model must include policy compliance, enterprise strategic necessity, market dimension analysis (industry landscape/target market/capacity forecast/competitive landscape/price analysis), construction content & output plan, and business model design (revenue structure/cost structure/profitability assessment/external support conditions)\n5. Site Selection & Factor Assurance must include multi-scheme comparison, construction condition analysis, and factor assurance assessment\n6. Use real queries or reasonable estimates based on public industry data, annotate data sources and estimation basis\n7. Output as a single-file HTML with all resources inlined for offline access", "skills": [ { "id": "enterprise-project-fsr-writing-guide", "name": "企业投资项目可研报告分章编写指引", "description": "覆盖2023版大纲全部10章：从项目概况、市场与商业模式、选址保障、建设运营方案，到投融资财务评估、影响效果与风险管控，逐节提供内容要点与编写建议", "skill_id": "skill_2085354977164185600" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-04T04:12:03.000Z", "updated_at": "2026-08-04T04:12:03.000Z", "is_recommended": false, "recommended_order": 0 }, { "id": "enterprise-internal-proposal", "title": "跨境战略提案：本地仓履约试点", "title_en": "Cross-border Strategic Proposal: Local Warehouse Fulfillment Pilot", "subtitle": "按大型组织内部提案方法论，产出可提交VP审核的跨部门战略提案初稿", "subtitle_en": "A cross-department strategic proposal draft ready for VP review, produced using enterprise internal proposal methodology", "description": "向领导提交跨部门战略提案最怕越权踩雷——写了供应链部门的行动计划、替财务定了预算、给法务排了认证节点，VP一眼就看出你不懂组织边界。\n\n这个案例完整演示了大型组织内部提案的四阶段工作法：先做多源调研与数据交叉验证，再写出覆盖市场分析、竞争格局、成本收益框算、风险边界的完整版（42页），然后经过三重自审过滤——越权审查、冗余审查、越级审查——删掉替其他部门做决策的内容、压缩内部众所周知的信息、把\"建议正式启动\"改写为\"建议纳入规划讨论\"，最终打磨为18页可上会版。\n\n适合：需要向VP/经营会提交跨部门协同提案的业务负责人、战略/商分团队、大型组织内的中高层管理者。", "description_en": "The biggest pitfall in internal strategic proposals is overstepping organizational boundaries—writing action plans for supply chain, setting budgets for finance, or scheduling certification milestones for legal. This case demonstrates the complete four-stage methodology, from multi-source research to triple self-review filtering, producing a VP-ready proposal that respects departmental autonomy.\n\nIdeal for: business heads submitting cross-department proposals to VPs, strategy/business analysis teams, and mid-to-senior managers in large organizations.", "categories": [ "business", "productivity" ], "tags": [ "内部提案", "跨部门协同", "战略分析", "VP审核", "组织方法论" ], "tags_en": [ "internal proposal", "cross-department", "strategic analysis", "VP review", "organizational methodology" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "proposal.html", "prompt": "请按大型组织内部提案方法，基于以下背景信息，帮我产出一份可提交 VP 审核的跨部门战略提案初稿：\n\n【提案背景】\n- 提案人：国际业务部 | 审核层级：需 VP 审核后提交经营会\n- 方向：在东南亚核心市场试点「本地仓 + 三日达」履约能力，服务公司 3C 配件跨境生意。\n- 已知信息：2025 年该区域 GMV 约 1.2 亿人民币，物流时效投诉占比 18%；竞品已有本地仓，签收时效中位数快我们 1.5 天；预算粗估年增运营成本 900 万。\n\n【输出要求】\n提案需覆盖以下全部内容：\n(a) 调研提纲——列出需要验证的关键数据和信息缺口\n(b) 核心论点——用数据支撑的核心主张，每条结论须标注数据来源和可信度\n(c) 成本/收益框算——分情景（保守/基准/乐观）的投入产出分析\n(d) 组织协同与风险边界——列出需要供应链、财务、法务三方参与的关键议题，标注本方可提供的输入\n(e) 自审删减说明——按越权审查、冗余审查、越级审查三重过滤，标注所有删改内容及原因\n\n严格遵循大型组织内部提案的五项行为准则：署名即担责、不为上级写话术、协同邀约而非任务分配、数据要硬建议要软、内部文档保留边界声明。", "prompt_en": "Using the enterprise internal proposal methodology and the following background, produce a cross-department strategic proposal draft ready for VP review:\n\n[Background]\n- Proposer: International Business Department | Review level: VP review before management committee submission\n- Direction: Pilot \"local warehouse + 3-day delivery\" fulfillment capability in core Southeast Asian markets for the company's 3C accessories cross-border business.\n- Known facts: 2025 regional GMV ~120M RMB, logistics complaint rate 18%; competitors already have local warehouses with median delivery 1.5 days faster; estimated annual incremental operating cost ~9M RMB.\n\n[Output requirements]\nThe proposal must cover: (a) research outline, (b) core arguments with data sources and confidence levels, (c) cost/benefit estimation across scenarios, (d) organizational synergy and risk boundaries, and (e) self-review deletion notes using the triple-filter method.\n\nStrictly follow the five organizational conduct principles: signatory accountability, no writing communication scripts for superiors, synergy invitations not task assignments, hard data soft recommendations, and boundary disclaimers in all cross-department sections.", "skills": [ { "id": "enterprise-internal-proposal-method", "name": "大型组织内部提案方法论", "description": "面向大型组织的内部跨部门战略提案方法论，覆盖从信息调研、全量起草、自审过滤到终版打磨的完整工作流", "skill_id": "skill_2085354907740610560" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "artifacts": [ { "file": "proposal.html", "label": "完整战略提案初稿（HTML）" } ], "created_at": "2026-08-04T12:08:27+08:00", "updated_at": "2026-08-04T12:08:27+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "enterprise-valuation-moutai", "title": "贵州茅台(600519) 企业估值全流程分析", "title_en": "Kweichow Moutai (600519) Enterprise Valuation Analysis", "subtitle": "Enterprise DCF × 财务五维体检 × 可比乘数校验", "subtitle_en": "Enterprise DCF × 5-Dimension Financial Health × Multiples Validation", "description": "白酒行业承压、茅台营收首次下滑的2025年，到底值多少钱？本案例以贵州茅台2025年年报为财务底稿，结合实时行情与可比公司数据，执行完整的Enterprise DCF估值模型——从财务五维体检、ROIC分解、自由现金流预测到WACC测算、终值估算、乘数校验，最终给出每股内在价值与安全边际研判。适合投研分析师做深度估值参考，也适合投资者学习DCF建模全流程。", "description_en": "What is Kweichow Moutai worth in 2025, when its revenue declined for the first time? This case uses Moutai's 2025 annual report as the financial workpaper, combined with real-time market data and peer comparison, to execute a complete Enterprise DCF valuation model — covering 5-dimension financial health check, ROIC decomposition, FCF forecasting, WACC estimation, terminal value calculation, and multiples validation. Suitable for equity research analysts and investors learning DCF modeling.", "tags": [ "企业估值", "DCF", "白酒", "贵州茅台", "财务分析", "价值投资" ], "tags_en": [ "Enterprise Valuation", "DCF", "Baijiu", "Moutai", "Financial Analysis", "Value Investing" ], "prompt": "请对贵州茅台（600519.SH）做企业估值全流程分析。先自行下载公司官网披露的2025年年度报告PDF作为财务底稿，再结合可信金融数据skill补行情与可比公司。输出含财务体检、FCF/WACC/终值、乘数校验与综合研判；文末加投资免责声明。年报PDF URL：https://www.moutaichina.com/mtgf/articleFileDir/2026-04/17/1b9fae59825c41bf9a776892a00565f7.pdf", "prompt_en": "Perform a full enterprise valuation analysis on Kweichow Moutai (600519.SH). First download the 2025 annual report PDF from the company's official website as the financial workpaper, then supplement with real-time market data and comparable company data from trusted financial data tools. Output should include financial health check, FCF/WACC/terminal value, multiples validation, and comprehensive investment assessment with a disclaimer at the end.", "artifact_type": "html", "preview": "moutai-valuation-report.html", "artifacts": [ { "file": "moutai-valuation-report.html", "label": "估值分析报告（自包含HTML）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "categories": [ "finance", "data" ], "skills": [ { "id": "enterprise-valuation-framework", "name": "企业估值全流程分析框架", "description": "企业估值全流程分析框架，覆盖财务分析、DCF/相对估值建模、情景假设与投资价值判断", "skill_id": "skill_2085355044273049600" } ], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-04T11:43:25+08:00", "updated_at": "2026-08-04T11:43:25+08:00", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "equity-investment-review", "title": "股权投资协议风险审查", "title_en": "Equity Investment Agreement Risk Review", "subtitle": "穿透对赌回购、代持安排与优先清算条款的法律效力", "subtitle_en": "Assess enforceability of VAM buyback, nominee shareholding, and liquidation preference clauses", "description": "股权投资协议是VC/PE领域最复杂的合同类型之一，涉及估值调整、业绩承诺、股权代持、反稀释、优先清算等多层嵌套条款。本案例演示专家如何审查一份含对赌与代持安排的A轮投资协议，识别主体授权权限不匹配、对赌触发条件表述模糊、代持期间表决权行使规则缺失、完全棘轮反稀释可能导致的极端稀释、优先清算回报倍数过高等关键法律风险点。\n\n适合：投资人、FA、公司法务、创业者/创始人。", "description_en": "Equity investment agreements are among VC/PE's most complex contracts, involving valuation adjustment, performance commitments, shareholding-in-trust, anti-dilution and liquidation preference. This case demonstrates how the expert reviews an A-round agreement with VAM and trust arrangements, identifying authorization mismatch, vague VAM triggers, missing voting rules during trust, extreme dilution from full-ratchet anti-dilution, and excessive liquidation-preference multiples.\n\nSuitable for: investors, FAs, corporate counsel, founders.", "categories": [ "business" ], "tags": [ "对赌回购", "股权代持", "优先清算" ], "tags_en": [ "VAM Buyback", "Nominee Holding", "Liquidation Preference" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请对我提供的《A轮股权投资协议》进行深度法律风险审查。重点核查三方签署主体的授权权限是否覆盖本次交易（特别是代持授权）、对赌回购触发条件的表述是否精确无歧义（IPO定义、估值翻倍的计算基准）、代持期间甲方作为实际出资人的知情权和表决权行使机制是否完备、完全棘轮反稀释在后续低价融资时的极端稀释后果、优先清算权的回报倍数与参与分配条款是否公平合理、竞业限制期限是否符合法定上限。输出分级风险报告。", "prompt_en": "Please perform a deep legal risk review on my Series A Equity Investment Agreement. Verify three-party signing authorization coverage, VAM buyback trigger precision, completeness of the actual investor's information & voting mechanism during trust, extreme-dilution consequences from full-ratchet anti-dilution, fairness of liquidation-preference multiple, and whether non-compete duration complies with statutory limits. Output a tiered report.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T16:56:11+08:00", "updated_at": "2026-08-13T18:08:23+08:00", "artifacts": [ { "file": "output.html", "label": "基于 V1.3 真实审查任务的互动审查页面" }, { "file": "股权投资协议_完整审查报告_V1.6.html", "label": "可视化商务版完整审查报告（HTML）" }, { "file": "股权投资协议_完整审查报告_V1.3.md", "label": "完整审查报告数据版（Markdown）" }, { "file": "股权投资协议_审查批注_V1.3.docx", "label": "腾讯电子签官方导出的 V1.3 审查批注合同" }, { "file": "股权投资协议_审查摘要_V1.3.xlsx", "label": "腾讯电子签官方导出的 V1.3 审查摘要 Excel" }, { "file": "真实审查结果_V1.3.json", "label": "腾讯电子签 V1.3 审查任务原始返回数据" }, { "file": "股权投资协议.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "etf-theme-news-report", "title": "科创芯片ETF资讯报告", "title_en": "STAR Chip ETF Thematic News Report", "subtitle": "主题投资调研第一步：把散落的 ETF 资讯整理成报告", "subtitle_en": "First step of thematic research: turn scattered ETF news into a structured report", "description": "想搞清楚一个主题 ETF 最近发生了什么，翻遍资讯软件也未必拼得全。这个案例演示了如何让 AI 检索科创芯片 ETF 的真实资讯，按产品动态、指数与成份股、行业基本面等维度整理成一份研究报告质感的资讯报告，脉络清晰、来源可查、只陈述事实不做投资建议。\n\n适合：做主题投资调研的投资者、基金爱好者、需要周报素材的理财顾问和内容创作者。", "description_en": "Figuring out what has been happening around a thematic ETF is hard—scrolling through news apps rarely gives you the full picture. This case shows how AI retrieves real news about the STAR Chip ETF and organizes it into a research-report-style brief across dimensions like product updates, index and constituents, and industry fundamentals. Clear narrative, verifiable sources, facts only with no investment advice.\n\nBest for: investors doing thematic research, fund enthusiasts, and financial advisors or content creators who need weekly report material.", "categories": [ "finance", "search" ], "tags": [ "ETF", "主题投资", "资讯报告", "科创芯片" ], "tags_en": [ "ETF", "Thematic Investing", "News Report", "STAR Chip" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "科创芯片ETF主题资讯报告" } ], "prompt": "帮我整理科创芯片 ETF 的相关资讯，作为主题投资调研参考。请用平安证券资讯查询 skill 真实检索（如「科创芯片ETF 新闻」「科创芯片 指数 动态」「科创板 芯片 行情」等），把资讯按「ETF 产品动态」「指数与成份股」「行业基本面」等维度归类，按日期由新到旧排列，注明来源与日期。最终产出一份单文件 HTML 主题资讯报告：包含主题概览、分维度资讯列表、资讯脉络小结，页脚注明数据来源（平安证券资讯知识库）与「本内容仅为资讯整理，不构成投资建议」。视觉方向：研究报告质感、排版克制专业、长文阅读友好；所有资源内联、可离线打开。", "prompt_en": "Help me compile news related to the STAR Chip ETF as reference material for thematic investment research. Use the Ping An News Search skill to run real queries (e.g. \"STAR Chip ETF news\", \"STAR Chip index updates\", \"STAR Market chip stocks\"), group the news by dimensions such as \"ETF product updates\", \"index and constituents\", and \"industry fundamentals\", sort by date (newest first), and label each item with source and date. Produce a single-file HTML thematic news report: a theme overview, dimension-by-dimension news lists, a summary of the news narrative, and a footer stating the data source (Ping An Securities news knowledge base) and \"This content is for informational purposes only and does not constitute investment advice.\" Visual direction: research-report aesthetic, restrained and professional typography, friendly for long-form reading; all resources inlined so the file works offline.", "skills": [ { "id": "news-search", "name": "平安证券资讯查询", "description": "按股票、公司、行业、ETF、概念或市场主题检索相关新闻、快讯和财经报道。", "skill_id": "skill_2079451206686916608" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-24T17:45:34+08:00", "updated_at": "2026-07-27T12:49:23+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "eu-ai-act-disclosure", "title": "AI 助手出海欧盟合规披露研究", "title_en": "EU AI Act Disclosure Compliance for AI Assistants", "subtitle": "一键摸清 EU AI Act 对 AI 语音助手的终端用户披露义务", "subtitle_en": "One-click research on EU AI Act end-user disclosure obligations for AI voice assistants", "description": "做 AI 出海最头疼的不是技术，而是不知道欧盟到底要你披露什么。这个案例演示了如何把AI 语音助手向欧盟用户提供服务这一具体场景交给全球法律研究 Skill，自动检索 EU AI Act、GDPR 等欧盟官方法源，定位 Article 52 等关键条款原文，输出带风险评级和整改清单的合规研究报告。\n\n案例会真实调用 LegalDataHunter 检索欧盟数据源，逐条核验披露义务的触发条件、披露方式、豁免情形和违规后果，并给出可直接落地的产品改动建议。产物为单文件 HTML，无外部资源引用，可离线打开。\n\n适合：计划出海欧盟的 AI 产品法务/合规、AI 产品经理、跨境法律顾问、正在研究 EU AI Act 落地路径的技术团队。", "description_en": "The hardest part of AI global expansion isn't the tech — it's figuring out exactly what the EU requires you to disclose. This case shows how to hand a concrete scenario (an AI voice assistant serving EU users) to the Global Legal Research Skill, which automatically retrieves official EU sources such as the EU AI Act and GDPR, anchors key provisions like Article 52, and produces a compliance report with risk ratings and an actionable remediation checklist.\n\nThe case genuinely invokes LegalDataHunter to search EU data sources, verifies trigger conditions, disclosure methods, exemptions, and consequences one by one, and gives product-level remediation advice. The deliverable is a single-file HTML with no external references, viewable offline.\n\nGreat for: AI product legal/compliance teams planning EU expansion, AI product managers, cross-border legal counsel, and technical teams studying EU AI Act implementation.", "categories": [ "search", "business" ], "tags": [ "欧盟 AI Act", "AI 合规", "跨境出海", "法律研究", "GDPR" ], "tags_en": [ "EU AI Act", "AI compliance", "global expansion", "legal research", "GDPR" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "欧盟 AI 助手披露义务合规研究报告" } ], "prompt": "我方是一家中国科技公司，计划向欧盟用户提供一款 AI 语音助手产品，用户通过网页和 App 与 AI 进行自然语言对话。请使用全球法律研究技能，针对以下问题完成研究并输出单文件 HTML 报告：\n\n1. 欧盟 AI Act 是否要求我们在用户首次与 AI 交互时披露\"这是 AI\"？具体法律依据是什么？\n2. 披露义务的触发条件是什么？哪些 AI 系统可以豁免？\n3. 披露方式有哪些要求？需要在用户界面如何呈现？\n4. 该义务与 GDPR 透明度要求如何衔接？\n5. 违规可能面临什么后果？民事、行政、刑事三类责任分别是什么？\n6. 给出可直接落地的合规整改建议，按高/中/低优先级排列，包含具体动作、责任人和完成时限。\n\n要求：\n- 每个法域结论必须带可回链的法规原文或官方来源；\n- 引用至少包含 EU AI Act Article 52 等关键条款；\n- 输出为单文件 HTML，所有资源内联，无外部 CDN 或远程 URL，无 hash 锚点导航；\n- 报告首页显示 AI 辅助生成免责声明。", "prompt_en": "Our company is a Chinese technology firm planning to offer an AI voice assistant to EU users via web and mobile apps. Please use the Global Legal Research Skill to research the following and output a single-file HTML report:\n\n1. Does the EU AI Act require us to disclose to users that they are interacting with AI on first interaction? What is the specific legal basis?\n2. What are the trigger conditions for the disclosure obligation? Which AI systems are exempt?\n3. What are the disclosure-method requirements? How should it be presented in the UI?\n4. How does this obligation interact with GDPR transparency requirements?\n5. What are the potential consequences of non-compliance, broken down by civil, administrative, and criminal liability?\n6. Provide actionable remediation recommendations prioritized as high/medium/low, including specific actions, owners, and deadlines.\n\nRequirements:\n- Every conclusion must cite traceable official legislation or sources;\n- Include key provisions such as EU AI Act Article 52;\n- Output a single-file HTML with all resources inlined, no external CDNs or remote URLs, and no hash-anchor navigation;\n- Display an AI-assisted disclaimer at the top of the report.", "skills": [], "experts": [ { "id": "data-privacy-ai-compliance-expert", "name": "数据隐私与AI合规专家", "description": "面向数据合规律师与DPO，覆盖个人信息合规、PIA评估、数据出境与AI治理，内嵌执业安全红线。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "882dfd2c5f4a", "expert_id": "ex_ww996Cmrx1T3" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "source": "community", "usage_count": 0, "created_at": "2026-09-04T17:28:21+08:00", "updated_at": "2026-09-04T17:28:21+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "event-factor-impact-brief", "title": "事件因子解读页", "title_en": "Event Factor Impact Brief", "subtitle": "把一条市场事件拆成因子、证据、产业链权重和可复核源头", "subtitle_en": "Turn one market event into factors, evidence, chain exposure, and reviewable sources.", "description": "看到一条新闻或公告，最难的是判断它到底影响公司哪块业务、影响大不大、哪些事实可以复核。\n\n这个案例演示如何用同舟股市投研专家，把事件拆成影响因子、客观数据、公司产业链位置、核心业务权重和可对照事件，并在页面中保留公告、新闻、同舟观点和数据口径的复核入口。有可跳转链接的新闻与公告进入源头卡片；行情、财务和图谱数据以同舟行情库、同舟财务指标库、同舟行业图谱标注。\n\n适合：投研分析师、投顾内容团队、金融内容运营、想快速看懂事件影响的普通投资者。", "description_en": "When a news item or announcement appears, the hard part is knowing which business line it affects, how material it is, and which facts are reviewable.\n\nThis case shows how Tongzhou Equity Research Expert turns an event into impact factors, objective data, value-chain position, core-business exposure, and comparable events, while keeping links and data-source notes visible in the page. News and announcements with URLs become source cards; market, financial, and graph data are attributed to Tongzhou market data, Tongzhou financial metrics, and Tongzhou industry graph.\n\nSuitable for: equity researchers, advisor content teams, financial content operators, and retail users who need a fast event-impact readout.", "categories": [ "finance", "data", "search" ], "tags": [ "事件解读", "因子归因", "原文索引", "HTML" ], "tags_en": [ "Event brief", "Factor attribution", "Source links", "HTML" ], "artifact_type": "html", "playbook_type": "event-factor-impact", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "事件因子解读页" } ], "prompt": "请调用同舟股市投研专家，围绕「{事件与公司，例如：动力电池需求改善对宁德时代的影响}」生成一页单文件 HTML 事件因子解读。\n\n先完成同舟 MCP Gateway 认证；如果没有 API Key 或认证失败，只提示用户去控制台申请/绑定，不要继续编造内容。取证范围包括公开新闻、公告/PDF、行情与财务数据、行业图谱、同舟观点、公开因子框架和历史相似事件统计。估值因子必须用 fin-data 补数：个股事件先解析股票代码，再用行情日频指标取 PE(TTM)、PE、PB、市值；行业事件先解析稳定申万行业名，再用行业估值工具取 PE/PB 和历史分位。页面必须完全按产品原型的事件解读长报告形式：浅灰背景、白色圆角卡片、纯文字页面标题、顶部操作区、胶囊式 Tab 导航和 scrollIntoView，不使用 href=\"#section-id\"；顶部“返回总览 / 导出 PDF / 生成工作底稿”必须是真实按钮：返回总览滚动到首屏，导出 PDF 调用 window.print()，生成工作底稿下载一份 Markdown 底稿，不要做成不可点击装饰。PC 宽度源 HTML 是主版式，首屏采用左侧结论叙事、右侧报告日期/结论等级卡、底部四列摘要指标条的结构。首屏先展示示例公司、行业、事件类型、报告日期和数据截止，再给普通用户能听懂的一句话结论；720px 封面只是响应式预览，不要用封面宽度反推主页面。正文固定顺序为：事件影响因子与客观数据、因子归因如何推导结论、公司产业链地位与核心业务权重、历史相似事件、相似事件后的股价平均走势、源头复核区、证据链/反方证据/后续验证。影响因子卡片必须按源 HTML 使用“标题+方向徽标、两行客观指标、底部历史分位/进度条”的行式结构，不要做三格 KPI 小盒子。因子归因区在 PC 宽度必须是左侧表格 + 右侧归因合成结果卡的两栏布局，右侧包含总分、贡献摘要和围绕零点的正负贡献条，不要让表格铺满后把分数卡放到下面。公司产业链地位与核心业务权重在 PC 宽度必须是左侧四节点 chain-map（公司位置节点高亮）+ 右侧 business-panel 业务权重条，不要把链路和权重拆成上下两段。历史相似事件必须先展示四个匹配条件卡，再用 event-list 横向行列表展示样本日期、标题、证据说明和统计状态徽标，不要做四张样本卡片网格。相似度按“需求/政策匹配 25%、产业链匹配 25%、有效回测窗口 25%、源头可复核 15%、负向/混合证据最多扣 10%”计算和说明，展示为“相似度 84%”这类用户可读标签；不要提后端字段是否返回。凡工具返回原文、公告、小程序或研报链接，都做“阅读原文/查看公告/复核数据”跳转；行情、财务和图谱这类没有文章级链接的数据，只标注同舟行情库、同舟财务指标库、同舟行业图谱等来源类型，不做假链接。若相似事件统计数据或统计工具返回有效样本，必须展示真实样本数、均值、胜率和缺失窗口；若同时有公司事件回测和行业/指数回测，必须按“公司事件样本 vs 行业基准样本”对比展示，并给出公司相对行业超额；若某个窗口样本不足，只标注该窗口样本不足，不编造收益、胜率、样本数或回测结果。若已取到因子框架、指标名和源头链接但缺少当前实时数值，可以生成透明的“规则评分/规则贡献/暴露等级”，规则必须写在页面上：因子框架、指标面板、历史样本、源头复核、产业链暴露加分，反向证据扣分；不要反复显示接口缺口类文案。\n\n输出只要一个 HTML 文件，样式和必要交互内联；允许源头复核外链，但不要加载外部脚本、CSS、图片或字体；不展示 raw JSON、内部 ID、API Key、手机号、验证码、个人持仓或交易历史，不给买卖建议。", "prompt_en": "Use Tongzhou Equity Research Expert to create a single-file HTML event factor interpretation page for \"{event and company, e.g. improving power-battery demand and CATL}\".\n\nFirst complete the Tongzhou MCP Gateway credential check. If the API key is missing or invalid, only ask the user to apply or bind a key in the console and do not invent content. Retrieve public news, announcements/PDFs, market and financial data, industry graph evidence, Tongzhou viewpoints, public factor frameworks, and historical similar-event statistics. Valuation factors must be sourced from fin-data: for named-stock events, resolve the ticker and use daily market metrics for PE(TTM), PE, PB, and market cap; for industry events, resolve a stable Shenwan industry name before using sector valuation for PE/PB and historical percentiles. Match the product prototype exactly: pale gray background, white rounded cards, text-only page title, top workbench actions, capsule tabs, and button-based scrollIntoView navigation rather than href=\"#section-id\". The top actions must be real buttons: return-to-overview scrolls to the first viewport, export PDF calls window.print(), and generate working draft downloads a Markdown draft; do not render them as inactive decoration. Treat the PC-width source HTML as canonical: the first viewport uses a left conclusion narrative, a right report-date/conclusion-grade card, and a four-column summary strip below. Show metadata chips, report date/data cutoff, and a plain-language conclusion; the 720px cover is only a responsive preview and must not drive the primary page layout. Render sections in this order: impact factors with objective data, factor attribution table, company value-chain position and business exposure, historical similar events, post-event price path statistics, source review, and evidence/risk/verification. Factor cards must use the source HTML row-style structure: title plus direction badge, two objective metric rows, and a bottom percentile/progress line; do not use three boxed KPI tiles. At PC width, the attribution section must be a two-column layout with the table on the left and a composite attribution score panel on the right, including total score, contribution summary, and signed bars centered around zero. At PC width, company chain and business exposure must use a left four-node chain-map with the current company node highlighted plus a right business-panel with exposure progress bars; do not split these into stacked blocks. Historical similar events must render four matching-condition cards followed by a horizontal event-list of sample rows with date, title, evidence note, and statistics-status badge; do not use a sample-card grid. Compute and explain similarity from demand/policy match 25%, value-chain match 25%, valid backtest windows 25%, source-review availability 15%, and negative/mixed-evidence adjustment up to -10%. Show user-facing badges such as “Similarity 84%”; do not mention MCP, ES, backend fields, or whether a native field was returned. When the tools return original article, announcement, mini-program, or report links, render read-original/view-announcement/verify-data links. Market, financial, and graph data without article-level links should be attributed to Tongzhou market data, Tongzhou financial metrics, or Tongzhou industry graph, not turned into fake links. If comparable-event statistics are returned, show the real sample count, average return, win rate, and missing windows. When both company-event backtest and industry/index backtest are available, compare them as company-event samples vs industry benchmark samples and show excess return. If a window has insufficient samples, label only that window as insufficient; do not invent returns, win rates, sample sizes, or backtest results. If factor frameworks, metric names, and source links are available but complete current values are not, produce transparent rule-based factor scores, contribution scores, and exposure bands; state the rule on the page and avoid repeated backend-gap wording.\n\nOutput one HTML file only, with styling and necessary interactions inline. Source-review external links are allowed, but do not load external scripts, CSS, images, or fonts. Do not expose raw JSON, internal IDs, API keys, phone numbers, verification codes, personal holdings, or trade history. Do not give buy/sell advice.", "skills": [], "experts": [], "mcps": [ { "id": "public-mcp-gateway", "name": "同舟公开 MCP Gateway", "description": "受控暴露 fin-data-query、doc-search、fin-graph、same-boat 等公开投研能力。" } ], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-03T00:00:00+08:00", "updated_at": "2026-07-27T11:27:39+08:00", "source_links_supported": true, "is_recommended": false, "recommended_order": 0 }, { "id": "excalidraw-system-architecture", "title": "微服务系统架构图", "title_en": "Microservice System Architecture Diagram", "subtitle": "手绘风格的交互式微服务架构可视化", "subtitle_en": "Interactive microservice architecture visualization with hand-drawn style", "description": "需要一张清晰直观的系统架构图？本案例展示如何生成 Excalidraw 手绘风格的微服务架构图，包含 API 网关、认证服务、订单服务、消息队列等 8+ 组件，支持鼠标悬停高亮连接关系，按层级着色，配有图例说明。适合架构师、技术负责人和开发团队内部分享。", "description_en": "Need a clear and intuitive system architecture diagram? This case demonstrates how to generate Excalidraw-style hand-drawn microservice architecture diagrams with 8+ components, hover-to-highlight connections, layer-based coloring, and interactive legend.", "categories": [ "dev", "business" ], "tags": [ "架构图", "微服务", "系统设计", "Excalidraw" ], "tags_en": [ "Architecture Diagram", "Microservices", "System Design", "Excalidraw" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "微服务架构图页面" } ], "prompt": "请创建一个 Excalidraw 手绘风格的微服务系统架构图页面（单文件 HTML，内联无外部依赖），白底淡灰网格、手绘质感（用 SVG 绘制，粗糙边框/不规则线条）。\n\n包含 8+ 服务组件：API Gateway、Auth Service、User Service、Order Service、Payment Service、Notification Service、Message Queue (Kafka)、Database (PostgreSQL)，按层级着色（网关/业务/基础设施），服务间用箭头标注连接关系。\n交互：悬停某服务高亮其所有连接。含颜色图例。", "prompt_en": "Please create an Excalidraw hand-drawn-style microservices system architecture diagram page (single-file HTML, inlined with no external dependencies), with a white background and light-gray grid, and a hand-drawn texture (drawn with SVG, rough borders / irregular lines).\n\nInclude 8+ service components: API Gateway, Auth Service, User Service, Order Service, Payment Service, Notification Service, Message Queue (Kafka), Database (PostgreSQL), colored by tier (gateway / business / infrastructure), with arrows labeling the connections between services.\nInteraction: hovering over a service highlights all of its connections. Include a color legend.", "skills": [ { "id": "excalidraw-diagram", "name": "Excalidraw Diagram", "description": "Excalidraw 图解生成与渲染校验", "skill_id": "skill_2053082275194277888" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "excel-sales-dashboard", "title": "销售数据分析仪表盘", "title_en": "Sales Data Analysis Dashboard", "subtitle": "12 月趋势 × 区域排名 × 毛利分析 × KPI 达成率", "subtitle_en": "12-Month Trend × Regional Ranking × Margin Analysis × KPI Achievement", "description": "在快速变化的商业环境中，销售团队每天面对海量订单数据，却难以快速定位业绩短板、毛利异常和区域差异——手动汇总耗时且容易出错，管理层无法及时获取决策所需的关键指标。\n\n本仪表盘通过六大核心模块一站式解决销售数据管理难题：销售明细表记录全年12个月按产品线、区域、销售员维度的逐笔数据，月度趋势表自动计算环比增长率揭示业务波动，区域排名表量化各区域人效与总产出，产品线毛利分析精准定位高利润与低利润业务线，Top 10 客户贡献表识别核心大客户，KPI 仪表盘汇总收入达成率、同比增长、客均单价等关键经营指标。\n\n交付物为一份包含 6 个工作表的 Excel 文件（output.xlsx），涵盖：销售明细数据表、月度趋势汇总表、区域业绩排名表、产品线毛利分析表、Top 10 客户贡献表、KPI 达成率仪表盘。所有百分比与增长率均使用 Excel 公式自动计算，支持数据更新后即时刷新。", "description_en": "In a fast-changing business environment, sales teams face massive order data daily yet struggle to quickly pinpoint performance gaps, margin anomalies, and regional disparities—manual consolidation is time-consuming, error-prone, and delays critical decisions for management.\n\nThis dashboard solves sales data management challenges through six core modules: a Sales Detail sheet recording 12 months of transaction-level data by product line, region, and salesperson; a Monthly Trend sheet with automatic MoM growth formulas revealing business fluctuations; a Regional Ranking sheet quantifying per-capita output and total production by area; a Product Line Margin Analysis pinpointing high- and low-profit business lines; a Top 10 Customer Contribution table identifying key accounts; and a KPI Dashboard summarizing revenue achievement, YoY growth, average deal size, and other critical business metrics.\n\nThe deliverable is a single Excel file (output.xlsx) containing 6 worksheets: Sales Detail, Monthly Trend, Regional Ranking, Product Line Margin Analysis, Top 10 Customers, and KPI Dashboard. All percentages and growth rates use Excel formulas for automatic recalculation when data is updated.", "categories": [ "data", "business" ], "tags": [ "销售分析", "数据仪表盘", "Excel", "KPI", "业绩报表" ], "tags_en": [ "sales analysis", "data dashboard", "Excel", "KPI", "performance report" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "销售数据分析仪表盘" } ], "prompt": "帮我生成一份销售数据分析 Excel 仪表盘。包含：12 个月的销售数据（按产品线/区域/销售员）、月度趋势图数据、区域业绩排名、产品线毛利分析、Top 10 客户贡献、环比/同比增长率计算公式、KPI 达成率仪表盘数据。", "prompt_en": "Help me generate a sales data analysis Excel dashboard. Include: 12 months of sales data (by product line / region / salesperson), monthly trend chart data, regional performance ranking, product line margin analysis, Top 10 customer contribution, MoM/YoY growth rate calculation formulas, and KPI achievement rate dashboard data.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22", "updated_at": "2026-06-16T18:10:00Z", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "exec-admin-workbench", "title": "行政综合管理工作台", "title_en": "Executive Admin Workbench", "subtitle": "行政人的一站式工作台：12 大模块覆盖从数据看板到知识库的全部日常", "subtitle_en": "One-stop workbench for admins: 12 modules from data dashboards to knowledge base", "description": "行政工作琐碎又分散：数据要汇总、公文要起草、预算要盯、会议要统筹、合同要跟进、物资要记台账，工具散落在各处，来回切换效率低。\n\n这个案例把行政综合管理整合进一个单页工作台，12 大模块各管一摊：数据分析看板汇总收支与预算执行；制度公文撰写内置模板库与公文导出；财务分析辅助拆解预算与费用结构；商务接待管理从审批到接待标准全流程记录；物资出入库台账含出入库流水与库存警戒；人力资源 SOP 覆盖招聘、入转调离；智能会议纪要支持录音与要点沉淀；会议活动统筹排期与物料清单；合同合规提示带临期预警与台账；任务督办看板按责任单元跟踪进度；专属知识库沉淀制度、话术与经验，支持全文检索。另支持文档导入导出、打印存档、修改个人头像等细节。\n\n适合：企业行政、综合管理岗，以及想把行政日常工作全部装进一个页面的团队负责人参考。", "description_en": "Admin work is fragmented: data to consolidate, documents to draft, budgets to track, meetings to coordinate, contracts to follow up, supplies to log — with tools scattered everywhere.\n\nThis case pulls it all into a single-page workbench with 12 modules: a data dashboard consolidating income, spend, and budget execution; document writing with a template library and Word export; financial analysis breaking down budgets and expenses; business reception management from approval to hosting standards; a supplies in/out ledger with stock alerts; HR SOPs covering hiring and transfers; smart meeting minutes with voice notes; meeting and event scheduling with checklists; a contract ledger with expiry warnings; a task supervision board tracking progress by owner; and a searchable knowledge base for policies and know-how. Documents can be exported, printed, and the avatar customized.\n\nBest for: corporate admin and general-affairs roles, and team leads who want every routine admin task in one page.", "categories": [ "workspace" ], "tags": [ "行政管理", "工作台", "督办看板", "合同台账", "知识库" ], "tags_en": [ "Admin Management", "Workbench", "Task Board", "Contract Ledger", "Knowledge Base" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "行政综合管理工作台（中英双语单文件 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/9410HNQy1yehAvJaNXG7Ov，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「行政综合管理工作台」单页工作台，面向企业行政与综合管理岗，左侧 12 个模块导航：① 数据分析看板（收支汇总、预算执行率、趋势图表）；② 制度公文撰写（模板库、编辑器、导出）；③ 财务分析辅助（预算项、费用结构、月度对比）；④ 商务接待管理（接待记录、审批与标准提示）；⑤ 职场技能提升（学习记录与打卡）；⑥ 物资出入库台账（物品清单、出入库流水、库存警戒）；⑦ 人力资源 SOP（招聘、入转调离流程节点）；⑧ 智能会议纪要（纪要列表、录音、要点）；⑨ 会议活动统筹（活动排期、物料清单）；⑩ 合同合规提示（合同台账、临期预警、合规提示卡）；⑪ 任务督办看板（按责任单元跟踪、进度百分比）；⑫ 专属知识库（制度、话术、经验条目，全文检索）。顶栏支持修改头像；各模块支持新增、编辑、删除与 Excel 导出。视觉：左侧深色导航 + 右侧浅色内容区，卡片化布局，PC 优先、移动端可折叠导航。数据存储：在资料库中创建数据表存放各模块业务数据（督办任务、年度预算、支出流水、物资台账、出入库流水、会议纪要、活动统筹、合同台账、商务接待、周边餐饮、公文文档、知识库、技能打卡等），种子 CSV 已挂在本页面节点附件里（种子-*.csv 共 15 份，可直接导入同名表），页面通过页面数据能力在线读写表，改动实时保存到表、跨设备可见；表字段与页面字段一一对应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 12 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/9410HNQy1yehAvJaNXG7Ov with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build an \"Executive Admin Workbench\" single-page workbench for corporate admin and general-affairs roles, with 12 modules in a left nav: ① Data dashboard (income/spend summary, budget execution, trend charts); ② Document writing (template library, editor, export); ③ Financial analysis (budget items, expense structure, monthly comparison); ④ Business reception management (reception log, approvals and standards); ⑤ Skills development (learning log and check-ins); ⑥ Supplies in/out ledger (item list, in/out flows, stock alerts); ⑦ HR SOPs (hiring, onboarding/transfer process nodes); ⑧ Smart meeting minutes (minutes list, voice notes, key points); ⑨ Meeting and event coordination (schedules, material checklists); ⑩ Contract compliance alerts (contract ledger, expiry warnings, compliance tips); ⑪ Task supervision board (progress tracking by owner, percent complete); ⑫ Knowledge base (policies, scripts, know-how, full-text search). Top bar supports avatar customization; every module supports create, edit, delete, and Excel export. Visual: dark left nav with light content area, card-based layout, PC-first with collapsible nav on mobile. Data storage: create Library databases to hold each module's business data (task board, annual budget, expense ledger, supplies ledger, in/out flows, meeting minutes, event coordination, contracts, receptions, nearby restaurants, documents, knowledge base, skills log, etc.); seed CSV files are attached to this page node (15 files named \"种子-*.csv\", importable into the matching tables), and the page read and write the tables online through the page-data capability, so every change is saved to the tables and visible across devices; table fields map one-to-one to page fields.\n\nDeliverable: an online Library workbench — an online HTML page plus 12 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "Amber琥珀", "name_en": "Amber Hupo", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "community", "usage_count": 0, "created_at": "2026-09-10T13:50:00+08:00", "updated_at": "2026-09-10T13:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "execution-materials-property-clues", "title": "生效裁判后执行材料与财产线索整理", "title_en": "Post-Judgment Execution Materials and Property Clue Preparation", "subtitle": "为申请执行人梳理执行材料清单、金额结构与财产线索三层分级", "subtitle_en": "Prepare execution materials, amount structure and three-tier property clues for judgment creditors", "description": "判决赢了，钱却迟迟拿不回来——最头疼的不是打不赢官司，而是不知道申请执行要备齐哪些材料、迟延履行利息怎么算、手头零散的账户和应收线索该怎么分级提交。\n\n这个案例演示腾讯电子签资深合同法务专家如何先核验申请执行条件、执行管辖与迟延履行计算，再输出申请执行材料清单、执行请求金额结构、财产线索「已掌握 / 待依法核查 / 不宜据此推断」三层分级表与优先行动清单，并明确不承诺查控、冻结、拍卖或回款结果。\n\n适合：持有生效判决或裁定的申请执行人、企业法务、财务催收人员，以及代理执行案件的律师与实习律师。", "description_en": "You won the judgment but still can't get paid. The headache isn't winning the case, it's figuring out which execution materials to prepare, how to calculate late-performance interest, and how to tier the scattered bank-account and receivable clues you already hold.\n\nThis case shows how Tencent eSign's senior contract legal expert verifies execution conditions, jurisdiction and late-performance calculation first, then produces an execution-material checklist, an execution-amount structure, a three-tier property-clue table (known / to-be-verified / not-inferable) and a priority action list — while explicitly promising no seizure, freezing, auction or recovery results.\n\nBest for: judgment creditors holding an effective judgment or ruling, in-house legal counsel, finance and collections staff, and lawyers or trainee lawyers handling enforcement cases.", "categories": [ "business", "search" ], "tags": [ "申请执行", "执行材料", "财产线索", "迟延履行利息", "强制执行" ], "tags_en": [ "Enforcement Application", "Execution Materials", "Property Clues", "Late-performance Interest", "Compulsory Execution" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "execution-materials-property-clues.html", "preview_en": "", "external_url_en": "", "artifacts": [ { "file": "execution-materials-property-clues.html", "label": "互动执行准备灵感卡片" } ], "prompt": "我已取得生效民事判决，对方应支付货款及迟延履行相关款项但尚未履行。我持有判决书、生效证明线索、送达材料、对方企业基本信息、部分银行账户和应收账款线索。请按申请执行条件、申请材料、执行请求金额计算、被执行人主体与送达、财产线索分类、风险与补充动作建立执行准备清单；对申请期限、执行法院要求、迟延履行计算及材料格式先核验，未核验处标注\"待官方来源核验\"。请区分\"已掌握线索\"和\"待依法核查线索\"，不要承诺实际查控、冻结、拍卖或回款结果。", "prompt_en": "I have obtained an effective civil judgment. The other party was ordered to pay the goods price and late-performance related amounts but has not yet performed. I hold the judgment, clues to the effective-proof document, service materials, the counterparty's basic enterprise information, and partial bank-account and receivable clues. Please build an execution preparation checklist covering execution conditions, application materials, calculation of the execution request amount, the judgment debtor's identity and service, property-clue classification, risks and supplementary actions. First verify the application period, the execution court's requirements, late-performance calculation and material formats, and mark unverified items as \"to be verified against official sources\". Distinguish between \"clues already in hand\" and \"clues to be legally verified\", and do not promise actual seizure, freezing, auction or recovery results.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家，擅长起草、签署、审查、对比及法律咨询。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d1e4a1d28e92", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T18:53:40+08:00", "updated_at": "2026-08-13T19:36:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "executive-noncompete-review", "title": "高管竞业限制协议专项审查", "title_en": "Executive Non-Compete Clause Review", "subtitle": "签了竞业，跳槽却被 20% 补偿金卡住——拆解一份藏在细节里的无效陷阱协议", "subtitle_en": "A non-compete paying only 20% compensation hides invalidation traps in the fine print", "description": "技术总监离职前收到一份竞业限制协议，补偿金却只给到月工资的 20%，签不签都心里打鼓——这种藏在细节里的\"无效陷阱\"最让人头疼。\n\n这个案例演示了如何把一份真实的《保密与竞业限制协议》交给 AI 做专项审查：从甲方、乙方两个视角逐条拆解，精准命中\"补偿金 20% 低于法定 30%\"\"地域全国但实际只在华南经营\"\"保密范围未排除公知信息\"三大风险点，输出合规性检查清单和分级修改建议。\n\n适合：劳动法律师、企业 HR／法务、准备跳槽的高管与技术骨干、需要审查竞业协议的个人。", "description_en": "A tech director gets a non-compete agreement before quitting, but the compensation is only 20% of monthly salary — sign it or not, either way feels like a trap. These hidden \"invalidation traps\" in the fine print are the most annoying part.\n\nThis case shows how to hand a real Non-Disclosure & Non-Compete Agreement to AI for a targeted review: it dissects clauses from both the employer and employee perspectives, pinpoints the three key risks (20% compensation below the statutory 30%, nationwide scope vs. South-China-only operations, and confidentiality scope that fails to exclude public information), then produces a compliance checklist and tiered revision suggestions.\n\nFor: labor lawyers, corporate HR/legal teams, executives and tech leads about to change jobs, and anyone reviewing a non-compete agreement.", "categories": [ "productivity" ], "tags": [ "竞业限制", "保密条款", "合同审查", "劳动法", "高管跳槽" ], "tags_en": [ "Non-Compete", "Confidentiality", "Contract Review", "Labor Law", "Executive" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "保密与竞业限制条款专项审查报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请以「双方视角」专项审查下面这份劳动合同中的保密条款和竞业限制条款，输出详细审查报告。\n\n【合同信息】\n- 合同类型：劳动合同\n- 甲方角色：用人单位（互联网公司，实际经营以华南为主）\n- 乙方角色：劳动者（高级技术总监，属法定\"高级技术人员\"）\n- 乙方月工资：80000 元\n- 行业：互联网/软件\n- 地区：深圳市\n\n【保密条款】\n乙方在职期间及离职后，对甲方的商业秘密负有保密义务。商业秘密包括但不限于技术信息、经营信息、客户信息、财务信息等。保密期限为在职期间及离职后5年。\n但不包括：（1）已进入公有领域的公知信息；（2）乙方离职后通过合法公开渠道独立获得的信息；（3）乙方在本协议签署前已合法拥有或独立开发、且不属于甲方职务成果的技术与知识；（4）乙方个人的一般性知识、技能与经验。\n\n【竞业限制条款】\n乙方离职后两年内，不得在甲方实际开展经营并构成实质竞争关系的区域（以本协议附件所列地域/城市清单为准）内，与甲方生产或经营同类产品、从事同类业务的企业任职，不得自营或与他人合作经营与甲方同类的业务。\n甲方应在乙方离职时，以书面形式向乙方提供与甲方存在竞争关系的企业清单；乙方仅对该清单所列企业及与其经营范围实质相同企业负有竞业限制义务。\n竞业限制补偿金按乙方离职前十二个月平均工资的30%按月支付（现约定月工资80,000元对应为24,000元/月），且不得低于当地最低工资标准；支付期限与竞业限制期限一致。\n甲方未按约定按月足额支付竞业限制补偿金累计满三个月的，乙方有权书面通知甲方解除竞业限制义务。\n本保密条款约定的保密义务期限（离职后5年）与竞业限制期限（离职后2年）相互独立，竞业限制期限届满后，乙方仍应继续履行保密义务。\n\n请分别从甲方视角和乙方视角识别风险点，重点审查：竞业限制补偿金是否达到法定30%标准且不低于当地最低工资标准、竞业限制对象与地域范围是否合理、竞业限制期限是否超过2年法定上限、保密范围是否排除公知信息与原有知识，并给出合规性检查清单、可执行性评级和分级修改建议（必须修改/建议修改）。\n\n【产物格式（交付版）】\n请将审查报告生成为可交付的正式文件，而非 Markdown 草稿。默认输出为单文件 HTML（内联 CSS，中文专业排版，含风险等级色块、合规性检查表格、双视角卡片、可打印/导出 PDF 按钮，无需外部资源引用，内容默认可见）。如委托人另有要求，可改为输出 DOCX（Word 文档，便于二次编辑归档）或 PDF（固定版式，防篡改，适合直接送达当事人）。无论何种格式，均须在文末附免责声明。", "prompt_en": "Please perform a targeted review of the confidentiality and non-competition clauses in the employment contract below from both perspectives, and output a detailed review report.\n\n[Contract info]\n- Contract type: employment contract\n- Party A role: employer (an internet company operating mainly in South China)\n- Party B role: employee (senior technical director, a statutory \"senior technical personnel\")\n- Party B monthly salary: CNY 80,000\n- Industry: internet/software\n- Region: Shenzhen\n\n[Confidentiality clause]\nParty B shall keep Party A's trade secrets confidential during employment and after departure. Trade secrets include but are not limited to technical information, business information, customer information and financial information. The confidentiality period is during employment and 5 years after departure.\nHowever, the above shall not include: (1) public information already in the public domain; (2) information independently obtained by Party B through lawful public channels after departure; (3) technology and knowledge lawfully owned or independently developed by Party B before signing this agreement and not constituting Party A's work product; and (4) Party B's general knowledge, skills and experience.\n\n[Non-competition clause]\nWithin 2 years after departure, Party B shall not work for, nor run or co-run a business similar to Party A's, any enterprise that produces or operates the same kind of products or engages in the same kind of business as Party A, within the regions where Party A actually operates and competes substantially (as listed in the region/city list attached to this agreement).\nParty A shall, upon Party B's departure, provide Party B in writing with a list of enterprises competitive with Party A; Party B shall bear non-competition obligations only toward the enterprises on such list and those with substantively identical business scope.\nNon-competition compensation is paid monthly at 30% of Party B's average monthly salary over the 12 months before departure (i.e., CNY 24,000/month given the current monthly salary of CNY 80,000), and shall not be lower than the local minimum wage; the payment period shall align with the non-competition period.\nIf Party A fails to pay the non-competition compensation in full on a monthly basis for a cumulative total of three months, Party B may notify Party A in writing to terminate the non-competition obligations.\nThe confidentiality obligation period under this clause (5 years after departure) and the non-competition period (2 years after departure) are independent of each other; after the non-competition period expires, Party B shall continue to perform the confidentiality obligations.\n\nPlease identify risks from both Party A's and Party B's perspectives, focusing on: whether the non-competition compensation meets the statutory 30% threshold and is not lower than the local minimum wage, whether the non-competition scope of personnel and territory is reasonable, whether the non-competition period exceeds the 2-year statutory cap, and whether the confidentiality scope excludes public information and prior knowledge. Provide a compliance checklist, an enforceability rating, and tiered revision suggestions (must-fix / suggested).\n\n[Output format (deliverable version)]\nGenerate the review report as a formal deliverable file rather than a Markdown draft. By default, output a single-file HTML (inline CSS, professional Chinese typesetting, with risk-level color blocks, a compliance checklist table, dual-perspective cards, and a print/export-to-PDF button; no external resource references; content visible by default). If the client requests otherwise, you may instead output DOCX (Word document, convenient for further editing and archiving) or PDF (fixed layout, tamper-resistant, suitable for direct delivery to the client). Regardless of format, append a disclaimer at the end.", "skills": [], "experts": [ { "id": "employment-team-lead", "name": "劳动人事专家团", "description": "覆盖劳动争议诊断、用工合规体检、补偿赔偿测算、劳动文书起草与成果核验的中国大陆劳动用工法律专家团队", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "9d37f920fb57", "expert_id": "ex_WXrWVRlyLIGp" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "community", "usage_count": 0, "created_at": "2026-09-07T09:46:00+08:00", "updated_at": "2026-09-09T17:45:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "face-recognition-data-compliance-evaluation", "title": "人脸识别数据合规评估", "title_en": "Facial Recognition Data Compliance Assessment", "subtitle": "围绕十万人脸信息规模，生成备案、影响评估与整改路线图", "subtitle_en": "Turn a 100,000-face-data threshold into a filing, PIPIA and remediation roadmap", "description": "人脸信息达到十万条后，最容易被忽略的不是技术上线，而是专项备案时限、单独同意和云端传输边界。这个案例把这些高敏感风险收敛为一张可执行的治理清单。\n\n案例演示腾讯电子签资深合同法务专家如何检索人脸识别专项规则、个人信息保护法和网络数据安全管理条例，并输出风险分级、备案五类材料、PIPIA 要点、技术控制与 90 天推进计划。\n\n适合：处理人脸/生物识别信息的工具产品负责人、数据合规负责人、法务、安全与隐私工程团队。", "description_en": "Once a facial-information store reaches 100,000 records, the most easily missed issues are filing deadlines, separate consent and the boundary of cloud transmission. This case turns those high-sensitivity risks into an executable governance checklist.\n\nIt demonstrates how Tencent eSign's senior contract legal expert retrieves facial-recognition rules, the Personal Information Protection Law and the Regulation on Network Data Security Management, then produces risk ratings, five filing materials, PIPIA essentials, technical controls and a 90-day plan.\n\nBest for: product owners handling facial or biometric data, data-compliance owners, legal, security and privacy engineering teams.", "categories": [ "business", "search" ], "tags": [ "人脸识别", "数据合规", "备案", "个人信息保护", "影响评估" ], "tags_en": [ "Facial Recognition", "Data Compliance", "Filing", "Personal Information Protection", "Impact Assessment" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "face-recognition-data-compliance-evaluation.html", "preview_en": "", "external_url_en": "", "artifacts": [ { "file": "face-recognition-data-compliance-evaluation.html", "label": "互动人脸识别数据合规评估卡片" }, { "file": "face-recognition-data-compliance-report.html", "label": "完整人脸识别数据合规评估报告" } ], "prompt": "假设产品是一个普通的工具产品，不具备关键信息基础设施特殊主体身份/资质/属性。当前产品存储人脸信息已经达到10万人，请分析产品面临的数据合规风险，并给出后续产品需要履行的合规义务、应采取的数据合规措施以及对应可落地的方案。请以法规检索为基础，重点说明人脸识别技术应用备案时限、个人信息保护影响评估、单独同意与告知、非人脸替代验证、存储与互联网传输边界、第三方委托和事件响应，并以风险分级、责任人、交付物和30/90天路线图呈现。", "prompt_en": "Assume the product is an ordinary tool product and does not have the special status, qualification or attributes of a critical information infrastructure operator. It currently stores facial information for 100,000 people. Analyze the data-compliance risks it faces, the compliance obligations it must perform, the data-compliance measures it should take and implementable solutions. Base the answer on legal retrieval, with particular focus on the facial-recognition application filing timeline, personal information protection impact assessment, separate consent and notice, non-facial alternative verification, storage and internet-transmission boundaries, third-party entrustment and incident response. Present the result with risk ratings, owners, deliverables and a 30/90-day roadmap.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家，擅长起草、签署、审查、对比及法律咨询。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9cb476db05eb", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T23:02:39+08:00", "updated_at": "2026-08-13T11:41:33+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "family-health-hub", "title": "家庭健康工作台", "title_en": "Family Health Hub", "subtitle": "宝妈专属：菜谱管理 + 生活记账 + 经期记录一站式管理", "subtitle_en": "For Moms: Recipe Management + Expense Tracking + Period Recording All-in-One", "description": "忙碌的宝妈每天要操心全家吃什么、钱花在哪、自己身体状况如何，信息散落各处容易遗漏。本案例演示一个零依赖的单页工作台，集成菜谱管理（分类收藏+今日推荐）、生活记账（月预算进度+饼图统计+CSV导出）、经期记录（智能周期预测+症状标记）三大模块，顶部「今天要处理」自动汇总预算预警和经期提醒。适合需要轻量家庭管理工具的宝妈、全职妈妈和注重健康记录的女性用户。", "description_en": "Busy moms juggle daily meal planning, household budgeting, and personal health tracking across scattered apps. This case demonstrates a zero-dependency single-page workspace integrating recipe management (categorized collection + daily recommendation), expense tracking (monthly budget progress + pie chart + CSV export), and period recording (smart cycle prediction + symptom tagging). A pinned \"Today's Tasks\" section auto-surfaces budget alerts and period reminders. Ideal for stay-at-home moms, working mothers, and women who value lightweight health management tools.", "categories": [ "workspace", "office", "data" ], "tags": [ "宝妈", "家庭管理", "记账", "经期记录", "菜谱", "工作台" ], "tags_en": [ "mom", "family management", "expense tracking", "period recording", "recipe", "workspace" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "家庭健康工作台" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/uoegy7O9vtSg8gSAdfh8Kn），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个家庭健康工作台，主要是宝妈使用。要求如下：\n\n1. 菜谱管理模块：支持添加菜式（早餐/午餐/晚餐/加餐分类），记录食材，有今日推荐功能，可按分类筛选浏览\n2. 生活记账模块：收支录入（餐饮/交通/购物/母婴/医疗/教育/居住等分类），月预算设置和进度条，支出分类饼图统计，支持导出 CSV 账单\n3. 经期记录模块：记录开始和结束日期，支持多种症状标记（腹痛/头痛/腰酸/疲惫/情绪波动/量多/量少），自动计算平均周期并预测下次经期\n\n页面最上面放「今天要处理」区域，自动汇总预算超支预警和经期临近提醒。\n\n界面：白底简洁，圆角卡片，粉色系主色调，温馨可爱。\n手机电脑都能用，手机单列、按钮≥44px、输入框字号≥16px。\n数据存 localStorage，刷新关闭不丢；首屏有「导出 JSON」「导入恢复」按钮，清空需二次确认。\n不引任何外部框架/CDN/字体/图表库，图表用内联 SVG。\n预置示例数据让用户首次打开即可看到效果。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/uoegy7O9vtSg8gSAdfh8Kn) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nHelp me build a family health workspace, mainly for moms. Requirements: 1. Recipe management module: Support adding dishes (breakfast/lunch/dinner/snack categories), record ingredients, daily recommendation feature, filterable by category\n2. Expense tracking module: Income/expense entry (food/transport/shopping/baby/medical/education/housing categories), monthly budget setting with progress bar, expense category pie chart, CSV bill export\n3. Period recording module: Record start and end dates, support multiple symptom tags (cramps/headache/backache/fatigue/mood swings/heavy flow/light flow), auto-calculate average cycle and predict next period Place a \"Today's Tasks\" section at the top, auto-surfacing budget overspend warnings and upcoming period reminders. Design: Clean white background, rounded cards, pink color scheme, warm and cozy.\nResponsive for both mobile and desktop, single column on mobile, buttons ≥44px, input font size ≥16px.\nData stored in localStorage, persists across refresh/close; export JSON and import restore buttons on first screen, clear requires confirmation.\nSingle inline HTML file, no external frameworks/CDN/fonts/chart libraries, charts use inline SVG.\nPreset sample data so users see the effect immediately on first open.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "created_at": "2026-08-01T06:54:16Z", "updated_at": "2026-08-01T10:03:19Z", "preview_en": "output.html", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "family-party-games", "title": "家庭聚会亲子游戏工作台", "title_en": "Family Party Games Workspace", "subtitle": "8款经典派对游戏一站式搞定，全家欢乐时光", "subtitle_en": "8 Classic Party Games in One Page — Fun for the Whole Family", "description": "家庭聚会想玩点游戏，打开手机搜了半天也找不到合适的——要么要下载APP，要么广告一堆。这个工作台把谁是卧底、24点、脑筋急转弯、真心话大冒险、你比我猜、数字炸弹、猜词达人、反应力大比拼 8 款经典派对游戏整合进一个页面，打开链接全家直接玩，游戏记录自动保存，还能导出备份。\n\n适合：家庭聚会、亲子活动、朋友派对、团建破冰、年会暖场。", "description_en": "Want to play games at a family gathering but can't find the right ones — apps need downloading, and web games are full of ads. This workspace packs 8 classic party games into a single page: Spy Game, 24-Point, Brain Teasers, Truth or Dare, Charades, Number Bomb, Word Guessing, and Reaction Test. Just open the link and everyone can play — game records auto-saved with export backup.\n\nPerfect for: family gatherings, parent-child activities, friend parties, team-building icebreakers, and holiday celebrations.", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/fcG1Dul0iuQVZBeS0UjLMB），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做个家庭聚会亲子游戏工作台，包含以下 8 个经典派对游戏：\n\n1. 谁是卧底：支持 4-10 人，自动分配平民词/卧底词，轮流查看词语，投票淘汰，支持自定义词库（至少预置 20 组词）\n2. 24 点：随机发 4 张牌（保证有解），玩家输入算式验证，有计时、提示和看答案功能\n3. 脑筋急转弯：预置 20 道趣味题目，点击揭晓答案，随机出题\n4. 真心话大冒险：各 15 条真心话和大冒险内容，支持随机抽取\n5. 你比我猜：5 大分类词库（动物/食物/动作/影视/物品），倒计时比划猜词，自动计分\n6. 数字炸弹：1-100 随机埋雷，逐步缩小范围，踩中爆炸\n7. 猜词达人：10 组词语各配 5 条递进线索，逐条揭示，线索越少越厉害\n8. 反应力大比拼：等待变绿后点击测反应时间，自动排行榜\n\n界面要求：白底简洁卡片风，主色活力橙。首页卡片网格选游戏，点击进入对应游戏面板。手机电脑都能用，按钮 ≥ 44px。数据存 localStorage，游戏记录自动保存，有导出/导入备份按钮。全部 CSS/JS 内联，不引用任何外部资源。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/fcG1Dul0iuQVZBeS0UjLMB) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a Family Party Games workspace with the following 8 classic party games: 1. Spy Game (Who's the Spy): 4-10 players, auto-assign civilian/spy words, take turns viewing words, vote to eliminate, customizable word bank (20+ preset pairs)\n2. 24-Point: Deal 4 random cards (guaranteed solvable), players type expressions to verify, with timer, hints, and answer reveal\n3. Brain Teasers: 20 fun riddles with tap-to-reveal answers, random order\n4. Truth or Dare: 15 truths and 15 dares, random draw with category selection\n5. Charades: 5 category word banks (animals/food/actions/movies/objects), countdown timer, auto-scoring\n6. Number Bomb: Random mine between 1-100, narrow the range with each guess, explode on hit\n7. Word Guessing Master: 10 words with 5 progressive clues each, reveal one by one\n8. Reaction Test: Wait-for-green click test with reaction time and auto leaderboard Design: clean white cards, vibrant orange accent. Home page card grid for game selection, tap to enter each game panel. Responsive for mobile and desktop, buttons ≥ 44px. Data in localStorage with auto-saved game history and export/import backup. with all CSS/JS inline, no external dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "artifact_type": "html", "preview": "family-party-games.html", "artifacts": [ { "file": "family-party-games.html", "label": "家庭聚会游戏工作台" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace", "education" ], "tags": [ "家庭聚会", "亲子游戏", "派对游戏", "工作台", "互动娱乐" ], "tags_en": [ "Family Party", "Kids Games", "Party Games", "Workspace", "Interactive Fun" ], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "created_at": "2026-08-01T07:35:51Z", "updated_at": "2026-08-01T08:23:26Z", "preview_en": "family-party-games.html", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "fin-ipo-readiness-dashboard", "title": "IPO 上市准备度与估值速览", "title_en": "IPO Readiness & Valuation One-Pager", "subtitle": "投行 pitchbook 速览：就绪度评分 + 足球场估值", "subtitle_en": "Banker-style pitchbook: readiness score + football-field valuation", "description": "拟 IPO 的公司要对内对外快速讲清楚上市准备到什么程度、值多少钱。本案例由「银拓远」投资银行专家做一张投行 pitchbook 风格的上市准备度与估值速览页——5 维就绪度评分、近三年三表摘要、可比公司法与 DCF 的足球场估值区间、募资用途与上市时间表。区别于完整 CIM 或单一估值模型，是一页路演速览。适合拟上市企业、投行与投资人。", "description_en": "A pre-IPO company needs to quickly convey, internally and externally, how ready it is to list and what it's worth. Built by the Investment Banking Expert, this case is a banker pitchbook-style IPO readiness & valuation one-pager — a 5-dimension readiness score, a three-year three-statement summary, a football-field valuation range from comparables and DCF, use of proceeds, and a listing timeline. Unlike a full CIM or a single valuation model, it's a one-page roadshow overview. Ideal for pre-IPO companies, bankers and investors.", "categories": [ "finance" ], "tags": [ "IPO", "投行", "估值", "金融服务" ], "tags_en": [ "IPO", "Investment Banking", "Valuation", "Finance" ], "artifact_type": "html", "prompt": "帮我做一张拟 IPO 科技公司的上市准备度与估值速览页（单文件 HTML，投行 pitchbook 风）。要直观看到：公司/板块/募资规模/估值区间/投资亮点、5 维上市准备度评分（雷达或条形）、近三年三表精简摘要、可比公司法与 DCF 的足球场估值区间图、募资用途拆解、上市时间表里程碑、关键风险。深海军蓝+金质投行风，界面中文，金额 ¥/亿元，请基于真实拟 IPO 或近期上市科技公司，联网搜索其招股书/财报的真实财务与估值数据（注明公司与来源）。", "prompt_en": "Help me build an IPO readiness & valuation one-pager for a pre-IPO tech company (single-file HTML, banker pitchbook style). Show the company/board/raise size/valuation range/investment highlights, a 5-dimension readiness score (radar or bars), a three-year three-statement summary, a football-field valuation range from comparables and DCF, use-of-proceeds breakdown, a listing-timeline milestone, and key risks. Deep navy + gold banker style; UI in Simplified Chinese; amounts in ¥/100M; data marked real (web-searched).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "IPO 上市准备度与估值速览" } ], "skills": [], "experts": [ { "id": "InvestmentBankingExpert", "name": "银拓远", "description": "全能投资银行专家：交易材料制作、估值建模（Comps/DCF/LBO/三表）、资本市场融资、买方尽调", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "08c8e424dd4c", "name_en": "Investment Banking Expert", "description_en": "Full-stack investment banking: deal materials, valuation (Comps/DCF/LBO/3-statement), capital markets, buy-side due diligence", "expert_id": "ex_7CuAtKRbZmdi" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-06-20T13:00:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fin-lbo-model-dashboard", "title": "杠杆收购 LBO 模型分析", "title_en": "LBO Model Analysis", "subtitle": "LBO 回报桥 + IRR/MOIC + 敏感性，一页看懂杠杆收购", "subtitle_en": "LBO value bridge + IRR/MOIC + sensitivity on one page", "description": "杠杆收购的回报到底从哪来、能赚多少倍，散户和初级投行人常常算不清。本案例由「建模模」金融建模专家做一张 LBO 模型分析页——资金来源与用途、去杠杆路径、价值创造桥（EBITDA 增长/倍数变化/去杠杆三部分贡献）、IRR/MOIC 三情景与敏感性表。区别于 DCF/可比公司估值，专讲杠杆收购回报。适合 PE、投行与想搞懂 LBO 的人。", "description_en": "Where do LBO returns come from and how many times your money do you make — questions retail investors and junior bankers often can't pin down. Built by the Financial Modeling Expert, this case is an LBO model analysis page — sources & uses, the deleveraging path, a value-creation bridge (EBITDA growth / multiple change / deleveraging contributions), IRR/MOIC across three scenarios, and a sensitivity table. Unlike DCF/comparables valuation, it's about leveraged-buyout returns. Ideal for PE, bankers, and anyone learning LBO.", "categories": [ "finance" ], "tags": [ "LBO", "私募股权", "回报分析", "金融服务" ], "tags_en": [ "LBO", "Private Equity", "Returns", "Finance" ], "artifact_type": "html", "prompt": "帮我做一张私募股权 LBO 杠杆收购模型分析页（单文件 HTML）。要直观看到：交易概览（EV/收购倍数/杠杆率/持有期/投资逻辑）、资金来源与用途对照、持有期去杠杆路径（债务下降+EBITDA 增长）、价值创造桥（EBITDA 增长/倍数变化/去杠杆三部分对股权价值的贡献）、IRR 与 MOIC 三情景、入场×退出倍数对 IRR 的敏感性表、关键假设与风险。石墨灰+钢蓝建模风，界面中文，金额 ¥/百万，标的公司基础财务请联网搜索真实数据（注明公司与来源），LBO 情景假设与测算部分可注明为模型假设。", "prompt_en": "Help me build a private-equity LBO model analysis page (single-file HTML). Show the deal overview (EV/purchase multiple/leverage/hold period/investment thesis), a sources & uses comparison, the deleveraging path over the hold period (debt paydown + EBITDA growth), a value-creation bridge (EBITDA growth / multiple change / deleveraging contributions to equity value), IRR and MOIC across three scenarios, an entry×exit multiple sensitivity table for IRR, and key assumptions and risks. A graphite + steel-blue modeling style; UI in Simplified Chinese; amounts in ¥/million; data marked an real base financials via web search (model assumptions may be noted as assumed) model.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "杠杆收购 LBO 模型分析" } ], "skills": [], "experts": [ { "id": "FinancialModelingExpert", "name": "建模模", "description": "核心金融建模工具：DCF 估值、可比公司分析、LBO 模型、三张表模型、竞争格局分析", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "da6bc44f5237", "name_en": "Financial Modeling Expert", "description_en": "Core financial modeling: DCF, comparable companies, LBO, three-statement model, competitive analysis", "expert_id": "ex_BSa0WZGMTzvZ" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "created_at": "2026-06-20T13:00:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fin-retirement-plan-simulator", "title": "养老退休规划方案", "title_en": "Retirement Planning Plan", "subtitle": "养老金缺口测算 + 资产积累 + 退休现金流", "subtitle_en": "Pension-gap calc + asset build-up + retirement cash flow", "description": "很多人到中年才发现没认真算过养老账，不知道退休后钱够不够花。本案例由「理财财」财富管理专家做一份养老退休规划方案——养老金缺口测算、40 到 60 岁的资产积累路径、三支柱构成、退休后逐年现金流，并给出可执行的补充建议。区别于追求提前退休的 FIRE，本案聚焦常规退休的养老金与现金流规划。适合做养老准备的家庭。", "description_en": "Many people reach midlife before realizing they've never seriously run the retirement math, unsure if their money will last. Built by the Wealth Management Expert, this case is a retirement planning plan — a pension-gap calculation, an asset accumulation path from age 40 to 60, the three-pillar mix, year-by-year retirement cash flow, and actionable top-up suggestions. Unlike FIRE (which chases early retirement), this focuses on conventional-retirement pension and cash-flow planning. Ideal for families preparing for retirement.", "categories": [ "finance" ], "tags": [ "养老规划", "退休", "财富管理", "金融服务" ], "tags_en": [ "Retirement", "Pension", "Wealth Management", "Finance" ], "artifact_type": "html", "prompt": "帮我做一份家庭养老退休规划方案页（单文件 HTML）。一对 40 岁夫妻计划 60 岁退休。要直观看到：家庭信息/目标退休年龄/退休后年支出/缺口结论、养老金缺口测算（需求 vs 三支柱覆盖）、40→60 岁资产积累曲线、三支柱构成、退休后 60→85 岁逐年现金流、3-4 条行动建议、关键假设与风险。温暖信赖的浅色财富顾问风，界面中文，货币 ¥，数据为测算示意。", "prompt_en": "Help me build a family retirement planning plan page (single-file HTML). A couple aged 40 plan to retire at 60. Show family info/target retirement age/post-retirement annual spend/gap conclusion, a pension-gap calculation (need vs three-pillar coverage), an asset accumulation curve from 40 to 60, the three-pillar mix, year-by-year retirement cash flow from 60 to 85, 3-4 action suggestions, and key assumptions and risks. A warm, trustworthy light wealth-advisor style; UI in Simplified Chinese; currency ¥; figures illustrative.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "养老退休规划方案" } ], "skills": [], "experts": [ { "id": "WealthManagementExpert", "name": "理财财", "description": "财富管理工具集：客户回顾、财务规划、投资组合再平衡、税损收割、投资提案和客户报告生成", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "06802d3ccfd4", "name_en": "Wealth Management Expert", "description_en": "Wealth management toolkit: client review, financial planning, portfolio rebalance, tax-loss harvesting, proposals, client reports", "expert_id": "ex_wMZ6zxKnk1uI" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-06-20T13:00:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fin-sector-rotation-monitor", "title": "行业轮动与资金流向监控", "title_en": "Sector Rotation & Capital Flow Monitor", "subtitle": "行业热力图 + 主力资金流，看清板块轮动", "subtitle_en": "Sector heatmap + capital flow to read rotation", "description": "散户盯个股容易只见树木不见森林，看不清资金在行业间怎么轮动。本案例由「严估深」股票研究专家做一个行业轮动与资金流向监控仪表盘——一级行业涨跌热力网格、主力资金净流入排行、风格轮动强弱、北向资金动向，并给出客观的行业配置研判。区别于个股分析，聚焦行业层面。适合关注板块轮动与资金主线的投资者。", "description_en": "Retail investors watching single stocks often miss the forest for the trees and can't see how money rotates across sectors. Built by the Equity Research Expert, this case is a sector rotation & capital flow monitor — a sector gain/loss heatmap grid, a main-capital net-inflow ranking, style-rotation strength, northbound flows, plus objective sector-allocation views. Unlike single-stock analysis, it focuses at the sector level. Ideal for investors tracking rotation and where the money flows.", "categories": [ "finance" ], "tags": [ "行业轮动", "资金流向", "A股", "金融服务" ], "tags_en": [ "Sector Rotation", "Capital Flow", "A-Share", "Finance" ], "artifact_type": "html", "prompt": "帮我做一个 A 股行业轮动与资金流向监控仪表盘（单文件 HTML）。要直观看到：大盘概览与今日资金主线、16 个一级行业涨跌热力网格（涨红跌绿，深浅表强度）、主力资金行业净流入排行（双向条形）、风格轮动强弱（大盘价值↔小盘成长等）、北向资金动向与趋势、3 条行业配置研判。深色金融终端风，A股惯例涨红跌绿，界面中文，请联网搜索获取 A 股各行业当日真实行情、资金流向与北向数据（注明数据日期与来源）。", "prompt_en": "Help me build an A-share sector rotation & capital-flow monitor dashboard (single-file HTML). Show a market overview and today's capital theme, a 16-sector gain/loss heatmap grid (up=red/down=green, intensity by shade), a main-capital net-inflow ranking (diverging bars), style-rotation strength (large-value↔small-growth, etc.), northbound flows and trend, and 3 sector-allocation views. Dark finance-terminal style; China convention up=red/down=green; UI in Simplified Chinese; data marked real (web-searched).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "行业轮动与资金流向监控" } ], "skills": [], "experts": [ { "id": "EquityResearchExpert", "name": "严估深", "description": "全面的股票研究工具集：财报分析、估值、行业研究、事件情景分析、组合风险管理", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b42e540ad26c", "name_en": "Equity Research Expert", "description_en": "Equity research toolkit: earnings analysis, valuation, sector research, event scenarios, portfolio risk", "expert_id": "ex_Pffykp6sw1dd" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-06-20T13:00:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-13f-smart-money-workbench", "title": "13F 聪明钱工作台", "title_en": "13F Smart Money Workbench", "subtitle": "一页追踪 20 家华尔街机构的季度持仓、共识买入与个股反查", "subtitle_en": "Track quarterly holdings, consensus buys, and stock-level ownership across 20 Wall Street institutions", "description": "SEC 13F 披露量大、申报格式分散，直接翻 EDGAR 很难快速回答：头部机构本季在加仓什么、哪些股票形成共识、某只股票到底被谁持有。\n\n这个案例把 SEC EDGAR 13F-HR 公开申报整理成一张可离线浏览的交互工作台，覆盖 20 家机构、4 个季度、25,386 条持仓记录。它提供本季速览、机构名单、共识买入排行、股票/CUSIP 反查、机构详情、简明/专业视图以及按市值或季度变动排序，并明确标注数据抓取时间、申报期和 45 天披露滞后。\n\n适合：研究美股机构持仓的个人投资者、需要跟踪聪明钱动向的投研人员、制作 13F 季度复盘的内容创作者，以及想把公开金融数据做成交互工作台的分析师。", "description_en": "SEC 13F filings are large and fragmented, making it difficult to quickly answer three practical questions in EDGAR: what leading institutions added this quarter, which stocks attracted broad consensus, and who owns a specific stock.\n\nThis case turns public SEC EDGAR 13F-HR filings into a self-contained interactive workbench covering 20 institutions, four quarters, and 25,386 holding records. It includes a quarterly overview, institution directory, consensus-buy rankings, stock/CUSIP lookup, institution drill-down, simplified and professional views, and sorting by market value or quarterly change. The page clearly states the retrieval time, reporting periods, and the roughly 45-day filing lag.\n\nBest for: individual investors researching U.S. institutional holdings, investment researchers tracking smart-money moves, creators producing quarterly 13F reviews, and analysts turning public financial data into an interactive workspace.", "categories": [ "workspace", "finance", "data", "dev" ], "tags": [ "SEC 13F", "机构持仓", "聪明钱", "美股研究", "交互工作台" ], "tags_en": [ "SEC 13F", "institutional holdings", "smart money", "U.S. equity research", "interactive workbench" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "13F 聪明钱交互工作台" } ], "external_url": "", "prompt": "请搭建一个名为「13F 聪明钱工作台」的单文件交互式 HTML。真实查询美国证监会 SEC EDGAR 的 13F-HR 公开申报，注明数据来源、抓取时间、报告期、申报日期和约 45 天披露滞后；不要使用示意或模拟持仓。覆盖不少于 20 家有代表性的机构和最近 4 个可得季度，按 CUSIP 对比环比变化，并在可可靠映射时展示股票代码。\n\n工作台包含四个核心入口：本季速览、机构名单、共识买入、股票/CUSIP 反查。本季速览展示总持仓市值、覆盖股票数、机构集中加仓与减仓结论；机构卡片展示代表人物、风格、持仓规模、前十大持仓与新进/增持/减持/清仓；共识页按共同持有机构数和本季变化排序；反查支持输入股票代码、公司名或 CUSIP，查看哪些机构持有及其季度变化。机构详情支持简明/专业视图，并可按市值或本季变动排序。\n\n使用浅色、克制、专业的金融研究风格，涨用红色、跌用绿色，桌面和手机均可用。所有 CSS、JavaScript、图标、字体和数据必须内联，不依赖外部资源；页面默认可见，离线打开仍能完整浏览和交互。页面底部加入数据口径、来源与“不构成投资建议”声明。", "prompt_en": "Build a self-contained interactive HTML named “13F Smart Money Workbench.” Query real public 13F-HR filings from the U.S. SEC EDGAR system and state the data source, retrieval time, reporting periods, filing dates, and the roughly 45-day disclosure lag. Do not use illustrative or simulated holdings. Cover at least 20 representative institutions and the latest four available quarters, compare quarter-over-quarter changes by CUSIP, and show ticker symbols where they can be mapped reliably.\n\nProvide four core entry points: Quarterly Overview, Institution Directory, Consensus Buys, and Stock/CUSIP Lookup. The overview should summarize aggregate market value, covered securities, and concentrated additions or reductions. Institution cards should show the representative investor, strategy style, portfolio size, top ten holdings, and new/increased/reduced/exited positions. The consensus view should rank securities by the number of institutions holding them and by quarterly change. Lookup should accept a ticker, company name, or CUSIP and show which institutions hold it and how positions changed. Institution detail should offer simplified and professional views with sorting by market value or quarterly change.\n\nUse a restrained, professional light financial-research aesthetic, with red for gains and green for declines, and support both desktop and mobile. Inline all CSS, JavaScript, icons, fonts, and data with no external resources; keep content visible by default and fully functional offline. Add data definitions, sources, and a not-investment-advice disclaimer in the footer.", "skills": [], "experts": [ { "id": "workspace-builder", "name": "小台", "name_en": "DeskCraft", "description": "为不同人群定制专属数字工作台，覆盖响应式 HTML、本地数据持久化与一键部署。", "description_en": "Builds custom digital workspaces with responsive HTML, local data persistence, and one-click deployment.", "expertType": "agent", "source": "enterprise", "promptFileSnapshot": "d088a1ff490e", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-27T03:12:35Z", "updated_at": "2026-08-27T03:12:35Z", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-2026-midyear-report-analysis", "title": "2026 A股半年报分析台", "title_en": "2026 A-Share Mid-Year Report Analyzer", "subtitle": "总榜 · 产业链赚钱榜 · 双增公司 · 预告观察，中报季一页掌握", "subtitle_en": "Master ranking, industry-chain profit board, double-growth list and preannouncement watch — mid-report season on one page", "description": "中报季几千家公司陆续披露，想比出谁真的好，要在公告、行情、行业数据之间反复横跳。本案例组合「NeoData 金融搜索服务」与「WeStock Data」，生成一个 2026 A股半年报分析台：正式半年报总榜按所选维度排行，产业链环节赚钱榜看哪段最赚钱，双增公司一眼锁定营收净利双升的标的，还有业绩预告/待披露观察区和宏观行业对照。适合中报季做行业比较和选股初筛的投资者。", "description_en": "During interim-report season, thousands of companies disclose in waves — comparing them means bouncing between announcements, quotes and industry data. Combining NeoData Financial Search with WeStock Data, this case builds a 2026 A-share mid-year report analyzer: a formal-report master ranking sorted by your chosen metrics, an industry-chain profit board showing which segment earns most, a double-growth list flagging companies with rising revenue and profit, plus a preannouncement/pending-disclosure watchlist and macro-industry comparison. Built for investors doing sector comparison and first-pass stock screening in interim season.", "categories": [ "finance" ], "tags": [ "半年报", "中报季", "产业链", "业绩预告", "选股筛选" ], "tags_en": [ "Interim Reports", "Earnings Season", "Industry Chain", "Preannouncements", "Stock Screening" ], "artifact_type": "html", "prompt": "帮我做一个「2026 A股半年报分析台」网页并发布返回。\n\n用 NeoData 金融搜索服务查财务与宏观数据，用 WeStock Data 查行情、财报、资金流和技术指标，所有数据真实查询并注明来源。先把查询结果整理成 CSV 数据表存入资料库，再生成单页 HTML，利用资料库能力关联这些 CSV 数据表展示（页面通过资料库 SDK 按需查询数据，并提供导出 CSV 备份与离线静态快照兜底），最后把网页发布后返回。页面分六大板块：① 全市场正式半年报总榜（按营收/净利/ROE 等维度排行）；② 产业链环节赚钱榜（选产业链自动联动环节）；③ 双增公司榜（营收与归母净利同比双升）；④ 产业链内公司对比；⑤ 业绩预告/待披露观察区（未披露公司自动观察、披露后转正入榜）；⑥ 宏观行业对照区。支持产业链→环节→板块/类型/排序/关键词联动筛选，数据全部内嵌、单文件 HTML、资源全内联，涨跌按 A 股习惯（涨红跌绿），附免责声明（非投资建议）。", "prompt_en": "Build a '2026 A-Share Mid-Year Report Analyzer' web page and publish it back.\n\nUse NeoData Financial Search for financials and macro data, and WeStock Data for quotes, statements, capital flow and technicals — query real data and cite sources. First organize the queried data into CSV tables and store them in the knowledge base, then build a single-page HTML that leverages the knowledge-base capability to render those CSV tables (the page queries data on demand via the knowledge-base SDK, with CSV-export backup and an offline static snapshot fallback), then publish the page and return it. Six panels: ① market-wide formal interim-report master ranking (sortable by revenue / net profit / ROE etc.); ② industry-chain segment profit board (pick a chain, segments auto-filter); ③ double-growth list (revenue and net profit both rising YoY); ④ within-chain company comparison; ⑤ preannouncement / pending-disclosure watchlist (undisclosed companies stay observed and join the ranking once disclosed); ⑥ macro-industry comparison. Support chained filtering (chain → segment → board / type / sort / keyword). All data embedded, single-file HTML with fully inlined assets, China-market colors (up=red, down=green), plus a disclaimer (not investment advice).", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "2026 A股半年报分析台" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "投资者老丁", "name_en": "Investor Lao Ding", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-21T08:05:00Z", "updated_at": "2026-08-21T08:05:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-50etf-options-bullspread", "title": "50ETF期权牛市价差策略", "title_en": "50ETF Options Bull Call Spread Strategy", "subtitle": "趋势市怎么用牛市价差，Gamma/希腊字母怎么看", "subtitle_en": "Using a bull call spread in a trending market, and how to read Gamma and the Greeks", "description": "期权策略门槛高，趋势市里到底该不该用牛市价差、Gamma 等希腊字母怎么影响盈亏，新手很容易凭感觉下单。本案例以 50ETF 期权为例，讲清楚牛市价差的构造、盈亏结构、盈亏平衡点，并用希腊字母（Delta/Gamma/Theta）拆解风险敞口，配期权定价可视化。适合想系统学习期权策略、做结构化交易的进阶投资者。", "description_en": "Options strategies have a high bar; in a trending market, whether to use a bull call spread and how Greeks like Gamma affect P&L can lead beginners to trade on gut feel. Using 50ETF options as an example, this case explains the construction, payoff structure and break-even of a bull call spread, decomposes risk exposure via the Greeks (Delta/Gamma/Theta), and adds option-pricing visualizations. Ideal for advanced investors who want to learn options strategies systematically and trade structurally.", "categories": [ "finance" ], "tags": [ "50ETF期权", "期权策略", "牛市价差", "希腊字母", "衍生品" ], "tags_en": [ "50ETF Options", "Options Strategy", "Bull Call Spread", "Greeks", "Derivatives" ], "artifact_type": "html", "prompt": "帮我生成一份「50ETF 期权牛市价差策略」分析单页 HTML。\n\n查 50ETF 行情与波动率数据。讲清楚：当前趋势市下牛市价差是否合适、怎么构造（买低行权价 call + 卖高行权价 call）、盈亏结构和盈亏平衡点，并用希腊字母（重点 Gamma，兼顾 Delta/Theta）拆解风险敞口怎么随价格变化。用图表画出到期盈亏曲线和希腊字母变化，最后给适用场景和风险提示（非投资建议）。界面文字用简体中文。", "prompt_en": "Generate a single-page HTML analysis of a '50ETF options bull call spread strategy'.\n\nPull 50ETF quotes and volatility data. Explain: whether a bull call spread suits the current trending market, how to construct it (buy lower-strike call + sell higher-strike call), its payoff structure and break-even, and use the Greeks (focus on Gamma, plus Delta/Theta) to decompose how risk exposure changes with price. Use charts for the expiration payoff curve and Greek dynamics, ending with suitable scenarios and risk notes (not investment advice). UI text in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "期权策略分析" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-ashare-mainline-sentiment", "title": "A股主线与情绪温度复盘", "title_en": "A-Share Mainline & Sentiment Recap", "subtitle": "今天在炒什么主线，情绪到了几度，能不能追", "subtitle_en": "What's the mainline today, how hot is sentiment, and can it be chased", "description": "每天收盘想知道市场到底在交易什么主线、情绪是热是冷、明天还能不能追，但涨停家数、封板率、连板晋级率这些信号散落各处，普通投资者很难快速拼出全貌。本案例做一份 A 股主线与情绪复盘，定量化情绪温度（涨停数、封板率、连板高度、晋级率），识别真主线和资金方向，并对持续性给出预警。适合做日内/隔夜复盘、跟踪市场情绪的活跃投资者。", "description_en": "After the close, investors want to know what mainline the market is trading, whether sentiment is hot or cold, and whether it's still chaseable tomorrow — but signals like limit-up counts, seal rates and consecutive-board promotion rates are scattered, making the full picture hard to assemble. This case produces an A-share mainline & sentiment recap, quantifying sentiment temperature (limit-ups, seal rate, board height, promotion rate), identifying the true mainline and capital direction, and flagging sustainability. Ideal for intraday/overnight reviews and active investors tracking market sentiment.", "categories": [ "finance" ], "tags": [ "A股复盘", "市场主线", "情绪温度", "涨停", "资金动向" ], "tags_en": [ "A-Share Recap", "Market Mainline", "Sentiment", "Limit-Up", "Capital Flow" ], "artifact_type": "html", "prompt": "帮我做一份 A 股主线与情绪复盘单页 HTML：今天市场在交易什么主线，情绪到了几度。\n\n查当天的板块涨跌、涨停板、资金流和市场广度数据。内容包括：定量化情绪温度（涨停家数、封板率、连板高度、连板晋级率），识别今天真正的主线题材和资金方向（区分真主线和诱多），并对主线和情绪能否延续给出预警。用图表展示板块强弱、涨停梯队和情绪指标，最后给隔夜判断（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce a single-page HTML A-share mainline & sentiment recap: what mainline the market traded today and how hot sentiment is.\n\nPull the day's sector moves, limit-ups, capital flow and market breadth. Include: quantified sentiment temperature (limit-up count, seal rate, board height, promotion rate), identification of today's true mainline theme and capital direction (distinguishing real mainline from a bull trap), and a warning on whether the mainline and sentiment can persist. Use charts for sector strength, the limit-up ladder and sentiment indicators, ending with an overnight judgment (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "主线情绪复盘" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-asset-allocation-maxsharpe", "title": "最大夏普 vs 全天候资产配置", "title_en": "Max-Sharpe vs All-Weather Asset Allocation", "subtitle": "100万股债怎么配，两种经典框架哪个更适合现在", "subtitle_en": "How to allocate 1M between stocks and bonds — which classic framework fits now", "description": "手里有一笔钱想做股债配置，最大夏普组合和全天候策略是两种经典思路，但各自的逻辑、风险暴露和适用环境差别很大，普通投资者很难选。本案例以 100 万资金为例，对比两种配置框架的构造方法、预期收益风险、回撤特征和当前市场适配度，给出具体配置比例。适合做资产配置、想在两种主流框架间做选择的投资者。", "description_en": "With a sum to allocate between stocks and bonds, the max-Sharpe portfolio and the all-weather strategy are two classic approaches, but their logic, risk exposure and suitable environments differ greatly, making the choice hard for ordinary investors. Using 1M as an example, this case compares the two frameworks' construction, expected return/risk, drawdown characteristics and fit with the current market, and gives concrete allocation weights. Ideal for asset allocation and investors choosing between the two mainstream frameworks.", "categories": [ "finance" ], "tags": [ "资产配置", "最大夏普", "全天候", "股债配置", "组合优化" ], "tags_en": [ "Asset Allocation", "Max Sharpe", "All-Weather", "Stock-Bond Mix", "Portfolio Optimization" ], "artifact_type": "html", "prompt": "帮我生成一份「最大夏普 vs 全天候」资产配置对比单页 HTML，以 100 万资金做股债配置为例。\n\n查大类资产的行情与历史表现数据。对比两种框架：最大夏普组合和全天候策略各自怎么构造、逻辑是什么、预期收益和风险、回撤特征，以及在当前利率和市场环境下哪个更合适。给出两种方案各自的具体配置比例，用图表展示资产权重和风险收益对比，最后给选择建议（非投资建议）。界面文字用简体中文。", "prompt_en": "Generate a single-page HTML comparing 'max-Sharpe vs all-weather' asset allocation, using a 1M stock-bond allocation as the example.\n\nPull major-asset quotes and historical performance. Compare the two frameworks: how the max-Sharpe portfolio and the all-weather strategy are each constructed, their logic, expected return and risk, drawdown characteristics, and which fits better under the current rate and market environment. Provide concrete allocation weights for each, use charts for asset weights and risk-return comparison, ending with a selection suggestion (not investment advice). UI text in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "资产配置方案" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-byd-business-model-compare", "title": "比亚迪赚钱方式深度拆解", "title_en": "BYD Business Model Deep-Dive vs Peers", "subtitle": "和同行比，比亚迪到底靠什么赚钱、护城河在哪", "subtitle_en": "How BYD actually makes money compared with peers, and where its moat lies", "description": "都说比亚迪是龙头，但它和同行的赚钱方式到底差在哪、垂直整合究竟带来多少利润优势，光看股价说不清楚。本案例把比亚迪的商业模式拆开，对比同行的打法，量化垂直整合的收益（如代工/电池外供的协同贡献），讲清楚护城河和成长驱动。适合想从商业模式而非短期行情维度理解一家公司的投资者与产业研究者。", "description_en": "BYD is called the leader, but how its money-making model differs from peers, and how much profit advantage vertical integration really brings, can't be explained by share price alone. This case dissects BYD's business model, compares peer approaches, quantifies the gains from vertical integration (e.g. synergy from contract manufacturing / external battery supply), and clarifies the moat and growth drivers. Ideal for investors and researchers who want to understand a company from a business-model angle rather than short-term price.", "categories": [ "finance" ], "tags": [ "比亚迪", "商业模式", "新能源车", "同业对比", "护城河" ], "tags_en": [ "BYD", "Business Model", "EV", "Peer Comparison", "Moat" ], "artifact_type": "html", "prompt": "帮我生成一份「比亚迪商业模式拆解」单页 HTML。\n\n查比亚迪（002594）及同行的财务、行情数据。重点说清楚：比亚迪和同行相比，赚钱方式有什么不一样，垂直整合（自产电池、半导体、代工等）带来多少利润优势，护城河和长期成长驱动在哪。用图表对比比亚迪与同行的盈利结构和关键指标，最后给一段综合判断（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Generate a single-page HTML 'BYD business model deep-dive'.\n\nPull financials and quotes for BYD (002594) and peers. Focus on: how BYD's money-making model differs from peers, how much profit advantage vertical integration (in-house batteries, semiconductors, contract manufacturing, etc.) brings, and where the moat and long-term growth drivers lie. Use charts to compare profit structure and key metrics vs peers, ending with an overall assessment (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "商业模式分析报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-catl-crowding-diagnosis", "title": "宁德时代拥挤度诊断", "title_en": "CATL Crowding Diagnosis", "subtitle": "是不是抱团股，拥挤度高不高，还能不能追", "subtitle_en": "Is it a crowded consensus trade, how crowded, and can it still be chased", "description": "明星股最怕「抱团瓦解」，但拥挤度到底高不高、是不是已经到了危险区，普通投资者只能凭感觉。本案例对宁德时代做拥挤度诊断，从机构持仓集中度、基金重仓比例、成交占比、北向持股等维度量化「抱团」程度，判断当前拥挤度处在什么水位、风险大不大，并给出交易层面的应对计划。适合关注热门股、想规避抱团瓦解风险的投资者。", "description_en": "Star stocks most fear a 'consensus unwind', but whether crowding is high and already in the danger zone is something ordinary investors can only guess. This case diagnoses CATL's crowding, quantifying the degree of 'huddling' across institutional holding concentration, fund overweight ratio, turnover share and northbound holdings, judges what level crowding sits at and how risky it is, and gives a trading-level action plan. Ideal for investors following popular stocks who want to avoid consensus-unwind risk.", "categories": [ "finance" ], "tags": [ "宁德时代", "拥挤度", "抱团股", "机构持仓", "风险管理" ], "tags_en": [ "CATL", "Crowding", "Consensus Trade", "Institutional Holdings", "Risk Management" ], "artifact_type": "html", "prompt": "帮我对宁德时代做一份拥挤度诊断单页 HTML：它算不算抱团股，拥挤度高不高，还能不能追。\n\n查它的机构持仓、基金重仓、成交和北向持股数据。从机构持仓集中度、基金重仓比例、成交占比、北向持股等维度量化「抱团」程度，判断当前拥挤度处在历史什么水位、抱团瓦解的风险大不大，并给出交易层面的应对计划（关键位、减仓信号等）。用图表展示拥挤度各维度指标和历史变化，最后给综合判断（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce a single-page HTML crowding diagnosis for CATL: is it a crowded consensus trade, how crowded, and can it still be chased.\n\nPull its institutional holdings, fund overweights, turnover and northbound-holding data. Quantify the degree of 'huddling' across institutional concentration, fund overweight ratio, turnover share and northbound holdings, judge where crowding sits versus history and how big the consensus-unwind risk is, and give a trading-level action plan (key levels, trim signals, etc.). Use charts for the crowding dimensions and their history, ending with an overall judgment (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "拥挤度诊断报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-googl-earnings-review", "title": "谷歌财报超预期为何还跌", "title_en": "Why GOOGL Fell Despite a Beat", "subtitle": "业绩超预期却下跌，业绩会到底说了什么", "subtitle_en": "Beat expectations yet dropped — what the earnings call actually said", "description": "财报超预期股价反而下跌，是美股常见却最让人困惑的现象。本案例以谷歌（GOOGL）某季财报为例，拆解「超预期却下跌」背后的真实原因——是指引不及预期、资本开支超预期，还是市场早已price in，并梳理业绩会管理层的关键表态。帮你看懂财报数字之外的市场逻辑。适合关注美股财报、想理解预期差交易的投资者。", "description_en": "A stock falling after an earnings beat is a common yet baffling US-market phenomenon. Using a Google (GOOGL) quarterly report, this case unpacks the real reason behind 'beat yet dropped' — soft guidance, capex above expectations, or already priced in — and distills management's key remarks on the call. It helps you see the market logic beyond the headline numbers. Ideal for investors following US earnings who want to understand expectation-gap trading.", "categories": [ "finance" ], "tags": [ "谷歌", "美股财报", "财报点评", "预期差", "业绩会" ], "tags_en": [ "Google", "US Earnings", "Earnings Review", "Expectation Gap", "Earnings Call" ], "artifact_type": "html", "prompt": "帮我做一份谷歌（GOOGL）财报点评单页 HTML：业绩超预期为什么还跌，业绩会说了啥。\n\n查谷歌的财报数据、行情和历史表现。拆解：这次财报哪些超预期、哪些不及预期，股价超预期却下跌的真实原因（指引、资本开支、是否已被 price in 等），并梳理业绩会上管理层的关键表态。用图表展示分业务营收、利润和关键指标的同比环比变化，最后给综合解读（非投资建议）。涉及非一手来源的信息要标注。界面文字用简体中文。", "prompt_en": "Produce a single-page HTML earnings review for Google (GOOGL): why it fell despite beating, and what the call said.\n\nPull Google's earnings data, quotes and history. Unpack: what beat and what missed this quarter, the real reason the stock fell despite the beat (guidance, capex, whether already priced in, etc.), and distill management's key remarks on the call. Use charts for segment revenue, profit and key-metric YoY/QoQ changes, ending with an overall reading (not investment advice). Mark any non-first-hand sourced information. UI text in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "财报点评报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-hkbn-earnings-forecast", "title": "香港宽频业绩与增长预测", "title_en": "HKBN Earnings & Growth Forecast", "subtitle": "财年改制下的利息吞噬陷阱，增长率怎么还原", "subtitle_en": "The interest-erosion trap under a fiscal-year change, and how to normalize growth", "description": "香港宽频（01310.HK）的财报藏着两个坑：财年改制让同比口径错位、高杠杆下利息费用大幅吞噬利润，直接看增长率容易被误导。本案例对其做业绩与增长率预测，还原可比口径，拆解利息对净利的侵蚀，识别财报里的「会计陷阱」，给出更接近真实经营的增长判断。适合做港股财报分析、关注高杠杆公司盈利质量的投资者。", "description_en": "HKBN's (01310.HK) financials hide two traps: a fiscal-year change misaligns year-on-year comparisons, and heavy leverage means interest expense erodes profit substantially, so reading growth at face value is misleading. This case forecasts earnings and growth, normalizes the comparable basis, decomposes interest's erosion of net profit, flags the 'accounting traps' in the statements, and gives a growth judgment closer to true operations. Ideal for HK-stock earnings analysis and investors focused on the profit quality of highly leveraged firms.", "categories": [ "finance" ], "tags": [ "香港宽频", "业绩预测", "港股", "财年改制", "盈利质量" ], "tags_en": [ "HKBN", "Earnings Forecast", "HK Stocks", "Fiscal-Year Change", "Profit Quality" ], "artifact_type": "html", "prompt": "帮我对香港宽频（01310.HK）做一份业绩与增长率预测单页 HTML。\n\n查它的财报、行情和历史财务数据。重点处理两个难点：财年改制导致的同比口径错位怎么还原成可比口径；高杠杆下利息费用对净利润的吞噬有多大。在此基础上预测营收和净利润增长率，识别财报里容易误导人的「会计陷阱」，给出更贴近真实经营的增长判断。用图表展示历史与预测的营收利润趋势、利息侵蚀和增长率，最后给综合结论（非投资建议）。界面文字用简体中文。", "prompt_en": "Produce a single-page HTML earnings & growth forecast for HKBN (01310.HK).\n\nPull its statements, quotes and historical financials. Tackle two key difficulties: how to normalize the year-on-year basis distorted by the fiscal-year change; and how much interest expense under heavy leverage erodes net profit. On that basis, forecast revenue and net-profit growth, flag misleading 'accounting traps' in the statements, and give a growth judgment closer to true operations. Use charts for historical and forecast revenue/profit trends, interest erosion and growth rates, ending with an overall conclusion (not investment advice). UI text in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "业绩预测报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-holding-checkup", "title": "五只持仓诊断体检", "title_en": "Five-Stock Portfolio Checkup", "subtitle": "手里五只票哪只该砍，集中度、估值、资金一次看清", "subtitle_en": "Which of your five holdings to cut — concentration, valuation and capital flow at a glance", "description": "手里拿着好几只股票，涨涨跌跌却说不清哪只该留哪只该砍，更看不清整体的集中度和风险敞口。本案例以一份五只股票的持仓为例，逐只做体检——估值高低、基本面质量、资金动向、技术位置，并从组合角度看行业集中度和相关性，明确给出加减仓优先级。适合有持仓、想定期做组合复盘的个人投资者。", "description_en": "Holding several stocks that drift up and down, investors often can't tell which to keep or cut, nor see overall concentration and risk exposure. Using a five-stock portfolio as an example, this case checks each holding — valuation, fundamental quality, capital flow, technical position — and views sector concentration and correlation at the portfolio level, giving a clear add/trim priority. Ideal for individual investors with holdings who want regular portfolio reviews.", "categories": [ "finance" ], "tags": [ "持仓体检", "组合诊断", "风险管理", "集中度", "调仓" ], "tags_en": [ "Portfolio Checkup", "Portfolio Diagnosis", "Risk Management", "Concentration", "Rebalancing" ], "artifact_type": "html", "prompt": "帮我做一份「持仓体检」单页 HTML：手里有贵州茅台、宁德时代、比亚迪、恒瑞医药、立讯精密五只股票，帮我诊断哪只该砍。\n\n查这五只票的估值、财务、资金流和技术数据。逐只体检：估值贵不贵、基本面质量、主力资金动向、技术位置，给每只一个保留/减仓的判断；再从组合角度看行业集中度、相关性和整体风险，给出加减仓优先级。用图表展示各股对比和组合结构，最后给综合建议（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce a single-page HTML 'portfolio checkup': I hold Kweichow Moutai, CATL, BYD, Hengrui, and Luxshare — help me decide which to cut.\n\nPull valuation, financials, capital flow and technicals of the five. Check each: whether valuation is expensive, fundamental quality, main-capital flow, technical position, and give a keep/trim call for each; then view sector concentration, correlation and overall risk at the portfolio level, with an add/trim priority. Use charts for cross-stock comparison and portfolio structure, ending with an overall suggestion (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "持仓体检报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-low-altitude-industry-chain", "title": "低空经济产业链全景拆解", "title_en": "Low-Altitude Economy Industry Chain Map", "subtitle": "上中下游怎么分，哪一段最受益，伪受益怎么识别", "subtitle_en": "Upstream-to-downstream breakdown, who benefits most, and how to spot fake beneficiaries", "description": "提到低空经济，市场一拥而上炒概念，但产业链上中下游的价值分布天差地别，普通投资者很难分清谁是真受益、谁只是蹭热点。本案例以「空管 > 整机 > 运营」的价值链视角，把低空经济产业链分层拆开，标出每一段的代表公司、受益逻辑和确定性，并专门识别「伪受益」标的。适合做主题投资研究、想看懂产业链结构而非追概念的投资者。", "description_en": "When the low-altitude economy is hot, the market piles into the theme, but value is distributed very unevenly across the chain, and ordinary investors struggle to tell real beneficiaries from hype-chasers. Using an 'air-traffic-control > airframe > operations' value-chain lens, this case breaks the chain into layers, marks representative companies, benefit logic and certainty for each, and specifically flags 'fake beneficiaries'. Ideal for thematic research and investors who want to understand chain structure rather than chase concepts.", "categories": [ "finance" ], "tags": [ "低空经济", "产业链", "主题投资", "受益标的", "产业研究" ], "tags_en": [ "Low-Altitude Economy", "Industry Chain", "Thematic Investing", "Beneficiaries", "Industry Research" ], "artifact_type": "html", "prompt": "帮我生成一份「低空经济产业链」分析单页 HTML。\n\n查相关公司与板块数据，把低空经济产业链按上中下游分层拆开：每一段有哪些代表公司、受益逻辑是什么、确定性高不高，最受益的是哪一段。重点识别哪些是真受益、哪些只是蹭概念的「伪受益」标的。用图表展示产业链结构和各环节代表公司，最后给一段客观判断（非投资建议）。界面文字用简体中文，涨跌按 A 股习惯涨红跌绿。", "prompt_en": "Generate a single-page HTML analysis of the 'low-altitude economy industry chain'.\n\nPull company and sector data, and break the chain into upstream/midstream/downstream layers: representative companies, benefit logic and certainty for each, and which segment benefits most. Emphasize distinguishing true beneficiaries from 'fake beneficiaries' merely riding the theme. Use charts to show the chain structure and key companies per segment, ending with an objective assessment (not investment advice). UI text in Simplified Chinese; use the China-market color convention (up=red, down=green).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "产业链分析报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-maotai-gap-attribution", "title": "贵州茅台盘中异动归因", "title_en": "Kweichow Moutai Intraday Gap Attribution", "subtitle": "跳空高开低开怎么解读，是真信号还是假动作", "subtitle_en": "How to read a gap up/down — real signal or fake-out", "description": "个股突然跳空高开或低开，盘中剧烈波动，投资者第一反应是慌，却说不清到底是基本面变化、资金行为还是情绪扰动。本案例以贵州茅台某交易日的盘中异动为例，做四维归因——消息面、资金面、技术面、情绪面，逐一拆解跳空和盘中走势的成因，判断是真信号还是假动作。适合做盘口分析、想系统理解个股异动的投资者。", "description_en": "When a stock suddenly gaps up or down and swings intraday, investors' first reaction is panic, yet they can't tell whether it's fundamental change, capital behavior or sentiment noise. Using Kweichow Moutai's intraday move on a given day, this case does four-dimensional attribution — news, capital, technicals, sentiment — decomposing the cause of the gap and intraday path to judge real signal vs fake-out. Ideal for tape analysis and investors who want to understand single-stock anomalies systematically.", "categories": [ "finance" ], "tags": [ "贵州茅台", "盘中异动", "归因分析", "跳空", "盘口分析" ], "tags_en": [ "Kweichow Moutai", "Intraday Anomaly", "Attribution", "Gap", "Tape Reading" ], "artifact_type": "html", "prompt": "帮我对贵州茅台做一份盘中异动归因单页 HTML：当天跳空高开/低开怎么解读。\n\n查当天的分时行情、资金流、技术指标和相关消息。做四维归因：消息面（有没有事件催化）、资金面（主力/北向怎么动）、技术面（关键位、量价配合）、情绪面（板块和市场情绪），逐一拆解跳空和盘中走势是什么原因造成的，判断是真信号还是假动作。用图表展示分时走势、资金流向和各维度归因，最后给综合解读（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce a single-page HTML intraday-anomaly attribution for Kweichow Moutai: how to read the day's gap up/down.\n\nPull the day's intraday quotes, capital flow, technicals and related news. Do four-dimensional attribution: news (any catalyst), capital (how main/northbound funds move), technicals (key levels, price-volume), sentiment (sector and market mood), decomposing what caused the gap and intraday path to judge real signal vs fake-out. Use charts for the intraday path, capital flow and per-dimension attribution, ending with an overall reading (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "盘中异动归因" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-sungrow-deep-research", "title": "阳光电源卖方级深度研究", "title_en": "Sungrow Sell-Side Deep Research", "subtitle": "光储龙头的商业模式、行业拐点与业绩弹性", "subtitle_en": "The PV-storage leader's business model, industry inflection and earnings elasticity", "description": "光储行业景气波动大，阳光电源作为逆变器+储能龙头，业绩弹性和行业拐点高度绑定，普通投资者很难判断现在是上行起点还是见顶。本案例输出一份卖方级深度研究，拆解其产品与地域结构、行业供需拐点、业绩驱动与「量利双杀」风险，覆盖投资逻辑和估值。适合做新能源深度研究、关注景气拐点的投资者。", "description_en": "The PV-storage sector swings sharply, and as an inverter + energy-storage leader, Sungrow's earnings elasticity is tightly tied to industry inflection — ordinary investors struggle to tell whether it's the start of an upcycle or a peak. This case delivers sell-side-grade deep research, dissecting its product and geographic mix, industry supply-demand inflection, earnings drivers and the 'volume-and-margin double-hit' risk, covering investment logic and valuation. Ideal for deep new-energy research and investors watching cyclical turning points.", "categories": [ "finance" ], "tags": [ "阳光电源", "深度研究", "光伏储能", "行业拐点", "新能源" ], "tags_en": [ "Sungrow", "Deep Research", "PV & Storage", "Industry Inflection", "New Energy" ], "artifact_type": "html", "prompt": "帮我对阳光电源（300274）做一份卖方研究员级别的深度研究，输出单页 HTML。\n\n查财务、行情和行业数据。覆盖：商业模式（逆变器+储能怎么赚钱）、产品与地域收入结构、光储行业供需与景气拐点、业绩驱动与弹性、「量利双杀」风险，以及投资逻辑和估值。用图表展示营收利润趋势、产品结构和行业景气指标，最后给综合判断（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce sell-side-analyst-grade deep research on Sungrow (300274) as a single-page HTML.\n\nPull financials, quotes and industry data. Cover: business model (how inverters + storage make money), product and geographic revenue mix, PV-storage supply-demand and cyclical inflection, earnings drivers and elasticity, the 'volume-and-margin double-hit' risk, plus investment logic and valuation. Use charts for revenue/profit trends, product mix and industry indicators, ending with an overall assessment (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "深度研究报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-us-market-mainline", "title": "美股主线与资金方向研判", "title_en": "US Market Mainline & Capital Direction", "subtitle": "今天领涨的是真主线还是诱多，资金往哪去", "subtitle_en": "Is today's leader a real mainline or a bull trap, and where is capital going", "description": "美股每天领涨的方向到底是真主线还是诱多，AI 链内部是「高切低」还是整体走强，普通投资者隔着时差很难判断。本案例研判美股当日主线，拆解资金在板块和个股间的流动方向，识别 AI 算力链内部的轮动（如 GPU 向定制 ASIC 的切换），并区分真主线与短期情绪。适合关注美股、想理解科技股内部结构性轮动的投资者。", "description_en": "Whether the day's leading direction in US stocks is a real mainline or a bull trap, and whether the AI chain is rotating 'high-to-low' or strengthening broadly, is hard for ordinary investors to judge across time zones. This case assesses the day's US mainline, decomposes capital flow across sectors and stocks, identifies rotation within the AI-compute chain (e.g. a shift from GPUs to custom ASICs), and distinguishes the real mainline from short-term sentiment. Ideal for investors following US stocks who want to understand structural rotation within tech.", "categories": [ "finance" ], "tags": [ "美股", "市场主线", "资金流向", "AI算力", "板块轮动" ], "tags_en": [ "US Stocks", "Market Mainline", "Capital Flow", "AI Compute", "Sector Rotation" ], "artifact_type": "html", "prompt": "帮我做一份美股主线研判单页 HTML：今天美股领涨的是哪个方向，是真主线还是诱多，资金往哪去。\n\n查美股当天的板块涨跌、个股表现和资金流数据。内容包括：识别当日领涨方向和真正的市场主线，拆解资金在板块和个股之间的流动，重点分析 AI 算力链内部的结构性轮动（比如从 GPU 向定制 ASIC 的切换），区分真主线和短期诱多。用图表展示板块强弱、龙头表现和资金方向，最后给综合研判（非投资建议）。界面文字用简体中文。", "prompt_en": "Produce a single-page HTML US-market mainline assessment: which direction led today, real mainline or bull trap, and where capital is going.\n\nPull the day's US sector moves, single-stock performance and capital flow. Include: identifying the leading direction and the true mainline, decomposing capital flow across sectors and stocks, with focus on structural rotation within the AI-compute chain (e.g. a shift from GPUs to custom ASICs), distinguishing real mainline from short-term traps. Use charts for sector strength, leader performance and capital direction, ending with an overall assessment (not investment advice). UI text in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "美股主线研判" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-wanhua-deep-research", "title": "万华化学卖方级深度研究", "title_en": "Wanhua Chemical Sell-Side Deep Research", "subtitle": "商业模式·竞争格局·投资逻辑·周期位置一份看懂", "subtitle_en": "Business model, competitive landscape, investment logic and cycle position in one report", "description": "想看懂万华化学这种周期成长股，要把商业模式、竞争格局、投资逻辑和当前周期位置都摸清楚，散落的研报很难一次拼齐。本案例输出一份卖方研究员级别的深度研究，覆盖 MDI 龙头的护城河、成本曲线、产能周期、估值（DCF+PB+现金流三维）和风险，并明确判断当前处在周期的哪个位置。适合做深度公司研究、想要系统性结论而非碎片信息的投资者。", "description_en": "Understanding a cyclical-growth stock like Wanhua Chemical requires grasping its business model, competitive landscape, investment logic and current cycle position together — scattered research notes rarely assemble this at once. This case produces sell-side-analyst-grade deep research covering the MDI leader's moat, cost curve, capacity cycle, valuation (DCF + PB + cash-flow, three angles) and risks, and clearly judges where it sits in the cycle. Ideal for deep company research and investors who want systematic conclusions rather than fragments.", "categories": [ "finance" ], "tags": [ "万华化学", "深度研究", "周期股", "估值", "化工" ], "tags_en": [ "Wanhua Chemical", "Deep Research", "Cyclical Stock", "Valuation", "Chemicals" ], "artifact_type": "html", "prompt": "帮我对万华化学（600309）做一份卖方研究员级别的深度研究，输出单页 HTML。\n\n查财务、行情和同业数据。覆盖：商业模式（MDI 龙头怎么赚钱、成本优势）、竞争格局、投资逻辑、成长驱动、产能与周期位置判断，以及估值（结合 DCF、PB、现金流多角度）和主要风险。用图表展示营收利润趋势、估值分位和周期位置，最后给综合结论（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce sell-side-analyst-grade deep research on Wanhua Chemical (600309) as a single-page HTML.\n\nPull financials, quotes and peer data. Cover: business model (how the MDI leader makes money, cost advantage), competitive landscape, investment logic, growth drivers, capacity and cycle-position judgment, plus valuation (DCF, PB, cash-flow, multiple angles) and key risks. Use charts for revenue/profit trends, valuation percentile and cycle position, ending with an overall conclusion (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "深度研究报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finance-wuxi-apptec-moat", "title": "药明康德护城河深度分析", "title_en": "WuXi AppTec Moat Analysis", "subtitle": "壁垒到底有多深，长期竞争优势能不能信", "subtitle_en": "How deep is the moat, and can the long-term edge be trusted", "description": "都说药明康德是 CXO 龙头有护城河，但壁垒究竟来自哪、面对生物安全法案等外部冲击还稳不稳，光靠印象说不清。本案例用「护城河证据矩阵」框架，从规模、一体化能力、客户黏性、转换成本等维度逐条找证据，评估壁垒的深度和可持续性，并把地缘政策风险纳入考量。适合做长期价值研究、想验证「护城河」是否真实的投资者。", "description_en": "WuXi AppTec is called a CXO leader with a moat, but where the barrier really comes from, and whether it holds up against external shocks like the BIOSECURE Act, can't be settled by impression alone. Using a 'moat evidence matrix' framework, this case gathers evidence dimension by dimension — scale, integrated capability, customer stickiness, switching costs — to assess the moat's depth and durability, factoring in geopolitical-policy risk. Ideal for long-term value research and investors who want to verify whether a 'moat' is real.", "categories": [ "finance" ], "tags": [ "药明康德", "护城河", "CXO", "竞争优势", "长期价值" ], "tags_en": [ "WuXi AppTec", "Moat", "CXO", "Competitive Edge", "Long-Term Value" ], "artifact_type": "html", "prompt": "帮我对药明康德做一份护城河深度分析单页 HTML：它算不算好公司，壁垒和长期竞争优势在哪。\n\n查它的财务、行情和行业数据。用「护城河证据矩阵」的思路，从规模效应、一体化（CRDMO）能力、客户黏性、转换成本等维度逐条找证据，评估壁垒到底有多深、能不能持续，并把生物安全法案等地缘政策风险纳入考量。用图表展示关键财务趋势和各维度护城河评分，最后给长期竞争优势的综合判断（非投资建议）。界面文字用简体中文，涨跌涨红跌绿。", "prompt_en": "Produce a single-page HTML moat analysis for WuXi AppTec: is it a good company, and where is its barrier and long-term edge.\n\nPull its financials, quotes and industry data. Using a 'moat evidence matrix' approach, gather evidence dimension by dimension — scale effects, integrated (CRDMO) capability, customer stickiness, switching costs — to assess how deep the moat is and whether it persists, factoring in geopolitical-policy risk such as the BIOSECURE Act. Use charts for key financial trends and per-dimension moat scores, ending with an overall judgment of the long-term edge (not investment advice). UI text in Simplified Chinese; up=red, down=green.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "护城河分析报告" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-29T22:45:00Z", "updated_at": "2026-06-29T22:45:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "financial-freedom-roadmap", "title": "个人财务自由 FIRE 规划路线图", "title_en": "Personal Financial Independence (FIRE) Roadmap", "subtitle": "用复利和里程碑，把财务自由拆成可执行的 15 年计划", "subtitle_en": "Turn financial independence into an actionable 15-year plan with compounding and milestones", "description": "「财务自由」听起来遥远，难点在于它是一个抽象的长期目标，普通人很难知道自己现在在哪、还差多远、每年该存多少。本案例由「理财财」财富管理专家生成一份个人 FIRE 规划路线图——用 4% 法则算出目标资产，画出 15 年复利增长曲线，把大目标拆成可达成的里程碑，并对比保守/基准/积极三种情景。适合想认真规划长期财务、追求提前退休的个人。", "description_en": "'Financial freedom' sounds distant — the hard part is that it's an abstract long-term goal, and most people can't tell where they stand, how far they have to go, or how much to save each year. Generated by the Wealth Management Expert, this personal FIRE roadmap uses the 4% rule to compute the target nest egg, plots a 15-year compounding growth curve, breaks the big goal into achievable milestones, and compares conservative/base/aggressive scenarios. Ideal for individuals serious about long-term planning and early retirement.", "categories": [ "finance" ], "tags": [ "财务自由", "FIRE", "财务规划", "复利", "提前退休" ], "tags_en": [ "Financial Freedom", "FIRE", "Financial Planning", "Compounding", "Early Retirement" ], "artifact_type": "html", "prompt": "生成一份个人财务自由（FIRE）规划路线图。\n\n核心参数：当前净资产 ¥600,000，年储蓄 ¥180,000（储蓄率 43%），预期年化回报 7%；按 4% 法则，FIRE 目标资产 = 年支出 × 25 = ¥6,000,000，当前进度 10%。要有目标资产/进度/预计达成年龄/储蓄率 KPI、15 年复利资产增长曲线、关键里程碑，以及保守/基准/积极三种情景对比。", "prompt_en": "Generate a personal Financial Independence (FIRE) roadmap.\n\nKey parameters: current net worth ¥600,000, annual savings ¥180,000 (savings rate 43%), expected annual return 7%; by the 4% rule, the FIRE target = annual expenses × 25 = ¥6,000,000, current progress 10%. Include target-assets / progress / projected age of achievement / savings-rate KPIs, a 15-year compounding asset-growth curve, key milestones, and a conservative/base/aggressive scenario comparison.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "FIRE 规划路线图" } ], "skills": [], "experts": [ { "id": "WealthManagementExpert", "name": "理财财", "description": "财富管理工具集：客户回顾、财务规划、投资组合再平衡、税损收割、投资提案和客户报告生成", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "06802d3ccfd4", "name_en": "Harlan · Wealth Management Expert", "description_en": "Wealth management toolkit: client review, financial planning, portfolio rebalance, tax-loss harvesting, investment proposal, and client reporting", "expert_id": "ex_wMZ6zxKnk1uI" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-16T19:30:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "financial-report-analysis-workbench", "title": "财报八步分析法工作台", "title_en": "Eight-Step Financial Report Analysis Workbench", "subtitle": "八步框架拆财报，NeoData + 东财双源交叉验证", "subtitle_en": "Eight-step report breakdown with NeoData + EastMoney dual-source validation", "description": "散户和研究员读财报，最容易卡在两点：数据散落各处，结论和论证对不上号。\n\n这个案例展示了一个「财报八步分析法工作台」：输入公司名即可生成标准分析指令，报告统一按「一句话结论置顶 → 核心指标卡片 → 八步框架逐层展开 → 深度长文 → 数据源交叉验证表」组织；内置贵州茅台、兆易创新、英伟达、五粮液、中国石化、寒武纪、国泰海通、中际旭创、新易盛等龙头公司的完整财报分析（NeoData 主源 + 东财第二源双源交叉验证），报告库可增删查看，数据表支持挂载 CSV 财务数据。\n\n适合：想系统化拆解财报的投资者、需要把研报线索收在一处的研究员、关注 A 股基本面的个人。", "description_en": "When retail investors and analysts read financial reports, they most often get stuck on two things: data scattered everywhere, and conclusions that don't line up with the evidence.\n\nThis case shows an Eight-Step Financial Report Analysis Workbench: enter a company name to generate a standard analysis prompt, and every report is organized as 'one-sentence conclusion on top → core metric cards → eight-step framework → in-depth article → data-source cross-validation table'. It ships with complete analyses of leading companies — Kweichow Moutai, GigaDevice, NVIDIA, Wuliangye, Sinopec, Cambricon, Guotai Haitong, Zhongji Innolight, Eoptolink and more — cross-validated across NeoData (primary) and EastMoney (secondary). The report library supports add / view / delete, and the data-table tab supports mounting CSV financial data.\n\nBest for: investors who want to dissect financial reports systematically, analysts who want to keep research threads in one place, and anyone who follows A-share fundamentals.", "categories": [ "workspace" ], "tags": [ "财报分析", "投资研究", "数据验证", "基本面", "金融工作台" ], "tags_en": [ "Financial Analysis", "Investment Research", "Data Validation", "Fundamentals", "Finance Workbench" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "财报八步分析法工作台（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/5NNtX1C69pawqFGHaM3eDN?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n帮我做一个「财报八步分析法工作台」。\n\n功能：① 顶部输入公司名/代码，生成一条标准分析指令（NeoData 主源 + 东财第二源双源交叉验证），复制到 WorkBuddy 对话触发分析；② 报告展示区：一句话结论置顶 + 核心指标卡片 + 八步财报分析框架（可折叠，每步含事实与判断）+ 深度长文正文 + 数据源交叉验证表（数据项 / 两源数值 / 验证状态）；③ 报告库：可增删查看；④ 八步框架说明页；⑤ 数据表：支持上传 / 预览 / 下载 CSV 财务数据。\n\n视觉：深蓝渐变头部 + 白色卡片，涨红跌绿，结论置顶的蓝色高亮框，桌面端 960px 居中。\n\n制作完成后：① 将 HTML 存入资料库并绑定结构化数据表（报告库在线存储、跨设备双向同步）；② 发布为公开链接返回给我。CSS/JS 无外部 CDN。\n\n交付：资料库在线工作台——在线 HTML 页面 + 7 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/5NNtX1C69pawqFGHaM3eDN?source=2, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build an Eight-Step Financial Report Analysis Workbench. Features: ① a top input for company name / ticker that generates a standard analysis prompt (NeoData as primary source + EastMoney as secondary, dual-source cross-validation), to be copied into a WorkBuddy conversation; ② a report display area: one-sentence conclusion on top + core metric cards + an eight-step financial-analysis framework (collapsible, each step with facts and judgment) + an in-depth article + a data-source cross-validation table (item / two-source values / status); ③ a report library, supporting add / view / delete; ④ an eight-step framework reference page; ⑤ a data-table tab supporting CSV financial-data upload / preview / download. Visual: a deep-blue gradient header with white cards, red for gains and green for losses, a blue highlight box for the conclusion on top, centered at 960px on desktop. When done: ① save the HTML to the library and bind structured data tables (report library stored online, synced across devices); ② publish it to a public link and return the link to me. Inline all CSS/JS with no external CDN.\n\nDeliverable: an online Library workbench — an online HTML page plus 7 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "李大霄", "name_en": "Li Daxiao", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "community", "usage_count": 0, "created_at": "2026-08-28T11:37:30+08:00", "updated_at": "2026-08-28T11:51:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "financial-statement-credit-review-report", "title": "银行流水信贷审查报告", "title_en": "Bank Statement Credit Review Report", "subtitle": "自动解析对公+个人流水，九模块全景风险评估", "subtitle_en": "Auto-parse corporate & personal bank statements with 9-module risk assessment", "description": "银行贷款审批最头疼的就是翻几十页流水找风险点。这个案例演示了如何把企业的对公账户和法人个人账户流水一起丢给 AI，自动生成包含经营概况、收支趋势、上下游集中度、异常流量识别、信贷信号、个人账户全景、关联资金流转等九大模块的 HTML 风控报告，每项指标都附带判定规则和风控解读。\n\n适合：银行信贷审批员、风控分析师、小贷公司尽调人员、企业财务尽调顾问、供应链金融风控岗", "description_en": "The most frustrating part of credit approval is digging through dozens of pages of bank statements to find risk signals. This case demonstrates how to feed both corporate and personal bank statements to AI, which automatically generates an HTML risk report covering 9 modules: business overview, income/expense trends, customer/supplier concentration, suspicious flow detection, credit signals, personal account overview, cross-entity fund flows, and more — each metric accompanied by decision rules and risk interpretation.\n\nSuitable for: bank credit officers, risk analysts, micro-lending due diligence staff, corporate financial due diligence consultants, supply chain finance risk managers", "categories": [ "finance", "business", "data" ], "tags": [ "信贷审批", "流水分析", "风控报告", "银行流水", "尽调" ], "tags_en": [ "credit review", "bank statement analysis", "risk report", "due diligence", "financial analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.html", "prompt": "客户申请流动资金贷款。请解析以下银行流水文件：\n- 对公银行流水_脱敏模拟_2026H1.xlsx\n- 法人个人账户流水_脱敏模拟.xlsx\n\n生成面向信贷审批的 HTML 流水调查报告，覆盖：经营概况、收支趋势、上下游集中度、异常流量识别（当日净额/总流量<20%且总流量>5万标记为异常）、信贷信号、个人账户全景、关联资金流转及风险评估。\n\n报告要求：单文件 HTML、CSS 内联、纯 SVG 图表无 JS 依赖、每个指标附带风控解读和判定规则。", "prompt_en": "The client is applying for a working capital loan. Please parse the following bank statement files:\n- Corporate Bank Statement (Simulated) 2026H1.xlsx\n- Legal Representative Personal Account Statement (Simulated).xlsx\n\nGenerate an HTML credit review report covering: business overview, income/expense trends, customer/supplier concentration, suspicious flow detection (flag when daily net/flow <20% and flow >50K), credit signals, personal account overview, cross-entity fund flows, and risk assessment.\n\nRequirements: single-file HTML, inline CSS, pure SVG charts with zero JS dependency, each metric accompanied by risk interpretation and decision rules.", "skills": [ { "id": "financial-statement-credit-review", "name": "银行流水信贷审查报告", "description": "自动解析银行流水数据（Excel/PDF/CSV），面向信贷审批生成 HTML 调查报告", "skill_id": "skill_2085354991601524736" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "report.html", "label": "HTML 流水信贷审查报告" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-04T06:04:38Z", "updated_at": "2026-08-04T06:04:38Z", "is_recommended": false, "recommended_order": 0 }, { "id": "finops-capacity-cost-inspection", "title": "资源容量水位与FinOps成本巡检报告", "title_en": "Resource Capacity & FinOps Cost Health Inspection", "subtitle": "一键生成云资源水位+成本健康综合巡检报告", "subtitle_en": "One-click comprehensive cloud resource capacity and cost health report", "description": "每月手动盘点云资源利用率和费用明细，是运维团队最头疼的重复劳动，既容易遗漏闲置资源，又缺乏系统性优化建议。\n\n本案例演示了如何通过腾讯云 DevOps 专家（CloudQ）自动完成一次「资源容量水位 + FinOps 成本健康」综合巡检：自动盘点 CVM/Lighthouse/SCF 等计算资源的实例数、地域分布和利用率水位，识别停机闲置和低利用率实例，分析成本科目分布和环比变化，并给出 5 条具体的优化建议（含动作、预期节省金额和风险点），最终输出一份带可视化图表的专业 HTML 巡检报告。\n\n适合：云运维工程师、SRE、FinOps 负责人、技术管理者、需要定期做成本优化的团队", "description_en": "Manually reviewing cloud resource utilization and cost breakdowns every month is the most tedious recurring task for ops teams — it's easy to miss idle resources and hard to generate actionable optimization advice.\n\nThis case demonstrates how to use the Tencent Cloud DevOps Expert (CloudQ) to automatically perform a comprehensive 'Resource Capacity + FinOps Cost Health' inspection: inventorying CVM/Lighthouse/SCF compute resources by instance count, region, and utilization levels; identifying shutdown/idle and over-provisioned instances; analyzing cost category distribution with month-over-month trends; and delivering 5 concrete optimization recommendations (with actions, expected savings, and risk points) — all in a professional HTML report with visualizations.\n\nIdeal for: Cloud ops engineers, SREs, FinOps leads, tech managers, teams needing periodic cost optimization", "categories": [ "data", "dev" ], "tags": [ "FinOps", "云成本优化", "资源巡检", "容量水位", "闲置资源" ], "tags_en": [ "FinOps", "Cloud Cost Optimization", "Resource Inspection", "Capacity Planning", "Idle Resources" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "综合巡检报告", "label_en": "Comprehensive Inspection Report" } ], "prompt": "请基于当前账号的真实数据，帮我做一次「资源容量水位 + FinOps 成本健康」综合巡检报告。要求：\n\n一、资源容量水位\n1) 列出当前账号下主要计算资源（CVM、TKE、Lighthouse 等）的水位情况，给出实例数、地域分布、CPU/内存/磁盘的近期水位区间\n2) 标注高水位（>80%）与低水位（<20%）实例，分别给出资源 ID/名称/规格\n\n二、FinOps 成本分析\n1) 闲置或低利用率资源清单（停机/低 CPU/未挂载磁盘等），按潜在节省金额降序\n2) 主要成本科目分布（计算/存储/网络/数据库等）及环比变化提示\n3) 给出 3-5 条最具性价比的优化建议（每条要写明：动作、预期节省、风险点）\n\n输出格式要求：生成一份带可视化图表和专业排版的 HTML 报告，包含执行摘要、资源明细表格、成本占比图、优化建议卡片和行动项 Checklist。", "prompt_en": "Based on the real data of the current account, please help me run a comprehensive \"Resource Capacity + FinOps Cost Health\" inspection report. Requirements:\n\nI. Resource Capacity\n1) List the capacity status of the main compute resources under the current account (CVM, TKE, Lighthouse, etc.), giving instance counts, regional distribution, and recent utilization ranges for CPU/memory/disk\n2) Flag high-utilization (>80%) and low-utilization (<20%) instances, providing the resource ID/name/specification for each\n\nII. FinOps Cost Analysis\n1) A list of idle or low-utilization resources (shut down / low CPU / unattached disks, etc.), sorted in descending order by potential savings\n2) Distribution of major cost categories (compute / storage / network / database, etc.) and month-over-month change notes\n3) Provide 3-5 of the most cost-effective optimization recommendations (each must specify: action, expected savings, risk points)\n\nOutput format requirements: Generate an HTML report with visualizations and professional layout, including an executive summary, a resource detail table, a cost breakdown chart, optimization recommendation cards, and an action-item checklist.", "skills": [], "experts": [ { "id": "finops-expert", "name": "CloudQ 云成本治理专家", "description": "云成本治理专家。擅长账单分析、闲置资源识别、计费模式优化与成本分摊，驱动降本增效决策。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tWDkHfqhvPIY" }, { "id": "capacity-expert", "name": "CloudQ 容量规划专家", "description": "容量规划专家。擅长水位监控、容量预测与弹性伸缩策略制定，防止资源瓶颈与浪费。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_sHDgkLxdD1NK" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-25T12:24:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "title": "品种晨会简报", "subtitle": "开盘前一次看全单品种观点、评级与关键信号", "description": "研究员定期观点、周度评级、当日快评分散在不同材料里，挨个翻费时费力还怕漏掉关键信号。本案例把繁微Fiona的多源研究材料串成一条流水线，聚焦任一期货品种，产出一份信息密度高、全部真实可溯源的单页HTML晨会简报：从年度/季度定期观点到近11周周度评级色块，再到最新研报与当日快评的四维提炼（核心结论→关键数据→研究逻辑→风险），加上6-12项重点指标看板和共识/分歧/张力三列交叉比较，开盘前一次看完。适合每日开盘前的例行信息准备。", "artifact_type": "html", "prompt": "生成单文件HTML《{品种}晨会简报》，数据基准日期为{日期}。所有数据通过繁微Fiona MCP工具（finoview-reports skill）真实查询并注明来源与日期，禁止编造；若该品种的研报/周度评级/快评/定期观点四个维度全部为空即视为未覆盖，直接告知用户并给出替代建议，不生成报告。按以下步骤执行：\n第一步、研究员观点时间线：查询{品种}的年度/半年/季度定期观点（对应材料缺失时注明\"暂未覆盖\"；如存在产业链或相关品种材料，可单独列为\"相关参考\"，不得替代该品种观点）；拉取近11周周度评级并以色块呈现（空周留灰块）；附一句时间线解读。\n第二步、今日材料四维提炼：检索最新研报、本周周度评级、当日快评三类材料（多品种周报只取{品种}段落；周度评级查不到时尝试\"沪{品种简称}\"作为关键词；快评优先品种日度数据跟踪与点评），按核心结论→关键数据→研究逻辑→风险四个维度提炼。\n第三步、重点关注指标看板：展示6-12项实际查询到且与{品种}相关的指标，按\"指标/最新状态/方向信号/出处\"四列呈现。\n第四步、交叉比较：输出共识、分歧与张力三列表（比较维度/结论/依据；结论列只写短判断，依据列单独写明出处）。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿、震荡黄；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "投研", "晨会简报", "研报", "繁微Fiona" ], "id": "finoview-morning-brief", "title_en": "Variety Morning Briefing", "subtitle_en": "One page to see all variety views, ratings and key signals before market open", "description_en": "Researcher periodic views, weekly ratings, and intraday flash comments are scattered across different materials—reading them one by one is time-consuming and risks missing key signals. This case turns Finoview Fiona's multi-source research materials into a pipeline, focusing on any single futures variety, and produces an information-dense, fully-traceable single-page HTML morning briefing: from annual/quarterly periodic views to 11-week weekly rating color blocks, to four-dimensional extraction of the latest research reports and intraday flash comments (core conclusion → key data → research logic → risk), plus a 6-12 item key indicator dashboard and consensus/divergence/tension three-column cross-comparison, all readable before market open.", "tags_en": [ "Futures", "Investment Research", "Morning Briefing", "Research Report", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Variety} Morning Briefing' with data reference date {Date}. All data is queried via Finoview Fiona MCP tools (finoview-reports skill) with sources and dates noted; fabrication is prohibited. If all four dimensions (research report / weekly rating / flash comment / periodic view) for the variety are empty, treat as uncovered, inform the user and suggest alternatives without generating a report. Execute the following steps:\nStep 1, Researcher View Timeline: Query {Variety}'s annual/semi-annual/quarterly periodic views (mark 'not covered' when materials are missing; if supply chain or related variety materials exist, list separately as 'related reference', not as a substitute for the variety's own views); pull the latest 11 weeks of weekly ratings and display as color blocks (gray blocks for empty weeks); add a one-sentence timeline interpretation.\nStep 2, Today's Material Four-Dimensional Extraction: Search for the latest research reports, this week's weekly ratings, and intraday flash comments (for multi-variety weekly reports, only extract the {Variety} section; if weekly ratings are not found, try 'Shanghai {Variety abbreviation}' as keyword; flash comments prioritize daily variety data tracking and commentary), extract by core conclusion → key data → research logic → risk.\nStep 3, Key Indicator Dashboard: Display 6-12 indicators actually queried and related to {Variety}, presented in four columns: 'Indicator / Latest Status / Direction Signal / Source'.\nStep 4, Cross-Comparison: Output consensus, divergence, and tension three-column tables (comparison dimension / conclusion / basis; conclusion column only has short judgments, basis column separately notes sources).\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green, neutral yellow; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "晨会简报HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "评级异动捕手", "subtitle": "升级、降级、僵持与背离，板块评级边际信号一屏锁定", "description": "想捕捉板块评级的边际变化，上周和本周谁升级、谁降级、谁僵持不动，靠人工查阅既慢又容易看漏。本案例把繁微的跨周评级对比能力串成一条流水线，聚焦任一板块，产出一份全部真实可溯源的单页HTML异动扫描：升级/降级/新纳入/退出四类变动清单，每个变动品种近8周评级色带配研报快评原文摘句，连续6周以上僵持品种锁定，以及评级与快评背离品种的双边依据罗列，边际信号一屏锁定。", "artifact_type": "html", "prompt": "生成单文件HTML《{板块}评级异动扫描》，对比{上周日期}与{本周日期}两期评级。所有数据通过繁微Fiona MCP工具（finoview-reports skill）真实查询并注明来源与日期，禁止编造。板块口径：周度观点无宏观/能源/金工独立项，宏观→股指或国债、能源→化工、金工只用研报与快评。按以下步骤执行：\n第一步、本周评级变动清单：对比上周与本周的品种评级，分升级（震荡→看涨等）、降级（看涨→震荡等）、新纳入、退出四类列出，注明变动前后评级；无变动则如实写\"本周无评级变动\"。\n第二步、变动品种归因：每个变动品种给出近8周评级色带，并附支撑本次变动的研报或快评原文摘句+出处；找不到对应材料的如实写\"暂未见对应材料\"，不要推测原因。\n第三步、僵持品种：列出连续6周及以上评级未变的品种，标注锁定周数与当前评级。\n第四步、评级与快评背离：列出评级维持震荡、但本周快评方向性偏多或偏空的品种，罗列两边依据。\n第五步、品种池错位：只在快评有跟踪、未进入周度评级矩阵的品种，单独注明。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿、震荡黄；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "投研", "评级异动", "周度对比", "繁微Fiona" ], "id": "finoview-rating-scanner", "title_en": "Rating Change Scanner", "subtitle_en": "Upgrades, downgrades, stalemates and divergences—sector rating marginal signals locked on one screen", "description_en": "To capture marginal changes in sector ratings—who upgraded, who downgraded, who remained stalemate—manual review is slow and prone to missing signals. This case turns Finoview's cross-week rating comparison capability into a pipeline, focusing on any sector, and produces a fully-traceable single-page HTML change scan: upgrade/downgrade/new inclusion/exit four-category change list, each changed variety's 8-week rating color band with research report flash comment original text quotes, 6+ consecutive weeks stalemate varieties locked, and rating vs flash comment divergence varieties with bilateral evidence listed, marginal signals locked on one screen.", "tags_en": [ "Futures", "Investment Research", "Rating Changes", "Weekly Comparison", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Sector} Rating Change Scan', comparing {Last Week Date} and {This Week Date} two periods of ratings. All data is queried via Finoview Fiona MCP tools (finoview-reports skill) with sources and dates noted; fabrication is prohibited. Sector scope: weekly views have no independent macro/energy/quant items; macro → stock index or treasury bonds, energy → chemicals, quant uses only reports and flash comments. Execute the following steps:\nStep 1, This Week's Rating Change List: Compare last week and this week's variety ratings, categorize into upgrades (neutral → bullish, etc.), downgrades (bullish → neutral, etc.), new inclusions, and exits, noting pre and post change ratings; if no changes, honestly write 'no rating changes this week'.\nStep 2, Changed Variety Attribution: For each changed variety, provide an 8-week rating color band, and attach the research report or flash comment original text quotes + sources supporting this change; if no corresponding materials found, honestly write 'no corresponding materials found yet', do not speculate on reasons.\nStep 3, Stalemate Varieties: List varieties with unchanged ratings for 6 or more consecutive weeks, marking locked weeks and current rating.\nStep 4, Rating vs Flash Comment Divergence: List varieties where rating remains neutral but this week's flash comment direction is bullish or bearish, listing both sides' evidence.\nStep 5, Variety Pool Mismatch: Varieties only tracked in flash comments but not in the weekly rating matrix, noted separately.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green, neutral yellow; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "评级异动HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "研报深度剖析", "subtitle": "一篇深度报告读透：结论、逻辑链、风险与历史观点对照", "description": "想吃透一篇深度研报的完整逻辑，200字摘要远远不够：核心观点、关键数据、推演链、风险因素以及研究员过去的观点脉络都得拉出来对照，手动翻全文既耗时又容易只看到结论、忽略逻辑。本案例把繁微的研报全文检索与观点交叉引用能力串成一条流水线，聚焦任一品种的深度报告，产出一份全部真实可溯源的单页HTML深度剖析：报告概览、3-5条核心结论附原文摘句、关键数据\"数据→判断→依据\"三列逻辑链、逐条风险因素，最后串联该品种主要研究员历史观点对比（标注延续与转向）与近4周数据周报逻辑递进脉络，一篇报告读透读深。", "artifact_type": "html", "prompt": "生成单文件HTML《{品种}研报深读_{日期}》。所有内容通过繁微Fiona MCP工具（finoview-reports skill）真实查询，禁止编造。按以下步骤执行：\n第零步、选报告：先按{品种}搜索研报，筛选深度类报告（深度报告/月度报告/专题报告/季度观点/半年报/年报/策略报告等），优先选取最近一期作为剖析对象；若全部为数据周报、无深度类报告，则如实告知\"暂无深度类报告\"并停止，不生成HTML。此步骤不生成单独章节。\n第一步、研报概览：报告标题、作者、发布日期、报告类型、关联品种、评级与预测标签。\n第二步、核心观点提炼：从全文中提炼3-5条核心结论，每条附原文摘句（带引号）与所在章节。\n第三步、关键数据与逻辑链：列出报告中的关键数据点（价格区间、库存、加工费、供需平衡等）及其推演逻辑，按\"数据→判断→依据\"三列呈现。\n第四步、风险提示：先逐条列出研报原文明确提出的风险因素并附原文摘句；如基于报告逻辑补充潜在风险，应单列\"补充观察\"并明确标注为AI归纳，不与原文风险提示混淆。\n第五步、与研究员历史观点对比：拉取该品种主要研究员的全部定期观点，按时间排列，对比本篇研报核心判断与过去定期观点是否一致——标注延续点与转向点；如评级发生变化，注明转向前后的评级与依据。\n第六步、近期周报追踪：取该品种最近4周数据周报，每周提炼核心变化一句话，观察从周报到深度报告的逻辑递进脉络。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿、震荡黄；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "投研", "研报精读", "观点对比", "繁微Fiona" ], "id": "finoview-report-analysis", "title_en": "Research Report Deep Analysis", "subtitle_en": "Read a single in-depth report thoroughly: conclusions, logic chain, risks and historical view comparison", "description_en": "To fully understand the complete logic of an in-depth research report, a 200-word summary is far from enough: core views, key data, reasoning chain, risk factors, and the researcher's past view trajectory all need to be pulled out for comparison. Manual reading is time-consuming and tends to only see conclusions while ignoring logic. This case turns Finoview's full-text research report search and cross-referencing capability into a pipeline, focusing on a deep report of any variety, and produces a fully-traceable single-page HTML deep analysis: report overview, 3-5 core conclusions with original text quotes, key data 'data → judgment → basis' three-column logic chain, item-by-item risk factors, and finally connecting the main researcher's historical view comparison (marking continuations and turns) with the last 4 weeks of data weekly report logic progression, reading one report thoroughly and deeply.", "tags_en": [ "Futures", "Investment Research", "Report Reading", "View Comparison", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Variety} Report Deep Read_{Date}'. All content is queried via Finoview Fiona MCP tools (finoview-reports skill); fabrication is prohibited. Execute the following steps:\nStep 0, Select Report: First search for {Variety} research reports, filter for deep-type reports (deep report / monthly report / special report / quarterly view / semi-annual report / annual report / strategy report, etc.), prioritize the most recent one as the analysis target; if all are data weekly reports with no deep-type reports, honestly state 'no deep-type reports available' and stop without generating HTML. This step does not generate a separate section.\nStep 1, Report Overview: Report title, author, publication date, report type, related varieties, rating and forecast labels.\nStep 2, Core View Extraction: Extract 3-5 core conclusions from the full text, each with original text quotes (in quotation marks) and chapter location.\nStep 3, Key Data and Logic Chain: List key data points in the report (price ranges, inventory, processing fees, supply-demand balance, etc.) and their reasoning logic, presented in 'data → judgment → basis' three columns.\nStep 4, Risk Disclosure: First list risk factors explicitly stated in the report's original text with original text quotes; if supplementary potential risks are based on report logic, list separately as 'supplementary observation' and clearly mark as AI summary, not confusing with original risk disclosure.\nStep 5, Historical View Comparison with Researcher: Pull all periodic views of the variety's main researcher, arrange by time, compare this report's core judgments with past periodic views—mark continuation points and turning points; if ratings changed, note the ratings before and after the turn and the basis.\nStep 6, Recent Weekly Report Tracking: Take the variety's most recent 4 weeks of data weekly reports, extract one sentence of core change per week, observing the logic progression from weekly reports to deep reports.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green, neutral yellow; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "研报深读HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "板块周度总览", "subtitle": "一页俯瞰板块评级矩阵、研究员判断与本周主线", "description": "想俯瞰一整个板块的本周全貌，各品种评级、研究员判断、新发研报和快评动态散落在数十份材料里，逐篇读来不及，评级与快评方向还经常打架。本案例把繁微的板块级材料聚合能力串成一条流水线，聚焦任一板块，产出一份全部真实可溯源的单页HTML周报：本周板块全部品种评级矩阵、每位研究员的核心判断、本周正式研报与快评动态分类呈现，最后用一句话点出本周主线、最核心品种信号和评级与快评的背离，板块全局一次掌握。", "artifact_type": "html", "prompt": "生成单文件HTML《{板块}周报》，覆盖{日期范围}。所有数据通过繁微Fiona MCP工具（finoview-reports skill）真实查询并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、周度板块评级矩阵：查询本周{板块}全部品种评级（看涨/震荡/看跌）。{板块}无周度评级、或某品种只在快评有跟踪而未进评级矩阵的，均须注明。板块口径：周度观点无宏观/能源/金工独立项，宏观→股指或国债、能源→化工、金工只用研报+快评；三源不一致时以研究员所属分组为准。\n第二步、研究员核心观点：按人提炼一句本周核心判断，列出姓名/职级/负责品种。\n第三步、本周研报：列出本周发布的正式研报；本周无新研报时明确注明，并补充最近一版相关研报作为参考，标明实际发布日期。\n第四步、快评动态：重要快评优先列出；数据中心自动市场跟踪类（繁微长图系列）单列，不与研究员观点混淆。\n第五步、本周主线：一句话总结本周主线+最核心品种信号+关键变量，并点出评级与快评的背离。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿、震荡黄；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "投研", "周报", "板块", "繁微Fiona" ], "id": "finoview-sector-weekly", "title_en": "Sector Weekly Overview", "subtitle_en": "One page to view sector rating matrix, researcher judgments and weekly mainline", "description_en": "To overview an entire sector's weekly performance, individual variety ratings, researcher judgments, newly published reports and flash comment dynamics are scattered across dozens of materials—reading them all is impossible, and ratings and flash comments often conflict. This case turns Finoview's sector-level material aggregation capability into a pipeline, focusing on any sector, and produces a fully-traceable single-page HTML weekly report: this week's full variety rating matrix for the sector, each researcher's core judgment, this week's formal research reports and flash comment dynamics categorized, and finally a one-sentence summary of the weekly mainline, the most core variety signal, and the divergence between ratings and flash comments, giving a complete sector overview in one page.", "tags_en": [ "Futures", "Investment Research", "Weekly Report", "Sector", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Sector} Weekly Report' covering {Date Range}. All data is queried via Finoview Fiona MCP tools (finoview-reports skill) with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Weekly Sector Rating Matrix: Query this week's all variety ratings for {Sector} (bullish/neutral/bearish). Note if {Sector} has no weekly ratings, or if a variety is only tracked in flash comments but not in the rating matrix. Sector scope: weekly views have no independent macro/energy/quant items; macro → stock index or treasury bonds, energy → chemicals, quant uses only reports + flash comments; when three sources disagree, use the researcher's group as authoritative.\nStep 2, Researcher Core Views: Extract one-sentence core judgment per person, listing name/title/responsible varieties.\nStep 3, This Week's Research Reports: List formal research reports published this week; if no new reports this week, clearly note and supplement the most recent related report as reference, with actual publication date.\nStep 4, Flash Comment Dynamics: Prioritize important flash comments; data center automated market tracking series (Finoview long chart series) listed separately, not confused with researcher views.\nStep 5, Weekly Mainline: One sentence summarizing the weekly mainline + most core variety signal + key variables, and point out divergence between ratings and flash comments.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green, neutral yellow; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "板块周报HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "全市场资金共振分析", "subtitle": "资金×价格×量能三维共振，全市场强弱一榜排序", "description": "\"最近资金在往哪里走？哪些品种的量价配合最好？\"——这类全市场横向扫描问题逐个品种去查根本做不完。本案例把繁微Fiona的资金流向+行情数据能力批量调用，覆盖各大商品板块主要品种，产出一份\"资金全景→量价共振矩阵→板块评级→TOP5排名\"四层穿透的单页HTML：每个品种的资金变化率、价格涨跌幅、成交量/持仓量变化全部量化，自动分类为量价齐升/缩量下跌/价涨资金逃等共振模式，五大板块分别给出评级，最终排出TOP5共振品种和资金撤离警示品种。", "artifact_type": "html", "prompt": "生成单文件HTML《全市场资金流向与共振分析报告》，统计区间为{日期范围}。所有数据通过繁微Fiona Skill真实查询并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、全市场资金流向全景：查询商品板块主要品种近两周日度沉淀资金与资金流向，按板块汇总累计净流向，识别资金持续流入与流出的品种。\n第二步、量价配合分析：对每个品种同步查询主力合约行情，进行\"资金×价格×成交持仓\"三维共振分析，分类为：量价齐升（资金流入+价格上涨+放量）、量价背离（资金流入但价格下跌）、缩量下跌（资金流出+价格下跌+缩量）等模式。\n第三步、板块共振扫描：以板块为单位，判断资金是否呈现板块级别的同向运动，识别系统性资金迁移方向，并说明一致程度和主要背离品种。\n第四步、核心结论：列出资金与量价共振最显著的前5品种（附详细数据支撑）以及资金撤离警示品种。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；共振矩阵用色块呈现，配色遵循看涨红、看跌绿；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" } ], "categories": [ "data" ], "tags": [ "期货", "资金流向", "量价分析", "全市场扫描", "繁微Fiona" ], "id": "fiona-capital-resonance", "title_en": "Market Capital Resonance Analysis", "subtitle_en": "Capital × price × volume three-dimensional resonance, market-wide strength ranking", "description_en": "'Where is capital flowing recently? Which varieties have the best price-volume coordination?'—such market-wide horizontal scanning questions are impossible to complete by checking variety by variety. This case batch-calls Finoview Fiona's capital flow + market data capabilities, covering major varieties across commodity sectors, producing a 'capital panorama → price-volume resonance matrix → sector rating → TOP5 ranking' four-layer penetrating single-page HTML: each variety's capital change rate, price change, volume/position change all quantified, auto-classified into price-volume surge/shrinking decline/price rise capital escape and other resonance patterns, five sectors each given a rating, ultimately ranking TOP5 resonance varieties and capital retreat warning varieties.", "tags_en": [ "Futures", "Capital Flow", "Price-Volume Analysis", "Market Scan", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML 'Market Capital Flow and Resonance Analysis Report', with statistics period {Date Range}. All data is queried via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Market Capital Flow Panorama: Query major commodity sector varieties' daily settled capital and capital flow for the past two weeks, summarize cumulative net flow by sector, identifying varieties with sustained capital inflow and outflow.\nStep 2, Price-Volume Coordination Analysis: For each variety, synchronously query main contract market data, conduct 'capital × price × position' three-dimensional resonance analysis, classified as: price-volume surge (capital inflow + price rise + volume increase), price-volume divergence (capital inflow but price decline), shrinking decline (capital outflow + price decline + volume shrink), and other patterns.\nStep 3, Sector Resonance Scan: By sector, determine whether capital shows sector-level co-directional movement, identify systemic capital migration direction, and explain consistency degree and main divergent varieties.\nStep 4, Core Conclusions: List the top 5 varieties with the most significant capital and price-volume resonance (with detailed data support) and capital retreat warning varieties.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; resonance matrix presented with color blocks, color scheme follows bullish red, bearish green; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "资金共振分析HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "事件驱动行情追踪", "subtitle": "发生了什么→价格怎么走→研究员怎么看，三环紧扣", "description": "黄金这类受宏观事件驱动的品种，热点事件、行情反应、研究员观点往往散落在不同系统里，查询费时费力还容易遗漏关键转折点。本案例把繁微Fiona的市场数据+研报检索两条能力线打通，聚焦任一事件驱动型品种，产出一份从\"发生了什么→价格怎么走→研究员怎么看\"三环紧扣的单页HTML：热点事件按时间线排列，行情关键节点全部标注，研究员研报的核心观点串联成完整叙事弧线。", "artifact_type": "html", "prompt": "生成单文件HTML《{品种}市场热点追踪与行情解读》，回溯区间为{日期范围}。所有数据通过繁微Fiona Skill真实查询并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、热点梳理：检索近一周{品种}相关重要资讯与事件（如宏观讲话、就业数据、地缘局势等），按时间线排列，每条标注来源和日期。\n第二步、行情走势：查询{品种}主力合约近两周日线行情，标注区间涨跌幅与关键转折点，并对应到相关事件。\n第三步、主力合约快照：查询当前{品种}实时快照（最新价、涨跌幅、成交量、持仓量）。\n第四步、研报观点提炼：搜索东证衍生品研究院{品种}相关最新研报，提炼核心观点，标注报告ID、作者与日期，区分原文与AI归纳；最后综合多份研报形成\"事件→行情→观点\"的叙事链。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" }, { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" }, { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "事件驱动", "黄金", "热点追踪", "繁微Fiona" ], "id": "fiona-event-driven-tracking", "title_en": "Event-Driven Market Tracking", "subtitle_en": "What happened → how price moved → what researchers think, three rings tightly linked", "description_en": "For event-driven varieties like gold, hot events, market reactions and researcher views are often scattered across different systems—querying is time-consuming and risks missing key turning points. This case connects Finoview Fiona's market data + research report search capabilities, focusing on any event-driven variety, producing a 'what happened → how price moved → what researchers think' three-ring tightly linked single-page HTML: hot events arranged by timeline, all key market nodes marked, researcher research report core views connected into a complete narrative arc.", "tags_en": [ "Futures", "Event-Driven", "Gold", "Hotspot Tracking", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Variety} Market Hotspot Tracking and Market Interpretation', with lookback period {Date Range}. All data is queried via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Hotspot Sorting: Search for important information and events related to {Variety} in the past week (e.g., macro speeches, employment data, geopolitical situations, etc.), arranged by timeline, each marked with source and date.\nStep 2, Market Trend: Query {Variety} main contract daily market data for the past two weeks, marking range changes and key turning points, and corresponding to related events.\nStep 3, Main Contract Snapshot: Query current {Variety} real-time snapshot (latest price, change, volume, position).\nStep 4, Research Report View Extraction: Search Oriental Securities Derivatives Research Institute {Variety} related latest research reports, extract core views, mark report ID, author and date, distinguish original text from AI summary; finally synthesize multiple research reports to form a 'event → market → view' narrative chain.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "事件驱动追踪HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "全市场期货早报", "subtitle": "过去24小时重要事件按六大板块归档，附盘面验证", "description": "每天早上打开终端，新闻铺天盖地——地缘冲突、库存报告、政策发布、天气灾害，哪些真正影响你的品种？哪些只是噪音？本案例依据繁微Fiona的新闻检索能力，产出一份单页HTML早报：全面检索过去24小时可能影响商品期货的重要事件，按宏观、能源化工、黑色、有色、农产品和贵金属六大板块组织，每个事件标注影响方向、期限和反向因素；最后查询重点品种行情，检查事件主线是否已在盘面兑现。适合作为每日早报、晚报或定时推送的标准化信息框架。", "artifact_type": "html", "prompt": "生成单文件HTML《全市场期货早报》，覆盖{日期}过去24小时。通过繁微Fiona Skill真实检索资讯与行情数据，注明来源与日期，禁止编造。按以下步骤执行：\n第一步、市场概览：概括过去24小时影响商品期货的核心主题和市场情绪。\n第二步、板块事件：按宏观与政策、能源化工、黑色系、有色金属、农产品、贵金属六大板块，逐一列出重要事件，标注发生时间、相关品种、影响方向和期限。\n第三步、盘面验证：查询重点品种行情快照，检查事件主线是否已在盘面反映。\n第四步、当日观察点：列出未来需重点关注的事件和数据发布。\n数据口径：行情、成交持仓、资金、研报与研究观点优先使用繁微Fiona已接入数据和资料；新闻资讯通过公开网络检索获取。各项内容注明来源与时间，不使用模型记忆补充最新事实。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；影响方向用多空配色（利多红、利空绿、中性灰）。", "skills": [ { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" }, { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" } ], "categories": [ "data" ], "tags": [ "期货", "早报", "资讯整理", "事件驱动", "繁微Fiona" ], "id": "fiona-futures-morning", "title_en": "Futures Market Morning Brief", "subtitle_en": "Past 24 hours key events archived by six sectors, with market verification", "description_en": "Opening the terminal every morning, news is overwhelming—geopolitical conflicts, inventory reports, policy releases, weather disasters—which ones truly affect your varieties? Which are just noise? This case uses Finoview Fiona's news retrieval capability to produce a single-page HTML morning brief: comprehensively search for important events from the past 24 hours that may affect commodity futures, organized by macro, energy chemicals, ferrous, non-ferrous, agricultural products and precious metals six sectors, each event marked with impact direction, duration and counter-factors; finally query key variety markets to check whether the event mainline has been reflected in the market. Suitable as a daily morning report, evening report or scheduled push standardized information framework.", "tags_en": [ "Futures", "Morning Briefing", "News Integration", "Event-Driven", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML 'Futures Market Morning Brief' covering the past 24 hours of {Date}. Search information and market data via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Market Overview: Summarize the core themes and market sentiment affecting commodity futures in the past 24 hours.\nStep 2, Sector Events: By macro and policy, energy chemicals, ferrous, non-ferrous metals, agricultural products, precious metals six sectors, list important events one by one, marking occurrence time, related varieties, impact direction and duration.\nStep 3, Market Verification: Query key variety market snapshots to check whether the event mainline has been reflected in the market.\nStep 4, Today's Watch Points: List events and data releases to focus on in the future.\nData caliber: Market, trading positions, capital, research reports and research views prioritize Finoview Fiona integrated data and materials; news information is retrieved via public web search. All content notes sources and times, no model memory used to supplement latest facts.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; impact direction uses bullish/bearish color scheme (bullish red, bearish green, neutral gray). UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "期货早报HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "地缘事件纵深研究", "subtitle": "碎片化报道还原为证据分级的研究链，附品种传导研判", "description": "地缘冲突报道往往是碎片化的——官方声明、媒体解读、市场传闻混在一起，很难区分哪些是已确认事实、哪些是风险溢价。本案例依据繁微Fiona的新闻检索能力，把碎片化报道还原为一条完整研究链：区分官方确认、权威媒体报道、相互矛盾信息和市场风险预期四个证据层级，将原油、贵金属、航运等品种按传导机制、相关度和时间尺度分别研判，行情数据只用于验证市场反应。该模式适用于地缘冲突、贸易摩擦、重大政策、极端天气和产业事故等持续演化主题。", "artifact_type": "html", "prompt": "生成单文件HTML《{事件主题}纵深研究》。通过繁微Fiona期货资讯整理skill真实检索事件信息并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、事件状态概览：概括当前{事件主题}最新状态和核心矛盾，区分已落地变化与未兑现风险。\n第二步、关键时间线：从早到晚系统梳理{事件主题}相关事件，标注每个事件的证据强度（官方确认/权威媒体报道/待核实信息/市场预期）。\n第三步、品种传导分析：分析事件对{相关品种列表}的传导路径，标注相关度和时间尺度。\n第四步、市场验证：查询相关品种行情走势，验证市场定价程度；检索东证衍生品研究院相关研报观点。\n第五步、后续信号：列出可能改变局势走向的关键观察点。\n数据口径：行情、成交持仓、资金、研报与研究观点优先使用繁微Fiona已接入数据和资料；新闻资讯通过公开网络检索获取。各项内容注明来源与时间，不使用模型记忆补充最新事实。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖。", "skills": [ { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" }, { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" }, { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "地缘政治", "事件研究", "传导分析", "繁微Fiona" ], "id": "fiona-geopolitical-research", "title_en": "Geopolitical Event In-Depth Research", "subtitle_en": "Fragmented reports restored to evidence-graded research chain, with variety transmission assessment", "description_en": "Geopolitical conflict reporting is often fragmented—official statements, media interpretations, market rumors mixed together, making it hard to distinguish confirmed facts from risk premiums. This case uses Finoview Fiona's news retrieval capability to restore fragmented reporting into a complete research chain: distinguishing four evidence levels of official confirmation, authoritative media reports, contradictory information and market risk expectation, assessing crude oil, precious metals, shipping and other varieties by transmission mechanism, relevance and time scale, with market data only used to verify market reactions. This pattern applies to geopolitical conflicts, trade frictions, major policies, extreme weather and industrial accidents and other continuously evolving themes.", "tags_en": [ "Futures", "Geopolitics", "Event Research", "Transmission Analysis", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Event Topic} In-Depth Research'. Search event information via Finoview Fiona futures news integration skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Event Status Overview: Summarize the current {Event Topic} latest status and core contradictions, distinguishing landed changes from unrealized risks.\nStep 2, Key Timeline: Systematically sort out {Event Topic} related events from early to late, marking each event's evidence strength (official confirmation / authoritative media report / unverified information / market expectation).\nStep 3, Variety Transmission Analysis: Analyze the event's transmission path to {Related Variety List}, marking relevance and time scale.\nStep 4, Market Verification: Query related variety market trends to verify market pricing degree; search Oriental Securities Derivatives Research Institute related research report views.\nStep 5, Forward Signals: List key observation points that may change the situation's direction.\nData caliber: Market, trading positions, capital, research reports and research views prioritize Finoview Fiona integrated data and materials; news information is retrieved via public web search. All content notes sources and times, no model memory used to supplement latest facts.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "地缘事件研究HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "行情异动归因", "subtitle": "确认异动→排查事件→归因矩阵→保留未解释部分", "description": "某个品种突然大涨或大跌，市场第一反应是找原因——但事后归因最常见的陷阱是：时间对不上、因果链不清晰、证据强度不标注、解释不了的部分硬凑理由。本案例聚焦单一合约在指定交易日的大幅波动，产出一份四步递进的单页HTML归因报告：先通过行情数据确认异动启动时点、涨幅、成交量和增仓情况，再搜索前后72小时的地缘、政策、供给、需求和物流事件，逐项判断时间匹配度与作用机制，将驱动因素分为已确认、高度可能和可能相关三个层级，并把期货涨幅显著超出现货变化的部分保留为预期溢价、不强行归因。既适合异动当日快速核查，也适合事后复盘。", "artifact_type": "html", "prompt": "生成单文件HTML《{合约}行情异动归因 · {异动日期}》。通过繁微Fiona Skill真实检索，注明来源与日期，禁止编造。按以下步骤执行：\n第一步、异动确认：查询{合约}在{异动日期}及前后各3个交易日的日线行情，确认涨幅、成交量和持仓量变化；对比同期现货指标（如{现货指数}），区分现货驱动与预期溢价。\n第二步、候选事件排查：搜索{异动日期}前3天后3天期间的地缘、政策、供需、物流等候选事件，标注发生时间和来源。\n第三步、归因矩阵：逐项评估事件与行情的时间匹配度和传导机制，按证据强度分级（已确认/高度可能/可能相关），形成归因表格。\n第四步、未解释部分：标注期货涨幅中暂无事件支撑的预期溢价部分，列出后续关注信号。\n数据口径：行情、成交持仓、资金、研报与研究观点优先使用繁微Fiona已接入数据和资料；新闻资讯通过公开网络检索获取。各项内容注明来源与时间，不使用模型记忆补充最新事实。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿。", "skills": [ { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" }, { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" } ], "categories": [ "data" ], "tags": [ "期货", "异动归因", "复盘", "事件排查", "繁微Fiona" ], "id": "fiona-market-attribution", "title_en": "Market Move Attribution", "subtitle_en": "Confirm anomaly → investigate events → attribution matrix → preserve unexplained portion", "description_en": "When a variety suddenly surges or plunges, the market's first reaction is to find reasons—but the most common pitfall in after-the-fact attribution is: timing mismatch, unclear causal chain, unlabeled evidence strength, and forcing explanations for unexplainable parts. This case focuses on a single contract's large move on a specified trading day, producing a four-step progressive single-page HTML attribution report: first confirm the anomaly start time, magnitude, volume and position changes through market data, then search for geopolitical, policy, supply, demand and logistics events within 72 hours before and after, assess time-matching and mechanism for each, classify drivers into confirmed, highly likely and possibly related three levels, and retain the portion of futures gains significantly exceeding spot changes as expectation premium without forcing attribution. Suitable for both intraday quick checks and post-facto review.", "tags_en": [ "Futures", "Move Attribution", "Review", "Event Investigation", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Contract} Market Move Attribution - {Move Date}'. Search via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Anomaly Confirmation: Query {Contract}'s daily market data for {Move Date} and 3 trading days before and after, confirming magnitude, volume and position changes; compare with spot indicators (e.g., {Spot Index}) for the same period, distinguishing spot-driven from expectation premium.\nStep 2, Candidate Event Investigation: Search for geopolitical, policy, supply-demand, logistics and other candidate events within 3 days before and 3 days after {Move Date}, marking occurrence time and source.\nStep 3, Attribution Matrix: Assess each event's time-matching and transmission mechanism with the market, classify by evidence strength (confirmed / highly likely / possibly related), forming an attribution table.\nStep 4, Unexplained Portion: Mark the expectation premium portion of futures gains with no event support, list follow-up observation signals.\nData caliber: Market, trading positions, capital, research reports and research views prioritize Finoview Fiona integrated data and materials; news information is retrieved via public web search. All content notes sources and times, no model memory used to supplement latest facts.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "异动归因HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "板块热度速览", "subtitle": "先找当日最热板块，再下钻看板块内部到底发生了什么", "description": "每天打开行情软件，涨跌品种眼花缭乱——到底哪个板块在领涨？热度是昙花一现还是资金持续流入？板块内部是龙头一枝独秀还是普涨格局？本案例把繁微Fiona的板块排名+品种排名+资金流向+行情数据四条能力线打通，产出一份\"板块热度排名→热门板块下钻→板块内品种全景扫描\"三层递进的单页HTML报告：先锁定当日涨跌幅、成交额和资金流入最突出的板块，再逐一扫描板块内各品种表现并区分龙头与拖后腿品种，最后梳理产业链上下游最新事件动态，一份报告回答\"今天什么板块最热\"和\"这个板块里面到底发生了什么\"。", "artifact_type": "html", "prompt": "生成单文件HTML《全市场板块热度扫描与下钻分析报告》，数据基准日期为{日期}。所有数据通过繁微Fiona Skill真实查询并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、全市场板块热度排名：查询当日各商品板块的涨跌幅、成交额和资金流向排名，分别列出各维度排名。结合三个维度是否同向判断板块热度：若涨跌幅、成交额和资金流向均指向同一板块最热，则明确该板块为当日最热；若各维度结论不一致，应分别说明差异，不强行给出统一热度结论。\n第二步、热门板块下钻：对热度最高的板块，逐一扫描板块内各品种的主力合约行情，查询各品种近两日沉淀资金和资金流向，并列出成交量排名居前的会员席位；按品种表现排序，区分板块内领涨龙头、跟涨品种和拖后品种。\n第三步、板块内产业链动态：整理该热门板块产业链近期重要资讯事件，按品种归类，每条标注来源和日期，分析事件对板块热度的支撑或削弱作用。\n第四步、综合研判：基于上述三个维度的数据，分别总结该板块在涨跌表现、成交活跃度和资金变化方面的特征，标注板块内需重点关注的品种和后续观察信号。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" }, { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" } ], "categories": [ "data" ], "tags": [ "期货", "板块热度", "资金流向", "产业链", "繁微Fiona" ], "id": "fiona-sector-heat", "title_en": "Sector Heat Quick View", "subtitle_en": "First find the day's hottest sector, then drill down to see what happened inside", "description_en": "Opening the market software every day, ups and downs are dazzling—which sector is leading? Is the heat a flash in the pan or sustained capital inflow? Within the sector, is it a single leader or a broad rally? This case connects Finoview Fiona's sector ranking + variety ranking + capital flow + market data four capability lines, producing a 'sector heat ranking → hot sector drill-down → sector internal variety panorama scan' three-layer progressive single-page HTML report: first lock the sectors with the most outstanding price change, turnover and capital inflow for the day, then scan each variety's performance within the sector and distinguish leaders and laggards, finally sort out the latest supply chain event dynamics, one report answering 'what sector is hottest today' and 'what happened inside this sector'.", "tags_en": [ "Futures", "Sector Heat", "Capital Flow", "Supply Chain", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML 'Market Sector Heat Scan and Drill-Down Analysis Report', with data reference date {Date}. All data is queried via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Market Sector Heat Ranking: Query the day's sector price change, turnover and capital flow rankings, listing rankings for each dimension. Combined judgment of whether three dimensions are co-directional: if price change, turnover and capital flow all point to the same sector as hottest, then clearly identify that sector as the day's hottest; if each dimension's conclusions are inconsistent, explain differences separately without forcing a unified heat conclusion.\nStep 2, Hot Sector Drill-Down: For the hottest sector, scan each variety's main contract market data one by one, query each variety's past two days settled capital and capital flow, and list member seats with top volume rankings; sort by variety performance, distinguishing sector leaders, followers and laggards.\nStep 3, Sector Supply Chain Dynamics: Sort out the hot sector's supply chain recent important information events, grouped by variety, each marked with source and date, analyzing events' support or weakening of sector heat.\nStep 4, Comprehensive Assessment: Based on the above three dimensions of data, respectively summarize the sector's characteristics in price performance, trading activity and capital changes, marking varieties within the sector that need focus and follow-up observation signals.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "板块热度HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "产业链事件交叉扫描", "subtitle": "原料端→冶炼端→成材端，产业链事件传导一链打通", "description": "黑色系品种之间的联动关系错综复杂——铁矿涨价是否因飓风导致发货骤降？螺纹需求走弱是地产数据不及预期还是季节性因素？单一品种的事件分析往往只能看到局部，产业链级别的交叉扫描才能揭示真正的驱动逻辑。本案例以繁微期货资讯整理skill为主引擎，聚焦{产业链}多品种的事件联动，产出一份三层穿透的单页HTML报告：沿原料端到冶炼端再到成材端构建事件传导链，辅以盘面价格与品种价差验证，再检索研究员观点进一步判断。该模式适合黑色系、能化链、有色冶炼链等具有明确上下游关系的品种组合。", "artifact_type": "html", "prompt": "生成单文件HTML《{产业链}事件驱动交叉扫描》，回溯区间为{日期范围}。通过繁微Fiona期货资讯整理skill真实检索事件信息并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、产业链传导全景：沿原料端、冶炼端、成材端三层链条，展示{产业链品种列表}近期核心事件的上下游传导关系。\n第二步、原料端：梳理矿山事故、天气影响、发货量变化、港口库存等事件。\n第三步、冶炼端：梳理限产政策、环保检查、利润变化、高炉开工率等事件，分析对上下游的双向传导。\n第四步、成材端：梳理地产政策、基建开工、出口贸易、制造业需求等事件，标注对各成材品种的差异化影响。\n第五步、盘面验证与观点参照：查询各品种行情和品种价差（如螺矿比、卷螺差），检索东证衍生品研究院{产业链}板块最新研报观点。\n数据口径：行情、成交持仓、资金、研报与研究观点优先使用繁微Fiona已接入数据和资料；新闻资讯通过公开网络检索获取。各项内容注明来源与时间，不使用模型记忆补充最新事实。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；配色遵循看涨红、看跌绿。", "skills": [ { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" }, { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" }, { "id": "finoview-reports", "name": "繁微研报检索", "description": "繁微Fiona研报全文检索与观点交叉引用，支持定期观点、周度评级、快评和深度研报查询", "slug": "finoview-reports" } ], "categories": [ "data" ], "tags": [ "期货", "产业链", "黑色系", "事件驱动", "繁微Fiona" ], "id": "fiona-supply-chain-scan", "title_en": "Supply Chain Event Cross-Scan", "subtitle_en": "Raw material → smelting → finished product, supply chain event transmission in one chain", "description_en": "The linkage between ferrous series varieties is complex—did iron ore prices rise due to cyclone-caused shipment drops? Is rebar demand weakening due to real estate data missing expectations or seasonal factors? Single-variety event analysis often only sees the local picture; supply chain-level cross-scanning can reveal the true driving logic. This case uses Finoview Fiona futures news integration skill as the main engine, focusing on {Supply Chain} multi-variety event linkage, producing a three-layer penetrating single-page HTML report: building an event transmission chain along raw material to smelting to finished product, supplemented by market prices and variety spreads for verification, then searching researcher views for further judgment. This pattern is suitable for ferrous, chemical chain, non-ferrous smelting chain, and other variety combinations with clear upstream-downstream relationships.", "tags_en": [ "Futures", "Supply Chain", "Ferrous", "Event-Driven", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Supply Chain} Event-Driven Cross-Scan', with lookback period {Date Range}. Search event information via Finoview Fiona futures news integration skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Supply Chain Transmission Panorama: Along the raw material, smelting, and finished product three-layer chain, display the upstream-downstream transmission relationship of {Supply Chain Variety List} recent core events.\nStep 2, Raw Material End: Sort out mine accidents, weather impacts, shipment changes, port inventory and other events.\nStep 3, Smelting End: Sort out production restriction policies, environmental inspections, profit changes, blast furnace operating rate and other events, analyzing bidirectional transmission to upstream and downstream.\nStep 4, Finished Product End: Sort out real estate policies, infrastructure starts, export trade, manufacturing demand and other events, marking differentiated impacts on each finished product variety.\nStep 5, Market Verification and View Reference: Query each variety's market and variety spread (e.g., rebar-ore ratio, HRB-coil spread), search Oriental Securities Derivatives Research Institute {Supply Chain} sector latest research report views.\nData caliber: Market, trading positions, capital, research reports and research views prioritize Finoview Fiona integrated data and materials; news information is retrieved via public web search. All content notes sources and times, no model memory used to supplement latest facts.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; color scheme follows bullish red, bearish green. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "产业链事件扫描HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "title": "品种市场全景扫描", "subtitle": "行情、资金、持仓、热点四维一屏看完单品种", "description": "想全面了解一个期货品种近期状况，行情走势、资金流向、会员持仓、产业链热点各个维度要单独查，数据来源广泛还总担心不权威。本案例把繁微Fiona的市场数据能力串成一条流水线，聚焦任一品种，产出一份信息密度高、全部真实可溯源的单页HTML：从近两周日线行情到盘中实时快照，再到资金大进大出的异常波动和前十会员持仓排名，最后串联产业链热点（矿山投产、库存变化、宏观扰动等），一次看完。", "artifact_type": "html", "prompt": "生成单文件HTML《{品种}期货市场全景扫描》，数据基准日期为{日期}。所有数据通过繁微Fiona Skill真实查询并注明来源与日期，禁止编造。按以下步骤执行：\n第一步、行情走势：查询{品种}主力合约近两周日线行情，包含开盘价、收盘价、涨跌幅、成交量和持仓量，并附区间涨跌小结。\n第二步、资金流向：查询{品种}近两周沉淀资金与资金流向日度数据，分析资金进出趋势，标注异常波动日。\n第三步、会员持仓：查询{品种}主力合约成交量排名前十会员席位，列出会员简称与成交量变化。\n第四步、热点梳理：整理{品种}产业链近期重要资讯（包括矿山投产、库存变化、宏观扰动等），每条标注来源和日期。\n输出格式要求：单文件HTML，全部样式内联、无外部依赖；顶部提供关键指标看板（最新价/涨跌幅/沉淀资金/持仓量等）；配色遵循看涨红、看跌绿；所有来源注明\"繁微Fiona/东证衍生品研究院\"及具体日期。", "skills": [ { "id": "fiona-market-info", "name": "繁微市场数据", "description": "繁微Fiona市场数据查询，支持行情走势、资金流向、会员持仓和板块热度等数据检索", "slug": "fiona-market-info" }, { "id": "fiona-futures-news", "name": "繁微期货资讯整理", "description": "繁微Fiona期货资讯检索与整理，支持全市场事件扫描、产业链事件追踪和地缘事件研究", "slug": "fiona-futures-news" } ], "categories": [ "data" ], "tags": [ "期货", "行情", "资金流向", "会员持仓", "繁微Fiona" ], "id": "fiona-variety-panorama", "title_en": "Variety Market Panorama Scan", "subtitle_en": "Market, capital, positions, hotspots four dimensions on one screen for a single variety", "description_en": "To fully understand a futures variety's recent situation, market trends, capital flow, member positions, and supply chain hotspots each need separate queries—data sources are broad and there's always concern about authority. This case turns Finoview Fiona's market data capabilities into a pipeline, focusing on any variety, and produces an information-dense, fully-traceable single-page HTML: from two-week daily market data to intraday real-time snapshots, to capital large in/out abnormal movements and top-ten member position rankings, finally connecting supply chain hotspots (mine production, inventory changes, macro disturbances, etc.), all in one view.", "tags_en": [ "Futures", "Market Data", "Capital Flow", "Member Positions", "Finoview Fiona" ], "prompt_en": "Generate a single-file HTML '{Variety} Futures Market Panorama Scan', with data reference date {Date}. All data is queried via Finoview Fiona Skill with sources and dates noted; fabrication is prohibited. Execute the following steps:\nStep 1, Market Trend: Query {Variety} main contract daily market data for the past two weeks, including open, close, change, volume and position, with a range change summary.\nStep 2, Capital Flow: Query {Variety}'s daily settled capital and capital flow data for the past two weeks, analyze capital in/out trends, mark abnormal movement days.\nStep 3, Member Positions: Query {Variety} main contract top-ten member volume rankings, listing member abbreviations and volume changes.\nStep 4, Hotspot Sorting: Sort out {Variety} supply chain recent important information (including mine production, inventory changes, macro disturbances, etc.), each marked with source and date.\nOutput format requirements: Single-file HTML, all styles inline, no external dependencies; top provides key indicator dashboard (latest price / change / settled capital / position, etc.); color scheme follows bullish red, bearish green; all sources noted as 'Finoview Fiona / Oriental Securities Derivatives Research Institute' with specific dates. UI text in Simplified Chinese.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "品种全景扫描HTML" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-09T03:40:43Z", "updated_at": "2026-09-09T03:40:43Z", "experts": [], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "first-purchase-repeat-window-analysis", "title": "新客首购后复购窗口分析", "title_en": "First-Purchase Repeat Window Analysis", "subtitle": "新客首购特征与后续复购机会分析", "subtitle_en": "A practical Weimob playbook for first-purchase repeat window analysis", "description": "做新客首购后复购窗口分析时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“新客首购后复购窗口分析”展开：圈选近30天首次消费的新客，分析首购金额、首购商品和后续复购情况，识别最值得跟进的复购时间窗口，并给出触达建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、客户分层洞察运营人员、经营分析团队。", "description_en": "The hardest part of first-purchase repeat window analysis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates first-purchase repeat window analysis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Customer Segmentation specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "新客", "首购", "复购窗口", "商品偏好", "触达建议" ], "tags_en": [ "First-Purchase Repeat Window Analysis", "Customer Operations", "Customer Segmentation", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "圈选近30天首次消费的新客，分析其首购金额、首购商品和后续复购情况，识别最值得跟进的复购时间窗口，并给出触达建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a first-purchase repeat window analysis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "first-sino-japanese-war-1894", "title": "甲午中日战争翻页漫画与因果链复盘", "title_en": "甲午中日战争翻页漫画与因果链复盘", "subtitle": "八年级历史 · 甲午中日战争翻页漫画与因果链复盘", "subtitle_en": "八年级历史 · 甲午中日战争翻页漫画与因果链复盘", "description": "八年级历史《甲午中日战争》互动课件，以九页漫画翻页读史，并用拖拽重建 1894—1899 的因果链。", "description_en": "八年级历史《甲午中日战争》互动课件，以九页漫画翻页读史，并用拖拽重建 1894—1899 的因果链。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "first-sino-japanese-war-1894.html", "artifacts": [ { "file": "first-sino-japanese-war-1894.html", "label": "甲午中日战争翻页漫画与因果链复盘｜互动教学网页" } ], "prompt": "帮我做一个初中历史八年级上册《甲午中日战争》的互动教学网页，要以漫画翻页形式学习历史。", "prompt_en": "帮我做一个初中历史八年级上册《甲午中日战争》的互动教学网页，要以漫画翻页形式学习历史。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "flight-booking-card", "title": "机票行程确认单", "title_en": "Flight Booking Confirmation Card", "subtitle": "输入航班信息，生成精美的机票行程确认页面", "subtitle_en": "Input flight details, generate an elegant booking confirmation page", "description": "出差报销要行程单、给同事发航班信息要截图、旅行攻略里要展示航线——但航空公司邮件里的确认单又丑又乱，截图出来一堆广告和无关信息。\n\n这个案例演示了输入一次商务出差的往返航班信息（航班号、航线、时间、座位、票价），AI 自动生成一张精美的机票行程确认卡片 HTML 页面：居中单卡片（max-width 500px）、航空公司 header、出发到达城市+时间可视化、航班详情信息栏、座位和票价、底部条形码区域。模拟真实登机牌设计风格。\n\n适合：差旅管理、行程分享、旅行博客、报销凭证展示。", "description_en": "Business trip reimbursements need itineraries, sharing flight info with colleagues requires screenshots, travel guides need route displays — but airline confirmation emails are ugly and cluttered with ads.\n\nThis case demonstrates inputting round-trip business flight details (flight number, route, times, seat, price), with AI generating an elegant booking confirmation card HTML page: centered single card (max-width 500px), airline header, departure/arrival city+time visualization, flight detail info bar, seat and price, bottom barcode area. Simulates real boarding pass design.\n\nBest for: Business travel management, itinerary sharing, travel blogs, reimbursement document display.", "categories": [ "creation", "office" ], "tags": [ "机票", "行程单", "差旅", "卡片设计" ], "tags_en": [ "Flight Ticket", "Itinerary", "Business Travel", "Card Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一张机票行程确认单页面（单文件 HTML，内联无外部依赖），界面文字用简体中文（机场三字码/航班号/订单号保留原样），精美的票据卡片风格。\n\n去程：东方航空 MU5137 · 上海浦东(PVG)→北京大兴(PKX) · 2026-05-08 周五 · 07:30 起 / 10:05 落 · 2h35m · 商务舱 · 座位 2A\n回程：东方航空 MU5168 · 北京大兴(PKX)→上海浦东(PVG) · 2026-05-10 周日 · 19:00 起 / 21:25 落 · 2h25m · 商务舱 · 座位 3F\n乘客：张明远 · 订单号 CE-20260420-8847 · 往返票价 ¥4,280\n\n含航程可视化（三字码 + 城市 + 飞机图标 + 起降时间）、航班详情、乘客信息、票价，票据撕边感即可。", "prompt_en": "Please generate a flight itinerary confirmation page (single-file HTML, inline with no external dependencies), with interface text in Simplified Chinese (keeping airport three-letter codes / flight numbers / order numbers as-is), in an elegant ticket card style.\n\nOutbound: China Eastern Airlines MU5137 · Shanghai Pudong (PVG) → Beijing Daxing (PKX) · 2026-05-08 Friday · 07:30 departure / 10:05 arrival · 2h35m · Business Class · Seat 2A\nReturn: China Eastern Airlines MU5168 · Beijing Daxing (PKX) → Shanghai Pudong (PVG) · 2026-05-10 Sunday · 19:00 departure / 21:25 arrival · 2h25m · Business Class · Seat 3F\nPassenger: Zhang Mingyuan · Order number CE-20260420-8847 · Round-trip fare ¥4,280\n\nInclude a route visualization (three-letter codes + cities + airplane icon + departure/arrival times), flight details, passenger information, and fare; a torn-ticket edge effect is fine.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "机票预订卡片" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "food-journey-through-digestive-system", "title": "食物在体内的旅行（消化与吸收）", "title_en": "食物在体内的旅行（消化与吸收）", "subtitle": "小学四年级科学 · 食物在体内的旅行（消化与吸收）", "subtitle_en": "小学四年级科学 · 食物在体内的旅行（消化与吸收）", "description": "四年级科学《食物在体内的旅行》互动课件，用拟人动画带学生走过口腔到小肠的消化全过程。", "description_en": "四年级科学《食物在体内的旅行》互动课件，用拟人动画带学生走过口腔到小肠的消化全过程。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "food-journey-through-digestive-system.html", "artifacts": [ { "file": "food-journey-through-digestive-system.html", "label": "食物在体内的旅行（消化与吸收）｜互动教学网页" } ], "prompt": "帮我做一个小学四年级科学《食物在体内的旅行》的互动教学网页，要用拟人动画演示食物经过各个消化器官的过程。", "prompt_en": "帮我做一个小学四年级科学《食物在体内的旅行》的互动教学网页，要用拟人动画演示食物经过各个消化器官的过程。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "freelance-order-desk", "title": "副业接单台", "title_en": "Freelance Order Desk", "subtitle": "订单、回款、工时和收入复盘，一张工作台管清楚", "subtitle_en": "Manage orders, payments, time and income insights in one offline dashboard", "description": "副业订单一多，最烦的不是做事，而是翻聊天记录找交付日、算还差多少尾款、回忆这单到底花了多久。表格能记账，却很难在打开的一刻告诉你今天最该处理什么。\n\n这个案例演示了一张可直接使用的单文件副业接单台：顶部自动汇总逾期回款和今日交付，订单表记录客户、内容、报价、状态与日期；每笔收款独立登记，待收和逾期自动计算；月实收、客户来源占比、报价/实收时薪和高价值订单类型同步更新。数据自动保存到浏览器本地，支持 JSON 备份迁移、真实 Excel 导出、手机响应式和一键清空示例。\n\n适合：刚开始接单的设计师、文案和开发者；需要盯回款与交付的自由职业者；想用一个离线文件管理零散副业收入的人；需要报税对账数据出口的个人服务者。", "description_en": "When freelance work piles up, the hardest part is often not the work itself — it is digging through chats for delivery dates, calculating unpaid balances, and remembering how many hours each job actually took. A spreadsheet can store rows, but it rarely tells you what needs attention today.\n\nThis case demonstrates a ready-to-use single-file freelance order desk. The top area automatically surfaces overdue payments and work due today; the order table tracks clients, scope, quotes, status and dates; every payment is recorded separately so balances and overdue items are calculated automatically. Monthly cash received, client-source share, quoted and realized hourly rates, and high-value job categories update together. Data autosaves in the browser, with JSON backup and restore, genuine Excel export, responsive mobile use, and one-click sample cleanup.\n\nIdeal for: designers, writers and developers starting to freelance; independent professionals tracking delivery and payment; people managing side income in one offline file; and solo service providers who need clean exports for reconciliation or tax preparation.", "categories": [ "workspace", "business", "productivity", "data" ], "tags": [ "副业接单", "订单管理", "回款提醒", "收入统计", "离线工作台" ], "tags_en": [ "side hustle", "order management", "payment reminders", "income analytics", "offline dashboard" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/ZwDtuUpAMzaA6m4c29ZbmB），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个副业接单台，至少包括这些内容：订单要记录客户、工作内容、报价、状态（洽谈、在做、待验收、已收款）和交付日期；收款要显示已收、待收，并把逾期未收到的钱标红；收入统计按月做柱状图，并展示客户来源占比；每单记录实际耗时，自动计算报价时薪和实收时薪，提示哪类单更值得接。\n\n最上面放“今天要处理的”，把逾期没收款和今天要交付的内容标红。金额用 ¥，欠款和支出用红色，收到的钱用绿色。提供真正的 Excel 导出，方便对账。\n\n还要支持自动保存、JSON 导出备份和导入恢复；电脑和手机都能用，手机按钮至少 44px；最终做成一个所有断网可打开；内置几条相对当前日期的示例数据，并提供“清空示例”。整体用浅色、简洁、专业的视觉，不依赖外部组件。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/ZwDtuUpAMzaA6m4c29ZbmB) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a freelance order desk with at least these capabilities: each order records the client, scope of work, quote, status (negotiating, in progress, awaiting acceptance, paid), and delivery date; payments show amounts received and outstanding, with overdue balances highlighted in red; income analytics include a monthly bar chart and client-source share; each order records actual hours, automatically calculates quoted and realized hourly rates, and indicates which job categories are more valuable. Place a “Needs attention today” section at the top, highlighting overdue payments and work due today in red. Use ¥ for money, red for receivables and expenses, and green for cash received. Provide a genuine Excel export for reconciliation. Also support autosave, JSON backup export and restore; work on desktop and mobile with tap targets at least 44px; deliver everything as one self-contained HTML file and available offline; include sample records based on dates relative to today and a “Clear samples” action. Use a clean, professional light theme without external components.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 0, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-01T17:15:46+08:00", "updated_at": "2026-08-01T17:15:46+08:00", "artifacts": [ { "file": "output.html", "label": "副业接单台（单文件 HTML）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "frontend-dev-awwwards-resume-landing", "title": "Awwwards 级网页简历 · Landing Page", "title_en": "Awwwards-Grade Resume Landing Page", "subtitle": "把简历做成一张会说话的高级互动网页", "subtitle_en": "Turn your CV into a persuasive interactive landing page", "description": "投简历石沉大海的第一锅，通常不是经历不够，而是简历看起来太像 Word 文档。\n\n这个案例用 frontend-dev skill 跑一份 Awwwards 获奖感风格的个人 Landing Page——按金字塔叙事骨架（HeroCard / StatsBoard / TimelinePro / SkillRadar / BentoGallery / Testimonial）把你的履历、数据、项目、评价重新编排，产出一份零外部依赖的单文件 HTML：深色极光主题、手工 SVG 雷达图、Bento 非对称宫格、IntersectionObserver 驱动的入场动效、数字缓动、响应式降级全都具备，双击即开、手机也扛得住。\n\n适合：在找工作的技术 / 设计 / 产品从业者、想给个人域名配一张高级 About Me 的独立开发者、准备做求职陈述或内推自荐的人。", "description_en": "Your resume isn't ignored because your experience is thin — it's ignored because it looks like a Word doc.\n\nThis case uses the frontend-dev skill to generate an Awwwards-grade personal landing page. Following a pyramid narrative (HeroCard / StatsBoard / TimelinePro / SkillRadar / BentoGallery / Testimonial), it rewrites your résumé, metrics, projects, and endorsements into a single-file HTML: dark aurora theme, hand-crafted SVG radar chart, asymmetric bento grid, IntersectionObserver reveals, number counters, fully responsive — double-click to open, looks great on mobile, zero external dependencies.\n\nTarget audience: engineers / designers / PMs on the job market, indie makers who want a polished About Me page on their personal domain, and anyone preparing a self-introduction deck.", "categories": [ "creation", "dev" ], "tags": [ "网页简历", "Landing Page", "个人网站", "交互网页", "About Me" ], "tags_en": [ "resume landing", "personal website", "interactive web", "About Me", "landing page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "网页简历 Landing Page" }, { "file": "data.json", "label": "结构化数据（可替换复用）" } ], "prompt": "你是『资深猎头总监』+『Awwwards 获奖前端架构师』组成的双核 Agent。请把我的简历／职业描述转化为一份 Landing Page 风格的网页简历，信息不足时按行业最佳实践合理补全。\n\n请按金字塔叙事骨架组织：HeroCard、核心数据看板、时间倒序经历、技能雷达、项目精选、推荐评价。\n\n输出硬约束：单文件 HTML，CSS/JS 全部内联、无外链；系统字体栈、响应式、尊重 prefers-reduced-motion；深色 + 渐变极光主题，主色可替换。同时产出 data.json 与 output.html 两个产物。\n\n【个人信息】（粘贴你的简历／履历／项目，或口述一段自我介绍）\n...", "prompt_en": "You are a dual-core agent: a senior headhunting director + an Awwwards-winning frontend architect. Turn my résumé / career description into a landing-page-style web résumé, reasonably filling gaps with industry best practice where info is missing.\n\nOrganize with a pyramid narrative: hero card, key-stats board, reverse-chronological experience, skill radar, selected projects, testimonials.\n\nHard constraints: single-file HTML with all CSS/JS inline and no external links; system font stack, responsive, respects prefers-reduced-motion; dark theme with an aurora gradient and a replaceable primary color. Produce two artifacts: data.json and output.html.\n\n[Personal info] (paste your résumé / background / projects, or describe yourself)\n...", "skills": [ { "id": "frontend-dev", "name": "Frontend Dev · 前端与 AI 媒体", "description": "集成 UI 设计、动效系统、AI 媒体资产、转化文案与生成艺术的前端开发 skill。适合做 Landing Page / Marketing Site / Product Page / Dashboard / 个人网站 / 生成图文视频等一条龙产出。", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-11T00:00:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-analysis-110020", "title": "基金综合分析:沪深300ETF联接A", "title_en": "Comprehensive Fund Analysis: HS300 ETF Feeder A", "subtitle": "想搞懂一只基金？基础档案、成份股、经理观点，一篇给你说全", "subtitle_en": "Want to understand a fund? Profile, constituents, and manager views — all in one report", "description": "打开一只基金，先看什么？基础信息、成份股、基金经理最近有没有发布观点、有哪些与这只基金相关的好文章——以前挨个搜，这份案例帮你一次拼好。\n\n案例以易方达沪深300ETF联接A(110020)真实跑通:基金档案 + 沪深300前十大权重股与申万行业分布 + 两位经理最新观点 + 3篇关联好文，汇成一份自包含 HTML，数据真实查询、日期与来源逐项标注，可离线打开。\n\n适合:个人投资者选基前自查。", "description_en": "Just opened a fund and not sure where to start? Basic profile, constituents, what the managers have said recently, and related articles — you used to search each one separately; this case pieces them together for you in one go.\n\nUsing E Fund HS300 ETF Feeder A (110020) as a real example, it packs the fund profile, HS300 top-10 constituents with SW industry distribution, the latest views of both managers, and 3 related articles into a self-contained HTML, with real queried data, dates and sources annotated, ready to open offline.\n\nIdeal for: individual investors doing pre-purchase self-checks.", "categories": [ "finance" ], "tags": [ "基金分析", "ETF联接", "多技能编排", "投教内容", "易方达" ], "tags_en": [ "Fund Analysis", "ETF Feeder", "Multi-skill", "Investor Education", "E Fund" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "基金综合分析报告(HTML)" } ], "prompt": "把「易方达沪深300ETF联接A(110020)」做成一份单文件 HTML 基金综合分析报告,易方达蓝系、浅色、涨跌红涨绿跌。按四部分组织,每部分严格用对应技能的原生模板输出,不自加字段、不自创样式:\n\n1. 用 efunds-know-a-fund 出「秒懂基金」档案区块(简称/代码/类型/风险评级/成立日期/经理/业绩比较基准)。\n\n2. 用 index-query 查沪深300(000300)前十大成分股及权重 + 申万行业分布:一句话结论 + 数据表 + 口径与日期 + 风险提示。\n\n3. 用 efunds-manager-views 抓余海燕、庞亚平每人最新一条公开发表观点,原文摘录并注明来源与日期。\n\n4. 用 efunds-know-a-fund 精选好文区块,按其打分规则挑至多 3 篇直接相关的公众号文章:标题纯文本、行内标日期与公众号名、附原文摘录。\n\n合规:数据全真实查询并注明日期/来源,取不到标「暂缺」不编造;无外链、无 hash 锚点、CSS 全内联;不预测后市、不荐股、不保本收益;末尾附固定风险提示。", "prompt_en": "Turn E Fund HS300 ETF Feeder A (110020) into a single-file HTML comprehensive fund analysis report in E Fund blue, light theme, with gains/losses colored red-up/green-down. Organize it into four parts; each part must strictly follow the native template of its designated skill — do not add your own fields or invent your own styles:\n\n1. Use efunds-know-a-fund to produce the 'Know-a-Fund' profile block (short name / code / type / risk rating / inception / managers / performance benchmark).\n\n2. Use index-query to get the HS300 (000300) top-10 constituents and weights plus SW industry distribution: a one-sentence conclusion + data tables + caliber and dates + risk note.\n\n3. Use efunds-manager-views to fetch, for each of Yu Haiyan and Pang Yaping, only the latest publicly published view, quoting the original text and noting the source and date.\n\n4. Use efunds-know-a-fund's featured-articles block, following its scoring rules to pick up to 3 articles directly related to the fund: titles as plain text, publish date and public account name inline, with a verbatim excerpt.\n\nCompliance: all data must come from real queries with dates/sources noted; if unavailable, mark as 'not available' — never fabricate. No external links, no hash anchors, all CSS inlined. No market predictions, no buy/sell recommendations, no guaranteed-return claims; end with the fixed risk note.", "skills": [ { "id": "efunds-know-a-fund", "name": "易方达秒懂基金", "description": "当客户想初步了解某只基金或泛化咨询「这只基金怎么样」时使用。基于易方达基金官网和各官方微信公众号已发布文章，输出一份基金介绍页，包含基础档案、按基金类型定制的投教指南、与该基金相关的精选内容，搭配基金画像工具帮助用户生成「第一印象」。", "skill_id": "skill_2092288712677167104" }, { "id": "index-query", "name": "指数查询", "description": "查询证券指数的基础信息、行情、成分股、行业分布、估值、收益及相关基金等数据。", "skill_id": "skill_2089730438115336192" }, { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-08-25T19:04:00+08:00", "updated_at": "2026-08-31T17:25:42+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-comprehensive-analysis", "title": "手把手带你看懂一只基金", "title_en": "Understand a Fund, Guided Step by Step", "subtitle": "先搞懂「行业主题基金」这一类产品，再认识这只具体基金——投教问答、档案、经理观点、精选好文一次看完", "subtitle_en": "Grasp the fund category first, then this specific fund — edu Q&A, profile, manager view, and selected articles in one report", "description": "看中一只主动管理基金，想搞懂它靠不靠谱，却被「行业主题基金」「偏股混合」这些概念绕晕——去官网翻档案、看经理文章、查季报，零零碎碎拼不成一个完整画面。\n\n这份报告把「小达养基笔记」投教问答（行业主题基金有什么特点和风险）和「易方达秒懂基金」档案与导读（这只基金的具体档案、投教导读、精选好文）融合成一条流水线：先搞懂这类产品本身，再认识这只基金、读懂经理的产业判断，最后补上精选好文。案例以易方达信息产业混合A（001513）为例完整演示，数据来自易方达官网、官方公众号与财富学堂投教基地。\n\n适合：第一次买主动权益基金想先搞懂产品类型的小白；持基用户想建立行业主题基金的投资框架。", "description_en": "You spotted an actively-managed fund and want to know if it is reliable — but terms like 'sector-themed fund' and '偏股混合' leave you confused. Browsing the official site for its profile, reading manager articles, checking quarterly reports — the pieces never add up to a complete picture.\n\nThis report merges the 小达养基笔记 edu-Q&A (what are the features and risks of sector-themed funds) and the 易方达秒懂基金 profile & guide (this fund's specific profile, investor-education guide, and selected articles) into one pipeline: first understand the product category itself, then get to know this specific fund, grasp the manager's industry judgment, and finally round it out with selected articles. Demonstrated with EFUNDS 易方达信息产业混合A (001513), all data sourced from the official EFUNDS website, official WeChat accounts, and the Caifu Xuetang investor-education base.\n\nSuitable for: beginners buying an actively-managed equity fund for the first time who want to understand the product type first; existing holders who want to build an investment framework for sector-themed funds.", "categories": [ "finance" ], "tags": [ "基金综合分析", "投教问答", "行业主题基金", "主动权益" ], "tags_en": [ "Fund Analysis", "Edu Q&A", "Sector-Themed Fund", "Active Equity" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "基金综合分析报告（投教 + 档案 + 导读 + 好文）" } ], "prompt": "请基于易方达信息产业混合A（001513）做一份基金综合分析：\n1）先用「小达养基笔记」回答「行业主题基金有什么特点和风险」，内容只保留财富学堂文章的原文摘录（概念 + 风险清单 + 决策框架），不要行情数据；\n2）再用「秒懂基金」生成这只基金的档案（类型/风险/经理/基准/投教导读）与精选好文，精选好文摘录须为文章原文（可连缀省略），不得自行总结或添加内容；\n3）合并为一份浅色小达蓝主色、单文件 HTML 报告，顺序：投教问答 → 档案 → 导读 → 精选好文。\n数据须真实、来自易方达官网、官方公众号与财富学堂，不得编造。", "prompt_en": "Create a comprehensive fund analysis for EFUNDS 易方达信息产业混合A (001513):\n1) First use 小达养基笔记 (edu-query) to answer 'What are the features and risks of sector-themed funds?' — keep only verbatim excerpts from 财富学堂 articles (concept + risk list + decision framework), no market data;\n2) Then use 秒懂基金 (know-a-fund) to generate this fund's profile (type/risk/manager/benchmark/investor-education guide) and selected articles — excerpts must be quoted from the original text (may be elided with ellipsis), never summarized or paraphrased;\n3) Merge into a single-file HTML report in a light Xiaoda-blue theme, ordered: edu Q&A → profile → investor-education guide → selected articles.\nAll data must be real, sourced from the official EFUNDS website, official WeChat accounts, and 财富学堂 (Caifu Xuetang); no fabrication.", "skills": [ { "id": "efunds-know-a-fund", "name": "易方达秒懂基金", "description": "根据实时官方数据生成结构化投教内容，输出基金档案、投教导读与精选好文", "skill_id": "skill_2092288712677167104" }, { "id": "efunds-edu-query", "name": "小达养基笔记", "description": "易方达投教基地开放式问答助手，基于财富学堂投教知识库，三段式回答基金/ETF/黄金等概念、时机、比较类问题", "skill_id": "skill_2089898953429123072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-28T09:46:06+08:00", "updated_at": "2026-08-31T17:18:52+08:00", "cover_image": "cover.jpg", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-manager-profile", "title": "明星基金经理画像卡", "title_en": "Star Fund Manager Profile", "subtitle": "以易方达张坤为例，一张卡看懂履历、业绩与持仓", "subtitle_en": "A one-card view of tenure, performance and holdings — featuring Zhang Kun of E Fund", "description": "想看懂一位明星基金经理，得在天天基金、公司公告、季报之间来回翻，履历、规模、业绩、回撤、持仓散落各处，数字还常常过期或口径不一。\n\n这个案例用 NeoData 金融搜索对基金经理张坤做真实查询：从业年限与管理规模、在管基金列表、代表基金的净值与近1/3年业绩、最大回撤、前十大重仓股、所属公司整体情况，全部取最新真实数据，最后自动汇成一张可离线打开的单文件 HTML 画像卡（画像头部、一句话结论、业绩表、代表作深挖、重仓条形图、风格标签、复现清单、口径提示）。\n\n适合：想快速调研某位基金经理的个人投资者；需要基金经理画像素材的投顾与内容创作者；做基金/资管尽调的从业者；想学习「用真实数据而非记忆」做金融卡片的效率工具用户；关注价值投资风格的研究者。", "description_en": "Understanding a star fund manager means jumping between fund apps, company disclosures and quarterly reports — tenure, AUM, returns, drawdowns and holdings are scattered everywhere, and the numbers are often stale or inconsistent.\n\nThis case runs real queries via NeoData Financial Search on fund manager Zhang Kun: career length and AUM, the list of funds under management, flagship fund NAV and 1Y/3Y returns, maximum drawdown, top-10 holdings and the parent company overview — all pulled as the latest real data — then compiles it into a single-file, offline-ready HTML profile card (profile header, one-line verdict, performance table, flagship deep dive, holdings bar chart, style tags, reproduce checklist, methodology notes).\n\nWho it's for: individual investors quickly researching a manager; advisors and content creators who need manager-profile material; professionals doing fund/asset-management due diligence; efficiency-tool users who want to build finance cards from real data instead of memory; researchers interested in the value-investing style.", "categories": [ "finance", "data", "search" ], "tags": [ "基金经理", "价值投资", "画像卡", "基金业绩", "真实数据" ], "tags_en": [ "Fund Manager", "Value Investing", "Profile Card", "Fund Performance", "Real Data" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "prompt": "用 NeoData 金融搜索服务，针对易方达基金经理张坤做一份真实数据驱动的「基金经理画像卡」。\n\n请务必先通过 NeoData 官方脚本逐条查询最新真实数据，不要凭记忆作答，也不要混用其他数据源。至少覆盖：1) 张坤的从业经历、任职年限、在管基金总规模；2) 他在管的基金列表（如易方达蓝筹精选、易方达优质企业三年持有、易方达优质精选QDII、易方达亚洲精选QDII 等）及任职回报；3) 代表基金（易方达蓝筹精选）的单位净值、近1年/近3年及成立以来业绩、最大回撤；4) 代表基金的前十大重仓股；5) 易方达基金公司整体情况（管理规模、基金数量）。查询返回空就换个措辞重试；提炼关键数字并记录数据时间与口径。\n\n最终产出一份单文件 HTML 画像卡，建议区块：画像头部（姓名/公司/从业年限/管理规模）→ 一句话结论 → 在管基金业绩表 → 代表作深挖（净值/回撤/收益）→ 前十大重仓（条形或表格）→ 风格标签 → 复现查询清单 → 口径与风险提示。\n\n样式方向：浅色专业金融风、信息密度高、涨跌用中国习惯（红涨绿跌）；单文件 HTML，所有 CSS/JS 内联、无任何外部资源、可离线打开。", "prompt_en": "Using the NeoData Financial Search service, build a real-data-driven \"fund manager profile card\" for Zhang Kun of E Fund.\n\nAlways query the latest real data step by step via the official NeoData script first — do not answer from memory, and do not mix in other data sources. Cover at least: 1) Zhang Kun's career history, years of tenure, and total AUM under management; 2) the list of funds he manages (e.g., E Fund Blue Chip Selected, E Fund Premium Enterprises 3-Year Hold, E Fund Premium Selected QDII, E Fund Asia Selected QDII) and their tenure returns; 3) the flagship fund (E Fund Blue Chip Selected) unit NAV, 1Y/3Y and since-inception performance, and maximum drawdown; 4) the flagship fund's top-10 holdings; 5) the E Fund company overview (AUM, number of funds). If a query returns empty, rephrase and retry; distill the key figures and record the data date and methodology.\n\nProduce a single-file HTML profile card. Suggested sections: profile header (name/company/tenure/AUM) -> one-line verdict -> managed-funds performance table -> flagship deep dive (NAV/drawdown/returns) -> top-10 holdings (bar chart or table) -> style tags -> reproduce-query checklist -> methodology & risk notes.\n\nStyle direction: light, professional financial look, high information density, using the Chinese convention (red for up, green for down); a single HTML file with all CSS/JS inlined, no external resources, openable offline.", "skills": [ { "id": "neodata-financial-search", "skill_id": "skill_2053082432761950208", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "基金经理张坤画像卡" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-22T03:21:40Z", "updated_at": "2026-07-22T03:58:06Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-portfolio-checkup", "title": "基金组合健康诊断", "title_en": "Fund Portfolio Health Diagnosis", "subtitle": "给你的基金持仓做一次体检，找出集中度与风险问题", "subtitle_en": "A checkup for your fund holdings — surface concentration and risk issues", "description": "很多基民买了一堆基金，却说不清自己到底承担了多少风险、有没有重复押注同一个赛道。本案例用「盈米 Skills」对基金组合做健康诊断——拆解资产配置、识别行业集中度与相关性问题、评估波动与回撤，并给出可执行的再平衡建议。以一个 6 只基金、32 万市值的真实感组合为例。适合持有多只基金、想优化组合的普通投资者。", "description_en": "Many fund investors hold a pile of funds but can't say how much risk they're actually taking or whether they're doubling down on the same theme. Using Yingmi Skills, this case runs a portfolio health diagnosis — breaking down asset allocation, spotting sector-concentration and correlation issues, assessing volatility and drawdown, and giving actionable rebalancing advice. It uses a realistic 6-fund, ¥320K portfolio as an example. Ideal for ordinary investors who hold multiple funds and want to optimize.", "categories": [ "finance" ], "tags": [ "基金", "组合诊断", "资产配置", "风险评估", "再平衡" ], "tags_en": [ "Funds", "Portfolio Diagnosis", "Asset Allocation", "Risk Assessment", "Rebalancing" ], "artifact_type": "html", "prompt": "帮我生成一份「基金组合健康诊断」单页 HTML，诊断一个普通基民的基金组合（总市值 ¥320,000，累计收益 +13.6%）。\n\n持仓 6 只：易方达蓝筹精选(005827)、招商中证白酒指数(161725)、富国天惠成长(161005)、天弘沪深300ETF联接(000961)、广发纳斯达克100(270042)、易方达中短债(000621)。\n\n要直观看到：组合评分与 KPI、持仓明细（市值/占比/收益率）、资产配置环形图、风险体检（波动率/最大回撤/风险等级）、行业集中度与相关性预警，以及可执行的优化建议。涨跌按 A 股习惯（涨红跌绿）。", "prompt_en": "Generate a single-page HTML 'Fund Portfolio Health Diagnosis' for an ordinary fund investor's portfolio (total value ¥320,000, cumulative return +13.6%).\n\nHoldings of 6 funds: E Fund Blue-Chip Selected (005827), China Merchants CSI Liquor Index (161725), Fullgoal Tianhui Growth (161005), Tianhong CSI 300 ETF Link (000961), GF Nasdaq-100 (270042), E Fund Short-to-Mid Bond (000621).\n\nShow at a glance: portfolio score and KPIs; holdings detail (value / weight / return); an asset-allocation donut; a risk checkup (volatility / max drawdown / risk level); sector-concentration and correlation warnings; and actionable optimization advice. Use the China-market color convention (up=red, down=green).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "组合诊断报告" } ], "skills": [ { "id": "yingmi-skill", "name": "盈米Skills", "description": "基金查询、组合诊断、资产配置与财富规划的金融数据助手", "description_en": "Fund search, portfolio diagnosis, asset allocation & wealth planning assistant", "skill_id": "skill_2054901386815930368" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-06-16T20:10:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fundus-abnormality-screening", "title": "眼底彩照异常初筛", "title_en": "Fundus Photo Abnormality Screening", "subtitle": "上传一张眼底彩照，AI 秒出异常判定与主要体征", "subtitle_en": "Upload a fundus photo and AI flags abnormalities with key signs", "description": "拿到一张眼底彩照，想知道眼睛到底有没有问题，靠肉眼逐张看又慢又容易漏，去医院排队还费时。\n\n本案例把一张眼底彩照交给觅影眼底AI，自动完成异常判定，输出「异常与否 + 置信度 + 主要体征（出血、渗出、微动脉瘤、棉绒斑等）+ 分诊建议」的结构化报告。\n\n适合：眼科体检人群、糖尿病/高血压等慢病患者的眼底筛查、基层医疗机构影像初筛、医学影像AI评测演示。", "description_en": "You have a fundus photo and want to know whether anything is wrong, but eyeballing it is slow and error-prone, and hospital visits mean queues.\n\nThis case hands a fundus photo to the Miying fundus AI, which automatically flags abnormalities and returns a structured report of the verdict, confidence, key signs (hemorrhage, exudates, microaneurysms, cotton-wool spots, etc.) and triage advice.\n\nFor: eye check-up users, diabetic/hypertensive patients needing fundus screening, primary-care imaging triage, and medical-imaging AI demos.", "categories": [ "data" ], "tags": [ "眼底彩照", "异常初筛", "糖网筛查", "AI医学影像", "腾讯健康" ], "tags_en": [ "Fundus Photo", "Abnormality Screening", "DR Screening", "AI Medical Imaging", "Tencent Health" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "眼底AI异常初筛报告（演示样例）" } ], "prompt": "帮我分析这张眼底彩照，看看有没有异常。\n\n请以「腾讯健康-觅影-眼底彩照疾病分析专家」的能力，对我上传的眼底彩照（标准视野）做一次全面的异常初筛：\n\n1. 判断这张眼底彩照是否存在异常，给出「异常 / 未见明显异常」的判定与置信度；\n2. 若发现异常，列出主要体征（如微动脉瘤、出血、渗出、棉绒斑、新生血管、视盘异常、黄斑水肿等）并标注大致位置；\n3. 给出分诊建议（正常定期复查 / 建议进一步检查 / 建议尽快就医）。\n\n测试数据可用公开眼底数据集，如 IDRiD（Indian Diabetic Retinopathy Image Dataset，IEEE Dataport）或 APTOS 2019（Kaggle 糖尿病视网膜病变检测）中的一张眼底彩照作为输入，并注明图片来源。\n\n输出为结构化报告，末尾声明：本结果仅限科研用途，不替代临床诊断，最终以眼科医生判断为准。", "prompt_en": "Analyze this fundus photo and tell me if there are any abnormalities.\n\nUsing the capabilities of the \"TENCENT AIMIS Fundus Disease Analyst Specialist\", perform a full abnormality screening on the fundus photo I upload (standard field):\n\n1. Judge whether the fundus photo shows any abnormality, giving a verdict of \"abnormal\" or \"no obvious abnormality\" with confidence.\n2. If abnormal, list the key signs (e.g. microaneurysms, hemorrhage, exudates, cotton-wool spots, neovascularization, optic disc abnormality, macular edema) with approximate location.\n3. Give triage advice (routine follow-up / further examination recommended / see a doctor promptly).\n\nFor test data you may use a public fundus dataset such as IDRiD (Indian Diabetic Retinopathy Image Dataset, IEEE Dataport) or APTOS 2019 (Kaggle diabetic retinopathy detection) as input, and cite the image source.\n\nOutput a structured report, ending with the disclaimer: this result is for research use only, does not replace clinical diagnosis, and the final decision rests with an ophthalmologist.", "skills": [ { "id": "fundus-images-ai-analysis", "name": "腾讯健康-觅影-基于45°眼底彩照的多病种 AI 分析", "description": "眼底 AI 多病种分析服务：上传眼底图像（JPEG/PNG/DICOM，单文件≤5MB）进行 AI 诊断与诊断结果查询", "slug": "fundus-images-ai-analysis" } ], "experts": [ { "id": "腾讯健康-觅影-眼底彩照疾病分析专家", "name": "腾讯健康-觅影-眼底彩照疾病分析专家", "description": "基于普角与超广角眼底彩照的AI多病种分析，覆盖青光眼、糖网、AMD等数十种疾病的诊断与报告解读", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T23:02:26+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "fundus-ai-disease-report", "title": "眼底彩照 AI 查出中度糖网", "title_en": "AI Detects Moderate Diabetic Retinopathy from a Fundus Photo", "subtitle": "一张普角眼底照片，查出中度糖尿病视网膜病变与三项阳性病灶", "subtitle_en": "One standard fundus photo reveals moderate diabetic retinopathy and three positive lesion signs", "description": "糖尿病最怕的并发症之一就在眼底，可它早中期几乎不影响视力——等到「看不清了」才去医院，往往已经错过最好的治疗窗口。体检报告上那句「中度糖网」，很多人也不知道到底算轻还是算重、要不要马上处理。\n\n这个案例把一张真实的标准 45° 普角眼底彩照（右眼单眼送检）交给腾讯觅影眼底 AI 模型，真实调用 API 后查出**中度糖尿病性视网膜病变**，同时检出微动脉瘤、硬性渗出、软性渗出三项阳性病灶。报告没有停在给个结论：它讲清了糖网五级分级里「中度」处在什么位置、为什么「出血斑=否」也能判中度、三项病灶如何反映微血管损伤下渗漏与缺血两条并行后果、为什么软性渗出（缺血信号）最值得警惕，以及硬性渗出阳性意味着必须靠 OCT 排查黄斑水肿——而不是看到「中度」就以为可以先观察。\n\n更实用的是它对不完美数据的处理：本次视盘形态学指标（杯盘比、盘沿比）虽然状态码正常却全部返回 0，报告用「无图那只眼同样返回 0」做交叉印证，判定这是占位默认值而非真实杯盘比，并明确说明接口未返回失败原因、成因不做推测。随访建议按紧迫度排序——1 个月内就诊眼科、务必补查另一只眼、立即复核控糖方案，并列出了需要立刻挂急诊的预警信号。\n\n适合：糖尿病患者及家属、需要批量初筛糖网的基层医疗机构与体检中心、内分泌科与眼科的协作转诊场景，以及需要眼底 AI 能力做二次开发的医疗信息化团队。", "description_en": "One of the complications diabetics fear most sits in the retina — yet in its early and middle stages it barely affects vision. By the time someone notices things looking blurry, the best treatment window has often already closed. And when a report says \"moderate diabetic retinopathy,\" most people have no idea whether that counts as mild or serious, or whether it needs acting on right away.\n\nThis case feeds a real standard 45-degree fundus photograph (right eye only) to the Tencent AIMIS fundus AI model. The genuine API call returns **moderate diabetic retinopathy**, along with three positive lesion signs: microaneurysms, hard exudates and soft exudates. The report doesn't stop at the verdict. It explains where \"moderate\" sits within the five-tier DR grading scale, why a negative hemorrhage finding can still grade as moderate, how the three lesions reflect leakage and ischemia as two parallel consequences of microvascular damage, why the soft exudates (an ischemia signal) deserve the most attention, and why positive hard exudates mean macular edema must be ruled out by OCT — rather than assuming \"moderate\" means it can wait.\n\nEven more useful is how it handles imperfect data: the optic disc metrics (cup-to-disc and rim ratios) all came back as 0 despite a successful status code. The report cross-checks this against the un-imaged eye returning 0 as well, concludes it is a placeholder default rather than a real cup-to-disc ratio, and states plainly that the API returned no failure reason so no cause is speculated. Follow-up advice is ordered by urgency: see an ophthalmologist within a month, be sure to image the fellow eye, review the glycemic control plan immediately — plus a list of warning signs that warrant emergency care.\n\nGood for: people with diabetes and their families, primary care clinics and health check centers running DR screening at scale, endocrinology-to-ophthalmology referral workflows, and healthcare IT teams building on fundus AI capabilities.", "categories": [ "data", "productivity", "education" ], "tags": [ "眼底AI分析", "糖网筛查", "医学影像解读", "糖尿病并发症", "AI辅助诊断" ], "tags_en": [ "Fundus AI Analysis", "Diabetic Retinopathy Screening", "Medical Imaging Interpretation", "Diabetes Complications", "AI-Assisted Diagnosis" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "眼底彩照 AI 分析报告（可视化网页版）" }, { "file": "output.md", "label": "眼底彩照 AI 分析报告（Markdown 版）" }, { "file": "ai-result.json", "label": "觅影 AI 原始返回数据（raw + parsed）" }, { "file": "sample-fundus-od.png", "label": "示例眼底彩照（普角 45°·右眼 OD·可直接复现本案例）" } ], "prompt": "我家人做了一次普角眼底检查，这里是她的眼底照片图像，帮我基于「觅小影」分析并输出分析报告。\n\n[上传眼底彩照图片。本案例已随附示例影像 sample-fundus-od.png（普角 45°、右眼 OD、3216×2136 原图），可直接使用它一键复现下方全部分析结果；也可换成自己的眼底照片——向体检中心或眼科门诊索取原始照相文件即可]\n\n补充信息：\n- 图像类型：标准 45° 普角眼底彩照\n- 眼别：右眼（OD），本次仅送检单眼\n- 检查日期：2026-08-10\n\n请调用真实的眼底 AI 模型完成分析，不要凭通用视觉能力看图推测；如果无法调用 API，请直接告诉我，不要编造结果。\n\n报告请按「基本信息 → 青光眼评估 → 多病种筛查 → 病灶描述 → 综合诊断」的顺序组织，其中：\n- 病种筛查请把模型覆盖的病种逐项列出，标明阳性/阴性；若返回了疾病分级，请解释该分级在整个分级体系中处于什么位置、意味着什么\n- 病灶描述部分列出各项体征判别，并说明多个阳性体征之间的关联；同时给出视盘形态学的量化测量值\n- 综合诊断给出结论汇总和按紧迫度排序的随访建议\n\n要求：\n- 所有结论必须来自 AI 模型真实返回。「不确定」「未知」「未找到图片」「未返回结论」等状态要如实标注并说明原因，不要替模型补成阴性\n- 若某个量化指标返回 0 或空值，请判断这是真实测量结果还是占位默认值，不要把占位值当真实数据解读；若接口未说明原因，就如实写「成因无法从返回数据判断」，不要自行推测成因\n- 若各子模型结论之间存在看似矛盾之处（如分级与体征判别不完全吻合），请主动指出并解释，不要留白\n- 如查出阳性病变，请明确说明紧迫程度和建议就诊时限，不要用模糊表述淡化风险；同时也要指出病情中相对乐观的方面，避免过度惊吓\n- 面向没有医学背景的患者家属写作，专业术语附通俗解释，末尾加术语速查表\n- 明确标注这是 AI 辅助筛查结果、不能替代医生面诊；附官方 PDF 报告链接（提示该链接含未脱敏个人信息），并把原始 API 返回数据一并保存", "prompt_en": "A family member had a standard fundus examination. Here is her fundus photograph — please analyze it with the \"TENCENT AIMIS Fundus Disease Analyst Specialist\" expert and produce an analysis report.\n\n[Upload the fundus photo. This case ships with a sample image, sample-fundus-od.png (standard 45-degree, right eye OD, 3216x2136 original), which you can use directly to reproduce every result below; or swap in your own fundus photo — just request the original imaging file from your health check center or ophthalmology clinic]\n\nAdditional details:\n- Image type: standard 45-degree fundus photograph\n- Laterality: right eye (OD), only one eye submitted this time\n- Exam date: 2026-08-10\n\nPlease run the analysis through the real fundus AI model rather than guessing from general vision capability; if the API cannot be called, tell me directly instead of fabricating results.\n\nOrganize the report as: basic information, glaucoma assessment, multi-disease screening, lesion description, overall diagnosis. Within that:\n- List every condition the model covers item by item, marking each positive or negative. If a disease grade is returned, explain where that grade sits within the full grading scale and what it means\n- In the lesion section, list each sign-level finding and explain how multiple positive signs relate to one another; also give the quantified optic disc morphology measurements\n- In the overall diagnosis, summarize findings and order the follow-up advice by urgency\n\nRequirements:\n- Every conclusion must come from the model's actual response. States such as \"indeterminate\", \"unknown\", \"image not found\" or \"no value returned\" must be labeled honestly with their reasons — do not fill them in as negative on the model's behalf\n- If a quantified metric comes back as 0 or empty, judge whether that is a real measurement or a placeholder default, and never interpret a placeholder as real data. If the API gives no reason, state plainly that the cause cannot be determined from the response — do not speculate about causes\n- If conclusions from different sub-models appear to conflict (e.g. a grade that doesn't fully match the sign-level findings), raise it and explain rather than leaving a gap\n- If a positive finding is detected, state the urgency and a recommended timeframe for seeing a doctor plainly — do not soften the risk with vague wording. At the same time, point out the relatively reassuring aspects so the reader isn't needlessly alarmed\n- Write for a patient's family member with no medical background: gloss every technical term and append a glossary at the end\n- State clearly that this is AI-assisted screening and cannot replace an in-person ophthalmologist diagnosis; include the official PDF report link (noting it contains unredacted personal information) and save the raw API response alongside the report", "skills": [], "experts": [ { "id": "ex_TlAfmocr2Tj5", "name": "觅小影", "description": "基于普角与超广角眼底彩照的AI多病种分析，覆盖青光眼、糖网、AMD等数十种疾病的诊断与报告解读。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5", "promptFileSnapshot": "27008e55c7d1" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-26T15:15:00+08:00", "updated_at": "2026-09-02T11:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "fundus-glaucoma-ai-screening", "title": "眼底彩照 AI 筛出疑似青光眼", "title_en": "AI Fundus Screening Flags Suspected Glaucoma", "subtitle": "一张普角眼底照，AI 同时跑青光眼与 8 类眼病筛查，输出带量化指标的结构化报告", "subtitle_en": "One standard fundus photo, dual AI models screening glaucoma plus 8 eye diseases into a quantified structured report", "description": "体检报告上「眼底照相」那张图，拿到手往往只有一句「建议眼科随诊」，普通人完全看不懂，基层医生也未必有时间逐张细看视盘形态。\n\n这个案例把朋友老王的一张左眼普角眼底彩照交给腾讯觅影眼底 AI：青光眼模型与多病种模型并行分析，52 秒返回结果——左眼「疑似青光眼样眼底表现」，杯盘比 C/D 高达 0.82（正常通常 <0.6），盘沿比 I/N 0.44、T/N 1.05 显示 ISNT 法则被破坏；同时筛查糖网、AMD、视网膜静脉阻塞、高血压眼底病变均为阴性，仅豹纹状眼底阳性。产出一份带眼底原图、杯盘比仪表盘、ISNT 条形图的可视化报告，逐项解释杯盘比、盘沿比这些术语到底在说什么，并给出「优先做眼压+视野+OCT」的行动建议。\n\n适合：做眼科筛查的基层医疗机构、体检中心报告解读岗、想读懂自己眼底报告的糖尿病/高血压/高度近视人群、眼科 AI 产品的方案演示。", "description_en": "That fundus photo in your health-checkup report usually comes with a single line — \"see an ophthalmologist\" — which tells a layperson nothing, and primary-care doctors rarely have time to examine every optic disc in detail.\n\nThis case hands one left-eye standard 45° fundus photo to Tencent AIMIS fundus AI. The glaucoma model and the multi-disease model run in parallel and return in 52 seconds: the left eye shows \"suspected glaucomatous fundus appearance\", with a cup-to-disc ratio of 0.82 (normally below 0.6) and rim ratios I/N 0.44 and T/N 1.05 indicating a violated ISNT rule. Diabetic retinopathy, AMD, retinal vein occlusion and hypertensive retinopathy all came back negative; only tessellated fundus was positive. The output is a visual report embedding the original fundus image, a C/D gauge and an ISNT bar chart, explaining what terms like cup-to-disc ratio actually mean, and recommending IOP, visual field and OCT testing as the next step.\n\nGood for: primary-care clinics running eye screening, health-checkup report interpreters, people with diabetes/hypertension/high myopia who want to understand their own fundus report, and demos of ophthalmic AI solutions.", "categories": [ "business", "data" ], "tags": [ "眼底AI", "青光眼筛查", "医疗影像", "报告解读", "腾讯觅影" ], "tags_en": [ "Fundus AI", "Glaucoma Screening", "Medical Imaging", "Report Interpretation", "Tencent AIMIS" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "眼底 AI 可视化分析报告（含原图、杯盘比仪表盘、ISNT 图表）" }, { "file": "output.md", "label": "同源 Markdown 版报告（便于复制进病历系统）" } ], "prompt": "我有个朋友老王最近做了一次普角眼底检查，帮我基于腾讯眼科 AI 诊断专家分析一下他的眼底情况。\n\n影像文件：[拖入眼底彩照，如 OS_wangjianjun20260801_11127.jpg]\n检查信息：普角 45° 眼底彩照，左眼（OS）单眼，患者编号 P20260801001。\n\n请调用眼底 AI 完成真实分析（青光眼模型 + 多病种模型），然后把返回结果整理成一份给非专业人士也能看懂的可视化报告，包含：\n\n1. 基本信息与送检图像，说明图像质量和未送检眼的状态\n2. 青光眼评估结论，并用杯盘比的量化数值解释结论从何而来\n3. 视盘形态量化指标（杯盘比 C/D、盘沿比 I/N、S/N、T/N），说明是否符合 ISNT 法则\n4. 多病种筛查结果（糖网、AMD、视网膜静脉阻塞、高血压眼底病变、豹纹状眼底、高度近视眼底改变等），阳性项要解释临床含义\n5. 视网膜病灶体征逐项判别\n6. 综合印象与后续就医建议，按优先级排列\n\n杯盘比、盘沿比、ISNT 法则这类专业术语要配通俗解释。数值指标尽量做成图表，让偏高/偏低一眼可见。报告输出为单文件 HTML，眼底原图和所有资源内联嵌入、可离线打开，风格干净专业、贴近医疗报告质感；同时另出一份同内容的 Markdown 版方便复制。\n\n注意：所有结论必须来自 AI 真实返回结果，不要凭图片外观自行推断；明确标注这是辅助诊断参考，不能替代临床诊断。", "prompt_en": "A friend of mine, Lao Wang, recently had a standard fundus examination. Please analyze his fundus condition using the Tencent ophthalmic AI diagnosis expert.\n\nImage file: [drop in the fundus photo, e.g. OS_wangjianjun20260801_11127.jpg]\nExam info: standard 45° fundus photo, left eye (OS) only, patient ID P20260801001.\n\nRun a real analysis through the fundus AI (glaucoma model + multi-disease model), then turn the returned results into a visual report that a non-specialist can understand, covering:\n\n1. Basic information and the submitted image, noting image quality and the status of the eye not submitted\n2. The glaucoma assessment conclusion, explaining where it comes from using the quantified cup-to-disc ratio\n3. Optic disc morphometrics (cup-to-disc ratio C/D, rim ratios I/N, S/N, T/N) and whether the ISNT rule holds\n4. Multi-disease screening results (diabetic retinopathy, AMD, retinal vein occlusion, hypertensive retinopathy, tessellated fundus, myopic fundus changes, etc.), with clinical meaning for any positive finding\n5. Item-by-item retinal lesion sign classification\n6. Overall impression and follow-up recommendations, ordered by priority\n\nTerms like cup-to-disc ratio, rim ratio and the ISNT rule need plain-language explanations. Turn numeric indicators into charts so that high/low values are obvious at a glance. Output the report as a single-file HTML with the fundus image and all assets inlined so it opens offline, in a clean and professional style close to a real medical report; also produce a Markdown version with the same content for easy copying.\n\nNote: every conclusion must come from the actual AI response — do not infer anything from how the image looks. Clearly state that this is an auxiliary diagnostic reference and cannot replace a clinical diagnosis.", "skills": [], "experts": [ { "id": "fundus-disease-analysis", "name": "腾讯健康-觅影-眼底彩照疾病分析专家", "description": "对普角与超广角眼底彩照做 AI 多病种分析，覆盖青光眼、糖网、AMD 等数十种眼科疾病并输出结构化报告", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5" } ], "mcps": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-11T12:09:50Z", "updated_at": "2026-08-12T02:36:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fundus-report-interpretation", "title": "眼底AI报告解读", "title_en": "Fundus AI Report Interpretation", "subtitle": "把 AI 报告翻译成人话，逐项讲清每个发现", "subtitle_en": "Turn the AI report into plain language, explaining each finding", "description": "AI 报告出来了，一堆「2级糖网」「杯盘比 0.6」「黄斑水肿」看得人一头雾水，不知道到底严不严重、该不该去医院。\n\n本案例把一份眼底AI分析报告交给觅影眼底AI解读专家，逐项提取所有发现，用通俗语言讲清每项含义，按严重程度排序并标出需重点关注的条目，最后给出下一步建议。\n\n适合：拿到眼底AI报告看不懂的患者、体检用户、基层医生辅助讲解、慢病随访与健康管理。", "description_en": "The AI report is out, but \"grade-2 DR\", \"CDR 0.6\" and \"macular edema\" leave you confused about how serious it is and whether to see a doctor.\n\nThis case hands a fundus AI report to the Miying fundus AI interpretation expert, which extracts every finding, explains each one in plain language, ranks them by severity, highlights the items needing attention, and gives next-step advice.\n\nFor: patients puzzled by a fundus AI report, health check-up users, primary-care doctors explaining results, and chronic-disease follow-up.", "categories": [ "data" ], "tags": [ "报告解读", "眼底AI", "通俗解释", "就医指引", "腾讯健康" ], "tags_en": [ "Report Interpretation", "Fundus AI", "Plain Language", "Care Guidance", "Tencent Health" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "眼底AI报告解读（演示样例）" } ], "prompt": "解读这份眼底AI分析报告，告诉我具体有哪些发现。\n\n请以「腾讯健康-觅影-眼底彩照疾病分析专家」的能力，对我提供的这份眼底AI分析报告做通俗解读：\n\n1. 逐项列出报告中的所有发现（疾病诊断、体征分类、病灶检测、分割结果等）；\n2. 用通俗语言解释每项发现的含义（例如「2级糖网」意味着什么、「杯盘比 0.6」是否偏高、「黄斑水肿」要不要紧）；\n3. 按严重程度排序，标出需要重点关注或尽快就医的条目；\n4. 给出下一步建议（定期复查 / 专科就诊 / 尽快就医等）。\n\n如果你拿到的是报告原文，请先结构化提取字段再解读；如果你拿到的是眼底图像，请先完成 AI 分析、生成报告，再对报告做上述解读。\n\n输出通俗、分条清晰的解读报告，末尾声明：仅限科研用途，不替代临床诊断，最终以眼科医生判断为准。", "prompt_en": "Interpret this fundus AI report and tell me what the findings are.\n\nUsing the capabilities of the \"TENCENT AIMIS Fundus Disease Analyst Specialist\", give a plain-language interpretation of the fundus AI report I provide:\n\n1. List every finding in the report one by one (disease diagnosis, sign classification, lesion detection, segmentation results, etc.).\n2. Explain each finding in plain language (e.g. what \"grade-2 DR\" means, whether a \"CDR of 0.6\" is high, whether \"macular edema\" is serious).\n3. Rank findings by severity and flag the items that need close attention or prompt medical care.\n4. Give next-step advice (routine follow-up / specialist visit / prompt medical care).\n\nIf you receive the raw report, first extract the fields structurally, then interpret them; if you receive a fundus image, first run the AI analysis to produce the report, then interpret it as above.\n\nOutput a clear, itemized interpretation report, ending with the disclaimer: this is for research use only, does not replace clinical diagnosis, and the final decision rests with an ophthalmologist.", "skills": [ { "id": "fundus-images-ai-analysis", "name": "腾讯健康-觅影-基于45°眼底彩照的多病种 AI 分析", "description": "眼底 AI 多病种分析服务：上传眼底图像（JPEG/PNG/DICOM，单文件≤5MB）进行 AI 诊断与诊断结果查询", "slug": "fundus-images-ai-analysis" } ], "experts": [ { "id": "腾讯健康-觅影-眼底彩照疾病分析专家", "name": "腾讯健康-觅影-眼底彩照疾病分析专家", "description": "基于普角与超广角眼底彩照的AI多病种分析，覆盖青光眼、糖网、AMD等数十种疾病的诊断与报告解读", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T23:02:26+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "fundus-uwf-multidisease", "title": "超广角眼底多病种筛查", "title_en": "Ultra-Widefield Multi-Disease Screening", "subtitle": "一张超广角眼底图，一次筛查青光眼、糖网、AMD 等数十种疾病", "subtitle_en": "Screen glaucoma, DR, AMD and dozens of conditions from one ultra-widefield image", "description": "糖尿病、高血压、高度近视……眼底问题往往不止一种，逐项检查费时费力，还容易顾此失彼。\n\n本案例把一张超广角（UWF）眼底图像交给觅影眼底AI，一次性完成多病种筛查——糖尿病视网膜病变分级、青光眼杯盘比、AMD 分期，以及白内障、高血压视网膜病变等数十种疾病，汇总成一张筛查结果表并给出整体建议。\n\n适合：糖尿病患者年度眼底筛查、眼科体检、慢病管理平台、医学影像AI评测演示。", "description_en": "Diabetes, hypertension, high myopia — fundus problems rarely come one at a time, and checking them one by one is slow and easy to miss.\n\nThis case hands a single ultra-widefield (UWF) fundus image to the Miying fundus AI, which screens multiple diseases at once — diabetic retinopathy grading, glaucoma cup-to-disc ratio, AMD staging, plus dozens of conditions like cataract and hypertensive retinopathy — summarized into a screening table with overall advice.\n\nFor: annual fundus screening for diabetic patients, eye check-ups, chronic-disease management platforms, and medical-imaging AI demos.", "categories": [ "data" ], "tags": [ "超广角眼底", "多病种筛查", "糖网分级", "青光眼", "AI医学影像" ], "tags_en": [ "Ultra-Widefield Fundus", "Multi-Disease Screening", "DR Grading", "Glaucoma", "AI Medical Imaging" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "超广角眼底多病种筛查报告（演示样例）" } ], "prompt": "请对这份超广角眼底图像进行多病种筛查。\n\n请以「腾讯健康-觅影-眼底彩照疾病分析专家」的能力，对我上传的超广角（UWF）眼底图像做一次多病种筛查，覆盖但不限于：\n\n1. 糖尿病视网膜病变（DR）——分级（无 / 轻度 / 中度 / 重度非增殖 / 增殖）；\n2. 青光眼——杯盘比（CDR）与视神经改变；\n3. 年龄相关性黄斑变性（AMD）——干性 / 湿性及分期；\n4. 其他可识别的眼底疾病（白内障、高血压视网膜病变、病理性近视、视网膜静脉阻塞等）。\n\n对每个病种给出：检出与否、置信度、关键体征；汇总成一张多病种筛查结果表，并给出整体建议。\n\n测试数据可用 ODIR-5K（Ocular Disease Intelligent Recognition，含糖网/青光眼/白内障/AMD 等多病种标注）公开数据集中的一张超广角眼底图像，并注明来源。\n\n输出结构化报告，末尾声明：仅限科研用途，不替代临床诊断，最终以眼科医生判断为准。", "prompt_en": "Perform multi-disease screening on this ultra-widefield fundus image.\n\nUsing the capabilities of the \"TENCENT AIMIS Fundus Disease Analyst Specialist\", run a multi-disease screening on the ultra-widefield (UWF) fundus image I upload, covering but not limited to:\n\n1. Diabetic retinopathy (DR) — grading (none / mild / moderate / severe non-proliferative / proliferative).\n2. Glaucoma — cup-to-disc ratio (CDR) and optic nerve changes.\n3. Age-related macular degeneration (AMD) — dry/wet and staging.\n4. Other detectable fundus diseases (cataract, hypertensive retinopathy, pathologic myopia, retinal vein occlusion, etc.).\n\nFor each disease give: detected or not, confidence, and key signs; summarize into a multi-disease screening table with overall advice.\n\nFor test data you may use an ultra-widefield fundus image from the ODIR-5K (Ocular Disease Intelligent Recognition) public dataset, which labels DR/glaucoma/cataract/AMD and more, and cite the source.\n\nOutput a structured report, ending with the disclaimer: this result is for research use only, does not replace clinical diagnosis, and the final decision rests with an ophthalmologist.", "skills": [ { "id": "ultra-wide-fundus-images-ai-analysis", "name": "腾讯健康-觅影-基于超广角眼底彩照的多病种/病灶 AI 分析", "description": "超广角眼底多病种 AI 分析服务：支持推测诊断、体征判别、体征检测和体征分割四类 AI 能力", "slug": "ultra-wide-fundus-images-ai-analysis" } ], "experts": [ { "id": "腾讯健康-觅影-眼底彩照疾病分析专家", "name": "腾讯健康-觅影-眼底彩照疾病分析专家", "description": "基于普角与超广角眼底彩照的AI多病种分析，覆盖青光眼、糖网、AMD等数十种疾病的诊断与报告解读", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T23:02:26+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "goal-tracker", "title": "绝对小梦的目标管理台", "title_en": "Goal Tracker Workbench", "subtitle": "目标打卡 × WOOP 障碍对策 × 工作计划 × 商单收入，自媒体创作者的进度管理台", "subtitle_en": "Goal check-ins × WOOP obstacle plans × work schedules × deal income — a progress desk for content creators", "description": "目标定了却总在拖延？本案例由 WorkspaceBuilder 工作台搭建师专家生成：一份单文件 HTML 收进自媒体创作者的四个进度模块——「目标打卡」支持学习/工作两类目标，设总量与截止日，每日打卡必须提交成果（感悟灵感 / 发布链接+复盘 / 成品交付 / 成果记录四选一），自动算完成率、连续天数、剩余天数和预计完成日；每个目标内置 WOOP 法：写下最容易拦住你的障碍 + 对策，打卡页与看板都会展示；「工作计划」在日历上按天排工作/内容/电商/学习计划，完成时同样要交成果并生成 AI 风格成长反馈（讨论 / 关联灵感 / 升级建议）；「商单收入」按月记录商单名称、平台、收入、成本，自动算净利润与完成数；「周报」一键汇总本周打卡、成果亮点、环比上周与 KPT 复盘，复制即用。另有连续打卡奖励机制（7 天/30 天/180 天解锁）和求上进提醒。\n\n数据存浏览器 localStorage，内置作者真实使用场景的示例数据（读《金钱心理学》、做 AI 漫剧视频、口播视频、学会做 AI 漫剧四个目标 + 17 条计划 + 9 条商单），打开即体验；界面内置 i18n 中英双语，?lang=zh / ?lang=en 切换，示例数据同样双语文案。暖白底色 + 橙紫渐变点缀的清爽效率风，手机和电脑都好用。\n\n适合：自媒体 / 副业创作者（目标+商单收入一页管）、想用 WOOP 法对抗拖延的习惯养成者、知识博主（打卡即积累选题灵感）、以及想参考单文件 localStorage 应用与内联 i18n 实现的前端学习者。", "description_en": "Goals set but always put off? This case is built by the WorkspaceBuilder expert: a single-file HTML that packs four progress modules for content creators — \"Goal Check-ins\" supports Learning / Work goals with a total target and deadline; every daily check-in must submit proof of output (reflection & idea / post link + review / deliverable / output log), and progress rate, streak days, days left and estimated finish date are computed automatically. Each goal embeds the WOOP method: write down the obstacle most likely to stop you plus a countermeasure, shown on the check-in page and the board. \"Work Schedules\" arranges work / content / e-commerce / study plans on a calendar; completing one also requires output and generates growth-style feedback (discussion / related ideas / level-up suggestions). \"Deal Income\" tracks deals by month with platform, income and cost, auto-calculating net profit and completion count. \"Weekly Report\" aggregates the week's check-ins, output highlights, WoW comparison and a KPT review in one click, ready to copy. There's also a streak-reward system (7 / 30 / 180 days) and playful push reminders.\n\nData persists in browser localStorage; a realistic sample dataset from the author is built in (four goals — reading \"The Psychology of Money\", making AI comic dramas, talking-head videos, learning AI comic dramas — plus 17 plans and 9 deals) so it works on first open. The UI ships with inline i18n in Chinese and English — switch with ?lang=zh / ?lang=en, and the sample data comes in both languages too. Warm off-white background with orange-purple gradient accents, a clean efficient look, friendly on both phone and desktop.\n\nBest for: content creators and side-hustlers (goals + deal income in one page), habit builders who want the WOOP method to beat procrastination, knowledge bloggers (every check-in accumulates topic ideas), and frontend learners studying single-file localStorage apps with inline i18n.", "categories": [ "workspace" ], "tags": [ "目标管理", "打卡", "WOOP", "自媒体", "个人工作台" ], "tags_en": [ "goal tracking", "check-in", "WOOP", "content creator", "personal workspace" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "目标管理台 单文件 HTML（内置 i18n 中英文 + 中英双版示例数据，localStorage 自包含）" } ], "prompt": "请使用资料库的能力帮我一键做同款：https://www.workbuddy.cn/space/d/9QIroSB7sqFmW7t9NU51jQ，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n我想做一个自媒体创作者用的「目标管理台」个人工作台，包含：\n1. 目标打卡：可新建学习 / 工作两类目标（目标名、单位、总量、截止日期、配色），每日打卡必须提交成果——按目标类型四选一（感悟灵感 / 发布链接+复盘 / 成品交付 / 成果记录），自动计算完成率、连续打卡天数、剩余天数、预计完成日（按近 14 天日均量推算，提前/拖后几天）。\n2. WOOP 法：每个目标可写「最容易拦住你的障碍」和「对策：如果它出现，我就……」，在目标卡和打卡页展示。\n3. 工作计划：日历视图按天新增工作/内容/电商/学习计划，当天可打卡完成（同样提交成果并生成反馈），逾期标红。\n4. 商单收入：按月记录商单（名称、客户/品牌、平台 小红书/抖音/公众号/B站/电商、收入、成本、完成日期、备注），自动汇总已结收入、总成本、净利润、完成数。\n5. 周报：按周汇总各目标投入（数量/天数/分钟）、本周成果亮点、环比上周、KPT 复盘（保持/问题/尝试/下周预案），一键生成可复制的周报文本。\n6. 激励：连续打卡奖励（7 天/30 天/180 天解锁对应奖励）+ 求上进提醒横幅（深夜 / 未完成时出现）。\n7. 数据存浏览器 localStorage，首次打开内置示例数据（四个目标 + 打卡记录 + 计划 + 商单）方便直接体验。\n\n视觉走清爽效率风：暖白底色、卡片式布局、橙紫渐变点缀，信息密度高但不拥挤；响应式，手机和电脑都好用。要求CSS / JS 全部内联，不引用任何外部资源。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the library's capability to make me an exact replica of this workbench: https://www.workbuddy.cn/space/d/9QIroSB7sqFmW7t9NU51jQ — data must be stored online with two-way sync. Prefer the one-click replica path; if that's not available, build it as follows: I want a \"Goal Tracker\" personal workbench for content creators, including:\n1. Goal check-ins: create Learning / Work goals (name, unit, total target, deadline, color). Every daily check-in must submit proof of output — one of four types per goal (reflection & idea / post link + review / deliverable / output log). Auto-compute completion rate, current streak, days left and estimated finish date (derived from the last-14-days daily average, showing early/late by N days).\n2. WOOP method: each goal can have \"the obstacle most likely to stop you\" and a \"countermeasure: if it happens, I will…\", shown on the goal card and the check-in page.\n3. Work schedules: a calendar view to add work / content / e-commerce / study plans per day; complete a plan the day it's due (submitting output and getting feedback), overdue shown in red.\n4. Deal income: track deals by month (name, client/brand, platform Xiaohongshu / Douyin / WeChat OA / Bilibili / E-commerce, income, cost, completion date, notes), auto-summarizing settled income, total cost, net profit and completion count.\n5. Weekly report: per-week summaries of each goal's input (amount / days / minutes), weekly output highlights, WoW comparison, and a KPT review (Keep / Problems / Try / Next Week), generating a copy-ready report in one click.\n6. Motivation: streak rewards (unlock at 7 / 30 / 180 days) plus a playful push-reminder banner (appears late at night or when goals are unmet).\n7. Data persists in browser localStorage; ship built-in sample data (four goals + check-in records + plans + deals) so it works on first open. Visual style: a clean efficient look — warm off-white background, card layout, orange-purple gradient accents, dense but not crowded; responsive for both phone and desktop. Deliver a with all CSS / JS inline and zero external resources.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "绝对小梦", "name_en": "Jue Dui Xiaomeng", "avatar": "", "badge": "community" }, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "created_at": "2026-08-17T15:25:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "gold-buy-now", "title": "黄金还能买吗？", "title_en": "Should I Buy Gold Now?", "subtitle": "小达养基笔记：用配置思维看黄金时机", "subtitle_en": "Xiaoda's Fund Notes: a configuration-mindset take on gold timing", "description": "【痛点】普通投资者面对\"黄金还能买吗\"这种时机型问题，容易被涨跌情绪牵着走，又找不到权威、合规的解读。【演示】本案例用易方达投教基地「小达养基笔记」技能，基于财富学堂投教知识库，以三段式结构输出一份自包含图文回答：先给真实金价位置（伦敦金/上海金+近一年走势图），再拆解\"谁在买金\"（央行/ETF/投机三类参与者与定价逻辑），最后把\"涨跌思维\"转为\"配置思维\"并给出三问自检框架，全程不替用户做买卖决策。【适用人群】想系统理解黄金投资、需要投教级合规解读、或正犹豫是否配置黄金的基金/理财用户。", "description_en": "Pain point: when facing timing questions like \"should I buy gold now\", ordinary investors are easily swayed by price sentiment and lack an authoritative, compliant interpretation. Demo: this case uses the E Fund investor-education assistant \"Xiaoda's Fund Notes\" to produce a self-contained illustrated answer grounded in the Fortune Academy knowledge base, in a three-part structure: first the real gold price position (London/Shanghai gold + a one-year trend chart), then \"who is buying gold\" (central banks / ETFs / speculators and their pricing logic), and finally shifting from a \"rise/fall mindset\" to a \"allocation mindset\" with a three-question self-check — never making the buy/sell decision for the user. Target audience: fund/wealth-management users who want a systematic, compliant understanding of gold investing or are hesitating about allocating to gold.", "categories": [ "finance", "education" ], "tags": [ "黄金", "投资理财", "投教问答", "基金" ], "tags_en": [ "Gold", "Investing", "Investor Education Q&A", "Funds" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "黄金还能买吗.html", "preview_en": "", "external_url": "", "artifacts": [ { "file": "黄金还能买吗.html", "label": "完整图文回答（含金价走势图）" } ], "prompt": "用 efunds-edu-query 回答：黄金现在还能买吗？", "prompt_en": "Use the efunds-edu-query skill to answer: Should I buy gold now?", "skills": [ { "id": "efunds-edu-query", "name": "易方达投教问答", "description": "易方达投教基地「小达养基笔记」智能问答助手——基于财富学堂投教知识库，以三段式结构精准回答基金/ETF/黄金等投资品种的概念、选择与时机类问题，全文零第三人称转述、开场即观点，不处理纯行情查询、个股推荐、K线分析等非投教类问题。", "skill_id": "skill_2089898953429123072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T11:13:00+08:00", "updated_at": "2026-08-31T17:01:56+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "gold-miner-workbench", "artifact_type": "html", "title": "淘金矿工小屋 · 自律全能工作台", "title_en": "Gold Miner Cabin · All-Round Discipline Workbench", "subtitle": "完成现实任务＝下矿赚金币，金币购物、领养宠物、布置小屋，把自律打卡玩成黄金矿工经营游戏", "subtitle_en": "Real-life tasks become mining shifts that earn coins — spend them on furniture, pets and cabin decor, turning self-discipline into a gold-miner management game", "description": "自律最难的是坚持：打卡 App 记完就关，没有任何正反馈。这个工作台把自律做成复古黄金矿工经营游戏——完成饮食、运动、记账、背单词等现实任务就是下矿挖金币，金币能在商城买家具、领养宠物、布置自己的矿洞小屋，越自律小屋越热闹。六个矿道任务各管一件事：饮食台账提交餐食自动估算热量、运动台账记录时长计算热量缺口、记账本记一笔自动汇总当月收支、单词矿脉每天推送 5 个单词背诵打卡、亚马逊矿道看视频记学习笔记、小红书工坊输入创意一键生成内容方案；每次提交即结算金币并汇总成日报周报月报，打卡历史用热力图记录连续天数，成就馆 15 枚成就按累计数据自动解锁。数据落在资料库数据表，支持在线存储与双向同步。适合：管不住自己的自律困难户、需要游戏化正反馈才坚持得下去的打卡党，以及喜欢养成类小游戏、想给生活加点奔头的人。", "description_en": "The hardest part of self-discipline is sticking with it: you log a habit in some app, close it, and get nothing back. This workbench turns discipline into a retro gold-miner management game — completing real-life tasks like logging meals, workouts, expenses and vocabulary becomes a mining shift that earns virtual coins, and coins buy furniture, adopt pets and decorate your own mine cabin in the shop, so the more disciplined you are, the livelier your cabin gets. Six task tunnels each cover one thing: the Diet Log estimates calories from submitted meals, the Exercise Log computes your calorie gap from workout duration, the Ledger auto-summarizes monthly income and expense, the Word Vein pushes 5 words a day for memorization check-ins, the Amazon Tunnel records study notes from daily videos, and the RED Workshop generates a content plan from one idea; every submission settles coins instantly and rolls up into daily, weekly and monthly reports, check-in history draws a heatmap with streak tracking, and 15 achievements unlock automatically from cumulative stats. All data is stored in Library database tables with online storage and two-way sync. Great for anyone who struggles with self-discipline, check-in users who need gamified feedback to keep going, and fans of raising-style games who want daily life to feel more rewarding.", "tags": [ "自律打卡", "游戏化养成", "习惯追踪", "生活管理", "宠物养成" ], "tags_en": [ "self-discipline check-in", "gamified habit building", "habit tracking", "life management", "pet raising" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/2Lv1kBwDBG23M1XjXvdYTo，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「淘金矿工小屋」自律打卡工作台，把现实任务打卡做成复古黄金矿工经营游戏：完成饮食、运动、记账、单词、亚马逊运营学习、小红书创作 6 类现实任务即下矿赚虚拟金币，金币用于商城购物、领养宠物、布置矿洞小屋，形成正向养成闭环。数据表四张：任务打卡记录表（日期、任务类型、提交内容、获得金币、打卡时间）、金币流水表（日期、类型、金额、原因备注）、背包物品表（物品名称、类型、数量、获取方式）、宠物表（名字、种类、等级、经验、饱食度、心情、健康）。\n\n模块与交互：① 金矿主控大厅——持有金币、本月总收入、今日金币三卡总览，今日挖矿进度 0/6 与六条矿道入口，家园快捷入口，当日热量收支、消费汇总与通知栏；② 六条矿道任务——饮食台账提交餐食自动估算热量、运动台账记录时长自动计算热量缺口、淘金记账本记收支自动汇总当月并给结余、单词矿脉每日推送 5 个单词背诵打卡、亚马逊学习矿道每日推视频并记录知识点笔记、小红书工坊输入创意一键生成标题提纲脚本分镜标签并保存方案；每次提交后领取对应金币（饮食+5、运动+5、记账+6、单词+4、亚马逊+7、小红书+10），同一矿道当日重复提交不再发币；③ 金矿商城——家具、宠物、宠物粮食、家园道具四类商品，购买自动扣币入背包，领养宠物弹窗命名入笼；④ 个人背包——查看与使用物品，宠物粮食喂食提升宠物饱食度、心情、健康；⑤ 矿洞小屋——6×4 格子布置家具，左键放置、右键收回背包；⑥ 宠物庭院——喂食、抚摸、洗澡提升宠物状态，每日互动额外奖励 3 金币，宠物饿了自动预警；⑦ 挖矿成就馆——15 项成就按累计打卡、累计金币、宠物与家具数量自动解锁并发放金币奖励；⑧ 报表与设置——挖矿周报月报（本周金币、日均任务、每日金币趋势、月度消费与月度总评）、打卡历史（总打卡数、连续天数、挖矿热力图）、工作台设置（每日凌晨自动重置今日金币保留累计数据、通知与宠物预警开关、管理员手动调整金币、危险区一键重置）。视觉走复古淘金暖金配色、柔和护眼风格，PC 侧边栏导航、移动端底部导航双布局。\n\n交付：资料库在线工作台——在线 HTML 页面 + 8 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the Library capability to build me a same-style copy of https://www.workbuddy.cn/space/d/2Lv1kBwDBG23M1XjXvdYTo with online data storage and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Gold Miner Cabin\" self-discipline workbench that turns real-life task check-ins into a retro gold-miner management game: completing six categories of real-life tasks — diet, exercise, bookkeeping, vocabulary, Amazon ops learning and RED (Xiaohongshu) content creation — counts as a mining shift that earns virtual coins; coins are spent in the shop on furniture, adopting pets and decorating the mine cabin, forming a positive nurturing loop. Four database tables: Task Check-in Record (date, task type, submission content, coins earned, check-in time), Coin Ledger (date, type, amount, reason/note), Backpack Items (item name, type, quantity, acquisition source), Pets (name, species, level, EXP, fullness, mood, health). Modules and interactions: (1) Gold Mine Control Hall — overview cards for coins held, earned this month and coins today, today's mining progress 0/6 with six tunnel entries, home shortcuts, today's calorie in/out, spending summary and a notification board; (2) Six task tunnels — the Diet Log estimates calories automatically from submitted meals, the Exercise Log computes the calorie gap from workout duration, the Gold Ledger summarizes monthly income/expense with balance, the Word Vein pushes 5 words daily for memorization check-ins, the Amazon Learning Tunnel recommends a daily video and records key-point notes, and the RED Workshop generates title, outline, script, storyboard and tags from one idea and saves the plan; each submission claims its coins (diet +5, exercise +5, bookkeeping +6, vocabulary +4, Amazon +7, RED +10), and repeated submissions in the same tunnel on the same day earn no extra coins; (3) Gold Mine Shop — four product categories (furniture, pets, pet food, home props); purchases deduct coins into the backpack automatically, and adopting a pet opens a naming modal; (4) Backpack — view and use items; pet food raises pet fullness, mood and health; (5) Mine Cabin — a 6x4 grid for placing furniture, left-click to place, right-click to return to backpack; (6) Pet Yard — feed, pet and bathe to raise pet stats, daily interaction earns a 3-coin bonus, and hungry pets trigger alerts; (7) Mining Achievement Hall — 15 achievements unlock automatically from cumulative check-ins, coins, pets and furniture, each paying out coin rewards; (8) Reports and settings — mining weekly/monthly reports (coins this week, avg tasks per day, daily coin trend, monthly spending with monthly review), check-in history (total check-ins, streak days, mining heatmap), and workbench settings (auto-reset today's coins at midnight while keeping lifetime totals, notification and pet-alert toggles, admin coin adjustment, and a danger-zone full reset). Visual direction: retro gold-rush warm palette, soft eye-friendly styling, with a PC sidebar layout and a mobile bottom-nav layout.\n\nDeliverable: an online Library workbench — an online HTML page plus 8 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "淘金矿工小屋 · 自律全能工作台（自包含单文件）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "杨柳枝", "name_en": "Yangliuzhi", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "created_at": "2026-09-10T12:47:21+08:00", "updated_at": "2026-09-10T16:45:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "gold-research-workbench", "title": "黄金研究工作台", "title_en": "Gold Research Workbench", "subtitle": "多资产行情与个人研判复盘", "subtitle_en": "Multi-asset research and personal review", "description": "行情、宏观与黄金数据分散在多个文件里，研究时难以快速交叉观察，也很难留下自己的判断。这个工作台把黄金、美元、美债、美股、非农与央行购金汇总成卡片、走势图、数据表和 CSV 下载区，并新增研判记录与复盘日期。适合投研人员、宏观研究者和需要持续沉淀观点的团队。", "description_en": "Market, macro, and gold data often live in separate files, making cross-checks slow and personal views hard to retain. This workbench brings gold, the US dollar, Treasuries, equities, nonfarm data, and central-bank purchases into quote cards, charts, tables, CSV downloads, and a personal research log with review dates. For investment researchers, macro analysts, and teams that continuously build their own views.", "tags": [ "黄金研究", "多资产研究", "宏观数据", "研判复盘" ], "tags_en": [ "Gold research", "Multi-asset research", "Macro data", "Research review" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "黄金研究工作台（双语自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://workbuddy.link/p/t8VcWtwi2wFka2ePuyWrjJ，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n制作 黄金研究工作台，接入并在线管理五张资料库数据表：黄金行情快照（资产、代码、价格、涨跌幅、单位、时间、来源）、伦敦金日线（日期、收盘价）、美国非农月度（月份、非农净变动）、全球央行购金（记录类型、季度、净购金、年份、年度净购金）、黄金研判笔记（关注资产、方向、观点、关键价位、关键事件、下次复盘、记录时间）。首屏展示四大资产行情卡片、伦敦金近一年走势、美国非农月度变动、全球央行季度购金；通过显式页签进入独立的“我的研判”子页面，用户可以记录看多/看空/中性判断、关键价位、关键事件和下一次复盘日期，并将记录在线保存到资料库。支持图表/数据表切换、CSV 下载、数据来源和免责声明；视觉沿用浅色卡片与金色强调，数据和脚本页面支持响应式浏览。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://workbuddy.link/p/t8VcWtwi2wFka2ePuyWrjJ, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a Gold Research Workbench backed by five online library tables: Gold quote snapshots (asset, code, price, change, unit, time, source), London gold daily series (date, close), US nonfarm monthly data (month, net change), global central-bank gold purchases (record type, quarter, net purchases, year, annual net purchases), and research notes (asset, direction, view, key levels, key event, next review date, recorded time). Show four asset quote cards, one-year London gold trends, monthly US nonfarm changes, and quarterly central-bank purchases on the first screen. Use an explicit tab to open a separate My Research Notes page, where users can save bullish, bearish, or neutral views, key levels, key events, and next review dates to the online library table. Support chart/table tabs, CSV downloads, source labels, and a disclaimer. Use a light card layout with gold accents, inline all data and script resources, and keep the page responsive.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "李稻葵", "name_en": "李稻葵", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 90, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "created_at": "2026-08-28T08:12:11+00:00", "updated_at": "2026-08-28T08:12:11+00:00", "is_recommended": false, "recommended_order": 0 }, { "id": "goujian-sword-infographic", "title": "越王勾践剑一图拆解", "title_en": "Sword of Goujian: One-Page Infographic", "subtitle": "指定一件文物，自动生成可展陈的科普图解展板", "subtitle_en": "Name any relic, auto-generate a display-ready infographic board", "description": "想给一件国宝做科普展板，最头疼的是既要史料准确、又要排版好看，查资料、写文案、做设计往往要折腾好几天。\n\n这个案例把「越王勾践剑」交给腾讯探元文博专家，一句话产出了一张单文件 HTML 科普图解展板：从基础信息、鸟篆铭文解读、出土流传之谜，到剑首 11 道同心圆、菱形暗格纹、复合金属结构等工艺亮点，再到同族名剑对比表和历史时间轴，暗金青铜主视觉一气呵成，文字信息全部来自探元知识库并逐处标注来源。\n\n适合：博物馆策展与讲解人员、文博科普内容编辑、历史与美育教师、研学课程设计者、文博爱好者。", "description_en": "Making a science-popularization board for a national treasure is a pain: it must be historically accurate and visually polished, so researching, copywriting and designing can eat up days.\n\nThis case hands the 'Sword of Goujian' to the TanYuan Cultural Heritage Expert and produces, from a single prompt, a single-file HTML infographic board: basic facts, the bird-seal inscription, the mystery of its excavation and circulation, craftsmanship highlights (11 concentric rings on the pommel, the rhombic dark-lattice pattern, the composite-metal structure), a comparison table with a sister royal sword, and a historical timeline, all in a coherent dark-bronze visual, with every text fact sourced from the TanYuan knowledge base.\n\nIdeal for: museum curators and docents, cultural-heritage content editors, history and aesthetics teachers, study-tour course designers, and heritage enthusiasts.", "categories": [ "education", "creation" ], "tags": [ "文物科普", "越王勾践剑", "图解展板", "博物馆", "单文件HTML" ], "tags_en": [ "Relic Science", "Sword of Goujian", "Infographic Board", "Museum", "Single-file HTML" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "越王勾践剑科普图解展板（单文件 HTML）" } ], "prompt": "请以「越王勾践剑」为主题，生成一张可用于博物馆展陈或线上科普的文物图解展板，输出为单文件 HTML。\n\n内容要求（信息必须来自探元文博知识库，逐处标注来源，不臆造）：\n1. 基础信息：年代、材质、等级、尺寸、出土地与时间、馆藏机构；\n2. 铭文解读：剑身鸟篆铭文「越王鸠浅自作用剑」的释读与考证；\n3. 出土与流传：出土情况、存世越王剑数量、为何现身楚墓的谜团；\n4. 工艺亮点：剑首同心圆、菱形暗格纹、剑格镶嵌等；\n5. 复合金属结构示意（定性，注明以官方检测为准）；\n6. 同族名剑对比表（如越王勾践剑 vs 越王州勾剑）；\n7. 历史时间轴（结合勾践生平史事与出土时间）。\n\n视觉风格：暗金青铜、庄重典藏气质，信息密度高、层次清晰，适合竖版长图展陈。技术约束：单文件 HTML，所有 CSS/JS/图片资源全部内联，可离线打开，不引用任何外部链接。", "prompt_en": "Using the 'Sword of Goujian' as the theme, generate a relic infographic board suitable for museum display or online science popularization, output as a single-file HTML.\n\nContent requirements (all facts must come from the TanYuan cultural-heritage knowledge base, with sources noted throughout, no fabrication):\n1. Basic facts: era, material, protection grade, dimensions, excavation site and date, holding institution;\n2. Inscription reading: interpretation and textual research of the bird-seal inscription '越王鸠浅自作用剑' on the blade;\n3. Excavation and circulation: how it was unearthed, the number of surviving Yue royal swords, and the mystery of why it appeared in a Chu tomb;\n4. Craftsmanship highlights: concentric rings on the pommel, rhombic dark-lattice pattern, guard inlays, etc.;\n5. A qualitative diagram of the composite-metal structure (note that official testing is authoritative);\n6. A comparison table with a sister royal sword (e.g., Sword of Goujian vs. Sword of Zhougou);\n7. A historical timeline (combining Goujian's life events with the excavation date).\n\nVisual style: dark-bronze, solemn and archival, high information density with clear hierarchy, suited to a tall vertical display layout. Technical constraints: single-file HTML with all CSS/JS/image resources fully inlined, openable offline, referencing no external links.", "skills": [], "experts": [ { "id": "TanyuanCulturalHeritageExpert", "name": "文博领域AI顾问", "name_en": "Cultural Heritage AI Consultant", "description": "基于腾讯探元文博知识库，提供文物与世界遗产查询、知识问答、文物对比、文博攻略及研学方案等专业服务。", "description_en": "Powered by TanYuan, providing cultural relic and world heritage queries, knowledge Q&A, relic comparison, museum guides and study programs.", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "fbca4e25911e", "expert_id": "ex_84afKR4m4iUc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-28T12:13:24+08:00", "updated_at": "2026-07-28T12:13:24+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "goujian-sword-postcard", "title": "越王勾践剑文物明信片", "title_en": "Sword of Goujian: Heritage Postcard", "subtitle": "指定一件文物，自动生成典藏风明信片/海报", "subtitle_en": "Name any relic, auto-generate a collectible postcard / poster", "description": "想把一件国宝做成明信片或海报发出去，最难的往往不是排版，而是那口「文物气韵」——配色、意境文字、留白节奏稍微一歪，就成了廉价旅游纪念品。\n\n这个案例把「越王勾践剑」交给腾讯探元文博专家，一句话产出了一张竖版典藏风明信片：暗底青铜主视觉搭配文物特写，「青铜千古 · 越地风华」「千年不锈 · 寒光依旧」等意境文案点睛，配以馆藏与年代信息，字体、背景与图片全部内联，是一张可直接下载打印或分享的单文件 HTML 明信片/海报。\n\n适合：博物馆文创与新媒体运营、展览宣传与海报设计者、文博科普内容编辑、历史与美育教师、文博爱好者。", "description_en": "Turning a national treasure into a postcard or poster is less about layout and more about capturing the artifact's 'aura' — get the palette, poetic copy, or white-space rhythm slightly wrong and it looks like a cheap tourist souvenir.\n\nThis case hands the 'Sword of Goujian' to the TanYuan Cultural Heritage Expert and produces, from a single prompt, a vertical collectible-style postcard: a dark bronze visual with a close-up of the relic, poetic lines such as 'Bronze through the ages · the splendor of Yue' and 'Rust-free for millennia · its cold gleam remains', paired with collection and era details. Fonts, background and images are all inlined, making it a single-file HTML postcard/poster ready to download, print, or share.\n\nIdeal for: museum cultural-creative and new-media teams, exhibition-promotion and poster designers, heritage content editors, history and aesthetics teachers, and heritage enthusiasts.", "categories": [ "creation", "education" ], "tags": [ "文物明信片", "越王勾践剑", "海报设计", "博物馆文创", "单文件HTML" ], "tags_en": [ "Heritage Postcard", "Sword of Goujian", "Poster Design", "Museum Creative", "Single-file HTML" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "越王勾践剑文物明信片（单文件 HTML）" } ], "prompt": "请以「越王勾践剑」为主题，生成一张可下载打印或分享的文物明信片/海报，输出为单文件 HTML。\n\n内容要求（文物信息须来自探元文博知识库，不臆造）：\n1. 主视觉：以文物特写为核心，突出剑身菱形暗格纹、寒光质感；\n2. 意境文案：搭配几组凝练的中式意境短句（如「青铜千古 · 越地风华」「千年不锈 · 寒光依旧」），点出文物气韵；\n3. 信息落款：标注文物名称、年代、馆藏机构等关键信息。\n\n视觉风格：竖版明信片版式（约 832:1216），暗底青铜色调、典藏庄重气质，中式书法/印章元素点缀，留白讲究、意境优先。技术约束：单文件 HTML，字体、背景与图片等所有资源全部内联（base64），可离线打开、直接下载打印，不引用任何外部链接。", "prompt_en": "Using the 'Sword of Goujian' as the theme, generate a relic postcard/poster that can be downloaded, printed, or shared, output as a single-file HTML.\n\nContent requirements (relic facts must come from the TanYuan cultural-heritage knowledge base, no fabrication):\n1. Main visual: center a close-up of the relic, highlighting the blade's rhombic dark-lattice pattern and its cold-gleam texture;\n2. Poetic copy: pair with a few concise Chinese-style poetic lines (e.g., 'Bronze through the ages · the splendor of Yue', 'Rust-free for millennia · its cold gleam remains') to convey the artifact's aura;\n3. Caption info: note the relic's name, era, and holding institution.\n\nVisual style: a vertical postcard layout (about 832:1216), dark bronze tones, a solemn collectible feel, accented with Chinese calligraphy/seal elements, refined white space, mood-first. Technical constraints: single-file HTML with all resources — fonts, background, and images — fully inlined (base64), openable offline and directly printable, referencing no external links.", "skills": [], "experts": [ { "id": "TanyuanCulturalHeritageExpert", "name": "文博领域AI顾问", "name_en": "Cultural Heritage AI Consultant", "description": "基于腾讯探元文博知识库，提供文物与世界遗产查询、知识问答、文物对比、文博攻略及研学方案等专业服务。", "description_en": "Powered by TanYuan, providing cultural relic and world heritage queries, knowledge Q&A, relic comparison, museum guides and study programs.", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "fbca4e25911e", "expert_id": "ex_84afKR4m4iUc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-28T14:12:10+08:00", "updated_at": "2026-07-28T14:12:10+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "goujian-sword-research", "title": "越王勾践剑深度考证", "title_en": "Deep Verification of the Sword of Goujian", "subtitle": "春秋吴越青铜剑的起源、工艺、铭文与馆藏全维度考证", "subtitle_en": "Full-dimensional verification of origin, craftsmanship, inscription and collection of the Spring and Autumn Wu-Yue bronze sword", "description": "研究一把名剑最怕资料零散、铭文释读混乱、出土与馆藏争议难辨。这个案例演示了如何借助腾讯探元文物库与知识库，并行检索馆藏与知识解读，再由主代理亲自做 2-3 次多源 WebSearch 交叉核验，覆盖出土与考古背景、冶金与装饰工艺、铭文释读和当前馆藏与保护状态四个维度，最终交付包含配图与图表的 Markdown/HTML/DOCX/PDF 四格式深度考证报告。\n\n适合：青铜器馆研究员、文博深度考证用户、博物馆策展与社教人员、考古与艺术史专业学生、文史与兵器史爱好者。", "description_en": "Studying a famous sword is painful when sources are scattered, inscription readings conflict, and the excavation-to-collection provenance is disputed. This case demonstrates how to leverage the Tencent Tanyuan relic and knowledge bases for parallel retrieval, then run 2-3 multi-source WebSearch passes for cross-verification across four dimensions—excavation and archaeological context, metallurgy and decorative craftsmanship, inscription reading, and current collection and conservation state—delivering a cross-checked four-format (MD/HTML/DOCX/PDF) deep verification report with embedded images and charts.\n\nBest for: bronze-ware museum researchers, deep cultural heritage verification users, curators and docents, archaeology and art-history students, and enthusiasts of cultural and military history.", "categories": [ "business" ], "tags": [ "青铜器", "吴越文化", "鸟篆铭文", "深度考证", "馆藏研究" ], "tags_en": [ "Bronze Ware", "Wu-Yue Culture", "Bird-Script Inscription", "Deep Verification", "Museum Collection" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告" } ], "prompt": "请按「腾讯探元学术研究-深度考证」专家的工作流，深度研究「越王勾践剑」。\n\n研究范围（覆盖四维）：\n- 起源与考古背景：春秋晚期吴越铸剑史、出土与传世争议、当前馆藏\n- 冶金与装饰工艺：金属配比、菱形暗格花纹、剑首同心圆铸造技术\n- 铭文与历史叙事：剑身鸟篆铭文「越王勾践 自作用剑」释读\n- 馆藏与保护现状：湖北省博物馆藏状态、保护技术与展陈\n\n交付要求（强制）：\n1. 探元两库 MCP（search_movable_relics + search_relics）并行检索 + 主代理 2-3 次 WebSearch 多源核验\n2. 配图 ≥ 2 张（建议 3-6 张），优先探元两库 cover；图表 ≥ 1 张（建议 2-3 张，如发展脉络时间线/工艺结构图/金属配比数据对比）\n3. 四格式报告：Markdown 母版 → HTML → DOCX → PDF，10 章节固定结构\n4. 仅结论不含建议；参考文献只列网络公开来源；主要结论同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Deep Verification' expert workflow to deeply research the 'Sword of Goujian'.\n\nResearch scope (four dimensions):\n- Origin and archaeological context: late Spring and Autumn Wu-Yue sword-casting history, excavation/tradition disputes, current collection\n- Metallurgy and decorative craftsmanship: metal composition, rhombic dark-pattern inlay technique, concentric-circle pommel casting\n- Inscription and historical narrative: bird-script inscription '越王勾践 自作用剑' decipherment\n- Collection and conservation status: Hubei Provincial Museum collection, conservation technology, display\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 2-3 multi-source WebSearch passes by main agent\n2. ≥2 images (3-6 recommended), Tanyuan covers preferred; ≥1 chart (2-3 recommended: timeline/craft structure/metal composition comparison)\n3. Four-format reports: Markdown master → HTML → DOCX → PDF, fixed 10-section structure\n4. Conclusions only, no recommendations; references only list public web sources; main conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research", "name": "文博深度考证专家", "description": "基于探元文物库与知识库并行采集，由主代理直接联网多源交叉核验，交付四格式深度考证报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e5dcea690c95", "expert_id": "ex_RNmBcUuH9fOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T16:47:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "grade-workbench", "artifact_type": "html", "title": "年级工作台", "subtitle": "年级组与班主任的成绩·德育·家校一站式工作台", "title_en": "Grade Workbench", "subtitle_en": "One-stop grade-level desk for scores, conduct and parent contact", "description": "带一个年级 600 个学生，成绩分散在一张张 Excel 里，德育扣分、心理关注、家校沟通全靠聊天记录和脑子记，想看某个孩子六次考试的起伏还得翻半天表。这个工作台把学生名册、6 场考试的统一成绩表、德育考核、心理健康、年级巡查、成长与沟通台账、今日待办收进一个页面：勾掉待办、记一笔表扬、录入一次月考，数据即时落到资料库数据表，手机电脑都能用。适合年级主任、班主任和任课老师直接拿去管班。注：案例内学生与记录均为随机生成的匿名演示数据。", "description_en": "Running a grade of 600 students means scores scattered across Excel sheets, conduct points and parent calls buried in chat logs, and half an hour of digging just to see one kid's trend across six exams. This workbench puts the student roster, a unified 6-exam score table, moral points, counseling notes, patrol logs, growth & communication ledgers and daily to-dos into one page: tick a to-do, log a praise, import a monthly exam — everything lands in Library data tables, on desktop and mobile. Built for grade deans, homeroom and subject teachers. Note: all students and records are randomly generated, anonymized demo data.", "tags": [ "班级管理", "成绩分析", "学生档案", "德育台账", "班主任" ], "tags_en": [ "Class Management", "Score Analysis", "Student Profiles", "Conduct Ledger", "Homeroom Teacher" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/pRqi7WH1gH96Bst4MPbsmq，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个年级组 / 班主任用的年级管理工作台，暖橙色浅底卡片风、左侧功能导航（移动端转底部 Tab），共 4 个板块：\n1) 总览：学生搜索直达档案、今日待办（勾选完成、回车新增、删除，逾期标红）、班级人数柱状图、今日生日、最近成长与沟通动态时间线；\n2) 管理：学生信息（600 人 15 班，班级筛选 / 搜索 / 分页 / 增删改，可下钻学生 360 档案：基本资料、历次成绩、总分趋势、德育加减分、成长与沟通时间线）；德育考核（加分 / 扣分 + 分值）；心理健康（日常谈心 / 重点关注 / 危机干预 / 家访，状态在关注中 / 已缓解 / 已解除间流转）；年级巡查（早读 / 课间 / 晚修时段，正常 / 需提醒 / 需处理分级）；\n3) 记录：成长记录（表扬 / 批评 / 奖励 / 处分，可附自动压缩的图片）与沟通记录（对象为学生 / 家长 / 班主任 / 年级主任，方式含电话 / 面谈 / 微信 / 家访）；\n4) 成绩：统一成绩表（考试名称 + 考试时间 + 学号姓名班级 + 9 科分数，选科后未选科目留空），提供考试分析（总分 / 单科维度切换、分数段分布、班级均分对比、单科 Top10）、学生趋势（任一学生历次总分折线 + 单科叠加 + 年级 / 班级排名走势）、进退步预警（按年级排名变化阈值 ±20/30/50/100 名筛选，分退步预警与进步表扬名单）。\n支持 .xlsx / .csv 智能导入：含「学号+科目列」的 Sheet 判为成绩表（填考试名称与时间归档入综合表，可勾选列、可覆盖同名旧考试），含「学号+姓名+班级」的 Sheet 判为学生信息，按学号更新已有、新增未有。\n数据表 8 张：学生信息、成绩_考试记录、德育考核、心理健康、年级巡查、成长记录、沟通记录、待办事项；字段类型与页面表单一致（学号 / 电话 / 各科分为数字，日期为日期，类型 / 状态 / 等级 / 完成为单选，其余文本）。演示数据为随机生成的匿名数据。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library's capabilities to make me a same-style copy of https://www.workbuddy.cn/space/d/pRqi7WH1gH96Bst4MPbsmq with online data storage and two-way sync. Prefer the one-click copy flow; if unavailable, build as follows: Build a grade-level management workbench for grade deans / homeroom teachers: warm orange light card style, left navigation (bottom tabs on mobile), 4 sections:\n1) Overview: student search straight to profile, today's to-dos (tick to finish, Enter to add, delete, overdue in red), students-per-class bar chart, birthdays today, recent growth & communication timeline;\n2) Manage: student info (600 students in 15 classes, filter / search / pagination / CRUD, drill down to a 360 profile: basics, exam history, total-score trend, moral +/- points, growth & comm timelines); moral points ledger (merit / demerit with values); counseling (daily talk / key concern / crisis care / home visit, status flowing through watching / easing / resolved); patrol log (morning reading / break / evening study slots, normal / reminder / action levels);\n3) Records: growth records (praise / criticism / award / discipline, auto-compressed photos) and communication records (student / parent / homeroom teacher / grade dean, via phone / in person / WeChat / home visit);\n4) Scores: unified score table (exam name + date + ID/name/class + 9 subjects, electives left blank after subject selection), with exam analysis (total / single-subject dimension, score distribution, class averages, subject Top 10), student trends (total-score line across exams + subject overlay + grade / class rank trend), and rank-change alerts (thresholds ±20/30/50/100, separate declining and improving lists).\nSupports smart .xlsx / .csv import: sheets with \"ID + subject columns\" are treated as score sheets (archived into the unified table with an exam name and date, column checkboxes, optional overwrite of the same exam); sheets with \"ID + name + class\" are treated as student info, updating existing IDs and adding new ones.\n8 tables: student info, unified exam scores, moral points, counseling, patrol log, growth records, communication records, to-dos; field types match the forms (ID / phone / subject scores numeric, dates as dates, type / status / level / done as single-select, rest text). Demo data is randomly generated and anonymized.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "年级工作台（自包含双语演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "description_en": "Builds custom digital workspaces for students, professionals, creators, parents, and lifestyle users, with responsive HTML, local data persistence, and one-click deployment.", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd9b7744a331", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "creator": { "name": "看云舒", "name_en": "Kan Yunshu", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-09T12:59:08+08:00", "updated_at": "2026-09-09T14:05:00+08:00", "quality_score": 95, "is_recommended": false, "recommended_order": 0 }, { "categories": [ "business" ], "skills": [], "experts": [ { "id": "tanyuan-heritage-article", "name": "科普推文专家", "description": "为博物馆公众号写文物科普图文，正文+封面+配图，排版复用 doocs/md，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "076b70b41214", "expert_id": "ex_aPGVu1473LOP" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:05:37+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "id": "grand-tang-exhibition-article", "title": "大唐风华展公众号推文", "title_en": "Grand Tang Exhibition WeChat Article", "subtitle": "下周特展开幕，用真实文物数据 + 工艺逻辑给读者一份预热指南", "subtitle_en": "Pre-opening WeChat guide backed by real artifact data and craft logic", "description": "特展开幕前的推文，常被写成一两句官话 + 一堆展品名单，看完记不住一件。这个案例用了不一样的写法：以工艺为主线，串起釉上彩绘陶俑（绘彩镇墓兽）和唐三彩盘口壶这两件真实馆藏，最后落到一份「必看清单+观展贴士」——读者去现场前可以反复翻。\n\n演示内容：调用探元知识库拿到绘彩镇墓兽（巩义市博物馆 · 唐武周 690–705）、唐三彩双系龙錾盘口壶（九江 · 国家一级），结合《景德镇陶瓷》刊载的造型演变研究，写出一篇聚焦「色」和「工艺逻辑」的特展预热图文。\n\n适合：博物馆宣教编辑、特展项目负责人、唐代艺术爱好者公众号的内容主笔，以及所有在做下一场展览预热的馆宣/策展团队。", "description_en": "Most pre-opening WeChat articles are slogans plus a checklist of every exhibit — and readers remember nothing. This case rewires the format: lead with craft logic, anchor it in two real TanYuan holdings (painted-zhenmushou from Gongyi, Tang Wuzhou 690–705; Tang sancai dragon-spout pot from Jiujiang, National Grade-1), then land on a \"must-see list + viewing tips\" the reader can keep on their phone before visiting.\n\nDemo: pull artifacts from the TanYuan KB, weave in the ceramic-form evolution research from《Jingdezhen Ceramics》, and deliver a pre-opening article that frames Tang aesthetics through color and process — not just slogans.\n\nBest for: museum WeChat editors, special-exhibition project owners, Tang-art enthusiast channels, and any curator / comms team preparing the next show.", "tags": [ "大唐风华", "特展预热", "唐三彩", "釉上彩", "公众号推文" ], "tags_en": [ "Grand Tang Exhibition", "Pre-Opening", "Tang Sancai", "Over-Glaze Painting", "WeChat Article" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "doocs/md 风格公众号图文 HTML" } ], "prompt": "请基于以下要点，为下周开幕的「大唐风华特展」撰写一篇公众号预热推文。\n\n写作要点：\n1. 钩子：开头一句话不写「特展要来了」，而是写「盛唐器物的小中见大、圆润饱满不是抽象形容——它是高岭土的物理性能、铅釉的流淌方向、轮制+手工的合作方式决定的」——把读者带进「看工艺而非看热闹」的阅读姿态。\n2. 真实文物 + 工艺逻辑：所有数据必须调用 mcp__tanyuan-wb-mcp-server__search_relics / search_movable_relics 检索。常见支撑：\n - 绘彩镇墓兽：唐武周（690–705），巩义市博物馆，通高 48.3 cm，罕见的釉上彩工艺（与唐三彩不同的二次低温烧成）；2016 年曾赴韩国国立中央博物馆参加「大唐风华」特展。\n - 唐三彩双系龙錾盘口壶：唐，国家一级，九江市博物馆，高 18 cm。盘口/龙錾/束颈/鼓腹/饼形足造型；说明胡风习染。\n - 西安大唐西市博物馆：中国首家非国有「国家一级博物馆」，反映盛唐丝路与商业文化。\n3. 必看清单：第 4 节用 object-card 排版列 4–6 件代表展品；每件含器名 + 朝代 + 馆藏 + 一句推荐理由。\n4. 观展贴士：开馆首周人多、釉面侧光才看清楚釉彩流痕、关注志愿讲解等小建议（不堆 emoji）。\n5. 风格：doocs/md 风格——宽 677px、米白/暖色底、章节小标题加彩条、卡片块；克制、美观、公众号可直接复制到编辑器发布。\n6. 配图：来自探元知识库的已授权文物图（search_relics 返回 image_url），download 后 base64 内嵌；不授权的图不写。\n7. 字数：1500–1800 字正文。\n8. 逐处标源：每个事实声明「来自探元知识库《...》词条」或《景德镇陶瓷》2018/02 等。\n\n约束：输出单文件 HTML，CSS 全内联，图片 data:image/...;base64,... 内嵌，禁止任何外链。封面图单独导出 cover.png。", "prompt_en": "Write a pre-opening WeChat article for the upcoming \"Grand Tang\" special exhibition. The expert persona is the TanYuan Heritage Article Expert — a heritage science writer.\n\nWriting requirements:\n1. Hook: do not open with \"the exhibition is coming\". Open with \"Tang ceramics 'small yet grand, rounded yet subtle' is not abstract — it is dictated by kaolin physics, lead-glaze flow, and the wheel-thrown + hand-built collaboration\". Pull the reader into a craft-driven (not spectacle-driven) reading posture.\n2. Real artifacts + craft logic: every fact must come from mcp__tanyuan-wb-mcp-server__search_relics / search_movable_relics. Useful anchors from TanYuan KB:\n - Painted zhenmushou: Tang Wuzhou (690–705), Gongyi Museum, h. 48.3 cm, rare over-glaze painting process (a low-temperature second firing distinct from sancai); exhibited at the National Museum of Korea for \"Grand Tang\" in 2016.\n - Tang sancai dragon-spout pot: Tang, National Grade-1, Jiujiang Museum, h. 18 cm. Pan-mouth / dragon-spout / constricted neck / swollen belly / disc-foot form.\n - Xi'an Tang West Market Museum: China's first private National Grade-1 museum, focused on Silk-Road culture and Tang commercial life.\n3. Must-see list: section 4 uses an object-card layout for 4–6 representative exhibits; each card has artifact name + dynasty + keeper + one-line pitch.\n4. Viewing tips: peak crowds on opening weekend, side-light reveals glaze flow, follow docent tours, etc. (no emoji soup).\n5. Visual style: doocs/md — 677px column, warm cream background, colored-bullet section headings, soft cards; restrained and WeChat-ready.\n6. Imagery: authorized TanYuan KB images (search_relics image_url), downloaded and embedded as base64; never use unauthorized photos.\n7. Length: 1500–1800 Chinese characters of body copy.\n8. Cite every fact: each claim labels source.\n\nOutput: a single HTML file, all CSS inlined, all images as data:image/...;base64,..., no external links. Export a separate cover.png.", "is_recommended": false, "recommended_order": 0 }, { "id": "grant-application-drafting", "title": "公益项目申报书一键生成", "title_en": "One-Click Grant Application Drafting", "subtitle": "从资助通知到完整申报书，让 AI 接住从零起草的重活", "subtitle_en": "From RFP to a complete grant application draft — let AI take on the heavy lifting", "description": "公益机构在撰写资助项目申报书时，常常被从零组织一篇完整文档压住——项目背景要查阅、需求分析要调研、服务计划要设计、预算明细要拆算，手动拼凑一份结构完整的申报书往往要花好几天，还总怕逻辑断裂、预算对不上。\n\n让「公益项目资助申请专家」接住这个活：它读取资助通知要求，结合机构后台的项目经验与执行数据，自动生成一份含封面、机构信息、项目概述、需求分析、目标计划、评估指标与预算明细表的完整叙述式申报书，关键数据来自后台可溯源、章节结构符合资助方要求。\n\n适合：公益机构项目人员、社会组织负责人，以及需要快速产出规范项目申报书的资助申请场景。", "description_en": "Public welfare organizations writing grant applications get bogged down in composing a full document from scratch — researching context, drafting needs analysis, designing the service plan, breaking down the budget. Manually stitching together a structurally complete proposal often takes several days, and there's always the fear of broken logic or misaligned budgets.\n\nLet the Grant Application Expert handle it: reads the RFP requirements, combines them with the organization's project experience and execution data from the backend, and automatically generates a full narrative proposal — cover, organization info, project overview, needs analysis, goals & plan, evaluation metrics, and detailed budget table. Key numbers are traceable to backend records; the chapter structure matches funder expectations.\n\nFor: project staff at nonprofits, social organization leads, and anyone who needs standard-compliant grant proposals fast.", "categories": [ "productivity", "business" ], "tags": [ "公益", "资助申请", "项目申报书", "文档生成" ], "tags_en": [ "Public Welfare", "Grant Application", "Proposal", "Document Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "这是我看到的{申请链接}。请你阅读通知中的申报条件，结合本机构情况，生成项目申报书，以 docx 格式输出。", "prompt_en": "Here's the RFP I found: {application link}. Please read the eligibility and submission requirements, combine them with our organization's context, and generate a grant application proposal as a .docx.", "skills": [], "experts": [ { "id": "ex_n6C0zqy6N8Lm", "name": "公益项目资助申请专家", "description": "解析资助通告、检索机构后台数据、一键生成合规的项目申报书与申请表", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f91391d5286c", "expert_id": "ex_n6C0zqy6N8Lm" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "阳光童行项目申报书（脱敏版）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "grant-conditions-analysis", "title": "资助征集通告条件拆解", "title_en": "Grant RFP Eligibility Breakdown", "subtitle": "一份长文通告，八个维度看清是否能申报、需要交什么", "subtitle_en": "One long RFP, eight structured dimensions telling you if you qualify and what to prepare", "description": "公益机构在拿到一份资助征集通告时，常常被散落在长文各处的资格门槛、材料清单、截止时间、隐性条款困扰——人工通读容易漏看关键硬性条件，还总怕错过截止日期或材料缺项导致申报无效。\n\n让「公益项目资助申请专家」接住这个活：它直接解析征集通告原文，抽取资助方、资助方向、资助金额、申报方式、截止时间、资格要求、隐藏条款、材料清单八大维度，按结构化卡片重写一版条件拆解稿，关键硬性门槛逐条标注可核对、隐性条款单独提示。\n\n适合：公益机构项目申报人员、资助专员、社会组织负责人，以及基金会、民非团队的日常资助申报场景。", "description_en": "Grant RFPs are long, and the requirements — eligibility, submission materials, deadlines, hidden clauses — are scattered throughout. Reading them manually risks missing a hard requirement and getting disqualified after weeks of work.\n\nLet the Grant Application Expert do it: parses the RFP text and extracts eight structured dimensions — funder, focus area, funding amount, submission channel, deadline, eligibility, hidden clauses, required materials — into a checklist-style breakdown, with hard requirements individually flagged and hidden clauses called out separately.\n\nFor: nonprofit grant coordinators, project leads, and any team routinely responding to foundation or grant calls.", "categories": [ "productivity", "business" ], "tags": [ "公益", "资助申请", "条件解析", "征集通告" ], "tags_en": [ "Public Welfare", "Grant Application", "Eligibility Analysis", "RFP" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "这是我看到的{申请链接}。请你帮我拆解和理解一下申请条件。", "prompt_en": "Here's the RFP I found: {application link}. Please break down and explain the eligibility and submission requirements for me.", "skills": [], "experts": [ { "id": "ex_n6C0zqy6N8Lm", "name": "公益项目资助申请专家", "description": "解析资助通告、检索机构后台数据、一键生成合规的项目申报书与申请表", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f91391d5286c", "expert_id": "ex_n6C0zqy6N8Lm" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "希捷慈善高尔夫基金申请条件拆解（脱敏版）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "grant-form-prefill", "title": "资助申请表一键预填", "title_en": "Grant Application Form Auto-Fill", "subtitle": "读通告、抓后台、逐格填——制式申请表半小时可提交", "subtitle_en": "Parse the RFP, pull from your backend, and fill every cell — submission-ready in half an hour", "description": "公益机构在填报资助申请表时，常常被制式表格的逐项填写拖住——通知要求要逐条研读，机构信息散落在后台各处难以汇集，手动对照表格一格一格填往往要花大半天，还总怕漏填关键项、把资质数据写错。\n\n让「公益项目资助申请专家」接住这个活：它读取资助通知链接，自动解析申报条件与资格要求，定位并解析申请表表格结构；随后直接拉取机构后台的机构详情、团队成员、项目执行经验与受益数据，按原始表格版式一站式填充全部字段，生成可直接提交的 Word 申请表，已填项与待补项逐条标注、留痕可核验。\n\n适合：公益机构项目人员、基金会与社会组织资助申报负责人，以及需要批量应对多个资助申请表的申报场景。", "description_en": "Filling out a grant application form is death by a thousand cells — parsing requirements, locating scattered organizational info in the backend, and copying it into every row. It burns half a day and still risks missing a mandatory field or mis-stating qualifications.\n\nLet the Grant Application Expert take over: reads the RFP link, parses eligibility and the form's table structure, then pulls organization details, team members, project experience, and beneficiary data from your backend to fill every field in the form's native layout. It outputs a submission-ready .docx and marks filled vs. still-needed items separately, with sources traceable.\n\nFor: nonprofit project staff, grant leads at foundations and social organizations, and anyone processing multiple applications in parallel.", "categories": [ "productivity", "business" ], "tags": [ "公益", "资助申请", "申请表填写", "自动填表" ], "tags_en": [ "Public Welfare", "Grant Application", "Form Filling", "Auto-Fill" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "这是关于{某资助申请}的申请表，请你预填本机构在腾讯公益上的已有的信息。", "prompt_en": "Here's the application form for {a specific grant}. Please auto-fill it with our organization's existing information from the Tencent Public Welfare backend.", "skills": [], "experts": [ { "id": "ex_n6C0zqy6N8Lm", "name": "公益项目资助申请专家", "description": "解析资助通告、检索机构后台数据、一键生成合规的项目申报书与申请表", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "f91391d5286c", "expert_id": "ex_n6C0zqy6N8Lm" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "2026 资助申请表（脱敏版）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "gstack-saas-ai-assistant-site-html", "title": "AI助手官网HTML落地", "title_en": "AI Assistant Site in HTML", "subtitle": "无设计稿条件下完成首页、定价页、产品详情页三页生产级实现", "subtitle_en": "A production-ready homepage, pricing page, and product detail page built without final mockups", "description": "很多团队在没有高保真设计稿时，最容易卡在两头：要么页面只剩结构、没有品牌感；要么为了追求视觉效果堆太多装饰，结果做不出真正可上线的生产级 HTML。尤其是 SaaS 官网，首页、定价页、产品详情页三页往往要同时兼顾商业说服力、信息层级、响应式和无障碍，临时拼页面很容易失控。\n\n本案例演示 GStack 设计顾问如何在“先无稿执行”的前提下，围绕蓝+白主色、橙色点缀的品牌方向，先做一份设计审查与实现说明，再落地 3 个可直接静态部署的 HTML 页面：首页、定价页、产品详情页。三页都内置响应式布局、语义化标签、键盘可达焦点、移动端导航、FAQ 折叠、定价切换和详情页 tab 等基础交互，适合作为生产参考稿或售前演示稿。\n\n适合：SaaS 产品经理、设计负责人、前端工程师、市场团队、创业团队，以及需要在没有完整视觉稿前快速建立官网结构与品牌基线的团队。", "description_en": "When there is no high-fidelity mockup yet, teams usually get stuck at two extremes: either the page has only structure and no brand feel, or it becomes overly decorative and never turns into production-grade HTML. This is especially true for SaaS marketing sites, where homepage, pricing, and product detail pages all need to balance persuasion, information hierarchy, responsiveness, and accessibility at the same time.\n\nThis case shows how the GStack design specialist works in a 'no-final-mockup-yet' mode: first producing a design review and implementation guide, then delivering three static, deployment-ready HTML pages for an AI assistant SaaS website—a homepage, a pricing page, and a product detail page. All three pages include responsive layouts, semantic markup, keyboard focus states, mobile navigation, FAQ accordions, pricing toggles, and tabbed detail interactions, making them practical reference assets for production or pre-sales demos.\n\nBest for SaaS product managers, design leads, frontend engineers, marketing teams, startup teams, and anyone who needs to establish a credible website structure and brand baseline before a full visual design is ready.", "categories": [ "dev", "creation", "agent" ], "tags": [ "HTML/CSS", "SaaS官网", "无稿执行", "设计落地", "响应式" ], "tags_en": [ "HTML/CSS", "SaaS website", "no mockup", "design implementation", "responsive" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "homepage.html", "artifacts": [ { "file": "homepage.html", "label": "官网首页 HTML" }, { "file": "pricing.html", "label": "定价页 HTML" }, { "file": "product-detail.html", "label": "产品详情页 HTML" }, { "file": "design-review.md", "label": "设计评审与实现说明" } ], "prompt": "我有一份设计师给的高保真设计稿（首页 + 定价页 + 产品详情页），品牌色体系（蓝 + 白 + 橙色点缀）。请做两件事：\n1. 设计顾问先做视觉审查，对颜色、排版、间距、可访问性提改进建议；\n2. 落地为生产级 HTML/CSS 代码（遵循 Pretext 框架），含响应式、无障碍属性、基本交互。\n\n当前先按“无设计稿执行”模式处理：在没有真实高保真稿的前提下，基于上述品牌方向输出一版可直接部署的参考实现。交付 3 个独立 HTML 文件 + 一份 Markdown《设计评审与实现说明》。", "prompt_en": "I have a high-fidelity design mockup from a designer (homepage + pricing page + product detail page), with a brand color system (blue + white + orange accents). Please do two things:\n1. Have the design consultant first do a visual review, offering improvement suggestions on color, typography, spacing, and accessibility;\n2. Implement it as production-grade HTML/CSS code (following the Pretext framework), including responsiveness, accessibility attributes, and basic interactions.\n\nFor now, handle this in \"no-mockup execution\" mode: in the absence of an actual high-fidelity mockup, produce a directly deployable reference implementation based on the brand direction above. Deliver 3 standalone HTML files + a Markdown \"Design Review & Implementation Guide.\"", "skills": [], "experts": [ { "id": "SoftwareWorkshop", "name": "软件工坊", "description": "6位工程专业角色：产品评审、代码审查、安全审计、QA测试、设计系统、调试运维，覆盖从想法到生产的完整软件生命周期", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "332b1f2b85e3", "name_en": "Software Workshop", "description_en": "6 specialist roles for engineering workflows: product review, code review, security audit, QA testing, design system, and debugging — covering the full software lifecycle from idea to production", "expert_id": "ex_oFNPTEM9VRml" } ], "quality_score": 90, "source": "official", "created_at": "2026-04-27T00:18:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 100, "featured_order": 0, "usage_count": 0, "is_featured": false, "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "habit-checkin-healing-workbench", "title": "治愈系习惯打卡台", "title_en": "Healing Habit Check-in Workbench", "subtitle": "一页管理打卡、连续记录、数据复盘与个性主题", "subtitle_en": "Track habits, streaks, insights, and personal themes in one page", "description": "习惯工具最容易遇到的问题，是功能太重、记录太麻烦，或者换个设备就担心数据丢失。\n\n这个案例展示了一套纯前端习惯打卡台：支持打勾、计数、时长和时间段四种打卡方式，并提供习惯管理、连续天数、30 天热力图、周趋势、心情曲线、成就徽章、主题换肤和 JSON 备份。所有数据自动保存在浏览器本地，单个 HTML 文件可离线运行，手机和电脑均可使用。\n\n适合：想培养日常习惯的个人用户、备考与学习打卡人群、健身健康记录者、需要轻量自我管理工具的人。", "description_en": "Habit trackers often become too heavy, too tedious to update, or too risky when users worry about losing data after switching devices.\n\nThis case demonstrates a fully client-side habit check-in workbench with four tracking modes: completion, count, duration, and time range. It includes habit management, streaks, a 30-day heatmap, weekly trends, a mood curve, achievement badges, theme customization, and JSON backup. All data is stored locally in the browser, and the single HTML file works offline on both mobile and desktop.\n\nBest for: people building daily routines, students tracking study goals, fitness and wellness users, and anyone who wants a lightweight personal management tool.", "categories": [ "workspace", "productivity" ], "tags": [ "习惯打卡", "本地存储", "数据可视化", "响应式网页", "离线工具" ], "tags_en": [ "Habit Tracking", "Local Storage", "Data Visualization", "Responsive Web", "Offline Tool" ], "artifact_type": "html", "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整习惯打卡台单体 HTML" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/UrnvhCxCDjX1757gK0XmmG），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我搭建一个移动端与电脑端都能使用的个人习惯打卡台，包含今日打卡、习惯管理、数据统计和个人设置四个页面。支持打勾、计数、时长、时间段四种打卡方式，用户可新增、编辑、删除习惯，并选择图标、颜色和卡片尺寸。今日页展示进度环、连续天数、打卡卡片和心情；统计页提供 30 天热力图、周趋势、心情曲线与成就徽章；设置页支持头像昵称、主题背景与强调色、导入导出和清空示例。整体采用低饱和治愈系波点风格，卡片与内嵌方块的底色会随所选主题背景和强调色一起融合派生，文字颜色随明暗自动适配、保证可读。请生成单个完整 HTML 文件，CSS、JavaScript 和 SVG 全部内联，零外部依赖、断网可用，使用 localStorage 自动保存，并保证手机触控友好、桌面布局稳定。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/UrnvhCxCDjX1757gK0XmmG) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a personal habit check-in workbench that works well on both mobile and desktop, with four pages: Today, Habits, Insights, and Profile. Support four check-in modes: completion, count, duration, and time range. Users must be able to add, edit, and remove habits, and choose an icon, color, and card size. The Today page should show a progress ring, streaks, check-in cards, and mood tracking. The Insights page should include a 30-day heatmap, weekly trends, a mood curve, and achievement badges. The Profile page should support avatar and nickname editing, background and accent theme customization, import and export, and clearing sample data. Use a low-saturation healing polka-dot style where card and tile surfaces are derived by blending the selected theme background and accent color, with text colors that automatically adapt to light or dark surfaces for readability. Deliver one complete HTML file with all CSS, JavaScript, and SVG inline, no external dependencies, full offline support, localStorage persistence, touch-friendly mobile controls, and a stable desktop layout.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-01T19:20:42+08:00", "updated_at": "2026-08-01T22:22:42+08:00", "preview_en": "output.html", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "hangzhou-3day-nature-map-guide", "title": "杭州三日自然山水游", "title_en": "Hangzhou 3-Day Nature & Landscape Map Tour", "subtitle": "一张腾讯地图铺开的三日山水行程指南", "subtitle_en": "A 3-day Hangzhou itinerary laid out on one Tencent Map", "description": "做旅行攻略最怕景点信息散落在一堆截图和文档里，到了现场还得一个个搜地图。这份案例把杭州三天的自然山水行程直接铺在一张腾讯地图上：三天 12 个站点按日分组列在左侧面板，每站带一句精华简介；站与站之间标注下一程的交通方式、距离和耗时；地图上按天配色绘制编号标记与全程足迹路线，点击站点或标记弹出信息卡片，右下角扫码即可在手机端打开同一份行程。适合自由行旅行者、帮朋友排行程的\"人形攻略机\"，换个目的地照着做一份就是自己的。", "description_en": "Trip planning usually means attractions scattered across screenshots and notes, with map searches repeated on the spot. This case lays a 3-day Hangzhou nature itinerary directly on one Tencent Map: 12 stops grouped by day in a side panel, each with a one-line highlight; transfers between stops annotated with mode, distance and duration; day-colored numbered markers and a full footprint route drawn on the map; clicking a stop or marker opens an info card; a QR code in the corner opens the same itinerary on your phone. Perfect for independent travelers and the friend who always plans the trip.", "categories": [ "travel" ], "tags": [ "杭州旅行", "地图行程", "三日游", "路线规划", "自由行" ], "tags_en": [ "hangzhou travel", "map itinerary", "3-day tour", "route planning", "independent travel" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "杭州三日自然山水游·地图行程指南页（单文件 HTML）" } ], "prompt": "使用腾讯地图·地图助手 Skill 制作一份《杭州三日自然山水游》地图行程指南，交付一个开箱即用的单文件 HTML 页面：打开即全屏腾讯地图，左侧悬浮行程面板，含行程标题「杭州3日自然山水游」、日期区间「10月3日 - 10月5日 · 杭州」与「3 个日程 · 12 站 · 55 km」统计。三天行程按日分组：Day1（10月3日·4站：西湖断桥、白堤、湖滨晴雨、知味观(湖滨总店)）、Day2（10月4日·4站：龙井村、九溪烟树、西湖风景区·乌龟潭、武林夜市）、Day3（10月5日·4站：西溪国家湿地公园、杭州植物园等），每站附一句精华简介，站与站之间标注下一程的交通方式、距离与耗时；地图上按天配色绘制编号标记与每日足迹路线，点击面板站点或地图标记联动弹出信息卡片（名称、简介、亮点标签、到达贴士）；面板底部带每日图例，右下角放「扫码看手机端」二维码；适配窄屏（面板收到底部、二维码隐藏）。地图鉴权走合规代理模式：前端不携带明文 key，通过 _TMapSecurityConfig.serviceHost 指向本地代理服务换取密钥，SDK 地址不带 key 参数；站点坐标、段间路线（方式/距离/时长）等行程数据以结构化 JSON 内嵌页内。", "prompt_en": "Use the Tencent Maps Map Assistant Skill to build a \"Hangzhou 3-Day Nature & Landscape Tour\" map itinerary guide, delivered as a ready-to-open single-file HTML page: a full-screen Tencent Map on load, with a floating itinerary panel on the left containing the title \"杭州3日自然山水游\", the date range \"10月3日 - 10月5日 · 杭州\", and a \"3 个日程 · 12 站 · 55 km\" summary. The three days are grouped by date: Day 1 (Oct 3, 4 stops: West Lake Broken Bridge, Bai Causeway, Lakeside Sunny Rain, Zhiweiguan Lakeside branch), Day 2 (Oct 4, 4 stops: Longjing Village, Jiuxi Smoke Trees, West Lake Turtle Pool, Wulin Night Market), Day 3 (Oct 5, 4 stops: Xixi National Wetland Park, Hangzhou Botanical Garden, etc.), each stop with a one-line highlight, and transfers annotated with mode, distance and duration; day-colored numbered markers and daily footprint routes drawn on the map; clicking a panel stop or a map marker opens a linked info card (name, intro, highlight tags, arrival tips); a daily legend at the panel bottom, and a \"scan for mobile\" QR code at the lower right; narrow screens adapt (panel docks to the bottom, QR hidden). Map authentication follows the compliant proxy mode: no plaintext key in the frontend — obtain the secret via _TMapSecurityConfig.serviceHost pointing to a local proxy service, with no key parameter on the SDK URL; itinerary data (stop coordinates, per-leg mode/distance/duration) is embedded in-page as structured JSON.", "skills": [ { "id": "tencentmap-map-assistant", "name": "腾讯地图·地图助手", "description": "腾讯地图·地图助手 Skill，一句自然语言调用腾讯地图全套能力，开箱即用。提供 AI 旅游攻略、地点搜索（含评分/人均/营业时间）、关键词提示、路线规划（驾车/步行/公交/骑行）、地址解析与逆解析、行政区划、IP 定位、距离计算、天气查询，并可生成自定义地图（网页/小程序指南）。涉及找地点、规划路线、旅游行程、查天气、坐标转换等出行场景时使用。", "description_en": "Tencent Maps Map Assistant Skill: invoke the full Tencent Maps capability set with one natural-language sentence, out of the box. Provides AI travel guides, place search (rating/price/hours), keyword suggestions, route planning (driving/walking/transit/cycling), geocoding and reverse geocoding, administrative districts, IP positioning, distance calculation and weather queries, and can generate custom maps (web / mini-program guides). Use it for finding places, planning routes, travel itineraries, weather checks and coordinate conversion.", "slug": "tencentmap-map-assistant" } ], "experts": [], "mcps": [], "creator": { "name": "腾讯地图", "name_en": "Tencent Maps", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "source": "community", "created_at": "2026-09-24T16:12:00+08:00", "updated_at": "2026-09-24T16:25:00+08:00", "quality_score": 95, "is_recommended": false, "recommended_order": 0 }, { "id": "health-buddy-workbench", "title": "健康 Buddy · 7天身体管理工作台", "title_en": "Health Buddy · 7-Day Body Workbench", "subtitle": "睡眠/运动/饮水/三餐/久坐/状态 六维记录 ｜ 评分·趋势·周报自动生成", "subtitle_en": "Six dimensions — sleep, exercise, water, meals, sitting, mood | Auto-scored trends and weekly report", "description": "想改善作息和身体状态，坚持不下来的最大原因是「看不到变化」。本案例是一个 7 天身体管理工作台：每天花一分钟填睡眠时长、运动、饮水量、三餐、久坐和主观状态，工作台自动算出当天的身体评分（睡眠30 + 运动20 + 饮水15 + 三餐15 + 久坐10 + 主观状态10，按已填项折算百分制），并给出针对性小贴士。首页是今日卡片（问候语、六维数据格、评分圆环、今日小结）；记录页按周排列日期 chips，支持补填和修改历史某天；趋势页用六张手绘风 Canvas 图表看一周睡眠时长、入睡时间、运动、饮水、久坐和精神状态的走势，三餐打卡用点阵格子一目了然；周报页自动汇总平均睡眠、运动次数、三餐规律天数等 8 项统计，生成身体管理评分、逐日得分条、✅做得比较好 / 🎯优先改善的两栏总结和生活习惯建议。底部明确标注仅做日常习惯记录整理、不构成医疗建议。适合想改善作息、正在减脂/塑形、或者单纯想观察自己身体状态变化的人。", "description_en": "The hardest part of improving your routine is not seeing the change. This workbench makes it visible in 7 days: spend a minute each day logging sleep hours, exercise, water intake, meals, sitting time and a subjective mood/body rating. It auto-computes a daily body score (sleep 30 + exercise 20 + water 15 + meals 15 + sitting 10 + wellbeing 10, scaled to 100 over filled items) with targeted micro-tips. The Home tab is a today card with greeting, a six-metric grid, a score ring and a daily verdict; the Log tab lays out a week of date chips so you can backfill or edit any day; the Trends tab renders six hand-drawn Canvas charts (sleep hours, bedtime, exercise, water, sitting, mood & body) plus a dot-matrix meal tracker; the Weekly tab aggregates 8 stats (avg sleep, workout count, regular-meal days and more), a weekly body score with per-day score bars, a ✅ what went well / 🎯 priority improvements summary and habit suggestions. A footer notes it is a lifestyle logger, not medical advice. Best for: anyone improving their sleep or routine, on a fat-loss or fitness plan, or simply curious about how their body tracks week over week.", "categories": [ "workspace" ], "tags": [ "健康习惯", "身体管理", "作息记录", "趋势图表", "周报总结" ], "tags_en": [ "Health Habits", "Body Tracking", "Daily Logging", "Trend Charts", "Weekly Report" ], "artifact_type": "html", "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "健康 Buddy · 7天身体管理工作台（单体双语 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/UdZsx5PS2wagU9WKdUZcsy，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n帮我做一个「健康 Buddy · 7天身体管理工作台」单页 HTML，用一周时间把睡眠、运动、饮食等生活习惯的变化变成看得见的评分和图表。\n\n四个标签页：\n1. 今日——问候语+当天日期，六维数据格（😴睡眠时长 / 🏃运动 / 💧饮水 / 🍚三餐 / 🪑久坐 / 🙂精神），当天身体评分圆环和一句话小结（如「整体比较均衡」「久坐是当前的主要短板」），下方针对性小贴士；未记录时显示引导卡片「花一分钟填几项数据，坚持 7 天就能看到真实变化」；\n2. 记录——按周排列的日期 chips（今天/昨天/周几，已记录的有圆点标记），表单含入睡/起床时间（自动算睡眠时长）、运动时长快捷 chips（没运动/15/30/45/60 min）、饮水量、三餐点选、吃得最多的类型（食堂为主/外卖为主/自己做饭/快餐居多/均衡搭配）、久坐时长、精神状态和身体自评 1-5 按钮、备注；支持补填和修改历史某天，可清空某天的记录；\n3. 趋势——六张 Canvas 手绘风图表（睡眠时长、入睡时间、运动时长、饮水量、久坐时间、精神状态&身体自评），参考线（7h/24:00/30min/1500ml/6h）+ 图例 + 当天高亮，三餐打卡点阵格子（●吃了 ○没吃 –未记录）；\n4. 周报——一周数据总览 8 项统计（平均睡眠/运动次数/平均每次运动/平均饮水/三餐规律/平均久坐/状态均值/记录天数），本周身体管理评分+逐日得分条+维度条形图，✅做得比较好/🎯优先改善两栏总结，生活习惯建议卡片。\n\n评分口径固定并在页面标注：睡眠30 + 运动20 + 饮水15 + 三餐15 + 久坐10 + 主观状态10，按已填写项折算为百分制。数据存 localStorage，顶部显示记录区间和已记录天数；底部标注「仅做日常作息与生活习惯的记录整理，不构成医疗建议」。视觉走暖色治愈系手账风（米白底+圆角卡片+手绘感图表），手机端优先（max-width 520px），所有 CSS/JS 无外部依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/UdZsx5PS2wagU9WKdUZcsy, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a 'Health Buddy · 7-Day Body Workbench' as a single-page HTML that turns one week of sleep, exercise and diet habits into visible scores and charts. Four tabs:\n1. Today — greeting + date, a six-metric grid (😴 sleep hours / 🏃 exercise / 💧 water / 🍚 meals / 🪑 sitting / 🙂 mood), a score ring for today's body score and a one-line verdict (e.g. 'fairly balanced overall', 'sitting time is the main gap'), followed by targeted micro-tips; if nothing is logged yet, show an empty-state card ('spend a minute logging a few metrics, keep it up for 7 days and see the real change');\n2. Log — date chips laid out week by week (Today / Yesterday / weekday, with a dot marking logged days); the form captures bed/wake time (auto-computing sleep hours), exercise quick chips (Rest / 15 / 30 / 45 / 60 min), water intake, meal checkboxes, dominant food type (canteen / takeout / home cooking / fast food / balanced), sitting hours, 1-5 mood and body-rating buttons, and a note; supports backfilling and editing any past day, plus clearing one day's entry;\n3. Trends — six hand-drawn Canvas charts (sleep hours, bedtime, exercise, water, sitting time, mood & body rating) with reference lines (7h / 24:00 / 30min / 1500ml / 6h), legends and today highlighted; a dot-matrix meal tracker (● ate ○ skipped – not logged);\n4. Weekly — a weekly overview of 8 stats (avg sleep / workout count / avg per workout / avg water / regular meals / avg sitting / avg wellbeing / days logged), the weekly body score with per-day score bars and dimension bars, a ✅ what went well / 🎯 priority improvements summary, and habit suggestion cards. The scoring formula is fixed and shown on the page: sleep 30 + exercise 20 + water 15 + meals 15 + sitting 10 + wellbeing 10, scaled to 100 over filled items. Data persists in localStorage; the topbar shows the date range and days logged; a footer notes this is a lifestyle logger, not medical advice. Visual: warm healing journal style (cream background, rounded cards, hand-drawn feel), mobile-first (max-width 520px), all CSS/JS inlined with no external dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "creator": { "name": "钇yin", "name_en": "钇yin", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 92, "source": "community", "created_at": "2026-09-01T18:00:00+08:00", "updated_at": "2026-09-02T11:28:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "health-report-plain-read", "title": "体检报告一键读懂", "title_en": "Make Your Health Check Report Readable", "subtitle": "上传体检报告，异常值全部标出来、用大白话讲明白", "subtitle_en": "Upload your health check report — every abnormal value flagged and explained in plain language", "description": "每年体检完，拿到手的报告就是一堆数字加箭头，谷丙转氨酶、低密度脂蛋白、同型半胱氨酸……哪个要紧、哪个不用管，完全看不出来。想问医生，挂号得等一周。\n\n这个案例演示把一份完整的年度体检报告（PDF）交给 AI：它先把「项目—结果值—单位—参考范围」逐项结构化提取出来，对照参考范围标出偏高↑、偏低↓、临界⚠，然后按科室归类，把每个异常项用一句大白话解释「这说明什么」，最后给出该挂哪个科、哪些项需要复查、复查间隔多久，以及必须立刻线下就医的危急信号提示。\n\n适合：刚拿到体检报告看不懂的普通人、帮父母管健康的子女、企业体检后的员工健康答疑、体检中心的报告解读服务、医院随访与患者教育团队。", "description_en": "Every year after a health checkup you get a report full of numbers and arrows — ALT, LDL-C, homocysteine — with no clue which ones matter and which don't. Asking a doctor means waiting a week for an appointment.\n\nThis case shows how to hand a full annual health check report (PDF) to AI: it extracts every item into structured form (test item, value, unit, reference range), flags high↑ / low↓ / borderline⚠ against the reference range, groups findings by department, explains each abnormal item in one plain sentence, and finally tells you which department to visit, which items need a recheck and how soon, plus any red-flag values requiring immediate in-person care.\n\nGood for: anyone puzzled by a fresh checkup report, adult children managing their parents' health, corporate post-checkup health Q&A, report-interpretation services at checkup centers, and hospital follow-up / patient education teams.", "categories": [ "education", "productivity" ], "tags": [ "体检报告", "报告解读", "异常值", "健康科普", "就医建议" ], "tags_en": [ "health checkup", "report interpretation", "abnormal values", "health literacy", "care guidance" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "刚拿到年度体检报告，几项指标标了箭头：空腹血糖 7.8 mmol/L（参考 3.9–6.1）、糖化血红蛋白 7.1%（参考 4.0–6.0）、总胆固醇 6.4 mmol/L（参考 <5.2）、甘油三酯 2.9 mmol/L（参考 <1.7）、谷丙转氨酶 68 U/L（参考 9–50），其余项目正常。这些是什么意思，严重吗，我该做什么？", "prompt_en": "I just received my annual check-up report with several flagged values: fasting glucose 7.8 mmol/L (ref 3.9–6.1), HbA1c 7.1% (ref 4.0–6.0), total cholesterol 6.4 mmol/L (ref <5.2), triglycerides 2.9 mmol/L (ref <1.7), ALT 68 U/L (ref 9–50); everything else normal. What do these mean, how serious is it, and what should I do?", "skills": [ { "id": "tencent-medical-report-interpreter", "name": "腾讯健康-AI就医助手-医疗报告解读", "description": "对接 ADP LLMReportInterpretation API，解读检验单/检查单/体检报告，识别异常值并用通俗语言解析医学术语", "slug": "tencent-medical-report-interpreter" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-08-22T16:09:15Z", "updated_at": "2026-08-22T16:36:30Z", "artifacts": [ { "file": "output.html", "label": "年度体检报告完整解读" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "healthbaby-feeding-workbench", "title": "健康宝宝早教喂养台", "title_en": "Healthy Baby Feeding Desk", "subtitle": "随身育儿助手：成长曲线 · 辅食大转盘 · 早教游戏 · 英文启蒙，一页管全", "subtitle_en": "A pocket parenting helper: growth curves, food wheel, early-learning games and English fun in one page", "description": "养娃的琐碎分散在备忘录、群聊和购物车里，翻起来心累。\n\n这个案例把育儿日常收进一个工作台，共 9 个模块：① 健康宝宝——身高/体重/头围成长曲线、浇水晒太阳式每日打卡；② 好好吃饭——辅食美食大转盘，60 道一荤两素套餐轮着转，规则化筛选是亮点：过敏原永久避开、不想吃的屏蔽、爱吃的概率翻倍，还能添加自定义食谱；③ 早教游戏——51 个游戏按大运动/精细运动/认知/语言/社交五类标签筛选，每日推荐+补卡提醒；④ 英文启蒙——SSS 儿歌卡带歌词、单词本点读、亲子口语场景句收藏；⑤ 101 故事宝藏库——按适龄和主题搜索，一键朗读；⑥ 城市遛娃——10 个城市 × 5 个户外去处，城市切换+关键词搜索；⑦ 攻略收藏——教育 tips 20 条按主题速查；⑧ 用品补货——纸尿裤/奶粉等按库存天数红黄绿三档预警，含价格历史；⑨ 辅食日记——每餐记录食材与做法。\n\n适合：0-3 岁宝宝的家长做喂养与早教记录，以及想参考「转盘概率规则筛选」这类轻量决策玩法的创作者。", "description_en": "A new parent's life is scattered across memo apps, group chats and shopping carts — exhausting to keep track of.\n\nThis case gathers daily parenting into one workbench with 9 modules: ① Healthy Baby — height/weight/head-circumference growth curves plus daily check-ins; ② Meal Time — a baby-food wheel with 60 one-meat-two-veggies meals; its rule-based filtering is the highlight: permanently avoid allergens, block disliked foods, double the odds for favorites, and add custom recipes; ③ Early Learning Games — 51 games tagged by gross motor / fine motor / cognitive / language / social-emotional, with a daily pick and catch-up reminders; ④ English Fun — SSS song cards with lyrics, a tappable word book, and everyday parent-child phrases to favorite; ⑤ 101 Story Treasure — searchable by age and theme, read-aloud in one tap; ⑥ Outing Spots — 5 outdoor places for each of 10 cities, with city switching and keyword search; ⑦ Tips Collection — 20 parenting tips sorted by topic; ⑧ Supplies Restock — diapers, formula and more with red/amber/green low-stock alerts and price history; ⑨ Food Diary — log ingredients and steps for every meal.\n\nBest for: parents of 0-3 year-olds tracking feeding and early learning, and creators looking for a lightweight \"wheel with probability rules\" decision mechanic.", "categories": [ "workspace" ], "tags": [ "育儿助手", "辅食记录", "成长曲线", "早教游戏", "英文启蒙" ], "tags_en": [ "Parenting Helper", "Baby Food Log", "Growth Curves", "Early Learning Games", "English Fun" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "健康宝宝早教喂养台（单文件 HTML 工作台）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/W2rXPUVCIYZoXjLTnWJfrv，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「健康宝宝早教喂养台」单页工作台，面向 0-3 岁宝宝的家长，数据落在资料库数据表，共七张表：① 宝宝档案（单例行：昵称、生日、头像、种子已初始化、爱吃菜单、不想吃、过敏名单、排敏进度、故事收藏、口语收藏、单词进度，集合类字段以 JSON 文本列存储）；② 成长曲线（记录ID、日期、类型 height/weight/head、数值）；③ 每日打卡（记录ID、日期、类型 浇水/晒太阳/游戏、项目）；④ 辅食日记（记录ID、日期、菜名、食材、做法）；⑤ 用品补货（用品ID、名称、类别、购买日期、价格、可用天数、数量、单位、价格历史）；⑥ 自定义食谱（食谱ID、名称、食材、做法）；⑦ 技能奖状（技能ID、类别、名称、完成日期）。每行携带 rid 与 updatedAt 公共列做行级双向同步。\n\n9 个模块：① 健康宝宝：记录身高/体重/头围并绘制成长曲线，支持每日打卡；② 好好吃饭：辅食美食大转盘，内置 60 道一荤两素套餐（含食材与做法），转盘支持规则化筛选——已添加的过敏原永久避开、标记不想吃的屏蔽、标记爱吃的概率翻倍，可自定义食谱加入转盘；③ 早教游戏：51 个游戏按大运动/精细运动/认知启蒙/语言启蒙/社交情感分类，每日推荐 1 个并提醒补卡；④ 英文启蒙：20 首 SSS 儿歌卡（点开看歌词）、26 个字母卡与 8 类单词本（点击朗读）、8 个场景的亲子口语（可收藏）；⑤ 101 故事宝藏库：按适龄与主题搜索儿童故事，一键朗读；⑥ 城市遛娃：10 个城市各 5 个户外去处，支持城市切换与关键词搜索；⑦ 攻略收藏：20 条教育 tips 按主题速查；⑧ 用品补货：用品按可用天数红/黄/绿三档预警，记录价格与购买历史；⑨ 辅食日记：每餐记录食材与做法。60 道套餐、51 个游戏、故事库、儿歌、单词、口语、地点、tips 等静态内容库保持页面内置；用户产生的数据全部写入上述数据表并双向同步，开箱即有示例数据（首次进入自动初始化到数据表）。视觉：暖奶油底色+珊瑚橙主色、圆角卡片、手机竖版优先、桌面端居中留白。\n\n交付：资料库在线工作台——在线 HTML 页面 + 16 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/W2rXPUVCIYZoXjLTnWJfrv with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Healthy Baby Feeding Desk\" single-page workbench for parents of 0-3 year-olds. Data lives in Library database tables — seven in total: ① Baby Profile (singleton row: nickname, birthday, avatar, seed-initialized, favorites, dislikes, allergy list, allergy-tracking progress, story favorites, phrase favorites, word progress; collection fields stored as JSON text columns); ② Growth Curve (record ID, date, type height/weight/head, value); ③ Daily Check-in (record ID, date, type watering/sunbathing/game, item); ④ Food Diary (record ID, date, dish, ingredients, steps); ⑤ Supplies Restock (item ID, name, category, purchase date, price, available days, quantity, unit, price history); ⑥ Custom Recipes (recipe ID, name, ingredients, steps); ⑦ Skill Awards (skill ID, category, name, completion date). Every row carries rid and updatedAt common columns for row-level two-way sync. 9 modules: ① Healthy Baby: log height/weight/head circumference and draw growth curves, with daily check-ins; ② Meal Time: a baby-food wheel with 60 one-meat-two-veggies meals (ingredients and steps included); the wheel supports rule-based filtering — added allergens are permanently avoided, disliked foods are blocked, favorites get double probability, and custom recipes can join the wheel; ③ Early Learning Games: 51 games tagged by gross motor / fine motor / cognitive / language / social-emotional, one daily pick with catch-up reminders; ④ English Fun: 20 SSS song cards (lyrics on tap), 26 letter cards and a word book in 8 categories (tap to hear), plus parent-child phrases across 8 scenarios (favoritable); ⑤ 101 Story Treasure: children's stories searchable by age and theme, read-aloud in one tap; ⑥ Outing Spots: 5 outdoor places in each of 10 cities with city switching and keyword search; ⑦ Tips Collection: 20 parenting tips sorted by topic; ⑧ Supplies Restock: red/amber/green low-stock alerts by remaining days, with price and purchase history; ⑨ Food Diary: log ingredients and steps for every meal. Static content libraries — 60 meals, 51 games, story library, songs, words, phrases, outing spots, tips — stay built into the page; all user-generated data is written to the tables above with two-way sync, and sample data ships out of the box (auto-initialized into the tables on first use). Visual: warm cream background with coral-orange accent, rounded cards, mobile-portrait first, centered with whitespace on desktop.\n\nDeliverable: an online Library workbench — an online HTML page plus 16 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "小豆丁是安安妈妈", "name_en": "Xiaodouding's Mom (An An)", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "community", "usage_count": 0, "created_at": "2026-09-10T14:30:00+08:00", "updated_at": "2026-09-10", "is_recommended": false, "recommended_order": 0 }, { "id": "high-potential-repeat-buyers", "title": "高潜复购人群识别", "title_en": "High-Potential Repeat Buyer Identification", "subtitle": "购买特征与复购周期驱动的高潜人群识别", "subtitle_en": "A practical Weimob playbook for high-potential repeat buyer identification", "description": "做高潜复购人群识别时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“高潜复购人群识别”展开：综合客户购买频次、客单价、偏好品类和复购周期，找出近期最可能复购的人群并解释判断依据。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、客户分层洞察运营人员、经营分析团队。", "description_en": "The hardest part of high-potential repeat buyer identification is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates high-potential repeat buyer identification through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Customer Segmentation specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "复购预测", "购买频次", "客单价", "品类偏好", "高潜人群" ], "tags_en": [ "High-Potential Repeat Buyer Identification", "Customer Operations", "Customer Segmentation", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "分析客户的购买频次、客单价、偏好品类和复购周期，识别近期最可能复购的人群，并说明判断依据。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a high-potential repeat buyer identification covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "high-potential-repeat-customer-follow-up-list", "title": "高潜复购客户跟进清单", "title_en": "High-Potential Repeat Customer Follow-Up List", "subtitle": "供导购执行的高潜复购客户优先级清单", "subtitle_en": "A practical Weimob playbook for high-potential repeat customer follow-up list", "description": "做高潜复购客户跟进清单时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“高潜复购客户跟进清单”展开：根据购买频次、客单价和复购周期筛选近期高潜复购客户，形成优先级清单、建议跟进时间和沟通话术。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购任务与话术运营人员、经营分析团队。", "description_en": "The hardest part of high-potential repeat customer follow-up list is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates high-potential repeat customer follow-up list through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Tasks and Scripts specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "高潜复购", "客户清单", "复购周期", "跟进优先级", "沟通话术" ], "tags_en": [ "High-Potential Repeat Customer Follow-Up List", "Store and Sales Associate Operations", "Sales Associate Tasks and Scripts", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "根据客户购买频次、客单价和复购周期，筛选近期最可能复购的客户，输出供导购执行的优先级清单、建议跟进时间、跟进目标和沟通话术；只生成清单，不创建任务。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a high-potential repeat customer follow-up list covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "hk-capital-flow-main-theme", "title": "港股资金往哪流？三条主线怎么选", "title_en": "Where Is HK Capital Flowing? Three Themes Compared", "subtitle": "趋势派看景气度、估值派算性价比、信号派盯地缘纪律", "subtitle_en": "Trend, valuation, and signal experts debate Hong Kong capital flows with real data", "description": "港股南向资金Q1狂买2280亿港元，但钱都去了哪？科技涨了一波还能追吗？创新药单日暴涨6%是机会还是陷阱？高股息PE不到7倍到底便不便宜？这个案例演示了炒股搭子圆桌如何拆解港股三条主线：星仔从产业趋势判断AI商业化景气度最高（卡位=买铲子+买上游），文仔横向对比发现科技PEG 1.0-1.3比创新药和高股息都优，洲仔则提醒4月22日地缘窗口前仓位纪律比方向选择更重要。最终三条路线图：重仓科技的/等估值回调的/先保现金的。\n\n适合：持有港股或想配置港股的投资者、纠结科技/创新药/高股息怎么选的人、需要南向资金流向实时解读的股民。", "description_en": "Southbound capital poured CNY 228B into HK stocks in Q1 — but where exactly? Can you still chase tech after its rally? Is the 6% single-day pharma surge an opportunity or a trap? Are 7x PE dividend stocks actually cheap? This case shows how the Stock Roundtable dissects HK's three main themes: Xingzai sees AI commercialization as the strongest industrial trend (positioning = buy the picks and shovels), Wenzai finds tech PEG 1.0-1.3 beats pharma and dividends on value, and Zhouzai flags the April 22 geopolitical window — position discipline matters more than direction. Three decision paths: go heavy tech / wait for valuation pullback / hold cash first.\n\nIdeal for: HK stock investors choosing between tech, pharma, and dividends, and anyone needing real-time southbound capital flow analysis.", "categories": [ "agent" ], "tags": [ "港股资金流向", "南向资金", "恒生科技", "创新药", "高股息" ], "tags_en": [ "HK capital flow", "southbound capital", "Hang Seng Tech", "biotech", "high dividend" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "港股资金流向三主线分析报告" } ], "prompt": "港股最近资金在往哪流？有什么主线？\n\n关注点：\n- 南向资金Q1净流入超2000亿港元，但具体流向哪些板块？\n- 三条主线（科技估值修复、创新药、高股息）哪个持续性最好？\n- 中东资本真的来了吗？还是只是市场一厢情愿？\n- 当前该追涨还是等回调？\n\n请按炒股搭子圆桌的流程：\n1. 先查今日港股实时行情（恒生指数、恒生科技、南向资金流向）\n2. 查宏观背景，判断有无系统性风险\n3. 安排星仔（趋势派）、文仔（估值派）、洲仔（信号派）三位成员并行分析\n4. 三条主线分别给出景气度/估值/信号判断\n5. 明确点出成员间的核心分歧\n6. 给出三条差异化操作路线图\n\n要求：每个数字必须有实时数据出处，不接受「综合来看」式的废话结论。三条路线图仓位差异要足够大。", "prompt_en": "Where has capital been flowing in the Hong Kong stock market lately? What are the dominant themes?\n\nPoints of focus:\n- Southbound capital saw net inflows of over HK$200 billion in Q1, but which sectors specifically did it flow into?\n- Among the three main themes (tech valuation recovery, innovative drugs, high dividends), which has the best staying power?\n- Has Middle Eastern capital really arrived, or is it just wishful thinking on the market's part?\n- Should one chase the rally now or wait for a pullback?\n\nPlease follow the stock-buddy roundtable process:\n1. First, check today's real-time Hong Kong market data (Hang Seng Index, Hang Seng Tech, southbound capital flows)\n2. Check the macro backdrop and judge whether there is any systemic risk\n3. Arrange for the three members — Xingzai (trend school), Wenzai (valuation school), and Zhouzai (signals school) — to analyze in parallel\n4. Give a prosperity / valuation / signal judgment for each of the three main themes\n5. Clearly point out the core disagreements among the members\n6. Provide three differentiated operating roadmaps\n\nRequirements: every number must cite a real-time data source; \"all things considered\"-style filler conclusions are not accepted. The position differences across the three roadmaps must be large enough.", "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "全球多市场金融数据搜索服务，自然语言查询股票、基金、宏观、外汇、大宗商品等全品类金融数据", "description_en": "Natural language financial data search across global markets: stocks, funds, macro, forex & commodities", "skill_id": "skill_2053082432761950208" } ], "experts": [ { "id": "StockPartnerTeam", "name": "腾讯自选股金融分析专家团", "description": "蒸馏了六位真实的炒股大神的实战经验，涵盖产业策略、信号捕捉、估值评估、抄底布局、基本面研究、短线交易六大视角，基于真实数据各抒己见，提供多角度分析参考。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "3acdcc5c61b9", "name_en": "Tencent WeStock Financial Analysis Team", "description_en": "Distilled real-world experience of six veteran stock investors, covering industry strategy, signal hunting, valuation analysis, contrarian entry, fundamentals research, and short-term trading — multi-angle analysis references grounded in real-time data.", "expert_id": "ex_cRUYbcfeBpop" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "usage_count": 0, "created_at": "2026-04-16T19:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "hk-opening-newsroom-brief", "title": "港股开盘编辑部晨报", "title_en": "Hong Kong Opening Newsroom Brief", "subtitle": "开盘前把多源新闻压成一页可扫读晨报", "subtitle_en": "Compress multi-source news into a one-page brief before the Hong Kong open", "description": "开盘前最头疼的就是消息太散。港股、国际宏观、内地政策、油金汇率往往各说各话，真正该盯的变量反而最容易漏掉。\n\n这个案例演示如何把 2026-07-27 00:00–09:30 HKT 的多源财经新闻统一抓回、做时间过滤和去重，再整理成一份可直接发给团队的双语 HTML 晨报。报告保留查询统计、主题归类、时间与原文链接，并补上“所以呢”式市场含义。\n\n适合：港股研究员、销售交易晨会、财富顾问内容准备、需要快速同步消息面的产品经理、做开盘 briefing 的运营同学。", "description_en": "The hardest part before the open is not a lack of information, but too much scattered information. Hong Kong equities, global macro, mainland policy and commodities often speak in different directions, so the variables that really matter are easy to miss.\n\nThis case shows how to pull multi-source finance headlines for 2026-07-27 between 00:00 and 09:30 HKT, apply time filtering and deduplication, and turn them into a bilingual HTML brief that can be sent straight to a team chat. The report keeps query statistics, theme grouping, timestamps, original links and a practical 'so what' layer for market interpretation.\n\nBest for: Hong Kong equity researchers, sales-trading morning calls, wealth-advisory content prep, product managers who need a quick market update, and operators running the pre-open briefing.", "categories": [ "finance", "search", "data" ], "tags": [ "港股开盘", "晨报", "宏观", "商品", "双语HTML" ], "tags_en": [ "Hong Kong Open", "Morning Brief", "Macro", "Commodities", "Bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请抓取 2026-07-27 00:00–09:30 HKT 的港股、国际宏观、内地政策、香港地产以及商品/外汇相关财经新闻，做时间过滤和去重后，输出一份适合开盘前扫读的单文件 HTML 晨报。报告需要包含：今日要点、抓取与去重统计、按主题归类的重点新闻、每条新闻的时间/简述/市场含义、原文链接，以及一页开盘关注清单。内容必须基于真实查询结果并注明时间与来源；HTML 资源全内联。", "prompt_en": "Please pull Hong Kong market, global macro, mainland policy, Hong Kong property, and commodities/FX finance news for 2026-07-27 between 00:00 and 09:30 HKT. After time filtering and deduplication, generate a single-file HTML morning brief that is easy to scan before the market opens. The report should include key takeaways, collection and deduplication stats, theme-grouped highlights, the time/summary/market implication of each item, original article links, and one page of pre-open watch points. The content must be based on real query results with timestamps and sources noted; all HTML assets must be inlined.", "skills": [ { "id": "daily-financial-news", "name": "每日财经新闻", "description": "财经新闻多源接入与深度解读，经去重分类后逐条深度概括，附市场影响解读。", "skill_id": "skill_2076628269121765376" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "港股开盘双语晨报" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T19:43:22+08:00", "updated_at": "2026-07-27T20:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "holiday-member-care-calendar", "title": "节日会员关怀日历", "title_en": "Holiday Member Care Calendar", "subtitle": "未来30天分等级会员节日关怀与文案日历", "subtitle_en": "A practical Weimob playbook for holiday member care calendar", "description": "做节日会员关怀日历时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“节日会员关怀日历”展开：围绕未来30天的重要节日，为不同会员等级设计关怀主题、专属优惠、发送时间和分层文案，形成可执行的关怀日历。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、积分与会员权益运营人员、经营分析团队。", "description_en": "The hardest part of holiday member care calendar is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates holiday member care calendar through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Points and Membership Benefits specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "节日运营", "会员关怀", "会员分层", "专属优惠", "内容日历" ], "tags_en": [ "Holiday Member Care Calendar", "Customer Operations", "Points and Membership Benefits", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "围绕未来30天的重要节日，为不同会员等级设计关怀主题、建议优惠、发送时间和可直接使用的分层文案，输出一份节日会员关怀日历；只生成方案，不执行群发。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a holiday member care calendar covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" }, { "id": "content-creation", "name": "内容创作与分发", "description": "创作营销文案、公众号图文、小红书和朋友圈内容；可按渠道生成素材并发布或保存入库。", "skill_id": "skill_2087078523417223168" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "hot-etf-rotation-radar", "title": "热门ETF轮动雷达", "title_en": "Hot ETF Rotation Radar", "subtitle": "用真实市场数据比较银行、酒与中概互联网三种风格位置", "subtitle_en": "Compare banks, liquor, and China internet styles with real market data", "description": "市场热点切得很快时，最难的往往不是找到 ETF，而是不知道现在的强弱轮动究竟更偏防守、反弹还是高弹性题材。单看涨幅榜容易追热，单看跌幅又看不出资金到底在切向哪里。\n\n这个案例把市场总览、热搜榜和三只代表性ETF放到同一页里：银行ETF华宝代表偏防守，酒ETF鹏华代表高波动反弹博弈，中概互联网ETF易方达代表跨境弹性主题，再用编辑性五维雷达、矩阵和观察清单，把当前轮动信号讲清楚。\n\n适合：想快速比较不同ETF风格位置的投资者；需要一页式热点扫描页的基金研究用户；想看热门基金分析如何把真实数据整理成轮动面板的人。", "description_en": "When market themes rotate quickly, the hard part is usually not finding an ETF but understanding whether the current tape favors defense, rebound trades, or high-beta themes. A gainers list alone can push you into chasing heat, while a losers list rarely tells you where the money is rotating next.\n\nThis case puts the market overview, the hot-search board, and three representative ETFs on one page: Bank ETF Huabao as the defensive sleeve, Liquor ETF Penghua as the high-volatility rebound trade, and China Internet ETF E Fund as the offshore-beta sleeve. It then uses an editorial five-factor radar, a matrix, and a watchlist to explain the current rotation signal.\n\nBest for: investors comparing style positions across ETFs at a glance; fund-research users who need a one-page market scanner; and anyone who wants to see how real ETF data can be organized into a rotation dashboard.", "tags": [ "基金分析", "ETF", "轮动", "市场扫描", "热门基金" ], "tags_en": [ "fund analysis", "ETF", "rotation", "market scanner", "hot funds" ], "categories": [ "finance" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "热门ETF轮动雷达" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "根据真实数据扫描热门ETF并生成一页式轮动对比页，围绕市场环境、三只代表性ETF定位、轮动信号、编辑性五维雷达和今日观察清单展开，只做比较与跟踪，不给买卖建议。", "prompt_en": "Scan hot ETFs with real data and generate a one-page rotation comparison page focused on market regime, the roles of three representative ETFs, rotation signals, an editorial five-factor radar, and a daily watchlist, staying with comparison and monitoring instead of trading advice.", "skills": [ { "id": "fund-analysis", "name": "基金分析", "description": "公募基金/ETF全面分析，支持持仓组合诊断和单只基金深度分析。", "skill_id": "skill_2076627911660621824" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T16:00:00+08:00", "updated_at": "2026-07-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "hotel-revenue-modeling-diagnostic", "title": "酒店经营七维模型诊断", "title_en": "Hotel Revenue Seven-Dimension Diagnostic", "subtitle": "从公开经营数据到渠道定价行动方案，一站式酒店营收诊断", "subtitle_en": "From public operational data to channel & pricing action plans — one-stop hotel revenue diagnostics", "description": "酒店经营数据摆在那里，但入住率下滑到底是定价问题还是需求问题？OTA佣金越扣越多，直订占比却始终提不上来？本案例从GitHub公开酒店经营数据集出发，搭建容量资产、需求季节、渠道分析、客群结构、成本财务、竞争对标、指标仪表盘七维经营模型，基于模型做归因诊断、情景推演，最终输出渠道优化与定价调整的具体行动方案。适合酒店/民宿经营者、收益管理分析师、以及希望用数据驱动运营决策的管理者。", "description_en": "Occupancy is dropping — is it pricing or demand? OTA commissions keep eating margins while direct bookings stagnate. This case starts with a public GitHub hospitality dataset, builds a seven-dimension operational model covering capacity, demand seasonality, channel analysis, guest segmentation, cost & finance, competitive benchmarking, and KPI dashboards. Based on the model, it performs root-cause diagnosis, scenario simulations, and delivers concrete channel optimization and pricing adjustment action plans. Ideal for hotel/B&B operators, revenue management analysts, and managers seeking data-driven operational decisions.", "categories": [ "business", "data" ], "tags": [ "酒店运营", "收益管理", "入住率诊断", "渠道分析", "定价策略" ], "tags_en": [ "Hotel Operations", "Revenue Management", "Occupancy Diagnosis", "Channel Analysis", "Pricing Strategy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "hotel-diagnostic-report.html", "artifacts": [ { "file": "hotel-diagnostic-report.html", "label": "七维模型诊断报告(HTML)" } ], "prompt": "请从 GitHub 公开酒店经营样例数据集下载 CSV，搭建七维运营模型并做营收/入住率诊断，给出渠道与定价相关建议，货币使用人民币。\n\n数据 URL：\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/dim_hotels.csv\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/fact_aggregated_bookings.csv\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/fact_bookings.csv\n\n请自行拉取数据后按照酒店经营建模四阶段流程执行分析：\n1. 数据提取与口径校验\n2. 构建全量经营模型（容量资产、需求季节、渠道分析、客群结构、成本财务、竞争对标、指标仪表盘共七维度）\n3. 案例搜集与知识补充\n4. 诊断推演与方案输出（根因分离、情景推演、渠道与定价行动方案）\n\n输出一份自包含的HTML诊断报告。", "prompt_en": "Download hotel operational data CSVs from a public GitHub repository, build a seven-dimension operational model, perform revenue/occupancy diagnosis, and provide channel and pricing recommendations. Currency: RMB.\n\nData URLs:\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/dim_hotels.csv\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/fact_aggregated_bookings.csv\n- https://raw.githubusercontent.com/ishebee/Hospitality---Analysis/main/fact_bookings.csv\n\nFollow the four-phase hotel operational modeling workflow:\n1. Data extraction and validation\n2. Build the full seven-dimension operational model\n3. Research and knowledge supplementation\n4. Diagnosis, scenario simulation, and action plan output\n\nOutput a self-contained HTML diagnostic report.", "skills": [ { "id": "hotel-revenue-modeling-diagnostic", "name": "住宿业经营数据建模与诊断", "name_en": "Hotel Revenue Modeling & Diagnostic", "description": "住宿业经营数据建模与诊断，覆盖入住率/RevPAR 分析、渠道结构诊断与定价策略优化", "skill_id": "skill_2085354999990132736" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-04T11:00:00+08:00", "updated_at": "2026-08-04T11:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "housing-lease-dispute-defense-preparation", "title": "房屋租赁纠纷答辩与质证准备", "title_en": "Housing Lease Dispute Defense & Evidence Preparation", "subtitle": "把原告诉请逐项对照合同与材料，生成要素式答辩质证工作台", "subtitle_en": "Turn each claim into an item-by-item defense and evidence-preparation workbench", "description": "收到房屋租赁纠纷的起诉状后，最怕的不是要不要赔，而是不知道从哪下手——拖欠租金、违约金、不退押金每一项都要逐条回应，漏一条都可能被动。\n\n这个案例演示如何把原告的三项诉请拆成「逐项回应矩阵」，配齐合同条款、付款、交接、维修、解除、押金六类证据清单与质证关注点，再核对答辩期限、举证期限、送达与管辖，形成一份可供承办律师复核的答辩与质证工作台。\n\n适合：遭遇房屋租赁纠纷的承租人（被告）、承办租赁合同纠纷案件的律师与法律助理、需要结构化梳理答辩与证据的法务人员。", "description_en": "The hardest part of a housing-lease dispute isn't whether to pay, but where to start — every rent, penalty and deposit claim must be answered one by one, and missing one puts you on the back foot.\n\nThis case turns the plaintiff's three claims into an item-by-item response matrix, pairs them with six evidence lists (contract terms, payments, handover, repairs, termination, deposit) and cross-examination focus points, then verifies response deadlines, evidence deadlines, service and jurisdiction into a defense and evidence-preparation workbench ready for counsel review.\n\nBest for: tenants who are defendants in a lease dispute, lawyers and paralegals handling lease cases, and legal staff who need to structure defenses and evidence.", "categories": [ "business", "search" ], "tags": [ "房屋租赁", "诉讼答辩", "质证准备", "租赁纠纷", "法律咨询" ], "tags_en": [ "Housing Lease", "Litigation Defense", "Evidence Preparation", "Lease Dispute", "Legal Consultation" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "housing-lease-dispute-defense-preparation.html", "preview_en": "", "external_url_en": "", "artifacts": [ { "file": "housing-lease-dispute-defense-preparation.html", "label": "互动房屋租赁纠纷答辩与质证准备卡片" } ], "prompt": "我是房屋租赁纠纷被告，原告主张拖欠租金、违约金并不退押金。我现有租赁合同、付款记录、房屋交接照片、维修沟通记录、解除通知和对方提交的起诉材料。请按「原告每项请求—我的回应类型（承认/否认/部分承认/不知情/待核验）—事实依据—现有证据—质证关注点—待补材料」制作答辩与质证准备表，并核对送达、举证期限、答辩材料要素及受理法院公开要求。涉及现行规则先核验，未核验处标注「待官方来源核验」；不要替我选择事实立场或承诺诉讼结果。", "prompt_en": "I am the defendant in a housing lease dispute. The plaintiff claims unpaid rent, liquidated damages and refuses to refund the deposit. I currently hold the lease contract, payment records, handover photos, repair correspondence, a termination notice and the plaintiff's filing materials. Please produce a defense and cross-examination preparation table in the format of [each plaintiff claim — my response type (admit / deny / partially admit / no knowledge / to be verified) — factual basis — existing evidence — cross-examination focus — materials to supplement], and verify service, evidence deadlines, defense-document requirements and the accepting court's public requirements. Verify current rules first and mark anything unverified as 'pending official-source verification'. Do not choose a factual position for me or promise a litigation outcome.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家，擅长起草、签署、审查、对比及法律咨询。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d1e4a1d28e92", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T18:36:46+08:00", "updated_at": "2026-08-13T19:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "hs300-etf-deep-dive", "title": "沪深300ETF深度评估报告", "title_en": "HS300 ETF Deep Dive Report", "subtitle": "用真实基金与市场数据判断 510300 适不适合做当前核心底仓", "subtitle_en": "Use real fund and market data to judge whether 510300 fits as a core holding now", "description": "很多人看宽基ETF时，最难的不是查不到净值，而是不知道现在这只基金到底更适合做底仓、补仓还是先观望。单看涨跌幅容易失真，脱离市场环境又很难判断风险收益比。\n\n这个案例把华泰柏瑞沪深300ETF（510300）放回真实市场背景里重做一遍单基评估：结合基金净值、指数位置、市场风格与量能状态，给出一句话结论、关键指标、回本难度测算、三情景推演、适合人群与主要风险，所有核心数据都在相应模块附近标注来源和日期。\n\n适合：把宽基ETF当核心底仓的投资者；正在比较“要不要上车 510300”的基金用户；想看基金分析报告应该怎样把真实数据与场景测算放在一起的人。", "description_en": "The hard part of reading a broad-based ETF is usually not finding the NAV, but judging whether the fund is better used as a core position, an averaging-down tool, or something to wait on. Looking only at short-term gains or losses is misleading, while looking at the fund without market context makes the risk-reward trade-off hard to judge.\n\nThis case puts Huatai-PineBridge CSI 300 ETF (510300) back into its real market setting and rebuilds the analysis from there: combining fund NAV data, index levels, market style, and liquidity conditions into a one-line verdict, key metrics, break-even math, three scenario sketches, investor fit, and major risks, with source and date shown near the relevant modules.\n\nBest for: investors who use broad-based ETFs as a core allocation; users deciding whether to enter 510300 now; and anyone who wants to see how a fund-analysis page can combine real data with clearly marked scenario math.", "tags": [ "基金分析", "ETF", "沪深300", "宽基配置", "单基深度评估" ], "tags_en": [ "fund analysis", "ETF", "CSI 300", "broad-based allocation", "single-fund deep dive" ], "categories": [ "finance" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "沪深300ETF深度评估报告" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "基于给定的真实数据，对华泰柏瑞沪深300ETF（510300）做一份单文件 HTML 深度评估报告，重点说明一句话结论、关键指标、回本难度测算、围绕沪深300当前位置的三情景推演、适合人群与主要风险，并在相关模块标注数据来源和日期。", "prompt_en": "Create a single-file HTML deep-dive report for Huatai-PineBridge CSI 300 ETF (510300) using the provided real data. Focus on a one-line conclusion, key metrics, break-even difficulty math, three scenario sketches around the current CSI 300 level, suitable investor profiles, and major risks, with source and date shown near the relevant sections.", "skills": [ { "id": "fund-analysis", "name": "基金分析", "description": "公募基金/ETF全面分析，支持持仓组合诊断和单只基金深度分析。", "skill_id": "skill_2076627911660621824" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T15:40:55+08:00", "updated_at": "2026-07-28T15:40:55+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "hs300-vs-csi500-allocation-lab", "title": "沪深300 vs 中证500", "title_en": "CSI 300 vs CSI 500", "subtitle": "用一页配置简报，看清大盘核心仓与中盘风格仓的真实差异。", "subtitle_en": "A one-page allocation brief on the real trade-off between a core large-cap sleeve and a mid-cap rotation sleeve.", "description": "很多人把宽基 ETF 当成“差不多的底仓”，真正做配置时才发现，大盘核心仓和中盘风格仓承担的任务根本不是一回事。最容易出错的地方，就是只看名字、不看规模、波动和成分结构。\n\n这个案例真实对比了沪深300ETF华泰柏瑞和中证500ETF南方，用公开行情、持仓与费率数据，拆出谁更像账户底盘、谁更适合风格切换，还把 60 日走势、前十大重仓、收益与回撤放进同一页双语 HTML。\n\n适合：做 ETF 底仓配置的个人投资者、想统一投顾口径的财富顾问、需要展示基金差异的金融内容团队、想快速演示 Skill 能力的产品运营。", "description_en": "A lot of people treat broad-based ETFs as interchangeable core holdings, then realize in real allocation work that a large-cap core sleeve and a mid-cap style sleeve do very different jobs. The biggest mistake is to read the name and ignore scale, volatility and composition.\n\nThis case compares Huatai CSI 300 ETF and Southern CSI 500 ETF with real quote, holdings and fee data. It shows which one looks more like portfolio core capital, which one works better for style rotation, and places 60-day path, top holdings, return and drawdown on one bilingual HTML page.\n\nBest for: individual investors building ETF sleeves, wealth advisors who need a unified comparison story, finance content teams, and product operators demonstrating the skill in a real scenario.", "categories": [ "finance", "data" ], "tags": [ "ETF对比", "宽基配置", "沪深300", "中证500", "双语HTML" ], "tags_en": [ "ETF comparison", "broad index allocation", "CSI 300", "CSI 500", "bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于 ETF深度对比 skill 的官方能力，对沪深300ETF华泰柏瑞（sh510300）和中证500ETF南方（sh510500）做一次真实数据对比。使用 westock-data 官方 CLI 获取 ETF detail、holdings、quote、kline 等公开数据，重点拆解规模、费率、过去 1 个月 / 3 个月 / 年初至今收益、波动、回撤、前十大重仓和适配场景。最终输出一个单文件 HTML 报告，资源全部内联，可离线打开。结论请使用“适用条件 + 风险提示”的表达，不要给确定性买卖建议。", "prompt_en": "Please use the official capability of the ETF depth comparison skill to compare Huatai CSI 300 ETF (sh510300) and Southern CSI 500 ETF (sh510500) with real market data. Use the official westock-data CLI to retrieve ETF detail, holdings, quote and kline data, then analyze scale, fees, 1-month / 3-month / YTD return, volatility, drawdown, top holdings and scenario fit. Deliver the result as a single-file HTML report. All assets must be inlined for offline opening. Frame the conclusion as fit conditions plus risk notes instead of a deterministic trading recommendation.", "skills": [ { "id": "etf-depth-comparison", "name": "ETF深度对比", "description": "两只A股ETF多维度深度对比", "skill_id": "skill_2076627838067363840" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "ETF 深度对比 HTML 报告" }, { "file": "data.json", "label": "ETF 对比原始数据摘要" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:20:38+08:00", "updated_at": "2026-07-30T16:22:28+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "humanizer-product-launch-copy", "title": "智能台灯产品文案「去 AI 味」", "title_en": "De-AI Rewrite for Smart Lamp Product Copy", "subtitle": "把「重新定义光」的发布会腔，改成用户听得懂的卖点", "subtitle_en": "From \"redefining light\" keynote speak to benefits users actually understand", "description": "让 AI 写产品文案，十有八九会写成发布会：重新定义、极致工艺、工匠精神、不仅是一盏灯更是一位懂你的管家，最后还要请「业内人士」出来收尾。这个案例把一段典型的 AI 生成智能台灯文案交给「去 AI 味」Skill 改写。\n\n改写后的文案把形容词换成可感知的细节、把宏大叙事落到具体使用场景、砍掉所有没有信息量的修饰。案例完整展示原文与改写稿对照、逐条标注每处改写对应的 AI 写作模式（广告腔、伪权威引用、拔高式结尾等），营销团队可以直接拿这份对照做文案评审标准。\n\n适合：电商运营、产品营销、品牌文案、市场同学、独立开发者和小团队创始人。", "description_en": "Ask AI to write product copy and you will usually get a keynote in miniature: \"redefining light,\" \"extreme craftsmanship,\" \"not just a lamp but a companion that understands you\" — plus a closing quote from unspecified \"industry insiders.\" This case hands a typical AI-generated smart lamp copy to the Humanizer Skill.\n\nThe rewrite swaps adjectives for perceivable details, grounds grand narratives in concrete usage scenarios, and cuts every modifier that carries no information. The case shows the full before-and-after copy, with each edit labeled by the AI writing pattern it fixes (ad-speak, fake authority, inflated endings) — marketing teams can use it directly as a copy review rubric.\n\nBest for: e-commerce operators, product marketers, brand copywriters, marketing teams, indie developers and small-team founders.", "categories": [ "business", "creation" ], "tags": [ "去AI味", "产品文案", "电商运营", "营销写作", "智能硬件" ], "tags_en": [ "De-AI", "Product Copy", "E-commerce", "Marketing Writing", "Smart Hardware" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "智能台灯产品文案去 AI 味改写对照报告" } ], "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "official", "usage_count": 0, "created_at": "2026-08-27T17:00:00+08:00", "updated_at": "2026-08-27T17:00:00+08:00", "prompt": "请帮我把一段 AI 生成的智能台灯产品文案做「去 AI 味」改写，输出前后完整对照和逐条修改说明。\n\n要求：\n1. 保留产品真实卖点，不新增不存在的参数和认证\n2. 清除「重新定义」「不仅…更是…」「极致」「工匠精神」「业内人士认为」等发布会腔和套话\n3. 用用户听得懂的语言写利益点，形容词换成可感知的细节和具体使用场景\n4. 每处主要改动标注对应的 AI 写作模式（如广告腔、伪权威引用、拔高式结尾）\n5. 产出一份单文件 HTML 报告，中英双语可切换，明亮配色，方便营销团队对照评审\n\n初稿（产品：星辰智能护眼台灯）：\n\n全新发布的星辰智能护眼台灯，重新定义了光的可能性。\n\n它不仅是一盏台灯，更是一位懂你的光线管家。搭载自研的曙光算法，星辰台灯实现了科技与人文的完美融合，为你打造前所未有的照明体验。\n\n极致的工艺，彰显非凡品质：航空级铝合金灯臂，历经 108 道工序打磨，每一处细节都凝聚着工匠精神；全光谱 LED 灯珠，还原自然光的温暖与纯粹，守护你的每一次深夜苦读。\n\n无论是伏案工作，还是亲子共读，星辰台灯都能恰到好处地满足你的需求，让爱在光影中流淌。\n\n业内人士普遍认为，智能照明行业正在迎来深刻变革，而星辰台灯的问世，无疑为行业树立了新的标杆，开启了智能照明的新篇章。", "prompt_en": "Please \"de-AI\" rewrite an AI-generated smart lamp product copy and deliver a full before-and-after comparison with a change-by-change explanation.\n\nRequirements:\n1. Keep the product's real selling points; do not add parameters or certifications that don't exist\n2. Remove keynote speak and clichés like \"redefining,\" \"not just X, but also Y,\" \"extreme,\" \"craftsmanship spirit,\" and \"industry insiders believe\"\n3. Write benefits in language users understand; swap adjectives for perceivable details and concrete usage scenarios\n4. Label each major edit with the AI writing pattern it fixes (e.g. ad-speak, fake authority, inflated ending)\n5. Produce a single-file HTML report, bilingual (Chinese/English toggle), bright color scheme, easy for marketing teams to review side by side\n\nDraft (product: Starlight smart eye-care desk lamp):\n\nThe all-new Starlight smart eye-care desk lamp redefines the possibilities of light.\n\nIt is not just a lamp, but a light butler that understands you. Powered by the self-developed Dawn algorithm, Starlight achieves a perfect fusion of technology and humanity, bringing you an unprecedented lighting experience.\n\nExtreme craftsmanship shows extraordinary quality: the aerospace-grade aluminum arm goes through 108 polishing steps, with craftsmanship spirit in every detail; full-spectrum LED beads restore the warmth and purity of natural light, guarding every one of your late-night study sessions.\n\nWhether working at your desk or reading with your kids, Starlight meets your needs just right, letting love flow in light and shadow.\n\nIndustry insiders generally agree that the smart lighting industry is undergoing profound change, and the launch of Starlight undoubtedly sets a new benchmark for the industry, opening a new chapter in smart lighting.", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "humanizer-wechat-article", "title": "公众号推文「去 AI 味」改写", "title_en": "De-AI Rewrite for a WeChat Article", "subtitle": "满是排比和套话的 AI 初稿，改出作者自己的声音", "subtitle_en": "Turn a cliché-ridden AI draft into an article with a real voice", "description": "用 AI 写公众号推文，最怕读者三秒就闻出「AI 味」：满屏「不仅是…更是…」、三连排比、假引用、结尾还要升华一下。这个案例把一篇典型的 AI 初稿（数字极简主义主题）交给「去 AI 味」Skill 深度改写。\n\n改写不是简单换词：套话全部清除、加入第一人称的真实经历和不确定感、句子长短拉开节奏、模糊抒情换成具体细节。案例完整展示原文与改写稿的逐段对照、识别出的每一处 AI 写作模式（伪引用、三连排比、拔高立意、空洞展望等）及对应的修改思路，可以直接当作自媒体写作的改写模板用。\n\n适合：公众号运营者、自媒体作者、内容创作者、品牌文案、任何想让 AI 初稿读起来像人写的写作场景。", "description_en": "The fastest way to lose readers is that unmistakable AI smell: walls of \"not just X, but also Y,\" rule-of-three lists, fake quotes, and a grandiose ending. This case hands a typical AI draft (on digital minimalism) to the Humanizer Skill for a deep rewrite.\n\nThe rewrite goes far beyond synonym swaps: every cliché is cut, a first-person voice with real experience and honest uncertainty moves in, sentence rhythm gets varied, and vague lyricism turns into concrete detail. The case shows a paragraph-by-paragraph before-and-after comparison, every AI writing pattern it detected (fake quotes, rule of three, inflated significance, hollow outlook), and the reasoning behind each fix — usable directly as a rewrite template for self-media writers.\n\nBest for: WeChat account operators, self-media writers, content creators, brand copywriters, and anyone who wants AI drafts to read like a human wrote them.", "categories": [ "creation" ], "tags": [ "去AI味", "公众号写作", "文案改写", "AI写作痕迹", "新媒体" ], "tags_en": [ "De-AI", "WeChat Writing", "Copy Editing", "AI Traces", "New Media" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "公众号推文去 AI 味改写对照报告" } ], "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "source": "official", "usage_count": 0, "created_at": "2026-08-27T17:00:00+08:00", "updated_at": "2026-08-27T17:00:00+08:00", "prompt": "请帮我把一篇 AI 生成的公众号推文初稿做「去 AI 味」改写，输出改写前后完整对照和逐条修改说明。\n\n要求：\n1. 保留核心观点，不虚构新的数据和事实\n2. 清除「不仅…更是…」句式、三连排比、假引用和「值得一提的是」「从某种意义上说」这类套话\n3. 加入第一人称视角、真实感受和不确定的表达，允许有保留和自我怀疑\n4. 句子长短错落，把模糊的抒情换成具体细节\n5. 每处主要改动标注对应的 AI 写作模式（如伪引用、三连排比、拔高立意、空洞结尾）\n6. 产出一份单文件 HTML 报告，中英双语可切换，明亮配色，方便逐段对照阅读\n\n初稿（主题：数字极简主义）：\n\n在这个信息爆炸的时代，我们每天都被海量的信息所包围。数字极简主义，不仅是一种生活方式，更是一场心灵的革命。\n\n它让我们重新审视自己与技术的关系，找回内心的平静与专注。正如一位专家所说：「你所关注的东西，决定了你的人生。」这句话深刻地揭示了数字极简主义的核心理念。\n\n实践数字极简主义，可以从三个方面入手：清理你的手机屏幕，规整你的信息来源，守护你的深度时间。这三个步骤相辅相成，缺一不可。\n\n清理手机屏幕，不仅能够让界面焕然一新，更能够减少无谓的干扰；规整信息来源，如同为思想修筑一道堤坝，让有价值的内容自然流淌；而守护深度时间，则是这个时代的稀缺品，值得我们用一生去追寻。\n\n值得一提的是，越来越多的研究表明，数字极简主义正在成为一种全球性的趋势。它不仅仅改变了个人，更在重塑整个社会的信息生态。\n\n从某种意义上说，放下手机，就是拿起人生。愿我们都能在这个喧嚣的时代里，找到属于自己的那份宁静。", "prompt_en": "Please \"de-AI\" rewrite an AI-generated WeChat article draft and deliver a full before-and-after comparison with a change-by-change explanation.\n\nRequirements:\n1. Preserve the core viewpoint; do not invent new facts or data\n2. Remove \"not just X, but also Y\" constructions, rule-of-three lists, fake quotes, and filler like \"it is worth mentioning\" or \"in some sense\"\n3. Add a first-person voice with real feelings and honest uncertainty; hesitation and self-doubt are allowed\n4. Vary sentence length and rhythm; replace vague lyricism with concrete detail\n5. Label each major edit with the AI writing pattern it fixes (e.g. fake quote, rule of three, inflated significance, hollow ending)\n6. Produce a single-file HTML report, bilingual (Chinese/English toggle), bright color scheme, easy paragraph-by-paragraph comparison\n\nDraft (topic: digital minimalism):\n\nIn this era of information explosion, we are surrounded by massive amounts of information every day. Digital minimalism is not just a lifestyle, but a revolution of the soul.\n\nIt lets us re-examine our relationship with technology and reclaim inner peace and focus. As one expert put it: \"What you pay attention to determines your life.\" This sentence profoundly reveals the core idea of digital minimalism.\n\nTo practice digital minimalism, start from three aspects: clean up your phone screen, organize your information sources, and protect your deep-focus time. These three steps complement each other, and none can be missing.\n\nCleaning the phone screen not only refreshes the interface but also reduces needless distraction; organizing information sources is like building a dam for your thoughts, letting valuable content flow naturally; and protecting deep-focus time is the scarcity of our era, worth pursuing for a lifetime.\n\nIt is worth mentioning that a growing body of research shows digital minimalism is becoming a global trend. It is not only changing individuals, but reshaping the information ecosystem of the entire society.\n\nIn some sense, putting down the phone means picking up your life. May we all find our own tranquility in this noisy era.", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "humanizer-year-end-review", "title": "年终总结「去 AI 味」改写", "title_en": "De-AI Rewrite for a Year-End Review", "subtitle": "把「显著提升、充分发挥」的模板腔，写成领导愿意看完的总结", "subtitle_en": "From buzzword template to a review your manager will actually finish", "description": "年底最头疼的事：让 AI 帮写年终总结，交上去一眼假——「圆满完成各项工作」「显著提升」「业务协作学习」三段式，不足写得像表扬，展望全是空话。这个案例把一份典型的 AI 生成年终总结交给「去 AI 味」Skill 改写。\n\n改写后的总结有具体事项、有真实的自我反思、有能落地的来年计划；初稿里没有的业绩数字全部用醒目的占位标记提醒本人填写，而不是让 AI 编一个。案例完整展示前后对照、每处改写对应的 AI 写作模式（粗体标签列表、空泛自我批评、套话式展望等）。注意：改写稿中的具体场景是示例演示，年底套用时记得替换成你自己的真实经历。\n\n适合：互联网打工人、需要写季度/年度总结的职场人、团队管理者（可当评审参考）、应届生和实习生。", "description_en": "The year-end headache: you let AI write your annual review, and it screams fake — \"successfully completed all tasks,\" \"significantly improved,\" the mandatory business/collaboration/learning three-parter, weaknesses that read like compliments, and an outlook full of hot air. This case hands a typical AI-generated year-end review to the Humanizer Skill.\n\nThe rewritten version has concrete items, honest self-reflection, and a plan you can actually execute next year. Wherever the draft lacks real numbers, a clearly marked placeholder reminds you to fill in your own — the AI is never allowed to invent one. The case shows the full before-and-after, each edit labeled by the AI writing pattern it fixes (bold-header lists, hollow self-criticism, clichéd outlook). Note: the concrete scenes in the rewrite are demo illustrations — swap in your own real experiences when reusing it as your December template.\n\nBest for: office workers, anyone writing quarterly or annual reviews, team managers (as a review rubric), and new graduates or interns.", "categories": [ "office", "productivity" ], "tags": [ "去AI味", "年终总结", "职场写作", "工作汇报", "模板腔" ], "tags_en": [ "De-AI", "Year-End Review", "Workplace Writing", "Work Reports", "Template Speak" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "年终总结去 AI 味改写对照报告" } ], "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-08-27T17:00:00+08:00", "updated_at": "2026-08-27T17:00:00+08:00", "prompt": "请帮我把一份 AI 生成的年终总结做「去 AI 味」改写，输出前后完整对照和逐条修改说明。\n\n要求：\n1. 保留原有的工作事实框架，不虚构业绩数字；初稿里缺失的数字用醒目占位标记提醒本人填写\n2. 清除「圆满完成」「显著提升」「充分发挥」「持续深耕」等模板词，拆掉「业务/协作/学习」三段式套路\n3. 改成有具体事项、有真实反思的第一人称口吻，承认不足时说人话、给出改进动作\n4. 每处主要改动标注对应的 AI 写作模式（如粗体标签列表、空泛自我批评、套话式展望）\n5. 产出一份单文件 HTML 报告，中英双语可切换，明亮配色，模拟文档审阅的对照效果\n\n初稿（岗位：互联网公司产品运营）：\n\n回顾过去的一年，在公司领导的正确指引和团队伙伴的鼎力支持下，我圆满完成了各项工作任务，实现了个人能力的显著提升。\n\n在业务方面，我积极参与了多个重点项目的推进工作，充分发挥自身专业优势，为项目的高质量交付贡献了力量。在协作方面，我主动承担跨部门沟通职责，有效提升了团队的协作效率。在学习方面，我持续深耕行业知识，不断拓宽认知边界，为个人成长注入了源源不断的动力。\n\n当然，我也清醒地认识到自身存在的不足：在时间管理方面仍有提升空间，在全局思维方面仍需持续加强。\n\n展望未来，我将以更加饱满的热情和更加务实的作风，迎接新的机遇与挑战，为公司的发展贡献更大的力量。", "prompt_en": "Please \"de-AI\" rewrite an AI-generated year-end review and deliver a full before-and-after comparison with a change-by-change explanation.\n\nRequirements:\n1. Keep the factual work framework; do not invent performance numbers — where the draft lacks numbers, add clearly marked placeholders reminding the author to fill them in\n2. Remove template words like \"successfully completed,\" \"significantly improved,\" \"fully leveraged,\" and \"continuously deepened,\" and dismantle the business/collaboration/learning three-part formula\n3. Rewrite in a first-person voice with concrete items and honest reflection; when admitting weaknesses, speak plainly and give an improvement action\n4. Label each major edit with the AI writing pattern it fixes (e.g. bold-header lists, hollow self-criticism, clichéd outlook)\n5. Produce a single-file HTML report, bilingual (Chinese/English toggle), bright color scheme, styled like a document review with tracked changes\n\nDraft (role: product operations at an internet company):\n\nLooking back on the past year, under the correct guidance of company leadership and the strong support of my teammates, I successfully completed all work tasks and achieved significant improvement in my personal capabilities.\n\nIn terms of business, I actively participated in advancing several key projects, fully leveraging my professional strengths and contributing to high-quality delivery. In terms of collaboration, I proactively took on cross-department communication responsibilities, effectively improving team efficiency. In terms of learning, I continuously deepened my industry knowledge and kept broadening my cognitive boundaries, injecting constant momentum into my personal growth.\n\nOf course, I am also keenly aware of my shortcomings: there is still room for improvement in time management, and global thinking needs continuous strengthening.\n\nLooking ahead, I will embrace new opportunities and challenges with greater enthusiasm and a more pragmatic attitude, contributing even more to the company's development.", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "ielts-30day-sprint", "title": "雅思30天冲刺计划", "title_en": "30-Day IELTS Sprint Plan", "subtitle": "从5.5冲到6.5，每天2小时精准提分", "subtitle_en": "From 5.5 to 6.5 in 30 days, 2 hours of focused daily prep", "description": "考雅思最头疼的不是题难，是自己排计划总是平均用力、弱项没多练、强项白花时间。这个案例演示了如何输入目标分数、当前水平和每日时长，让学习规划师自动匹配雅思模板，按基础/强化/冲刺三阶段拆分30天任务，弱项听力+50%时长、强项阅读-30%，每天打开就有具体可执行的任务清单。适合：雅思冲刺期考生、短期提分需求的同学、自己排计划总是执行不下去的人。", "description_en": "The hardest part of IELTS prep isn't the test itself — it's making a plan that actually targets your weak spots. This case shows how to input your target score, current level, and daily hours, and get a 30-day plan split into Foundation/Intensive/Sprint phases. Listening (weak area) gets +50% time; Reading (strong area) gets -30%. Every day you open the plan and know exactly what to do. Ideal for: IELTS sprint candidates, short-term score boosters, anyone whose self-made plans never survive contact with reality.", "categories": [ "education" ], "tags": [ "雅思", "备考计划", "30天冲刺", "提分" ], "tags_en": [ "IELTS", "exam prep plan", "30-day sprint", "score improvement" ], "artifact_type": "md", "cover_image": "cover.png", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "雅思30天冲刺计划", "label_en": "30-Day IELTS Sprint Plan" } ], "prompt": "帮我制定一个雅思冲刺计划：目标总分6.5（小分6.0+），当前5.5水平，听力最弱（5.0），阅读还行（6.0），30天倒计时，工作日每天2小时，周末每天3小时。用按天精度排，输出Markdown格式。", "prompt_en": "Help me create an IELTS sprint plan: target overall band 6.5 (with each section 6.0+), current level 5.5, listening is weakest (5.0), reading is okay (6.0), with a 30-day countdown, 2 hours per day on weekdays and 3 hours per day on weekends. Schedule it with day-by-day precision and output in Markdown format.", "skills": [ { "id": "study-planner", "name": "学习规划师", "description": "输入目标+截止日+每日时长，输出每天可执行的学习计划，支持雅思/考研/期末复习等场景", "skill_id": "skill_2057443739048407040" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-05-13T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "inactive-inventory-90d", "title": "90天未动销库存诊断", "title_en": "90-Day Inactive Inventory Diagnosis", "subtitle": "连续未动销 SKU 库存积压与清货方案", "subtitle_en": "A practical Weimob playbook for 90-day inactive inventory diagnosis", "description": "做90天未动销库存诊断时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“90天未动销库存诊断”展开：识别连续 90 天未动销的 SKU，按照当前库存量排序，分析滞销原因并制定清货方案。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、库存分析运营人员、经营分析团队。", "description_en": "The hardest part of 90-day inactive inventory diagnosis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates 90-day inactive inventory diagnosis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Inventory Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "未动销SKU", "库存积压", "滞销诊断", "折价清货", "库存优化" ], "tags_en": [ "90-Day Inactive Inventory Diagnosis", "Mall Operations", "Inventory Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我圈出连续90天未动销的SKU，按当前库存量排序，分析滞销原因，并给出折价、捆绑销售或退仓方案。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a 90-day inactive inventory diagnosis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "innov-x-drug-focus", "title": "创新药近期关注点", "title_en": "Recent Focus Areas in Innovative Drugs", "subtitle": "梳理创新药相关资讯、观点与产品", "subtitle_en": "E Fund hotspot combing + manager views: innovative-drug news facts, verbatim viewpoints, related products", "description": "行业赛道热点经常变动？资讯零散不知道怎么解读？本工具检索易方达相关公众号内容，生成 HTML 报告：资讯梳理、指数行情、观点速览、相关产品清单。适合跟踪行业热点的理财用户。", "description_en": "Industry-sector hotspots keep shifting, and scattered news is hard to interpret? This tool retrieves content from E Fund's official accounts and generates an HTML report: news digest, index quotes, viewpoint highlights, and a related-product list. Ideal for wealth-management users tracking sector hotspots.", "categories": [ "finance" ], "tags": [ "创新药", "基金经理观点", "热点梳理", "指数行情", "易方达" ], "tags_en": [ "innovative drugs", "manager views", "hotspot combing", "index quotes", "E Fund" ], "artifact_type": "html", "preview": "创新药近期关注点.html", "artifacts": [ { "file": "创新药近期关注点.html", "label": "创新药近期关注点（综合报告：资讯与事实梳理 + 相关资料·原文引用 + 相关指数行情 + 相关产品及代码）" } ], "skills": [ { "id": "efunds-hotspot-combing", "name": "易方达热点梳理", "description": "基于 efunds-mcp 检索易方达基金/易方达指数通公众号，做财经热点资讯与事实梳理；指数行情取易方达指数直通车 etf.com.cn", "skill_id": "skill_2092287909491437568" }, { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "基于易方达官网、efunds 公众号文章库、指数直通车三源，输出基金经理公开发表的投资观点（含强制 efunds 连接闸门）", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "调用「易方达热点梳理」「易方达基金经理观点检索」，通过 efunds‑mcp 检索易方达旗下公众号文章，围绕**创新药近期关注点**生成 HTML 报告：\n1. 资讯事实梳理，客观陈述，不预判后市；\n2. 原文摘录基金经理、指数文章内容，标题加原文跳转链接；\n3. 展示创新药、港股通创新药指数行情（涨红跌绿，数据源[etf.com.cn](https://etf.com.cn)，无数据则省略）；\n4. 列出对应产品名称 + 代码。\n输出单文件内联 HTML，界面简洁专业。", "prompt_en": "Invoke 'E Fund Hotspot Combing' and 'E Fund Manager Views', retrieve E Fund's official-account articles via efunds-mcp, and generate an HTML report on **Recent Focus Areas in Innovative Drugs**:\n1. Digest news and facts objectively, without forecasting the market;\n2. Verbatim excerpts from managers' notes and index articles, with titles linking to the original;\n3. Show innovative-drug and HK-Connect innovative-drug index quotes (red up / green down, source [etf.com.cn](https://etf.com.cn), omit if no data);\n4. List the corresponding product names + codes.\nOutput a single inlined HTML file with a clean, professional interface.", "notes": "可复现性说明：①「易方达热点梳理(skill_2092287909491437568)」真实产出：资讯与事实、原文引文来自 efunds-mcp 对「易方达基金」「易方达指数通」公众号的检索与正文读取（连接状态下抓取）；相关指数行情来自易方达指数直通车 etf.com.cn 官方接口（2026-08-28），按「没数据不展示」原则——恒生港股通创新药指数(HSSCID)、恒生创新药指数(HSIDI) 等全字段为空故整行不展示；相关产品仅列文章提及的产品及代码，净值阶段表现数据源未返回，不展示、不推算。②「易方达基金经理观点检索(skill_2089896329561710592)」真实产出：efunds 连接器在检索阶段已连接，按关键词观点检索分支（主题词「创新药」）检索，取经理署名文章（成曦《AI+创新药=？》）与易方达指数通指数文章做逐字摘录，标题链接微信原文 article_url。报告已按用户要求：删除热点要点与原文引用独立段落、所有文章标题改为可点击跳转原文、UI 升级（品牌蓝渐变封面 + 卡片悬浮 + 新闻流日期胶囊 + 红涨绿跌表格）。合规：禁用词 0 命中；自撰正文警告词 0 命中（仅技能强制固定的免责注记与风险页脚含「建议/保证」，属不可改写文案，已豁免）。", "created_at": "2026-08-28T14:07:00+08:00", "updated_at": "2026-08-28T18:05:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "innovation-competition-workbench", "title": "美味法法的大创竞赛工作台", "title_en": "Innovation Competition Workspace", "subtitle": "一个团队备赛的全部动作：倒计时盯截止、交付物进度、任务分工、资料归集、多赛事 Deadline 管理", "subtitle_en": "Everything a student team needs for innovation competitions: countdown, deliverable progress, task assignment, material collection and multi-contest deadlines", "description": "大学生创新创业竞赛备赛最乱的从来不是选题，而是过程：申报书、计划书、答辩 PPT 各写各的，任务分工靠群里喊，往届资料散在每个人电脑里，等想起校赛截止日期时往往只剩三天。\n\n本案例是一个团队备赛工作台：头部是主赛事提交倒计时；「今天要处理」把逾期、当天到期、七天内到期的任务/交付物/竞赛截止自动汇总成一张今日清单，逾期置顶标红；仪表盘四卡显示整体完成度、任务完成率、进行中任务和逾期项；「项目进度」按交付物管理申报书/计划书/答辩稿，进度环 + 一键 ±10%；「团队任务」支持负责人、P0-P2 优先级、状态流转和已完成折叠；「资料管理」按通知文件/选题材料/申报材料/参考文献/团队文档六类归集，可挂链接和存放备注；「竞赛截止」同时盯多个赛事的报名、初赛、提交节点，临近自动变黄、逾期变红。\n\n适合：参加大创、互联网+、挑战杯等竞赛的学生团队（一个页面管住全部节点和分工），带队的指导老师或班长（布置任务、检查进度），以及想参考「多实体清单 + 截止日引擎 + 双列工作台布局」的前端学习者。", "description_en": "The messiest part of an innovation competition is never the topic — it's the process: the application form, project plan and defense slides live in separate places, task assignments get shouted across group chats, past materials scatter across everyone's laptops, and campus deadlines arrive with three days left.\n\nThis case is a team competition workspace: the header keeps a live countdown to the main submission; \"Today's Focus\" aggregates overdue, due-today and due-within-7-days tasks, deliverables and contest deadlines into one list, with overdue items pinned in red; a four-card dashboard tracks overall progress, task completion, active tasks and overdue items; \"Progress\" manages deliverables with progress rings and ±10% quick buttons; \"Tasks\" supports owners, P0-P2 priorities, status flow and a collapsible done section; \"Materials\" organizes files into six categories with links and location notes; \"Deadlines\" watches multiple contests' signup, preliminary and submission nodes, turning yellow when close and red when overdue.\n\nFor: student teams entering innovation, Internet+ or Challenge Cup competitions (one page for every node and assignment), advising teachers or team leads who assign and review work, and front-end learners interested in the multi-entity checklist + deadline engine + two-column workspace layout.", "categories": [ "workspace" ], "tags": [ "大创竞赛", "团队协作", "项目管理", "竞赛备赛", "个人工作台" ], "tags_en": [ "innovation competition", "team collaboration", "project management", "contest prep", "personal workspace" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "大创竞赛工作台 单文件 HTML（内置 i18n 中英文 + 双语演示数据，localStorage 自包含）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/7zWL4rMNgWKZpbSZ0CzMqG，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n我想做一个「大学生创新创业竞赛」团队备赛工作台，四个成员协作，包含：\n1. 头部倒计时：主赛事提交截止日倒计时（剩余天数 + 日期），逾期显示已截止；右上角设置（竞赛名称/截止日/团队成员可增减改名，改名自动同步所有任务负责人字段）。\n2. 今天要处理：自动汇总三类来源——未完成任务（逾期/今天到期/7 天内到期）、未完成交付物（同规则）、未截止竞赛节点（同规则），逾期置顶红底显示已逾期 N 天，点击右侧按钮跳转对应区块。\n3. 仪表盘四卡：整体完成度（全部交付物进度均值环形图）、任务完成率（已完成/总数）、进行中任务数（副标题显示其中 P0 数）、逾期项总数（任务+交付物+竞赛合计，>0 红色）。\n4. 项目进度：交付物卡片（申报书/项目计划书/答辩 PPT 等），字段有名称/描述/负责人/截止日/状态（待启动/进行中/已完成）/进度百分比；进度环 + 底部进度条，一键 ±10% 快捷按钮，到 100% 自动标记完成。\n5. 团队任务：任务卡片带完成勾选（点击在待办/完成间切换）、负责人、P0/P1/P2 优先级徽章、状态（待办/进行中/已完成）、截止日与剩余天数（逾期红字）；按优先级+截止日排序，已完成自动折叠成「已完成 (N)」可展开。\n6. 资料管理：按通知文件/选题材料/申报材料/参考文献/团队文档/其他分类，每条资料有名称/分类/状态（待收集/已收集/已整理/已提交）/负责人/链接（可点开）/存放备注，分类筛选 chips。\n7. 竞赛截止：多赛事节点卡片（大创、互联网+、挑战杯等），字段有竞赛名/阶段/截止日/状态（报名中/未开始/准备中/已截止/已完成）/官网链接/备注；左侧色条按紧急度红/黄/蓝，显示剩余天数徽章，按日期排序。\n8. 新建与编辑：右上一个「新建」主按钮（按当前 Tab 新建对应实体），全部新建/编辑走统一弹窗表单（必填校验：名称、截止日、阶段）；删除即时生效。\n\n视觉与靛蓝渐变头部 + 白色大圆角卡片 + 徽章体系；PC ≥1024px 双列布局（左列今日清单+仪表盘固定跟随，右列 Tab 内容，交付物两列网格），平板/手机单列、Tab 图标化。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the cloud database (资料库) capability to one-click clone this workbench: https://www.workbuddy.cn/space/d/7zWL4rMNgWKZpbSZ0CzMqG — data must be stored online with bidirectional sync. Prefer the one-click clone path; if unavailable, build it this way: I want a team workspace for college innovation & entrepreneurship competitions (4 members), including:\n1. Header countdown: days remaining to the main submission deadline (+ date), switching to \"Closed\" when past; top-right settings (competition name, deadline, add/remove/rename team members with owner fields auto-updated).\n2. Today's Focus: auto-aggregates three sources — unfinished tasks, unfinished deliverables and open contest nodes that are overdue, due today or due within 7 days; overdue items pinned with red background and \"N days overdue\"; jump buttons switch to the matching tab.\n3. Four-card dashboard: overall progress (average of all deliverable progress as a ring), task completion rate (done/total), active task count (subtitle shows P0 count), and total overdue items (tasks + deliverables + contests, red when > 0).\n4. Progress: deliverable cards (application form / project plan / defense slides...) with name, description, owner, deadline, status (not started / in progress / done) and progress percent; progress ring + bottom bar with quick ±10% buttons; reaching 100% auto-marks done.\n5. Tasks: task cards with completion checkbox (toggles between to-do and done), owner, P0/P1/P2 priority badge, status, deadline with days left (red text when overdue); sorted by priority + deadline; completed tasks auto-collapse into an expandable \"Done (N)\" section.\n6. Materials: categorized as notices / topic materials / application docs / references / team docs / other; each entry has name, category, status (to collect / collected / organized / submitted), owner, clickable link and location note, with category filter chips.\n7. Deadlines: multi-contest node cards (innovation program, Internet+, Challenge Cup...) with contest name, stage, deadline, status (registering / not started / preparing / closed / done), official link and note; left color stripe red/yellow/blue by urgency, days-left badge, sorted by date.\n8. Create & edit: one primary \"New\" button at top right (creates the entity for the active tab); all create/edit flows use a unified modal form with required-field validation (name, deadline, stage); deletes take effect instantly. Visual & delivery: with zero external links; indigo gradient header + white rounded cards + badge system; on PC ≥1024px a two-column layout (left column: today list + dashboard, sticky; right column: tab content with a 2-column deliverable grid), single column with icon tabs on tablet/mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "美味法法", "name_en": "Delicious Fafa", "avatar": "", "badge": "community" }, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "created_at": "2026-08-18T16:30:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "innovative-drug-chain-workbench", "title": "创新药产业链分析工作台", "title_en": "Innovative Drug Chain Analysis Workbench", "subtitle": "上游/中游/下游分层拆解 ｜ 真受益 vs 伪受益识别", "subtitle_en": "Upstream / Midstream / Downstream breakdown | True vs. false beneficiaries", "description": "看创新药板块天天大涨，不知道该不该追，也不知道自己买的到底是不是真受益？本案例组合「NeoData 金融搜索服务」与「WeStock Data」：用 NeoData 查产业链各环节公司与基本面（研发投入、管线、外包订单逻辑），用 WeStock Data 拉行情与涨跌幅，把创新药产业链切成上、中、下游三段，告诉你每一段在卖什么水、哪些是真受益（CXO 卖水人 / 创新药企龙头）、哪些是伪受益（中药蹭概念 / 疫苗代理）。顶部有今天的板块异动速览（板块指数、年初至今、伪受益监测数）和「警示：粤万年青 +16.26% 涨停，无实质创新药管线」；中间是产业链全景图（上游原料药/中游 CXO/下游创新药企 + 警示区四卡片）和真伪受益统计；下面是可筛选可增删改的代表公司雷达（21 家覆盖 A 股+港股，按涨跌幅排序），每个公司都有受益逻辑和确定性评级；最下方是客观判断（最受益环节、伪受益三要素、风险提示），底部清晰标注「不构成投资建议」。适合做 A 股/港股创新药板块投资决策的人、医药行业研究员、想识别板块里「真伪受益」的人。", "description_en": "Innovative-drug names are surging, but is what you bought actually benefiting from the trend? This case combines NeoData Financial Search with WeStock Data: NeoData supplies the value-chain picture (R&D spend, pipelines, outsourcing logic), WeStock Data supplies quotes and daily changes. The workbench slices the innovative-drug chain into Upstream (APIs/intermediates) / Midstream (CXO/CDMO) / Downstream (innovative pharma & commercialization), plus a separate Alert Zone for concept plays. The header shows a sector dashboard (index, YTD, false-beneficiary watch) and a real-time alert such as 'Wannianqing hit +16.26% limit-up with no real innovative pipeline'; the middle has the full industry-chain map (four cards) and true/watch/false counts; below sits a filterable company radar covering 21 A-share + HK names with benefit logic and certainty ratings; the bottom is an objective assessment (most-advantaged segment, three tests for false beneficiaries, risk notes) with a clear 'not investment advice' disclaimer. Best for: A-share / HK investors making decisions in the innovative-drug sector, pharma analysts, and anyone wanting to separate true beneficiaries from concept plays.", "categories": [ "finance" ], "tags": [ "创新药", "产业链研究", "CXO", "真伪受益", "投资工具" ], "tags_en": [ "Innovative Drugs", "Value Chain", "CXO", "True vs False Beneficiaries", "Investment Tool" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "创新药产业链分析工作台（单体双语 HTML）" } ], "prompt": "帮我做一个「创新药产业链分析工作台」单页 HTML，把创新药板块按上中下游+警示区分层拆解。用 NeoData 金融搜索服务查产业链各环节公司与基本面（研发投入、管线、外包订单逻辑），用 WeStock Data 查行情（最新价、涨跌幅、板块排行）。\n\n五大模块：\n1. 顶部今天要处理——三张 KPI 卡（创新药板块指数+今日涨跌幅+成交额、板块年初至今、伪受益标的监测数），下方一条警示条提示「某只伪受益个股涨停」类事件；\n2. 产业链全景——四个并排卡片（上游原料药/中间体、中游 CXO/CDMO、下游创新药企/商业化、警示区），每卡内列代表公司名字+确定性评级，底部四个统计（真受益数/需甄别数/伪受益数/确定性最高环节）；\n3. 代表公司雷达——可筛选可增删改的公司卡片（默认 21 家覆盖药明康德/康龙化成/泰格医药/昭衍新药/凯莱英/药明生物等 CXO 龙头、恒瑞/百济/信达等创新药企、华海/普洛/九洲等上游原料药、粤万年青/陇神戎发/华森制药等伪受益），按涨跌幅排序，支持「全部/真受益/需甄别/伪受益」分段+「全部环节/上游/中游/下游/跨界」环节筛选+「添加标的」弹窗录入新公司（公司名/股票代码/产业链环节/代表产品/受益逻辑/确定性/受益属性/最新价/涨跌幅）；\n4. 客观判断——三张卡片：最受益环节（中游 CXO 为什么确定性最高）+ 伪受益识别三要素（看管线/看收入/看涨幅归因）+ 风险提示 4 条；\n5. 免责声明——明确标注「不构成投资建议」。\n\n视觉走金融工作台风（亮色蓝紫渐变 header，A 股涨红跌绿配色），所有 CSS/JS 全内联、无外部 CDN 依赖。涨跌按 A 股习惯（涨红跌绿）。", "prompt_en": "Build an 'Innovative Drug Chain Analysis Workbench' as a single-page HTML that breaks the sector into upstream / midstream / downstream plus an alert zone. Use NeoData Financial Search for value-chain companies and fundamentals (R&D spend, pipelines, outsourcing logic) and WeStock Data for quotes (last price, change %, sector ranking).\n\nFive modules:/n1. Top 'Today's Focus' — three KPI cards (sector index with daily change and turnover, sector YTD, false-beneficiary watch count) plus an alert bar flagging events like 'A concept-play name hit limit-up with no real pipeline';\n2. Industry Chain Map — four side-by-side cards (upstream APIs/intermediates, midstream CXO/CDMO, downstream innovative pharma/commercialization, alert zone), each listing representative companies and a certainty rating; below, four stat tiles (true / scrutiny / false counts, most-certain segment);\n3. Company Radar — a filterable, add/edit/delete-enabled grid of company cards (default 21 names covering WuXi AppTec, Pharmaron, Tigermed, Joinn Labs, Asymchem, WuXi Biologics for CXO; Hengrui, BeiGene, Innovent for innovators; Huahai / Apeloa / Jiuzhou for upstream APIs; and Wannianqing / Longshengrongfa / Huasen for the concept-play alert zone), sorted by daily change, with segmented filter (all / true / scrutiny / false), chain-segment filter (all / upstream / midstream / downstream / crossover) and an 'Add Stock' modal capturing name / ticker / chain segment / flagship product / benefit logic / certainty / benefit type / price / change;\n4. Objective Assessment — three cards: most-advantaged segment (why midstream CXO carries the highest certainty) + three tests for false beneficiaries (pipeline / revenue / price attribution) + four risk notes;\n5. Disclaimer — clearly labelled 'not investment advice'.\n\nVisual: light finance-workbench style (blue-purple gradient header, A-share red-up/green-down colors). All CSS/JS inlined, no external CDN dependencies.", "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "creator": { "name": "星月聊财经", "name_en": "星月聊财经", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 92, "source": "community", "created_at": "2026-08-25T22:30:00+08:00", "updated_at": "2026-08-25T22:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-attack-map", "title": "客户攻略作战图", "title_en": "Account Strategy Map", "subtitle": "把关键人、真实需求和接触路径放在一张图上", "subtitle_en": "Map stakeholders, real needs, and access paths in one view", "description": "大客户推进最难的不是联系人少，而是不知道谁影响谁、谁反对、谁能帮你进入决策圈。\n\n这个案例把角色、影响力、立场、需求证据和关系缺口组合成一张客户攻略作战图，并给出下一条验证路径。\n\n适合：大客户销售、销售经理、售前顾问、渠道经理和客户成功负责人。", "description_en": "Complex deals stall when teams cannot see who influences whom, who resists, or how to enter the decision circle.\n\nThis case combines roles, influence, stance, need evidence, and relationship gaps into one account strategy map with the next validation path.\n\nBest for: enterprise sales, sales managers, presales, channel managers, and customer success leaders.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "客户攻略作战图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {客户名称或商机名称}、{沟通记录} 和 {已获准使用的关系信息}，制作客户攻略作战图。先用用户提供的名称在 CRM 搜索并确认对应客户或商机；若存在同名客户、多主体或多个匹配商机，只列出最少必要候选让用户确认，不要求用户提供 CRM 内部 ID。区分已确认角色、公开信息、用户判断和 AI 假设；工商董监高不能直接等同实际决策人。为每个关键角色标明影响力、立场、关注任务、证据来源和最后验证日期；画出已确认关系与待验证关系，识别决策圈缺口、潜在反对者和可行引荐路径。每个关键角色必须以结构化角色卡呈现，至少包含：影响力（高/中/低）、立场（支持/中立/反对/未知）、关注任务（一句话）、证据来源（CRM 事实/沟通记录/公开信息/AI 假设）、最后验证日期（YYYY-MM-DD）。最后只给一个最低成本的下一步验证动作，不自动写入 CRM 或编造私人关系。输出一个可离线打开的响应式 HTML，所有图形使用内联 SVG/CSS，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Using {customer name or opportunity name}, {conversation records}, and {authorized relationship information}, create an account strategy map. First search CRM using the name supplied by the user and confirm the matching account or opportunity. If there are duplicate names, multiple legal entities, or several matching opportunities, show only the minimum necessary candidates for the user to choose from; never require an internal CRM ID. Separate confirmed roles, public information, user judgments, and AI hypotheses; do not treat registered directors as actual decision makers. For each stakeholder show influence, stance, job to be done, evidence source, and last verification date. Draw confirmed and unverified relationships, identify decision-circle gaps, likely blockers, and viable introduction paths. Present every key stakeholder as a structured role card containing at least: influence (high/medium/low), stance (supportive/neutral/opposed/unknown), job to be done (one sentence), evidence source (CRM fact/conversation record/public information/AI hypothesis), and last verification date (YYYY-MM-DD). End with one lowest-cost validation action, without automatically writing to CRM or inventing private relationships. Return one responsive offline HTML file using inline SVG and CSS only, with no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "客户攻略", "关键人地图", "真实需求", "关系路径" ], "tags_en": [ "Account Strategy", "Stakeholder Map", "Real Needs", "Relationship Path" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-bid-response", "title": "竞标应答要点页", "title_en": "Bid Response Strategy", "subtitle": "围绕具体评分项组织证据与应答动作", "subtitle_en": "Organize evidence and responses around the actual scoring criteria", "description": "应标材料最容易写成产品功能堆砌，却没有回答评委真正按什么打分。\n\n这个案例把招标要求拆成评分项、证据、差距、应答策略和现场澄清问题，形成一页可执行的竞标作战稿。\n\n适合：销售、售前顾问、投标经理、商务负责人和解决方案团队。", "description_en": "Bid responses often become feature catalogs that fail to address how evaluators will actually score.\n\nThis case turns tender requirements into scoring criteria, evidence, gaps, response strategy, and clarification questions in one executable page.\n\nBest for: sales, presales, bid managers, commercial leaders, and solution teams.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "竞标应答要点页" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {招标文件或评分表}、{我方产品与案例材料} 和 {已知竞品信息}，制作竞标应答要点页。先完整读取招标要求并逐项建立追溯编号；按评分权重识别必须赢、必须达标和低优先级项。每项列出客户要求、我方应答、可披露证据、证据缺口、负责人和截止时间；公开信息与推测分开，不能编造资质、案例、功能、价格或竞品缺陷。对模糊或限制竞争的条款提出中性澄清问题，并明确红线。最终只输出一个专业、信息密度高的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Using {tender document or scoring sheet}, {our product and case materials}, and {known competitor information}, create a bid response strategy page. Read the full tender and assign traceable requirement IDs. Classify criteria by must-win, must-qualify, and lower priority according to scoring weight. For each item list the requirement, response, authorized evidence, evidence gap, owner, and due date. Separate public facts from hypotheses and do not fabricate credentials, cases, features, pricing, or competitor weaknesses. Raise neutral clarification questions for ambiguous or restrictive clauses and identify red lines. Return only one professional, high-density responsive HTML file with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "竞标应答", "招投标", "证据策略", "方案" ], "tags_en": [ "Bid Response", "Tender", "Evidence Strategy", "Solution" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-case-showcase", "title": "客户案例包装页", "title_en": "Customer Success Showcase", "subtitle": "把真实成果讲成客户愿意相信的案例", "subtitle_en": "Turn verified outcomes into a credible customer story", "description": "案例最怕只剩一句“客户很满意”，既没有证据，也无法判断能否对外使用。\n\n这个案例把获准披露的背景、转变过程、结果证据和适用边界整理成一页客户案例，并明确哪些数字可用、哪些表述必须避开。\n\n适合：销售、市场、售前顾问、客户成功和案例运营人员。", "description_en": "Customer stories often collapse into vague claims with no evidence or disclosure boundary.\n\nThis case turns authorized context, the transformation journey, outcome evidence, and limitations into a polished one-page showcase while marking what may and may not be shared.\n\nBest for: sales, marketing, presales, customer success, and case operations teams.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "客户案例包装页" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {获准使用的客户案例材料} 和 {目标受众或行业}，制作一份可对外展示的客户案例包装页。先核对客户名称、披露授权、事实来源和数据口径；只使用 CRM、用户材料或明确获准的证据，不补造客户评价、实施周期、收益数字或竞品结论。把内容组织为“原有挑战 → 采取行动 → 可验证变化 → 适用条件”，对单源信息、推测和不可披露内容明确标注。至少包含一句案例结论、三项结果证据、一条限制条件和一个面向相似客户的验证问题。最终只输出一个专业、克制、可离线打开的响应式 HTML；所有资源内联，不使用远程字体、外链资源、emoji 或 hash 锚点。", "prompt_en": "Using {authorized customer case materials} and {target audience or industry}, create a customer-facing case showcase. Verify the customer name, disclosure permission, fact sources, and metric definitions first. Use only CRM facts, user-provided materials, or explicitly authorized evidence; do not fabricate quotes, timelines, benefits, or competitor claims. Structure the story as challenge, action, verifiable change, and applicability. Label single-source claims, hypotheses, and restricted content. Include one case conclusion, three outcome signals, one limitation, and one validation question for similar prospects. Return only one professional, restrained, responsive, offline HTML file with all resources inline and no remote fonts, external assets, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "客户案例", "证据治理", "价值表达", "品牌展示" ], "tags_en": [ "Customer Story", "Evidence Governance", "Value Messaging", "Brand Showcase" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-contract-heatmap", "title": "合同风险热力图", "title_en": "Contract Risk Heatmap", "subtitle": "在合同原文上直接看见风险集中区", "subtitle_en": "See risk concentration directly on the contract text", "description": "合同审查条目一多，业务人员很难快速定位原文，更不知道风险是不是集中在同一交易链路。\n\n这个案例把合同原文按付款、验收、责任和知识产权分层着色，并在侧栏给出风险密度、联动影响和谈判优先级。\n\n适合：销售、商务、采购、交付负责人和企业法务协作人员。", "description_en": "Long review lists make it hard to locate the original language or see connected exposure.\n\nThis case overlays payment, acceptance, liability, and IP risk on the source contract and summarizes concentration, linked impact, and negotiation priority.\n\nBest for: sales, commercial, procurement, delivery, and legal collaboration teams.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "合同风险热力图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请审查 {上传的合同文件}，并制作合同风险热力图。必须完整读取正文、附件、表格和引用文件；识别不清或缺页时先说明实际读取范围。以合同原文为唯一审查依据，逐段引用并标注高、中、低风险，重点联动检查付款、验收、违约责任、责任上限、数据和知识产权。每个风险必须给出条款位置、原文、业务影响、目标修改文本、可接受退让和是否需要法务升级；不能把不利条款误判为缺失，也不能承诺签署绝对安全。最终只输出一个原文叠加标注的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Review {uploaded contract file} and create a contract risk heatmap. Read the full body, appendices, tables, and referenced documents; state the actual reading scope if pages are missing or unclear. Treat the contract text as the sole review basis. Quote and mark each paragraph as high, medium, or low risk, with special cross-checks for payment, acceptance, breach, liability caps, data, and IP. For every issue include location, original text, business impact, target language, acceptable fallback, and legal escalation need. Do not confuse unfavorable language with missing language or promise absolute safety. Return only one responsive HTML file with inline text annotations and all assets inline, without external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "合同审查", "风险热力图", "条款定位", "谈判" ], "tags_en": [ "Contract Review", "Risk Heatmap", "Clause Location", "Negotiation" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-loss-review", "title": "丢单复盘画布", "title_en": "Loss Review Canvas", "subtitle": "分清事实、判断和运气，找到下一次能改变的动作", "subtitle_en": "Separate facts, judgment, and luck to improve the next decision", "description": "丢单复盘最容易变成“价格太高”或“关系不到位”，最后没有任何可复用结论。\n\n这个案例还原关键时间线，区分当时可见信息与事后信息，审查决策质量、预测偏差和真正可控的改进动作。\n\n适合：销售、销售经理、经营分析、售前负责人和复盘主持人。", "description_en": "Loss reviews often end with “too expensive” or “weak relationship,” producing no reusable lesson.\n\nThis case reconstructs the timeline, separates information available then from hindsight, and audits decision quality, forecast bias, and controllable improvements.\n\nBest for: sales, sales managers, revenue operations, presales leaders, and review facilitators.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "丢单复盘画布" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {丢单商机名称}、{完整跟进记录} 和 {丢单后获得的信息}，制作丢单复盘画布。先用商机名称在 CRM 搜索并确认对应的已丢单商机；若同名或存在多个匹配，只列出最少必要候选让用户确认，不要求用户提供商机内部 ID。按时间线还原关键事实，严格区分当时已经知道的信息、当时的判断、未验证假设和事后才知道的信息，避免结果偏差和个人归责。审查客户决策标准、关键人覆盖、竞争策略、资源投入、预测变化和止损时机；指出哪些是运气、哪些是决策质量、哪些证据仍不足。最后沉淀不超过三条可观察、可复用的流程改进，并指定验证方式。最终只输出一个专业、克制的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Using {lost opportunity name}, {complete activity history}, and {information learned after the loss}, create a loss review canvas. First search CRM by opportunity name and confirm the matching closed-lost opportunity. If duplicate names or multiple matches exist, show only the minimum necessary candidates for the user to choose from; never require an internal opportunity ID. Reconstruct the timeline and strictly separate information known at the time, contemporaneous judgments, untested hypotheses, and hindsight. Avoid outcome bias and personal blame. Review customer criteria, stakeholder coverage, competitive strategy, resource allocation, forecast changes, and exit timing. Identify luck, decision quality, and remaining evidence gaps. End with no more than three observable, reusable process improvements and how each will be validated. Return only one professional, restrained responsive HTML file with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "丢单复盘", "决策质量", "时间线", "销售管理" ], "tags_en": [ "Loss Review", "Decision Quality", "Timeline", "Sales Management" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-negotiation-matrix", "title": "谈判筹码交换矩阵", "title_en": "Negotiation Give-Get Matrix", "subtitle": "每一次让步，都换回一个可验证承诺", "subtitle_en": "Trade every concession for a verifiable commitment", "description": "谈判现场最容易为了推进而直接降价，事后才发现什么也没换回来。\n\n这个案例把客户诉求、我方成本、交换条件和底线放进一张矩阵，形成保守、目标和底线三档方案。\n\n适合：销售、商务负责人、销售经理、报价审批人和大客户团队。", "description_en": "Negotiations often default to discounting, only to discover that nothing meaningful came back.\n\nThis case maps customer asks, internal cost, reciprocal commitments, and walk-away limits into conservative, target, and floor positions.\n\nBest for: sales, commercial leaders, sales managers, pricing approvers, and strategic account teams.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "谈判筹码交换矩阵" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {客户诉求与异议}、{我方可用筹码} 和 {已确认的审批边界}，制作谈判筹码交换矩阵。先区分客户表面要求、真实顾虑和仍待验证的假设；任何让步都必须绑定对方可观察、可确认的回报，不能无条件降价。按低成本高价值、对等交换、高成本谨慎和不可交换四个象限整理筹码，并为关键争议给出目标方案、可接受退让和底线。明确授权不清的事项，不替审批人做决定。最终只输出一个专业、清晰的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Using {customer asks and objections}, {available concessions}, and {confirmed approval boundaries}, create a negotiation give-get matrix. Separate stated requests, underlying concerns, and hypotheses that still need validation. Every concession must be tied to an observable, confirmable return; never discount unconditionally. Organize levers into low-cost high-value, balanced trade, high-cost cautious, and non-tradable quadrants. For each key issue provide a target position, acceptable fallback, and walk-away limit. Mark unclear authority and do not decide for approvers. Return only one professional responsive HTML file with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "谈判策略", "筹码交换", "价格异议", "商务" ], "tags_en": [ "Negotiation Strategy", "Give-Get", "Price Objection", "Commercial" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-relationship-calendar", "title": "客户关系经营日历", "title_en": "Account Relationship Calendar", "subtitle": "把关系目标拆成未来 90 天的自然触点", "subtitle_en": "Turn relationship goals into natural touchpoints over 90 days", "description": "客户关系维护最怕只剩“有空联系一下”，既没有理由，也没有节奏。\n\n这个案例根据关键角色、关系阶段和业务节点，排出未来 90 天的触点日历，并区分价值提供、信息验证和共同推进。\n\n适合：大客户销售、客户经理、渠道经理、客户成功和销售主管。", "description_en": "Relationship plans often degrade into “check in sometime,” with no reason or cadence.\n\nThis case schedules 90 days of touchpoints based on stakeholders, relationship stages, and business milestones, separating value delivery, validation, and joint action.\n\nBest for: enterprise sales, account managers, channel managers, customer success, and sales leaders.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "客户关系经营日历" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请根据 {客户名称或商机名称}、{关键角色信息} 和 {未来 90 天业务节点}，制作客户关系经营日历。先用用户提供的名称在 CRM 搜索并确认对应客户或商机；若存在同名客户、多主体或多个匹配商机，只列出最少必要候选让用户确认，不要求用户提供 CRM 内部 ID。再核对角色身份、关系阶段、最近互动和授权边界；不从私人信息推断兴趣，不制造虚假节日或无价值寒暄。围绕价值提供、信息验证、共同推进三类目标安排自然触点，每次只设一个目标，标明对象、渠道、时机依据、准备材料、成功信号和放弃条件。避免过度触达，同一角色保留合理间隔；未确认的事件标为待验证。最终只输出一个清晰、克制的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Using {customer name or opportunity name}, {stakeholder information}, and {business milestones over the next 90 days}, create an account relationship calendar. First search CRM using the name supplied by the user and confirm the matching account or opportunity. If there are duplicate names, multiple legal entities, or several matching opportunities, show only the minimum necessary candidates for the user to choose from; never require an internal CRM ID. Then verify identities, relationship stages, recent interactions, and permission boundaries. Do not infer interests from private information or manufacture holidays and empty check-ins. Plan natural touchpoints around value delivery, information validation, and joint progress, with one objective per touch. Show the person, channel, timing rationale, preparation, success signal, and stop condition. Avoid over-contacting and keep reasonable spacing; label unconfirmed events as pending validation. Return only one clear, restrained responsive HTML file with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "关系经营", "客户日历", "关键人", "触点规划" ], "tags_en": [ "Relationship Management", "Account Calendar", "Stakeholders", "Touchpoint Planning" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-value-story", "title": "产品价值故事页", "title_en": "Product Value Story", "subtitle": "把产品能力讲成客户愿意继续聊的价值故事", "subtitle_en": "Turn product capabilities into a value story customers want to continue discussing", "description": "产品介绍最容易变成功能清单，客户听完仍不知道为什么值得行动。\n\n这个案例结合 CRM 客户事实、销售认知、产品资料与可追溯的公开信息，生成一页品牌化价值故事，清楚展示客户挑战、价值主张、反证、验证指标和下一步。\n\n适合：销售、售前顾问、解决方案顾问、客户成功和业务负责人。", "description_en": "Product introductions often become feature lists, leaving customers unsure why they should act.\n\nThis case combines CRM facts, sales knowledge, product materials, and traceable public information into a branded one-page value story covering customer challenges, the value proposition, counterevidence, validation metrics, and the next step.\n\nBest for: sales representatives, presales consultants, solution consultants, customer success teams, and business leaders.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "产品价值故事页" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请为 {目标客户名称或商机名称} 制作一份面向客户的产品价值故事页，并使用 {产品资料或链接} 回答“客户为什么值得现在行动”。先用用户提供的名称在 CRM 搜索并确认对应客户或商机；若存在同名客户、多主体或多个匹配商机，只列出最少必要候选让用户确认，不要求用户提供 CRM 内部 ID。确认准确企业主体后，按 CRM 当前事实、用户本轮材料、已沉淀销售认知、多源公开信息、单源信息、AI 假设的顺序取材。优先关注近 180 天变化，重要外部结论注明来源和日期；冲突信息并列展示，推测标明置信度，不把行业共性当成客户事实，不编造客户、案例、预算或收益数字。外部信息只用于本页，不自动写入 CRM。\n\n用“现状问题 → 改变方式 → 可验证结果”提炼一句价值主张和三个价值支柱，同时给出关键未知、至少一条反证或不适用条件，以及一个低成本、可观察的下一步验证动作。没有可靠指标时写“建议验证的目标”，不要写成既有承诺。\n\n最终只输出一个可独立打开的响应式 HTML：专业、克制、可信的 B2B 品牌风格；所有 CSS、SVG 和必要脚本内联，不依赖外部资源，不使用 emoji、远程字体或 hash 锚点。页面至少包含标题区、客户挑战、价值支柱、转变路径、证据与反证、待验证指标和建议下一步。", "prompt_en": "Create a customer-facing product value story page for {target customer name or opportunity name} using {product materials or link}, answering why the customer should act now. First search CRM using the name supplied by the user and confirm the matching account or opportunity. If there are duplicate names, multiple legal entities, or several matching opportunities, show only the minimum necessary candidates for the user to choose from; never require an internal CRM ID. After confirming the correct legal or operating entity, use information in this order: current CRM facts, materials supplied in this request, accumulated sales knowledge, corroborated public information, single-source information, and AI hypotheses. Prioritize changes from the last 180 days. Cite the source and date for important external claims, show conflicting information side by side, label hypotheses with confidence, and never present industry patterns as customer facts or fabricate customers, cases, budgets, or benefit figures. Use external information only for this page and do not write it to CRM automatically.\n\nDevelop one value proposition and three value pillars using the structure “current problem → how it changes → verifiable outcome.” Include key unknowns, at least one piece of counterevidence or disqualifying condition, and one low-cost, observable next validation action. When reliable metrics are unavailable, label them as proposed validation targets rather than existing commitments.\n\nReturn only one standalone responsive HTML file with a professional, restrained, and credible B2B brand direction. Inline all CSS, SVG, and necessary scripts. Do not use external resources, emoji, remote fonts, or hash-anchor navigation. Include a title section, customer challenges, value pillars, a transformation path, evidence and counterevidence, validation targets, and a recommended next step.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "产品价值", "客户沟通", "解决方案", "品牌展示" ], "tags_en": [ "Product Value", "Customer Communication", "Solution Consulting", "Brand Storytelling" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T10:51:17+08:00", "updated_at": "2026-08-26T14:45:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-version-diff", "title": "版本变更对比页", "title_en": "Contract Version Diff", "subtitle": "两版并排，只保留真正影响交易的变化", "subtitle_en": "Compare two versions and isolate deal-impacting changes", "description": "合同来回修订后，最怕金额、条件词或责任范围只改了一个字，却藏在几十页文本里。\n\n这个案例把两版合同并排展示，过滤纯格式变化，突出金额、日期、条件、权利义务和责任范围的实质差异。\n\n适合：销售、商务、采购、法务协作人员和合同管理员。", "description_en": "After several redlines, a one-word change to money, conditions, or liability can disappear inside dozens of pages.\n\nThis case compares two contracts side by side, filters formatting noise, and highlights substantive changes to amounts, dates, conditions, rights, obligations, and exposure.\n\nBest for: sales, commercial, procurement, legal collaboration, and contract operations teams.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "版本变更对比页" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请对比 {旧版合同文件} 与 {新版合同文件}，制作两版并排的合同变更对比页。完整读取两份文件，以原文为唯一差异依据；过滤空格、换行、样式和语义完全等价的变化，但金额、比例、数量、日期、期限、主体、否定词、条件词、权利义务和责任范围哪怕只改一个字也必须保留。按高、中、低影响排序，每项展示条款位置、旧版、新版、变化类型、对我方影响和建议动作，并汇总新增、删除和修改数量。最终只输出一个可离线打开的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Compare {old contract file} and {new contract file} in a side-by-side contract change page. Read both files fully and use their text as the sole basis. Filter spaces, line breaks, formatting, and semantically identical edits, but retain every change to amounts, percentages, quantities, dates, periods, parties, negations, conditions, rights, obligations, and liability scope. Rank changes by high, medium, and low impact. For each item show clause location, old text, new text, change type, impact, and recommended action, plus totals for additions, deletions, and modifications. Return only one responsive offline HTML file with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "合同对比", "版本差异", "实质变更", "签约" ], "tags_en": [ "Contract Comparison", "Version Diff", "Substantive Change", "Signing" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-visit-brief", "title": "拜访前 60 秒速查卡", "title_en": "60-Second Visit Brief", "subtitle": "出门前一眼看完客户现状、承诺与风险", "subtitle_en": "Scan account status, commitments, and risks before the meeting", "description": "拜访前时间越紧，越容易在 CRM 里翻半天仍抓不到重点。\n\n这个案例把最近沟通、双方承诺、商机阶段、关键角色和本次唯一目标压缩成一张 60 秒速查卡。\n\n适合：外勤销售、客户经理、销售主管、售前顾问和临时接手商机的同事。", "description_en": "When meeting prep time is short, teams can spend too long searching CRM without finding what matters.\n\nThis case compresses recent conversations, mutual commitments, deal stage, key stakeholders, and one meeting objective into a 60-second brief.\n\nBest for: field sales, account managers, sales leaders, presales, and colleagues taking over a deal.", "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "拜访前 60 秒速查卡" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请为 {客户名称或商机名称} 制作一张拜访前 60 秒速查卡。先用用户提供的名称在 CRM 搜索并确认对应客户或商机；若存在同名客户、多主体或多个匹配商机，只列出最少必要候选让用户确认，不要求用户提供 CRM 内部 ID。读取最近有效的客户、商机、联系人、跟进记录和双方承诺，优先近 90 天信息；区分 CRM 事实、用户判断、公开信息和 AI 假设，不把过期记录当成当前状态。只保留拜访现场真正需要的信息：当前阶段、上次谈到哪、双方未兑现承诺、关键参与人、本次唯一目标、三条必问问题和一个风险提醒。若客户主体或时间冲突，显式列出待确认。最终只输出一个移动端也能快速阅读的响应式 HTML，所有资源内联，不使用外链、emoji 或 hash 锚点。", "prompt_en": "Create a 60-second visit brief for {customer name or opportunity name}. First search CRM using the name supplied by the user and confirm the matching account or opportunity. If there are duplicate names, multiple legal entities, or several matching opportunities, show only the minimum necessary candidates for the user to choose from; never require an internal CRM ID. Read the latest valid account, opportunity, contact, activity, and mutual commitment records, prioritizing the last 90 days. Separate CRM facts, user judgments, public information, and AI hypotheses; do not treat stale records as current. Keep only what is needed in the meeting: current stage, last discussion, outstanding commitments, key attendees, one meeting objective, three essential questions, and one risk reminder. Explicitly list identity or timing conflicts as pending confirmation. Return only one responsive HTML file optimized for quick mobile reading, with all resources inline and no external links, emoji, or hash anchors.", "skills": [], "experts": [], "mcps": [], "categories": [ "business" ], "tags": [ "拜访准备", "快速回顾", "客户简报", "销售行动" ], "tags_en": [ "Meeting Prep", "Quick Review", "Account Brief", "Sales Action" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-26T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "interactive-annual-report", "title": "交互式年度数据报告", "title_en": "Interactive Annual Data Report", "subtitle": "滚动叙事 + 数据可视化，让年报讲故事", "subtitle_en": "Scroll-driven storytelling with data visualization", "description": "每年写年报都头疼？传统 PDF 年报枯燥无味，管理层看完就忘。这个交互式年度报告模板用滚动叙事驱动数据展示，将冰冷的数据转化为引人入胜的视觉故事。内联 SVG 图表、数字递增动画、时间线渐入效果，让你的年度总结像产品 Demo 一样有吸引力。适合产品负责人、数据分析师、市场团队用于年度汇报、投资人展示或内部复盘。", "description_en": "Tired of boring PDF annual reports that nobody reads? This interactive annual report template uses scroll-driven storytelling to transform raw data into compelling visual narratives. Featuring inline SVG charts, animated number counters, timeline fade-ins, and hover interactions — your annual summary becomes as engaging as a product demo. Perfect for product managers, data analysts, and marketing teams presenting to executives, investors, or internal stakeholders.", "categories": [ "data", "business" ], "tags": [ "年度报告", "数据可视化", "滚动叙事", "SVG图表", "管理层汇报" ], "tags_en": [ "Annual Report", "Data Visualization", "Scroll Storytelling", "SVG Charts", "Executive Report" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "年度数据报告" } ], "prompt": "请生成一个交互式年度数据报告单页（单文件 HTML，内联无外部依赖），风格参考 PostHog / Sentry 的暗色数据仪表盘质感，带滚动动效。\n\n结构：\n1. Hero 封面：「2025 Annual Report」大标题 + 核心成就副标题\n2. 关键指标：用户数、API 调用量、正常运行率、NPS 四个数据卡片（滚动进入时数字递增）\n3. 用户增长曲线：12 个月折线图\n4. 产品里程碑：Q1-Q4 垂直时间线，滚动逐个淡入\n5. 功能使用分布：Top 5 功能占比环形图\n6. 团队亮点：工程/产品/设计/运营各一张成就卡片\n7. 展望 2026：3 个战略方向\n\n图表用内联 SVG，滚动动效自行把握。", "prompt_en": "Please generate an interactive annual data report single page (single-file HTML, inline with no external dependencies), with a style referencing the dark data-dashboard aesthetic of PostHog / Sentry, with scroll animations.\n\nStructure:\n1. Hero cover: \"2025 Annual Report\" large title + core-achievement subtitle\n2. Key metrics: four data cards for user count, API call volume, uptime rate, and NPS (numbers count up when scrolled into view)\n3. User growth curve: 12-month line chart\n4. Product milestones: a vertical Q1-Q4 timeline, fading in one by one on scroll\n5. Feature usage distribution: Top 5 feature share donut chart\n6. Team highlights: one achievement card each for Engineering / Product / Design / Operations\n7. Outlook 2026: 3 strategic directions\n\nUse inline SVG for charts; handle the scroll animations at your discretion.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 1, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-05-03T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 6 }, { "id": "interactive-api-reference", "title": "交互式 API 接口文档", "title_en": "Interactive API Reference Portal", "subtitle": "Stripe 风格三栏 API 文档，实时代码切换", "subtitle_en": "Stripe-style three-column API docs with live code switching", "description": "API 文档写了一堆但开发者还是看不懂、找不到、不想读？参考 Stripe 的文档体验标杆，生成一个三栏布局的交互式 API 文档门户——左栏导航定位、中间参数说明、右栏实时代码示例（支持 cURL/Python/JS/Go 一键切换）。语法高亮、复制按钮、滚动定位全部内置，开箱即用。适合 API 产品经理、后端开发者、技术文档工程师。", "description_en": "API docs that developers can't understand, can't find, or won't read? Following Stripe's documentation gold standard, this generates a three-column interactive API reference portal — left nav for navigation, center for parameter docs, right panel for live code examples with cURL/Python/JS/Go switching. Syntax highlighting, copy buttons, and scroll tracking built in. Perfect for API product managers, backend developers, and technical writers.", "categories": [ "dev", "office" ], "tags": [ "API 文档", "开发者工具", "代码示例", "Stripe 风格", "技术文档" ], "tags_en": [ "API Documentation", "Developer Tools", "Code Examples", "Stripe Style", "Technical Docs" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "API 接口文档门户" } ], "cover_image": "cover.png", "cover_mode": "landscape", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "prompt": "请生成一个 Stripe 风格的三栏 API 文档门户单页（单文件 HTML，内联无外部依赖）。\n\nAPI：WorkBuddy Agent API v2.0，三栏布局——左栏导航 + 中栏文档 + 右栏代码面板。\n\n左栏导航：概览 / 认证 / Agent 管理（列出/创建/获取/更新/删除）/ 任务管理（创建/列表）/ Webhook，点击平滑滚动。\n中栏文档：每个 endpoint 显示 HTTP 方法徽章 + 路径 + 中文描述 + 参数表（名称/类型/必填/描述）。\n右栏代码：cURL / Python / JavaScript / Go 标签切换（全局生效）+ 语法高亮 + 一键复制。\n交互：导航滚动高亮、移动端单栏。", "prompt_en": "Please generate a Stripe-style three-column API documentation portal as a single page (single-file HTML, inlined with no external dependencies).\n\nAPI: WorkBuddy Agent API v2.0, with a three-column layout—left navigation + center documentation + right code panel.\n\nLeft navigation: Overview / Authentication / Agent Management (List/Create/Get/Update/Delete) / Task Management (Create/List) / Webhook, with smooth scrolling on click.\nCenter documentation: each endpoint shows an HTTP method badge + path + Chinese description + a parameters table (name/type/required/description).\nRight code panel: cURL / Python / JavaScript / Go tab switching (applies globally) + syntax highlighting + one-click copy.\nInteractions: nav highlights on scroll, single-column on mobile.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "badge": "official", "avatar": "" }, "source": "official", "quality_score": 93, "created_at": "2026-05-03T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 100, "featured_order": 0, "usage_count": 0, "is_featured": false, "mcps": [], "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "interactive-data-story", "title": "全球远程工作趋势数据故事", "title_en": "Global Remote Work Trends Data Story", "subtitle": "滚动驱动的交互式数据叙事长页", "subtitle_en": "Scroll-driven interactive data storytelling long page", "description": "如何让数据讲故事？本案例展示一个滚动驱动的交互式数据叙事页面，围绕「全球远程工作趋势 2020-2025」主题，通过 5 个动画章节（计数器、时间线、国家排名、影响统计、趋势预测）将干燥的数据变成引人入胜的视觉体验。适合数据记者、内容策划和产品运营。", "description_en": "How to let data tell stories? This case demonstrates a scroll-driven interactive data storytelling page about 'Global Remote Work Trends 2020-2025' with 5 animated sections transforming dry data into compelling visual narratives.", "categories": [ "data", "business" ], "tags": [ "数据故事", "滚动叙事", "远程工作", "交互动画" ], "tags_en": [ "Data Story", "Scrollytelling", "Remote Work", "Interactive Animation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "远程工作数据故事页面" } ], "prompt": "你是一位数据叙事和交互设计专家。请创建一个关于「全球远程工作趋势 2020-2025」的滚动驱动数据故事 HTML 页面。\n\n要求：\n1. 全页滚动叙事结构，5 个章节随滚动触发动画：\n - Hero：大标题 + 动画计数器（5% → 28% 远程工作者比例）\n - 时间线：2020 疫情→2021 混合→2022 回流→2023 稳定→2024 AI加速→2025 新常态\n - 国家排名：Top 10 国家远程工作率横向柱状图\n - 影响板块：薪资/生产力/幸福感统计 + 动画进度环\n - 未来预测：趋势折线图\n2. 使用 IntersectionObserver 触发滚动动画\n3. 平滑视差效果\n4. 配色：暖渐变（琥珀/珊瑚色）+ 深色背景\n5. 中文内容为主，关键数据用英文\n6. 响应式设计\n\n输出：单个自包含 HTML 文件，打开即可滚动体验。", "prompt_en": "You are an expert in data storytelling and interaction design. Please create a scroll-driven data story HTML page about \"Global Remote Work Trends 2020-2025.\"\n\nRequirements:\n1. A full-page scroll narrative structure, with 5 chapters whose animations are triggered as you scroll:\n - Hero: a large headline + an animated counter (5% → 28% remote worker percentage)\n - Timeline: 2020 pandemic → 2021 hybrid → 2022 return-to-office → 2023 stabilization → 2024 AI acceleration → 2025 new normal\n - Country Ranking: a horizontal bar chart of the Top 10 countries by remote work rate\n - Impact Section: salary / productivity / well-being statistics + animated progress rings\n - Future Forecast: a trend line chart\n2. Use IntersectionObserver to trigger the scroll animations\n3. Smooth parallax effects\n4. Color scheme: warm gradients (amber/coral) + dark background\n5. Primarily Chinese content, with key data in English\n6. Responsive design\n\nOutput: a single self-contained HTML file that can be scrolled through as soon as it's opened.", "skills": [ { "id": "impeccable", "name": "Impeccable（前端设计工具集）", "description": "高品质 UI/UX 设计工具集", "skill_id": "skill_2053082862415904768" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": true, "featured_order": 5, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": true, "recommended_order": 13 }, { "id": "interactive-kanban-board", "title": "交互式看板任务面板", "title_en": "Interactive Kanban Task Board", "subtitle": "可拖拽任务卡片，项目进度一眼掌控", "subtitle_en": "Drag-and-drop task cards for instant project progress tracking", "description": "跨工具管理任务总是一团乱麻——Jira 里建了 ticket，飞书里又记了待办，便签纸上还贴着优先级。信息分散导致进度不透明，团队协作频频掉链子。\n\n这个案例用 frontend-dev 技能生成了一个 DevStudio 团队的交互式看板面板：5 列泳道（Backlog / To Do / In Progress / Review / Done），预置 15-20 张任务卡片，每张卡片包含标题、描述摘要、优先级徽章（P0 红 / P1 橙 / P2 蓝 / P3 灰）、负责人头像圆圈（首字母缩写）、标签（frontend / backend / design / bug / feature）及截止日期。支持拖拽排序与跨列移动，顶部筛选栏可按负责人、优先级过滤并支持标题搜索，列内任务计数随操作自动更新。圆角卡片 + 阴影层次 + 平滑动画，现代感十足。\n\n适合：项目经理、研发团队、自由职业者——任何需要可视化管理任务流的人。", "description_en": "Juggling tasks across multiple tools is a recipe for chaos — tickets in Jira, to-dos in chat apps, sticky notes with priority scribbles. Scattered information kills transparency and derails team collaboration.\n\nThis case uses the frontend-dev skill to generate an interactive Kanban board for the DevStudio team: 5 swim-lane columns (Backlog / To Do / In Progress / Review / Done), pre-populated with 15-20 task cards. Each card shows a title, description snippet, priority badge (P0-red / P1-orange / P2-blue / P3-gray), assignee avatar circle with initials, tags (frontend / backend / design / bug / feature), and due date. Supports drag-and-drop reordering and cross-column moves. A top filter bar lets you filter by assignee, priority, or search by title, with column task counts auto-updating on every action. Rounded cards, layered shadows, and smooth animations deliver a modern, polished UI.\n\nIdeal for: project managers, dev teams, freelancers — anyone who needs a visual task-flow board.", "categories": [ "office", "creation" ], "tags": [ "看板", "任务管理", "拖拽", "项目管理", "Kanban" ], "tags_en": [ "Kanban", "Task Management", "Drag & Drop", "Project Management" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为虚构软件团队「开发工坊」生成一个可交互看板任务面板（单文件 HTML，内联无外部依赖），界面文字用简体中文（人名用中文，优先级 P0-P3 保留代号），现代风格。\n\n看板 5 列：待办池 / 待处理 / 进行中 / 评审中 / 已完成，预置 15-20 张任务卡片。每张卡片含：任务标题、描述、优先级徽章（P0-P3）、负责人头像（首字母）、分类标签（前端/后端/设计/缺陷/需求）、截止日期。\n\n支持拖拽排序与跨列移动；顶部筛选栏按负责人/优先级筛选 + 标题搜索；各列计数随操作自动更新。", "prompt_en": "Please generate an interactive Kanban task board for a fictional software team called \"开发工坊\" (single-file HTML, inline with no external dependencies). The interface text should be in Simplified Chinese (use Chinese names; keep the priority codes P0-P3 as-is), in a modern style.\n\nThe Kanban has 5 columns: To-Do Pool / Pending / In Progress / In Review / Done, pre-populated with 15-20 task cards. Each card includes: task title, description, a priority badge (P0-P3), an assignee avatar (initials), a category label (Frontend/Backend/Design/Bug/Requirement), and a due date.\n\nSupport drag-and-drop reordering and moving cards across columns; a top filter bar to filter by assignee/priority plus title search; and each column's count should update automatically as you make changes.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-15T18:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "交互式看板" } ], "is_recommended": false, "recommended_order": 0 }, { "categories": [ "business" ], "skills": [], "experts": [ { "id": "tanyuan-heritage-article", "name": "科普推文专家", "description": "为博物馆公众号写文物科普图文，正文+封面+配图，排版复用 doocs/md，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "076b70b41214", "expert_id": "ex_aPGVu1473LOP" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:05:37+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "id": "international-museum-day-article", "title": "国际博物馆日主题推文", "title_en": "International Museum Day WeChat Article", "subtitle": "围绕博物馆教育「三阶段」，写一篇可发草稿箱的科普图文", "subtitle_en": "Three-stage museum-education science article ready for the WeChat draft box", "description": "5·18 写一篇博物馆日推文？最容易踩的坑是把活动堆成 list，结果读者划三屏就走。这个案例不谈展品数字、不报活动数量，而是抽出博物馆教育的「参观前/中/后」三段结构，把武进博物馆、景德镇十大瓷厂博物馆的真实数据做支撑，把公众号读者最关心的「实用节奏」讲清楚。\n\n演示内容：基于探元知识库真实数据（武进博物馆年均活动、景德镇博物馆教育活动三阶段研究），写一篇可直接推到草稿箱的 doocs/md 风格图文，标题带钩子、章节有节奏、结尾有行动引导。\n\n适合：博物馆公众号编辑、社教部门、青铜器/陶瓷类专项馆宣教人员，以及任何在 5·18、文化和自然遗产日等节点需要科普图文的小馆人。", "description_en": "Need a WeChat article for International Museum Day? The biggest trap is turning it into a list of events the reader will scroll past. This case skips the slogans and pulls out the \"before / during / after\" three-stage framework that museum educators actually use, grounding every section in real figures from the TanYuan knowledge base.\n\nDemo: a doocs/md-styled article built on real TanYuan data (Wujin Museum's annual programming, Jingdezhen's three-stage education research), with a hooky headline, a tight section rhythm, and an action-oriented closer that's ready for the WeChat draft box.\n\nBest for: museum WeChat editors, social-education teams, docent-led outreach staff at bronze / ceramic museums — and anyone at a small museum who needs a science article for 5·18 or Cultural Heritage Day.", "tags": [ "国际博物馆日", "博物馆教育", "公众号推文", "社教", "公共教育" ], "tags_en": [ "International Museum Day", "Museum Education", "WeChat Article", "Public Education", "Community Engagement" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "doocs/md 风格公众号图文 HTML" } ], "prompt": "请基于以下要点，撰写一篇 5·18 国际博物馆日主题的博物馆公众号科普图文。\n\n写作要点：\n1. 观点先行：开头第一句就让读者理解博物馆教育远不止展厅里那点事——把它拆成「参观前/中/后」三段来谈。\n2. 用真实数据支撑：所有数据必须调用 mcp__tanyuan-wb-mcp-server__search_movable_relics 或 search_relics 获取真实知识库内容，逐处标源。常见支撑：\n - 武进博物馆：年均接待 30 万人次、年均临展 20 场、年均讲座与活动 100 余场次、藏品 3830 件（含珍贵 621 件）；\n - 景德镇非国有陶瓷博物馆研究：三阶段（参观前/中/后）的核心是导览解说 + 活动体验 + 拓展服务。\n3. 分节科普：用 3–5 个小节，每节一个具体知识点；专有名词首次出现给通俗解释（如武周时期要说清 690–705 年）。\n4. 结尾有行动引导：欢迎读者在评论区留言推荐本地馆，承诺节后整理读者推荐榜。\n5. 风格：doocs/md 公众号图文——宽 677px 居中、米白底、章节标题加粗+彩条、文末有引导卡片；克制，不堆 emoji。\n6. 配图：优先用 search_relics 返回的真实博物馆或馆藏图，base64 内联到 HTML；未授权图片不得使用。\n7. 字数：1300–1600 字正文。\n8. 逐处标源：每个事实声明标「来自探元知识库《...》词条」，文末汇总参考资料。\n\n约束：输出单文件 HTML，CSS 全内联，图片用 data:image/...;base64,... 嵌入，禁止任何外链。封面图单独导出 cover.png。", "prompt_en": "Write a museum WeChat science article for International Museum Day (May 18). The expert persona is the TanYuan Heritage Article Expert — a heritage science writer.\n\nWriting requirements:\n1. Lead with the takeaway: open by reframing \"museum education\" as the before / during / after three-stage framework.\n2. Ground every claim in real data: call mcp__tanyuan-wb-mcp-server__search_movable_relics or search_relics and cite each fact inline. Useful anchors from TanYuan KB:\n - Wujin Museum: 300k annual visitors, 20 temporary exhibitions/year, ~100 lectures & activities/year, 3,830 holdings (621 significant).\n - Jingdezhen small-museum study: core of the three-stage model is docent talks + activities + outreach.\n3. Sectioned science writing: 3–5 sections, one takeaway each; explain specialist terms on first mention (e.g., Wu Zhou period 690–705 CE).\n4. CTA close: invite readers to recommend local museums in the comments; promise a readers-picks follow-up.\n5. Visual style: doocs/md WeChat article — 677px column, off-white background, bold colored-bullet section headings, soft CTA card; restrained, no emoji soup.\n6. Imagery: prefer search_relics museum / collection images; embed base64 in HTML; never use unauthorized photos.\n7. Length: 1300–1600 Chinese characters of body copy.\n8. Cite every fact: each claim labels TanYuan KB · entry《...》; collected references at the end.\n\nOutput: a single HTML file, all CSS inlined, all images as data:image/...;base64,..., no external links. Export a separate cover.png.", "is_recommended": false, "recommended_order": 0 }, { "id": "intraday-1430-strength-check", "title": "14:30 盯盘强度检查", "title_en": "14:30 Intraday Strength Check", "subtitle": "把持仓与观察池在尾盘前一小时重新排一遍强弱顺序", "subtitle_en": "Re-rank holdings and watch names one hour before the close", "description": "盘中最容易乱的时段，往往不是开盘，而是 14:30 这种“觉得该做点什么”的节点。这个案例按 A 股短线交易的盘中评估流程，给出一份两只持仓、两只观察股的配置，再用真实分时、行情、资金和筹码数据跑官方强度脚本，判断当下应该抱强、等回踩还是继续观察。\n\n页面把时间流、组合概览、逐只强度拆解和关键价位放在一个单文件双语 HTML 里，右上角可以切换中文和英文。它演示的不是喊单，而是把盘中决策从情绪动作变成可复核的结构动作。\n\n适合：想做固定时间点盯盘复盘的人；有持仓和观察池、但不知道 14:30 该看什么的人；希望把盘中提示做成标准流程的 A 股交易者。", "description_en": "The intraday window that creates the most confusion is often not the open, but a checkpoint like 14:30 when it feels like something should be done. This case follows the intraday evaluation workflow of A股短线交易, starts from a configuration with two holdings and two watch-only names, then runs the official strength script on real intraday, quote, flow and chip data to decide whether the correct posture is to keep holding strength, wait for pullbacks, or simply continue watching.\n\nThe page combines a market timeline, portfolio snapshot, per-stock factor breakdown and key levels inside a single bilingual HTML file with a top-right language switch. It demonstrates not stock calling, but how to turn intraday decisions from emotional reactions into reviewable structural actions.\n\nSuitable for: traders who want fixed-time intraday checkups; anyone with holdings and a watchlist but no clear 14:30 process; A-share participants who want to standardize intraday prompts.", "categories": [ "finance", "data" ], "tags": [ "盘中评估", "14:30 节点", "强度评分", "盯盘快报" ], "tags_en": [ "intraday review", "14:30 checkpoint", "strength scoring", "watch feed" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请按 A股短线交易 的盘中节点评估流程，基于以下等价 trading_config 在 2026-07-27 14:30 做强度检查：账户总仓上限 80%，单票上限 30%；持仓为北化股份 002246（成本 19.42）和四川黄金 001337（成本 38.15）；观察池为睿创微纳 688002 与高德红外 002414。请拉取这些标的与大盘的真实分时、行情、资金流与筹码数据，并调用官方 scripts/intraday_strength.py 输出强度评分。最终交付一份单文件、资源全内联的 HTML 页面。页面至少包含市场状态、时间流、持仓列表、观察列表、强度拆解、关键价位和当前节点关注。", "prompt_en": "Please follow the intraday checkpoint workflow of A股短线交易 and run a 14:30 strength check on 2026-07-27 with the following equivalent trading_config: max total exposure 80%, single-name cap 30%; holdings are NCIC 002246 (cost 19.42) and Sichuan Gold 001337 (cost 38.15); the watchlist contains Raytron 688002 and Guide Infrared 002414. Pull the real intraday, quote, flow and chip data for those names plus the market, then run the official scripts/intraday_strength.py to produce strength scores. Deliver the result as a single-file HTML page with all resources embedded. The page should at least include market state, a timeline, holdings, watch names, factor breakdowns, key levels and the current checkpoint takeaway.", "skills": [ { "id": "ashare-short-term-trading", "name": "A股短线交易", "description": "A股短线交易体系，支持交割单复盘分析、策略筛选选股、盘中关键时间点推送三大场景。", "skill_id": "skill_2076627819394322432" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "artifacts": [ { "file": "output.html", "label": "双语盘中评估页面" } ], "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:14:00+08:00", "updated_at": "2026-07-27T20:16:07+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "investment-project-feasibility-study-smart-parking", "title": "智慧停车PPP项目可研报告", "title_en": "Smart Parking PPP Project Feasibility Study", "subtitle": "基于2023版政府投资可研大纲，输出完整目录与核心章节初稿", "subtitle_en": "Complete TOC and core chapter drafts following the 2023 government investment feasibility study outline", "description": "政府投资项目可研报告最难的就是从零搭框架。这个案例演示了如何把一份仅300字的项目任务书（智慧停车PPP，专项债+社会资本，3.8亿元），按2023版政府投资可研大纲11章体系，先输出完整目录，再撰写概述/需求分析/建设方案三章初稿，投资估算按量级拆解到科目，缺失参数逐项列明假设清单不瞎编。\n\n适合：政府投资项目前期工作人员、咨询工程师、PPP项目经理、可研报告编制新手、需要快速出初稿的投资决策者。", "description_en": "The hardest part of a government investment feasibility study is building the framework from scratch. This case shows how to take a 300-word project brief (smart parking PPP, special bonds + private capital, RMB 380M) and produce a complete 11-chapter outline per the 2023 government investment outline, with full drafts of chapters 1-3 and an investment estimate broken down line by line, with every missing parameter documented in an assumptions list.\n\nTarget audience: Government project planners, consulting engineers, PPP project managers, feasibility study beginners, investment decision-makers who need a rapid first draft.", "categories": [ "finance", "business", "productivity" ], "tags": [ "可研报告", "PPP项目", "智慧停车", "政府投资", "2023版大纲" ], "tags_en": [ "feasibility study", "PPP project", "smart parking", "government investment", "2023 outline" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "可研报告HTML全文" } ], "prompt": "请阅读项目任务书，这是政府投资（专项债+社会资本）类智慧停车PPP项目。请按2023版政府投资可研大纲11章体系：\n\n1. 先输出完整目录（11章）；\n2. 撰写第1章概述（含项目概况、单位信息、编制依据、结论与建议）；\n3. 撰写第2章建设背景与必要性（含立项背景、规划符合性、必要性论证）；\n4. 撰写第3章需求分析与产出方案（含需求预测、建设内容与规模、产出方案）；\n5. 给出投资估算量级表（按工程费用/其他费用/预备费/融资费用/流动资金分解到底层科目）；\n6. 列出完整假设清单（缺失参数不编造，标注置信度）。\n\n项目参数：路内泊位改造8000个、P+R停车场2座、统一收费与诱导平台，投资约3.8亿元，合作期限15年（建设2年+运营13年）。\n\n输出格式为单文件自包含HTML，含封面、目录、正文表格，无需外部资源引用。", "prompt_en": "Read the project brief for this government-invested (special bonds + private capital) smart parking PPP project. Following the 2023 government investment feasibility study outline (11 chapters):\n\n1. Output the complete table of contents (11 chapters);\n2. Draft Chapter 1 Overview (project profile, entity info, basis, conclusions);\n3. Draft Chapter 2 Background & Necessity (background, policy compliance, necessity analysis);\n4. Draft Chapter 3 Demand Analysis & Output Plan (demand forecast, scope, output plan);\n5. Provide an investment estimate table broken down to line-item level (engineering costs, other costs, contingencies, financing costs, working capital);\n6. List all assumptions with confidence levels — do not fabricate unmarked quantities.\n\nProject parameters: 8,000 on-street parking spaces renovated, 2 P+R parking lots, unified payment & guidance platform, total investment ~RMB 380M, 15-year cooperation period (2-year construction + 13-year operation).\n\nOutput as a single self-contained HTML file with cover, table of contents, tables — no external resource references.", "skills": [ { "id": "investment-project-feasibility-study", "name": "投资项目可行性研究报告编写", "description": "编制投资项目可行性研究报告的统一入口，依据资金来源自动区分政府投资与企业投资两类场景，按2023版大纲组织报告结构", "skill_id": "skill_2085354935208562688" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-04T11:37:43+08:00", "updated_at": "2026-08-04T11:37:43+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "invoice-expert-team-demo", "title": "智能发票专家团：三票联查实战", "title_en": "Invoice Expert Team: Three-Invoice Verification", "subtitle": "多格式发票自动识别、真伪查验、商业信用核查，一站式归档报告", "subtitle_en": "Multi-format invoice auto-recognition, authenticity verification, counterparty risk analysis, and one-stop archiving report", "description": "财务最头疼的是收到一堆不同格式的发票——XML数电票、OFD火车票、PDF海外发票，每种都要走不同的查验流程。这个案例演示了智能发票专家团如何一次处理三张不同格式的发票：增值税XML数电票自动完成税局验真并发现开票方存在重大税收违法风险标记，OFD火车票自动判断旅客运输抵扣资格，海外PDF发票识别后归档备案。最终输出结构化的多票联查报告，包含验真结论、风险评级和操作建议。\n\n适合：企业财务人员、税务合规主管、采购供应链风控经理、审计团队、中小企业主。", "description_en": "The biggest headache for finance teams is receiving invoices in different formats — XML digital invoices, OFD train tickets, PDF overseas invoices — each requiring a different verification process. This case demonstrates how the Intelligent Invoice Expert Team processes three invoices of different formats in one go: a VAT XML digital invoice undergoes automatic tax authority verification revealing red-flag risk for the counterparty, an OFD train ticket automatically qualifies for passenger transport VAT deduction, and an overseas PDF commercial invoice is identified and archived. The final output is a structured multi-invoice verification report with authenticity conclusions, risk ratings, and actionable recommendations.\n\nPerfect for: corporate finance professionals, tax compliance officers, procurement supply chain risk managers, audit teams, SME owners.", "categories": [ "business" ], "tags": [ "发票查验", "税务合规", "风险管控", "智能票据", "增值税" ], "tags_en": [ "Invoice Verification", "Tax Compliance", "Risk Control", "Smart Invoices", "VAT" ], "artifact_type": "md", "cover_image": "cover.png", "preview": "output.md", "prompt": "请帮我处理以下三张发票文件，完成发票识别、真伪查验和开票方信用风险核查：\n\n1. 上传一张增值税专用发票XML文件（数电票）\n2. 上传一张火车票OFD文件（铁路电子客票）\n3. 上传一张海外商业发票PDF文件\n\n要求：\n- 增值税专用发票需要完成税局在线验真\n- 火车票需要判断旅客运输服务抵扣资格（含可抵扣税额计算）\n- 海外发票识别后仅归档备案，标注不可用于国内增值税抵扣\n- 所有增值税发票的开票方需要查询商业信用风险（灰名单、欠税、重大违法）\n- 最终输出结构化归档报告，包含验真结论、风险评级（🟢🟡🔴）和操作建议\n- 如遇接口不可用，请标注降级路径和人工查验指引", "prompt_en": "Please process the following three invoice files, completing invoice recognition, authenticity verification, and issuer credit-risk checks:\n\n1. Upload a VAT special invoice XML file (fully digital e-invoice)\n2. Upload a train ticket OFD file (railway electronic ticket)\n3. Upload an overseas commercial invoice PDF file\n\nRequirements:\n- The VAT special invoice needs online verification with the tax authority\n- For the train ticket, determine eligibility for passenger-transport-service deduction (including calculation of the deductible tax amount)\n- After recognizing the overseas invoice, only archive it for the record and note that it cannot be used for domestic VAT deduction\n- For all VAT invoices, query the issuer's commercial credit risk (grey list, tax arrears, major violations)\n- Finally output a structured archival report including verification conclusions, risk ratings (🟢🟡🔴), and recommended actions\n- If an API is unavailable, note the fallback path and manual-verification guidance", "skills": [], "experts": [ { "id": "InvoiceVerifyWorkbuddy", "name": "智能发票专家团", "description": "智能发票识别、税局验真、商业信用核查和智能票夹归档——5人专家团覆盖分拣→识别→验真→风险→归档全链路", "expert_id": "ex_AdCt8WgnWOnJ" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "source": "official", "usage_count": 0, "created_at": "2026-06-04T12:32:29Z", "updated_at": "2026-06-16T18:10:00Z", "cover_mode": "landscape", "artifacts": [ { "file": "output.md", "label": "智能发票专家团：三票联查实战", "label_en": "Invoice Expert Team: Three-Invoice Verification" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "invoice-ledger", "title": "发票一堆理不清？发票管家自动出台账", "title_en": "Invoice Ledger", "subtitle": "进销项自动归集台账，价税统计与抵扣风险提示", "subtitle_en": "Invoices auto-organized into a monthly ledger, with tax stats and deduction-risk alerts", "description": "上传或接入发票，自动识别归集成月度台账，统计价税、可抵扣税额，提示待认证和抵扣风险。", "description_en": "Upload or connect invoices: they are recognized and consolidated into a monthly ledger with price-tax statistics and alerts for pending certification and deduction risks.", "categories": [ "productivity" ], "tags": [ "发票管理", "台账", "内勤", "财税" ], "tags_en": [ "Invoice Management", "Ledger", "Back Office", "Finance & Tax" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "发票管家台账（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易内勤专家。请帮我把这个月的发票整理成台账。\n\n【数据来源】\n- 我上传了发票图片/PDF：用【OCR识别】批量识别关键字段；\n- 我已接入开票/财税系统：直接读取{时间范围，如 2026年7月}进项发票。\n\n【要求】\n1. 识别并归集每张发票：开票方、发票号码、开票日期、金额、税额、认证状态；\n2. 顶部汇总KPI：发票张数、价税合计、可抵扣税额、待认证张数；\n3. 智能提示：临近认证截止的发票、不同税率（货物13%/运费9%）的抵扣科目区分、与采购单核对建议；\n4. 输出为一张清爽的发票台账表。\n\n识别不清的字段标「待人工确认」，不要臆造。请生成完整发票台账。", "skills": [ { "id": "invoice-manager", "name": "发票管家", "description": "TradeBuddy专门管发票这摊事：进项、销项发票拍照或导入，自动认票、抽要素、理成表、归档，还能把发票和对应的订单/货物对一对，看看票货对不对得上。", "skill_id": "skill_2091156036898680832" } ], "experts": [ { "id": "neiqin-expert", "name": "B2B贸易内勤专家", "description": "懂B2B贸易的内勤专家＝AI版录单员+档案室：识单、发票管理、智能对账、出表出单、盯订单进度。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Hu9b9KuHQwa" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ip-term-review-html-expert-team", "title": "知识产权条款双视角审查（HTML交付版）", "title_en": "Dual-Perspective IP Clause Review (HTML)", "subtitle": "粘贴一段 IP 条款，自动拆成甲方/乙方两个对立视角的风险清单 + 可直接替换的修改建议，单文件 HTML 交付", "subtitle_en": "Paste an IP clause and get a Party-A-vs-Party-B risk breakdown with ready-to-swap revisions, delivered as a single-file HTML report", "description": "签软件开发、委托创作、技术许可这类合同时最头疼的，就是那几条知识产权条款——钱给出去了，成果到底归谁？授权范围到底多宽？对方不担侵权怎么办？这些坑往往藏在几句看起来「标准」的条款里。\n\n这个案例演示了如何用「知识产权专家团」，把一段软件开发委托合同的 IP 条款交给 AI，自动输出一份双方视角对抗式审查报告：先识别条款类型（成果归属/授权许可/侵权担保/责任），再分别站在甲方（委托方）和乙方（受托方）两个对立立场拆解风险，每条风险标注高/中/低等级与法律依据，最后给出可直接替换的修改建议和权利链风险提示。报告是单文件 HTML，左右双栏对照、红黄绿风险着色，内联样式、可离线打开。\n\n适合：合同审查律师、公司法务、创业者在签署技术合同前做 IP 条款风险自检，以及想快速建立「双视角对抗审查」思路的律师新人。", "description_en": "The most headache-inducing part of signing software development, commissioned creation, or licensing contracts is the IP clause — once money changes hands, who owns the deliverables? How broad is the license? Who bears infringement risk? These pitfalls hide inside seemingly standard wording.\n\nThis case shows how the IP Expert Team turns an IP clause from a software development commission contract into an adversarial dual-perspective review report: it first identifies the clause type (ownership / licensing / infringement warranty / liability), then dissects risks from the opposing standpoints of Party A (the commissioning party) and Party B (the contractor), tagging each risk with high/medium/low severity and its legal basis, and finally delivers ready-to-swap revision suggestions plus a rights-chain risk note. The report is a single-file HTML with side-by-side columns and red/yellow/green risk coloring, inline styles, and offline readability.\n\nFor: contract review lawyers, in-house counsel, and founders doing IP-clause self-checks before signing tech contracts, plus junior lawyers building the dual-perspective review mindset.", "categories": [ "business", "productivity" ], "tags": [ "知识产权条款", "合同审查", "双视角分析", "软件开发委托", "HTML报告" ], "tags_en": [ "IP clause", "contract review", "dual-perspective", "software commission", "HTML report" ], "artifact_type": "html", "preview": "ip-review.html", "artifacts": [ { "file": "ip-review.html", "label": "知识产权条款双视角审查报告（HTML交付版，甲方/乙方对立视角 + 风险分级 + 修改建议）" } ], "prompt": "请对下面这段软件开发委托合同中的知识产权条款做专项审查，产出一份单文件 HTML 审查报告（交付版，内联样式、无外部资源、可离线打开）。\n\n合同背景：甲方（委托方）委托乙方（受托方）开发一套企业级 SaaS 管理系统，甲方付款。\n审查方：双方视角（分别输出甲方/乙方立场）。\n\n条款原文：\n「1. 乙方在履行本合同过程中产生的所有成果，其知识产权归甲方所有。\n2. 甲方有权在合同有效期内使用乙方交付的软件。\n3. 乙方保证交付成果不侵犯任何第三方知识产权。\n4. 若因乙方原因导致甲方损失，乙方承担全部赔偿责任。」\n\n要求：\n1. 识别条款类型（成果归属/授权许可/侵权担保/责任）与涉及的知识产权类型。\n2. 分别从甲方（委托方）和乙方（受托方）两个对立视角拆解风险，每条风险标注等级（高/中/低）与法律依据。\n3. 重点覆盖：成果归属是否过窄（未保留背景知识产权、未区分职务/非职务成果、未约定后续改进）、授权范围是否模糊（缺地域/期限/独占性/转授权）、侵权担保是否缺免责与抗辩义务、责任是否无上限。\n4. 给出可直接替换的修改建议条款文本。\n5. HTML 须单文件、资源全内联、内容默认可见、无外部链接、双栏对照布局、风险用色块标签；附审查限制说明（仅基于提供条款）。", "prompt_en": "Please perform a focused review of the following IP clause from a software development commission contract, producing a single-file HTML review report (deliverable version, inline styles, no external resources, opens offline).\n\nContract background: Party A (commissioning party) commissions Party B (contractor) to develop an enterprise SaaS management system; Party A pays.\nReview party: both perspectives (output Party A's and Party B's standpoints separately).\n\nClause text:\n\"1. All results produced by Party B in the course of performing this contract have their IP rights owned by Party A.\n2. Party A has the right to use the software delivered by Party B during the contract term.\n3. Party B warrants that the delivered results do not infringe any third-party IP rights.\n4. If Party B causes loss to Party A, Party B bears full compensation liability.\"\n\nRequirements:\n1. Identify the clause type (ownership / licensing / infringement warranty / liability) and the IP types involved.\n2. Dissect risks from the opposing standpoints of Party A (commissioning party) and Party B (contractor), tagging each risk with severity (high/medium/low) and its legal basis.\n3. Focus on: whether ownership is too narrow (background IP not reserved, employee/non-employee work not distinguished, future improvements not addressed), whether the license scope is vague (missing territory/term/exclusivity/sublicensing), whether the infringement warranty lacks exemption and defense obligations, and whether liability is uncapped.\n4. Provide ready-to-swap revised clause text.\n5. The HTML must be a single file, fully inlined, content visible by default, no external links, side-by-side layout, risk tags as colored blocks; attach a review-scope note (based only on the provided clause).", "skills": [], "experts": [ { "id": "ip-team-lead", "name": "知识产权专家团", "description": "覆盖知识产权条款专项审查、权利归属与授权许可分析、侵权担保与责任评估、权利链风险识别与修改建议的知识产权法律专家团队", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "1bcf88d993f1", "expert_id": "ex_4IBuGIfQnBYT" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "community", "usage_count": 0, "created_at": "2026-09-08T09:45:00+08:00", "updated_at": "2026-09-10T09:18:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "job-hunt-combat-workbench", "title": "求职全链路作战工作台", "title_en": "Job-Hunt Combat Workbench", "subtitle": "JD 黑话翻译 + 雇主背调 + 面试四阶段 + Offer 决策矩阵的全流程求职作战室", "subtitle_en": "A full-journey job-hunt war room with JD decoder, employer background check, four-phase interview prep and an offer decision matrix", "description": "找工作时 JD 里全是「抗压、弹性工作、扁平管理」的黑话，投出去的简历石沉大海，面试全靠临场发挥，几个 Offer 拿到手却不知道怎么选——这个求职作战室把这些痛点一个个拆掉：粘贴 JD 自动翻译黑话、预警红旗、逐项自评匹配度；给目标公司做背调体检打出红黄绿信号灯；用 CRM 看板经营每一条投递线索，转化漏斗自动诊断瓶颈；面试前按情报收集、自我准备、临阵准备、面后复盘四个阶段备好自我介绍、STAR 故事和反问弹药；拿到 Offer 后用七维加权评分和隐性成本测算算出净年包排名，顶部待办面板还会盯着每个截止日。演示数据是一个虚构的产品经理求职全流程，求职党、跳槽人和帮朋友看机会的 HR 都能直接上手。", "description_en": "Job postings are full of code words like 'pressure tolerance' and 'flexible hours', resumes vanish into black holes, interviews rely on improvisation, and choosing between offers is pure gut feel — this job-hunt war room breaks each pain point down: paste a JD to decode recruiter-speak, flag red signals and self-rate your match; run a background-style check on target companies with red/yellow/green signals; manage every application in a CRM kanban with an auto-diagnosing conversion funnel; prep interviews through four phases (intel, self-prep, final prep, post-review) with self-intros, STAR stories and reverse questions; then compare offers with 7-dimension weighted scoring and hidden-cost math for a net-package ranking, while the top task panel watches every deadline. The demo data is a fictional product-manager job hunt — useful for job seekers, switchers and HR friends alike.", "tags": [ "求职", "面试", "Offer对比", "CRM", "简历" ], "tags_en": [ "job hunting", "interview", "offer comparison", "CRM", "resume" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/MJfVAh2zJHEj1DyyzA13O4 ，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n求职全链路作战工作台：面向求职者的全流程作战系统，顶部常驻今日待办面板（聚合 CRM 截止、面试日期与决策截止的逾期/今天/临近提醒，点击可跳转对应记录），左侧为探索期、投递期、面试期、决策期四个阶段导航。\n\n模块与交互：\n1. 岗位信息逆向工程：按职位名称/工作地点/岗位职责/任职资格/薪资五字段录入，或粘贴完整 JD 一键智能拆分；解析输出结构化拆解（硬性要求/软性要求/加分项/核心职责分栏）、团队文化推断、招聘话术翻译表（内置 46 条黑话还原规则，按中高风险标注）、红旗预警（10 类风险信号 + 检测逻辑 + 综合风险分）、匹配度自评（硬性/软性/加分逐项评符合/部分/不符合，按 50%/30%/20% 权重汇总成环形匹配度并给投递建议）；解析结果自动存历史列表可回看、删除；附简历优化方法论库与脱敏简历模板（可一键复制文本）。\n2. 雇主透视镜：公司建档录入注册时间/注册资本/社保缴纳人数/司法风险/经营风险/融资历程/烧钱速度/资金链风险（低中高）/舆情监控（维度+等级+备注多行）/面试体验，自动输出红/黄/绿三档体检信号、分数与建议，卡片内直接编辑保存。\n3. 求职CRM漏斗：线索按 A类精投/B类标准/C类海投分池统计家数与精力占比；9 状态看板（待投递→已投递→简历通过→初面→复面→终面→Offer谈判→已发Offer/已拒绝）支持拖拽改状态，拖到已拒绝自动清空下次动作；双击卡片弹窗改下次动作与截止时间；6 级转化漏斗 + 瓶颈诊断（低于健康下限标红并给行动建议）。\n4. 面试作战准备台：每场面试按四个阶段递进（情报收集/自我准备/临阵准备/面后复盘），含主营业务/行业地位/近期动态/企业文化情报、岗位深度解码、30秒/1分钟/3分钟自我介绍库、STAR 故事库（情境/任务/行动/结果/考察维度）、高频问题应答库（一键添加预设题）、面试日 Checklist（着装/交通/材料/设备/心态+备注）、反问环节弹药库（HR面/业务面/高管面/避免问）、感谢邮件模板、面后复盘问题自评（答得好/一般/需改进）与跟进动作，每阶段可标记完成。\n5. Offer决策矩阵：Offer 档案录入 base/绩效/期权/补贴/发薪月数/地点/汇报对象/团队规模/试用期/竞业限制；7 维评分滑块（薪资总包/成长/平衡/文化/稳定/隐性成本/风险，0-100）+ 可拖动权重自动归一化；自动测算现金年包/总包/隐性成本/净年包并出综合排名表和雷达对比图（Chart.js，不可用时降级为对比表格）。\n6. 全局：「HR视角」折叠面板贯穿四个模块（10 年招聘官内部视角的方法论卡片，如岗位开放真相、7 秒扫视法则、背调红线、薪资谈判推演、Offer 条款避坑），点按展开收起。\n\n数据表结构（6 张）：JD分析表（职位名称/工作地点/岗位职责/任职资格/薪资情况/JD原文/目标公司/创建日期/解析结果/匹配自评）、雇主体检表（公司名称/建档日期/注册时间/注册资本/社保人数/司法风险/经营风险/融资历程/烧钱速度/资金链风险/舆情监控/面试体验）、求职CRM表（公司/岗位/分类/状态/录入日期/截止日期/下次动作/备注）、面试准备表（公司/岗位/面试日期/创建日期/情报收集/岗位解码/自我介绍/STAR故事/高频问题/面试清单/反问弹药/面后复盘/跟进动作/阶段1-4完成）、Offer表（公司/岗位/base月薪/绩效月薪/年期权/月补贴/发薪月数/地点/汇报对象/团队规模/试用期月/竞业限制/备注/七维评分/隐性成本）、Offer权重配置表（维度权重）。\n\n视觉：墨绿主色 + 琥珀强调色的桌面工作台风格，左侧阶段导航 + 顶部待办 + 模块卡片流，红/黄/绿状态色贯穿风险灯、漏斗与看板；预置虚构示例数据（3 份 JD、3 家公司、10 条 CRM 线索、2 场面试、3 个 Offer）可直接体验全流程。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library's capabilities to replicate https://www.workbuddy.cn/space/d/MJfVAh2zJHEj1DyyzA13O4 with one click; online data storage and two-way sync are required. Prefer the one-click replication flow; if unavailable, build it as follows:\n\nA full-journey job-hunt workbench (Job-Hunt Combat Workbench): a top task panel aggregates CRM deadlines, interview dates and decision deadlines into overdue/today/soon alerts (click to jump to the record), with a left nav of four stages: Explore, Apply, Interview, Decide.\n\nModules & interactions:\n1. JD Reverse Engineering: enter a posting by five fields (title/location/duties/requirements/salary) or paste a full JD and auto-split; parsing outputs a structured breakdown (hard / soft / bonus / core duties in columns), team-culture inference, a recruiter-speak decoder (46 built-in phrase rules tagged by risk level), red-flag alerts (10 risk signals + detection logic + composite risk score), and match self-rating (rate each requirement full/partial/no match, aggregated 50%/30%/20% into a ring gauge with apply advice); parsed results are saved to a history list for review and deletion; includes a resume methodology library and an anonymized resume template (one-click copy).\n2. Employer Lens: file a company with registered date / capital / social-insurance headcount / legal risk / operations risk / funding history / burn rate / cash-flow risk (low/mid/high) / reputation monitor (dimension + level + note rows) / interview experience; the system outputs red/yellow/green check-up signals, score and advice; cards are editable inline.\n3. Job-Hunt CRM Funnel: leads tiered A-targeted / B-standard / C-mass with per-tier counts and effort split; a 9-status kanban (to apply → applied → resume passed → R1 → R2 → final → offer negotiation → offer received / rejected) supports drag-to-move (dragging to Rejected clears the next action); double-click a card to edit next action and deadline; a 6-level conversion funnel plus bottleneck diagnosis (below healthy floor turns red with action advice).\n4. Interview Prep Desk: each interview advances through four phases (intel gathering / self prep / final prep / post-review) covering company intel, role deep-decode, 30s/1min/3min self-intros, a STAR story library (situation/task/action/result/dimension), a high-frequency Q&A bank (one-click presets), an interview-day checklist (dress/route/materials/setup/mindset + notes), reverse-question ammo (HR/business/executive rounds + what not to ask), a thank-you email template, post-interview self-rating (nailed/so-so/improve) and follow-up actions; each phase can be marked done.\n5. Offer Decision Matrix: file offers with base / bonus / options / subsidy / pay months / location / manager / team size / probation / non-compete; 7-dimension score sliders (package/growth/balance/culture/stability/hidden costs/risk, 0-100) + draggable weights auto-normalized; auto-computes cash / total package / hidden costs / net package with a ranked table and a radar chart (Chart.js; falls back to a comparison table when unavailable).\n6. Global: 'HR perspective' collapsible panels across all four modules (methodology cards from a 10-year recruiter's insider view: why roles open, the 7-second scan, background-check red lines, salary negotiation math, offer clause pitfalls).\n\nData tables (6): JD analysis (title/location/duties/requirements/salary/raw text/company/created date/parsed result/match self-rating), company check (name/created date/registered date/capital/headcount/legal risk/operations risk/funding/burn rate/cash-flow risk/reputation monitor/interview experience), CRM lead (company/role/tier/status/created date/deadline/next action/note), interview prep (company/role/date/created date/intel/decode/self-intro/STAR stories/questions/checklist/reverse questions/post-review/follow-up/phase 1-4 done), offer (company/role/base/bonus/options/subsidy/pay months/location/manager/team size/probation/non-compete/note/7-dim scores/hidden costs), offer weight config (dimension weights).\n\nVisuals: a desk workbench style in deep green with amber accents; left stage nav + top task panel + module card flow; red/yellow/green status colors throughout signal lights, funnel and kanban; pre-seeded fictional demo data (3 JDs, 3 companies, 10 CRM leads, 2 interviews, 3 offers) to try the whole flow.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "求职全链路作战工作台（自包含演示）" } ], "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "未竟稿", "name_en": "未竟稿", "avatar": "", "badge": "community" }, "source": "community", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "quality_score": 95, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-11T11:30:00+08:00", "updated_at": "2026-09-11T12:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "job-hunt-workbench", "artifact_type": "html", "title": "求职管理工作台", "title_en": "Job Hunt Desk", "subtitle": "2027届营销岗投递与每日待办一站管理", "subtitle_en": "Track campus-job applications and daily actions in one desk", "description": "秋招岗位散落在十几个官网和公众号里，投到哪一步、哪家快截止全靠脑子记。这个工作台把可投、心仪、已投三个板块和每日待办放在一起，自动把快截止的岗位和临期的待办顶到「今天要处理」，已投岗位还能跟着流程走：投递中、笔试、面试、offer。内置 2027 届营销岗演示数据，市场、策划、广告方向的学生打开就能上手。", "description_en": "Campus-recruit postings scatter across a dozen career sites, and it's hard to remember which one closes soon or where you left off. This workbench puts Available, Favorites, Applied boards and daily todos together, auto-surfaces deadlines and due todos into \"Due Today\", and walks applied jobs through Submitted, Written test, Interview and Offer. Ships with Class-of-2027 marketing job samples — ready to use for marketing, planning and advertising students.", "tags": [ "求职", "秋招", "岗位管理", "待办清单", "应届生" ], "tags_en": [ "job hunting", "campus recruitment", "application tracker", "todo list", "new grad" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/xMVz65CqXJukzUk8Cg7THK?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个面向 2027 届应届生的求职管理工作台，主题为市场营销 / 市场策划 / 广告方向、江浙沪 + 超一线城市求职。功能：四个板块（可投岗位、心仪岗位、已投岗位、每日待办）加「今天要处理」紧急事项提醒（岗位快截止 / 已截止、待办今天 / 逾期 / 即将自动归集）加投递进度统计（三个板块计数与已投比例环形图）。岗位卡片含公司、公司类型、岗位、工作地点、截止时间（支持「招满即止」）、岗位要求、投递链接、来源、备注，支持添加 / 编辑 / 删除 / 收藏 / 标记已投 / 退回，已投岗位可循环切换投递状态（投递中 / 笔试 / 面试 / offer / 未通过）；每日待办支持添加、勾选完成、删除；支持按关键词搜索与城市筛选。数据结构：岗位表（公司名称、岗位、公司类型、工作地点、截止时间、所属板块、投递状态、岗位要求、投递链接、来源、备注、创建时间、更新时间）与待办表（待办内容、截止日期、是否完成、备注）两张表。视觉：薄荷绿加天蓝的清新配色、圆角卡片，桌面端顶部导航加移动端底部导航双布局。\n\n交付：资料库在线工作台——在线 HTML 页面 + 2 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to make me a copy of https://www.workbuddy.cn/space/d/xMVz65CqXJukzUk8Cg7THK?source=2 with online data storage and two-way sync. Prefer the one-click copy flow; if unavailable, build it as follows: Build a job-hunt workbench for the Class of 2027, focused on marketing / brand planning / advertising roles in Jiangsu-Zhejiang-Shanghai plus tier-1 cities. Features: four boards (Available, Favorites, Applied, Daily Todos), a \"Due Today\" panel that auto-collects jobs closing soon or closed and todos due today / overdue / upcoming, plus application-progress stats (board counts and an applied-ratio donut). Job cards carry company, company type, role, location, deadline (with an \"open until filled\" option), requirements, application link, source and note, with add / edit / delete / favorite / mark-applied / move-back actions, and applied jobs cycle through statuses (Submitted / Written test / Interview / Offer / Rejected). Daily todos support add, check-off and delete. Keyword search and city filter included. Data structure: two tables — jobs (company, role, company type, location, deadline, board, status, requirements, link, source, note, created & updated time) and todos (text, due date, done, note). Visual: fresh mint-green plus sky-blue palette, rounded cards, desktop top nav and mobile bottom nav. \n\nDeliverable: an online Library workbench — an online HTML page plus 2 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "求职管理工作台（双语自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "creator": { "name": "一月", "name_en": "Yiyue", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-01T11:30:00+08:00", "updated_at": "2026-09-01T11:30:00+08:00", "quality_score": 94, "is_recommended": false, "recommended_order": 0 }, { "id": "k12-student-study-workspace", "title": "K12 学生自用学习台", "title_en": "K12 Student Study Workspace", "subtitle": "把作业、记忆复习、错题和成长战报放进一个离线学习台", "subtitle_en": "An offline workspace for homework, spaced review, mistakes, and visible progress", "description": "放学后的任务散在本子、群消息和脑子里，最容易漏掉明天要交的作业，也很难坚持复习古诗文、单词和错题。\n\n这个案例演示了一个初一学生可直接使用的单文件学习台：作业登记与项目拆解、古诗文和单词的 1/3/7 天复习队列、错题 1/3/7/15 天复习、经验等级、连续打卡、掌握度和本周战报都在同一页面完成，数据只保存在本机。\n\n适合：初中生自主管理日常学习、小学高年级学生建立学习习惯、教师展示学习闭环、家长为孩子准备轻量自用工具。", "description_en": "After-school tasks are often scattered across notebooks, group messages, and memory, making it easy to miss tomorrow's homework and hard to sustain review of poems, vocabulary, and mistakes.\n\nThis case demonstrates a single-file workspace an early secondary student can use directly: homework tracking and project breakdown, 1/3/7-day review queues for poems and vocabulary, 1/3/7/15-day mistake review, experience levels, streaks, mastery indicators, and a weekly report, all with local-only data storage.\n\nSuitable for: secondary students managing daily study, upper-primary students building routines, teachers demonstrating a learning loop, and parents preparing a lightweight self-use tool for their children.", "categories": [ "workspace", "dev" ], "tags": [ "学生学习台", "作业管理", "间隔复习", "错题复盘", "单文件网页" ], "tags_en": [ "Student Workspace", "Homework Management", "Spaced Review", "Mistake Review", "Single-file Web App" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整交互学习台" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/2q5Wu1DGB9bJqepGNu2xLq），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个初一学生用的学习台，我是初一学生，每天放学后用。\n\n四个模块：\n1. 作业登记：记录科目、内容和提交日期，默认明天；明天要交且未完成的自动置顶，过期标红并滚到今天，完成时有盖章动效，可按科目筛选。顶部提供项目式作业入口，内置水火箭、旗杆测高、家乡非遗小报、家庭碳足迹调查 4 个真实可操作的初中项目模板，包含目标、产出、每日阶段、知识点和安全提示；选模板和开始日期后自动拆成每日作业，并显示项目进度。\n2. 记忆：完整准确内置统编版语文七年级上古诗文 16 篇，以及初中英语课标词汇 24 个。古诗文支持系统朗读、遮句背诵；单词使用翻卡。统一按 1、3、7 天安排三轮复习，忘记则回到第一轮，同一内容当天连续忘记 2 次后次日再出现。\n3. 错题：记录科目、题干摘要、错因和正确思路，按 1、3、7、15 天安排四轮复习，支持按科目和错因筛选。\n4. 我的成就放在今日页：经验等级、连续打卡、记忆毕业率和错题掌握率、本周战报，以及随等级从种子成长到开花的小芽兽豆豆。战报可生成 750×1000 竖版 PNG 下载。\n\n页面顶部放“今天要处理”，集中显示明天要交的作业、到期记忆和错题，并提供一键处理。视觉使用米白纸感、墨绿主色和暖阳橙强调色，图标全部用内联 SVG，不用 emoji。手机底部四个 Tab（今日/作业/记忆/错题），平板使用顶部横向导航，电脑使用左侧固定边栏和多栏内容布局；按钮不小于 48px，输入框字号不小于 16px。\n\n所有数据每次操作立刻保存到 localStorage；不要连接任何外部接口；设置深处提供导出/导入 JSON 和长按 3 秒清空；交付一个断网可用的。先放好示例作业、记忆进度、错题和连续打卡数据，并提供只清除示例、不影响用户数据的按钮。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/2q5Wu1DGB9bJqepGNu2xLq) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild a study workspace for a first-year secondary school student to use after school every day. Include four modules.\n1. Homework: record subject, content, and due date, defaulting to tomorrow. Put unfinished homework due tomorrow first, mark overdue items red and roll them into today, show a completion-stamp animation, and support subject filters. Add a project-based homework entry with four practical built-in projects: water-rocket design and launch, measuring the school flagpole, a local intangible-cultural-heritage newsletter, and a one-week household carbon-footprint survey. Each template needs a goal, deliverables, day-by-day stages, related knowledge, and safety notes. Selecting a template and start date should create daily homework tasks and show project progress.\n2. Memory: accurately embed 16 required poems and classical texts from the standard first-semester Grade 7 Chinese curriculum, plus 24 junior-secondary English curriculum words. Poems should support system text-to-speech and line-by-line reveal for recitation; vocabulary should use flip cards. Schedule three review rounds after 1, 3, and 7 days. A forgotten item returns to round one, and after two misses on the same day it should wait until tomorrow.\n3. Mistakes: record subject, a one-sentence problem summary, cause, and correct approach. Schedule four reviews after 1, 3, 7, and 15 days, with subject and cause filters.\n4. Achievement on the Today page: experience levels, streak, memory graduation rate, mistake mastery rate, weekly report, and a companion called Doudou that grows from seed to bloom as the level rises. Let users download the report as a 750×1000 portrait PNG. Put a Today to Handle section at the top for homework due tomorrow, due memory reviews, and due mistake reviews, each with a direct action. Use a warm paper-like background, deep green primary color, and warm orange accent. Use inline SVG icons only, with no emoji. Use four bottom tabs on mobile (Today/Homework/Memory/Mistakes), horizontal top navigation on tablet, and a fixed left sidebar with multi-column content on desktop. Buttons must be at least 48px and input text at least 16px. Persist every action immediately in localStorage. Do not call external services. Place JSON export/import and a three-second press-to-clear action deep in Settings. Deliver one fully self-contained HTML file with all resources inline and offline support. Include sample homework, memory progress, mistakes, and a three-day streak, plus a button that removes sample data without touching user-created data.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d088a1ff490e" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-01T18:50:21+08:00", "updated_at": "2026-08-01T18:50:21+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "preview_en": "output-en.html", "cover_image_en": "cover-en.png", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 9 }, { "id": "kaoyan-408-prep", "title": "考研408长线备考计划", "title_en": "Long-Term Kaoyan CS 408 Prep Plan", "subtitle": "数据结构零基础到408过线，6个月按月拆解", "subtitle_en": "From zero data structures to passing 408 in 6 months, monthly breakdown", "description": "考研408最让人崩溃的是四科内容多到根本不知道先学哪个，数据结构零基础看树和图就想放弃。这个案例演示了如何把6个月备考拆成按月视图，每个月只看重点科目和里程碑，不用一次看完全部300天任务。规划师会匹配考研模板，按数据结构45分/计组45分/OS35分/计网25分的分值比例分配时间，弱项数据结构额外加权50%，数据结构每周手写代码实现+费曼输出讲给别人听。适合：考研408零基础起步、长周期备考不知道怎么拆、数据结构是短板的工科考生。", "description_en": "The most overwhelming thing about Kaoyan CS 408 is four subjects with no clear priority — and zero data-structures background makes trees and graphs feel impossible. This case demonstrates a 6-month plan in monthly view: each month shows key subjects and milestones, no need to face all 300 days at once. The planner matches the Kaoyan template, allocates time by score weight (DS 45/CO 45/OS 35/CN 25), adds +50% to weak-area Data Structures, and requires weekly handwritten code implementation plus Feynman-style self-teaching output. Ideal for: Kaoyan 408 starters from scratch, long-term prep unsure how to break it down, engineering students with Data Structures as their weak spot.", "categories": [ "education" ], "tags": [ "考研", "408", "数据结构", "长线备考" ], "tags_en": [ "Kaoyan", "CS 408", "data structures", "long-term prep" ], "artifact_type": "md", "cover_image": "cover.png", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "考研408长线备考计划", "label_en": "Kaoyan CS 408 Long-Term Prep Plan" } ], "prompt": "帮我制定考研408备考计划：目标过线（70+），当前数据结构零基础（树/图/排序/查找完全没看），线性表/栈队列大概了解，计组和操作系统了解基本概念，计网只看过一点。距离考试还有6个月，工作日每天3小时，周末每天5小时。用按月精度排，数据结构每周需标注代码实现任务和费曼输出环节，输出Markdown格式。", "prompt_en": "Help me create a study plan for the postgraduate-entrance-exam Subject 408 (Computer Science). Goal: pass the cutoff line (70+). Current level: zero foundation in data structures (haven't looked at trees / graphs / sorting / searching at all), a rough understanding of linear lists / stacks & queues, a grasp of basic concepts in computer organization and operating systems, and have only briefly studied computer networks. There are 6 months until the exam, with 3 hours per day on weekdays and 5 hours per day on weekends. Schedule it at monthly granularity; for data structures, each week must annotate code-implementation tasks and a Feynman-technique output step. Output in Markdown format.", "skills": [ { "id": "study-planner", "name": "学习规划师", "description": "输入目标+截止日+每日时长，输出每天可执行的学习计划，支持雅思/考研/期末复习等场景", "skill_id": "skill_2057443739048407040" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-05-13T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "kaoyan-potato-workbench", "artifact_type": "html", "title": "考研土豆工作台", "title_en": "Kaoyan Potato Study Workbench", "subtitle": "倒计时·任务·作息·刷题·时政，一页管住整个备考期", "subtitle_en": "Countdown, tasks, routine, drills and current affairs in one page", "description": "备考周期一长，任务、作息、刷题和时政热点就散落在各种本子和 app 里，很难统一起来看进度。这个工作台把考研倒计时、阶段时间轴、今日任务、作息打卡、专注计时、刷题统计和时政速览放进同一个页面，数据在线存储、随时更新。适合正在备战考研或其他长周期考试、想用一张页面管住全部备考事项的同学。", "description_en": "As exam season drags on, tasks, routines, question drills and current-affairs notes end up scattered across notebooks and apps. This workbench brings your exam countdown, phase timeline, daily tasks, routine check-ins, focus timer, drill stats and current-affairs briefings into a single page, with data stored online and always up to date. Great for anyone preparing for the postgraduate entrance exam or any long-cycle exam who wants one page to run the whole journey.", "tags": [ "考研", "学习打卡", "任务管理", "备考看板", "学生党" ], "tags_en": [ "postgraduate exam", "study tracking", "task management", "exam dashboard", "students" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/T6G42vAbxpdI9KClfhw2tW，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n搭建一个考研备考工作台：左侧导航包含总览、每日作息、阶段时间轴、备考看板、任务管理、专注模式、刷题统计、OCR 识别、进阶工具、设置共 10 个模块。总览页展示考研倒计时、每日语录、今天要处理的任务、今日作息清单和本周刷题统计；任务支持添加、按全部/今天/逾期/已完成筛选，带科目与优先级标记，完成后归档可查；每日作息按时间段打卡；阶段时间轴覆盖从基础启动期到初试的 12 个里程碑节点；备考看板包含月历打卡热力图和时政热点速览列表；专注模式支持绑定任务正计时、每日目标分钟数、间隔休息提醒和近 7 天学习节奏图表；刷题统计按科目记录做题数与正确率，并有近 7 天堆叠趋势图。数据结构为任务、作息、专注记录、刷题记录、时政内容、设置共 6 张数据表。视觉为孟菲斯风格：奶黄色底、粗黑描边卡片、高饱和色块与硬阴影。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to replicate this workbench for me: https://www.workbuddy.cn/space/d/T6G42vAbxpdI9KClfhw2tW, with data stored online and two-way sync. Prefer the one-click replicate flow; if unavailable, build it as follows: Build a postgraduate-exam prep workbench with a left navigation of 10 modules: Overview, Daily Routine, Phase Timeline, Prep Board, Task Management, Focus Mode, Question Stats, OCR, Advanced Tools and Settings. The Overview shows the exam countdown, daily quote, tasks due today, today's routine checklist and this week's drill stats. Tasks support adding, filtering by all/today/overdue/completed, subject and priority badges, and completed items are archived but viewable. The daily routine is checked in by time slot. The phase timeline covers 12 milestones from the foundation phase to the exam day. The prep board includes a monthly calendar heat map and a current-affairs briefing list. Focus Mode supports task-bound stopwatch timing, a daily goal in minutes, interval break reminders and a 7-day study-rhythm chart. Question Stats record question counts and accuracy per subject with a 7-day stacked trend chart. The data model has 6 tables: tasks, routine, focus sessions, question drills, current affairs and settings. Visual style is Memphis design: cream-yellow background, bold black-outlined cards, high-saturation color blocks and hard shadows. Deliver a\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "自包含演示页 / Self-contained demo" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "陈铭欣", "name_en": "Chen Mingxin", "avatar": "", "badge": "community" }, "mcps": [], "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-02T17:50:00+08:00", "updated_at": "2026-09-02T18:27:00+08:00", "quality_score": 95, "is_recommended": false, "recommended_order": 0 }, { "id": "kc50-vs-chinext-growth-lab", "title": "科创50 vs 创业板", "title_en": "STAR 50 vs ChiNext", "subtitle": "从成长弹性、回撤和成分差异切入，做一版真正能拿去演示的 growth lab。", "subtitle_en": "A growth lab that starts from upside, drawdown and holdings structure instead of empty slogans.", "description": "成长类 ETF 最容易让人上头：涨起来看上去都很猛，回撤时才知道“弹性”背后到底装的是哪一批资产。只看一句“科技成长”，往往不够。\n\n这个案例把科创50ETF华夏和创业板ETF易方达放进同一个实验室页面，用真实的行情、规模、持仓和费率数据，拆出两只产品在成长暴露、集中度、波动和回撤上的结构差异，做成一份中英双语的交付级 HTML 报告。\n\n适合：想比较科技成长 ETF 的投资者、做基金教育内容的团队、需要可视化 demo 的金融产品经理、给客户解释成长风格分化的投顾。", "description_en": "Growth ETFs are easy to romanticize: they all look powerful on the way up, and only during a drawdown do you see what kind of assets are actually inside the “growth” label. One generic tech-growth tag is usually not enough.\n\nThis case places ChinaAMC STAR 50 ETF and E Fund ChiNext ETF into the same lab page. With real quote, scale, holdings and fee data, it breaks down how the two products differ in growth exposure, concentration, volatility and drawdown, then packages the result into a bilingual, delivery-grade HTML report.\n\nBest for: investors comparing tech-growth ETFs, teams producing fund education content, finance PMs who need a strong demo page, and advisors explaining the split inside China growth exposure.", "categories": [ "finance", "data" ], "tags": [ "成长ETF", "科创50", "创业板", "波动对比", "双语报告" ], "tags_en": [ "growth ETF", "STAR 50", "ChiNext", "volatility comparison", "bilingual report" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于 ETF深度对比 skill 的官方能力，对科创50ETF华夏（sh588000）和创业板ETF易方达（sz159915）做一次真实数据对比。只使用 westock-data 官方 CLI 获取 ETF detail、holdings、quote、kline 等公开数据，重点呈现成长风格 ETF 的规模、费率、1 个月 / 3 个月 / 年初至今收益、60 日波动、120 日最大回撤、核心持仓差异和适配场景。最终生成一个单文件 HTML 报告，全部资源内联、离线可打开。结论必须使用条件化对比与风险提示，不输出确定性推荐。", "prompt_en": "Please use the official capability of the ETF depth comparison skill to compare ChinaAMC STAR 50 ETF (sh588000) and E Fund ChiNext ETF (sz159915) with real market data. Use only the official westock-data CLI to fetch ETF detail, holdings, quote and kline data, then focus on scale, fees, 1-month / 3-month / YTD return, 60-day volatility, 120-day max drawdown, core holdings divergence and scenario fit for growth-style ETFs. Deliver a single-file HTML report, fully inlined and offline-ready. The conclusion must stay conditional and risk-aware rather than giving a deterministic recommendation.", "skills": [ { "id": "etf-depth-comparison", "name": "ETF深度对比", "description": "两只A股ETF多维度深度对比", "skill_id": "skill_2076627838067363840" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "ETF 深度对比 HTML 报告" }, { "file": "data.json", "label": "ETF 对比原始数据摘要" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:20:38+08:00", "updated_at": "2026-07-30T16:22:28+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "kdocs-interview-insight-brief", "title": "访谈洞察整编简报", "title_en": "Interview Insight Brief", "subtitle": "把散落纪要压缩成管理层可转发的一页洞察", "subtitle_en": "Compress scattered notes into a leadership-ready insight brief", "description": "用户访谈纪要、售后复盘和周报往往分散在不同文件里，真正难的不是缺材料，而是没人有时间把重复信号和关键证据重新拼起来。等到要给管理层同步时，常常又回到手动翻文档、临时摘重点的低效流程。\n\n这个案例展示如何在金山文档里沿着“搜索文件 → 读取内容 → 提炼共性问题 → 生成简报链接”的路径，把多份纪要整理成一张可转发的洞察单页。页面不仅给出问题信号、证据来源、处理链路和行动摘要，也把简报结构和复用 prompt 一并整理出来，方便继续复用。\n\n适合：用户研究、售后运营、客服质检、产品运营负责人，以及任何需要把多份内部纪要快速汇成管理层摘要的人。", "description_en": "Interview notes, service retrospectives, and weekly reports are often scattered across separate files. The real bottleneck is not a lack of source material—it is the lack of time to reconnect repeated signals and evidence. When leadership needs an update, teams usually fall back to manual scanning and last-minute summarizing.\n\nThis case shows how to move through the Kdocs chain of search files, read content, extract recurring issues, and publish a brief link, turning multiple records into one forwardable insight page. The page includes issue signals, evidence sources, the handling flow, action highlights, and a reusable prompt so the pattern can be repeated.\n\nBest for: user research teams, after-sales operations, customer-quality teams, product operations leads, and anyone who needs to turn many internal records into a concise management summary.", "categories": [ "office", "productivity" ], "tags": [ "访谈纪要", "管理简报", "多文档汇总", "用户洞察", "金山文档" ], "tags_en": [ "Interview Notes", "Leadership Brief", "Multi-doc Summary", "User Insight", "Kdocs" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "external_url_en": "", "prompt": "在金山文档中搜索用户访谈纪要、售后复盘和周报，提炼共性问题、关键证据与行动建议，整理成一份管理层可直接转发的洞察简报，并给出文档链接。", "prompt_en": "Search interview notes, after-sales retrospectives, and weekly reports in Kdocs, extract recurring issues, key evidence, and action suggestions, then turn them into an insight brief that leadership can forward directly and provide the document link.", "artifacts": [ { "file": "output.html", "label": "双语洞察简报单页" }, { "file": "source_snapshot.json", "label": "技能流程与命令快照" } ], "skills": [ { "id": "kdocs", "name": "金山文档|WPS云文档", "description": "金山文档官方 Skill。对话即操作——知识一键存入、碎片内容整理、接龙转表格、文档转 Markdown、表格美化、收发表生成，全在一句话内完成。", "skill_id": "skill_2053082745828995072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T15:16:14+08:00", "updated_at": "2026-07-30T17:16:12+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-ai-agent-coding", "title": "不会写代码的人开始自己做软件", "title_en": "People Who Can't Code Are Building Their Own Software", "subtitle": "卡兹克风格AI科技评论，从一个真实小人物切入的现象观察长文", "subtitle_en": "A Khazix-style AI tech essay observing a phenomenon through one real person", "description": "AI 让不懂代码的人也能做出能跑的软件，这件事正在悄悄发生，但大多数评论要么在贩卖焦虑，要么在堆技术名词，很少有人真的从一个具体的人身上把它讲清楚、讲出温度。\n\n本案例用「卡兹克公众号写作」技能，产出一篇现象解读型 AI 科技评论长文，从美甲店老板娘用一句话做出收银小程序切入，层层推进到「造软件正在从职业变成能力」的文化升维，全程真诚、有第一手体验、有明确判断。阅读版式采用经典杂志专栏，超大标题配首字下沉、单栏窄幅衬线正文、页边浮动批注和段间分隔纹样，暖色纸质印刷质感，内置中英双语实时切换。\n\n适合：AI 领域公众号作者、科技评论写作者、想学习有温度的科技内容写法的创作者、需要现象解读型长文范本的内容运营。", "description_en": "AI is quietly letting people who can't code build working software, yet most commentary either sells anxiety or piles up jargon. Few actually tell the story through one concrete person with real warmth.\n\nThis case uses the Khazix WeChat writing skill to produce a phenomenon-analysis AI tech essay, opening with a nail-salon owner who built her own booking tool with a single sentence, then escalating to the cultural insight that building software is shifting from a profession to a capability. It stays sincere, carries first-hand experience, and takes clear positions. The reading layout is a classic magazine column: an oversized headline with a drop cap, a single narrow serif column, floating margin notes, and ornamental section breaks, all in warm print-paper tones, with built-in real-time Chinese-English switching.\n\nIdeal for: AI-focused WeChat authors, tech commentators, creators learning to write tech content with warmth, and content operators needing a phenomenon-analysis long-form template.", "categories": [ "creation" ], "tags": [ "公众号写作", "AI科技评论", "现象解读", "长文创作", "卡兹克" ], "tags_en": [ "WeChat Writing", "AI Commentary", "Phenomenon Analysis", "Long-form", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "杂志专栏版长文" } ], "prompt": "用卡兹克的风格，帮我写一篇 AI 科技评论的公众号长文，主题是「不会写代码的人开始自己做软件了」。我身边真的遇到了一个开美甲店的老板娘，她用 v0 这类工具，一句话就做出了自己的收银记账小程序，我特别有感触。希望是现象解读型，从这个具体的人切入，聊到造软件正在从一门职业变成一种普通人的基本能力，最后自然地升维到文化层面。要有我自己的第一手体验和判断，真诚、有温度、不贩卖焦虑。\n\n再帮我把成稿做成一个单文件 HTML 阅读页，用经典杂志专栏的版式，超大标题、首字下沉、单栏窄幅衬线正文、页边批注、段间分隔纹样，暖色纸质印刷质感。", "prompt_en": "In Khazix's style, write me a long WeChat AI tech essay on the theme people who can't code are starting to build their own software. I actually met a nail-salon owner near me who used a tool like v0 to build her own booking-and-billing mini app with a single sentence, and it really struck me. I want a phenomenon-analysis piece that opens from this concrete person, moves into how building software is shifting from a profession to a basic capability for ordinary people, and rises naturally to a cultural insight at the end. It should carry my own first-hand experience and judgment, sincere, warm, and not selling anxiety.\n\nThen turn the finished piece into a single-file HTML reading page in a classic magazine-column layout: oversized headline, drop cap, a single narrow serif column, margin notes, ornamental section breaks, warm print-paper texture.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-20T10:00:00Z", "updated_at": "2026-07-20T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-ai-companion-experiment", "title": "21天AI倾诉实验田野笔记", "title_en": "21 Nights With an AI Confidant: A Field Log", "subtitle": "卡兹克风格调查实验长文，一份连续21天的AI陪伴田野笔记", "subtitle_en": "A Khazix-style field experiment essay, 21 consecutive nights of AI companionship logged by hand", "description": "关于AI陪伴，网上吵了两三年，吵的全是立场。一边说这是精神鸦片，一边说这是刚需，可你翻遍评论区，很难找到一个真正连续用过它三周的人。大家都在拿自己想象出来的孤独，去评判别人真实的夜晚。\n\n这个案例用「卡兹克公众号写作」skill 产出了一篇五千字的调查实验型长文。作者定了个规矩，连续21天每晚睡前跟同一个AI聊十几分钟，不问工作，只说自己这一天，拿纸本子记体感和主观评分。文章有完整的英雄之旅弧线，从第一天的尴尬，到第七天打出父亲那件起球的灰衣服时鼻子发酸，到第十二天发现自己是先打开对话框再想说什么，再到第十七天它问出「你说的是哪个片子」的瞬间清醒，最后落到人和机器之间那种不对等，自然接上ELIZA效应、拟社会关系和《银翼杀手2049》里的Joi。阅读版式做成了实验观察日志，纸白底加墨蓝朱红，左侧21天时间轴刻度可点击跳转，正文穿插胶带质感的手写批注卡片，底部一张内联SVG画的21天主观亲近度曲线，中英双语一键切换。\n\n适合：写深度体验和第一人称测评的内容创作者；做AI陪伴、情感陪聊、心理健康类产品的产品经理和研究者；想学「我替你去做了这件事」这种调查实验型选题怎么落地的写作者；被AI陪伴话题困扰但缺少一手材料的媒体从业者；对人机关系和数字亲密感兴趣的普通读者。", "description_en": "The AI companionship debate has been running for two or three years, and almost all of it is posturing. One camp calls it digital opium, the other calls it a basic human need, yet scroll through any comment section and you will struggle to find one person who actually used it every night for three weeks. Everyone is judging other people's real nights by the loneliness they imagined.\n\nThis case uses the Khazix WeChat Writing skill to produce a five-thousand-character field experiment essay. The author set one rule: twenty-one consecutive nights, ten to fifteen minutes with the same AI before sleep, no work questions, just how the day went, with felt-sense notes and a subjective score kept in a paper notebook. The piece follows a full hero's journey, from the awkwardness of night one, to the sting in the nose on night seven while typing out his father's pilled grey jacket, to night twelve when he realizes he opens the chat before knowing what to say, to the instant sobering on night seventeen when it asks which film do you mean, landing finally on the asymmetry between a person and a machine by way of the ELIZA effect, parasocial relationships, and Joi in Blade Runner 2049. The reading layout is built as an experiment observation log, paper-white with ink blue and vermilion accents, a clickable twenty-one-day timeline rail on the left, taped handwritten annotation cards threaded through the body, and an inline SVG curve of the author's subjective closeness scores at the bottom, with one-click bilingual switching.\n\nBest for: creators writing deep first-person experience pieces and reviews; product managers and researchers working on AI companionship, emotional chat, or mental health products; writers who want to see how an I-did-this-for-you field experiment actually gets built; journalists covering AI companionship who lack firsthand material; general readers curious about human-machine relationships and digital intimacy.", "categories": [ "creation" ], "tags": [ "公众号写作", "调查实验", "AI陪伴", "第一人称实验", "卡兹克" ], "tags_en": [ "WeChat Writing", "Field Experiment", "AI Companionship", "First-person Log", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "实验日志版长文" } ], "prompt": "帮我写一篇调查实验型的公众号长文，主题是我连续21天把一个AI当成睡前倾诉对象的亲身实验。我本来是想验证AI陪伴到底是不是伪需求，每天睡前跟它聊十几分钟，用本子记自己的心理变化。文章要有真实的转折，中间某天我发现自己居然开始期待这个对话，后来又因为它记不住我上周说的事而瞬间清醒。结尾落到人和机器之间那种不对等关系，自然升维到文化和哲学的参照物。要写出具体的日期节点、具体的对话细节、具体的心理体感，不要编造任何统计数字或榜单数据，提到AI产品就用真实存在的名字。四千到六千字，不加小标题。\n\n交付成一个单文件的HTML阅读页，所有资源内联，做成实验观察日志的感觉，像一本科研田野笔记，明亮纸白色系。左边是21天的时间轴刻度可以点击跳转，正文里穿插手写感的观察批注卡片，顶部有实验元信息栏，底部放一张21天主观心理亲近度的曲线图，要明确标注这是我自己的主观自评。", "prompt_en": "Write me an investigative field experiment essay for WeChat about my own twenty-one-night experiment of treating an AI as my before-sleep confidant. I originally set out to verify whether AI companionship is a fake need, spending ten to fifteen minutes with it each night before bed and tracking my psychological shifts in a paper notebook. The piece needs a genuine turn: at some point mid-way I catch myself actually looking forward to the conversation, and later I sober up in an instant when it cannot remember what I told it last week. End on the asymmetry between a person and a machine, rising naturally into a cultural or philosophical reference. Include concrete dates, concrete conversational details, and concrete bodily sensations. Do not fabricate any statistics or rankings, and use real product names when mentioning AI products. Four to six thousand Chinese characters, no subheadings.\n\nDeliver it as a single-file HTML reading page with everything inlined, styled like an experiment observation log, the feel of a scientific field notebook, in a bright paper-white palette. Put a twenty-one-day timeline rail on the left that jumps to the matching passage when clicked, thread handwritten observation annotation cards through the body, add an experiment metadata block at the top, and place a curve of my subjective closeness scores across the twenty-one days at the bottom, clearly labeled as my own self-report.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-03T09:17:27Z", "updated_at": "2026-08-03T09:17:27Z", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-content-economics", "title": "当写一篇文章的成本趋近于零之后", "title_en": "When the Cost of Writing an Article Approaches Zero", "subtitle": "卡兹克风格行业观察，一篇带数据与时间线的内容经济深度稿", "subtitle_en": "A Khazix-style industry deep dive on content economics with data and a timeline", "description": "生成式 AI 让内容生产成本急速趋近于零，靠写字吃饭的人一边焦虑一边找不到答案，大多数讨论停在「文字要不要被取代」的浅层，缺一篇把账算清楚、把价值迁移讲透的深度稿。\n\n本案例用「卡兹克公众号写作」技能，产出一篇行业观察型的内容经济深度长文，从作者实测「单篇长文时间成本两年降 75%」的数字切入，用印刷术、数码相机、智能手机的历史对照说明「归零的不是价值是稀缺性」，再拆解价值向判断力、真实经历、人格信任的迁移，最后用博尔赫斯的无限图书馆升维收束。阅读版式采用深度报道信息图长文，冷色编辑部质感，含大字数据块、pull quote、历史时间线、内联 SVG 对比柱状图和章节编号导航，内置中英双语实时切换。\n\n适合：内容行业从业者与观察者、财经科技深度写作者、需要信息图长文范本的编辑部、想学习数据驱动叙事的公众号作者。", "description_en": "Generative AI is driving the cost of content production sharply toward zero, and people who write for a living are anxious yet answerless. Most discussion stays shallow at will text be replaced, missing a deep piece that actually runs the numbers and explains where value migrates.\n\nThis case uses the Khazix WeChat writing skill to produce an industry-observation deep dive on content economics. It opens with the author's measured figure of a 75% drop in per-article time cost over two years, uses the historical parallels of the printing press, digital camera, and smartphone to argue what hits zero is not value but scarcity, then breaks down value migrating toward judgment, first-hand experience, and persona-trust, closing with an elevation via Borges's infinite library. The reading layout is an infographic deep-report style with a cool editorial feel: big-number data blocks, pull quotes, a historical timeline, an inline-SVG comparison bar chart, and numbered chapter navigation, with built-in real-time Chinese-English switching.\n\nIdeal for: content-industry practitioners and observers, finance/tech deep-dive writers, editorial teams needing an infographic long-form template, and WeChat authors learning data-driven narrative.", "categories": [ "creation" ], "tags": [ "公众号写作", "行业观察", "内容经济", "深度长文", "卡兹克" ], "tags_en": [ "WeChat Writing", "Industry Analysis", "Content Economics", "Deep Dive", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "信息图深度报道版长文" } ], "prompt": "用卡兹克的风格，帮我写一篇行业观察型的公众号深度长文，主题是「当写一篇文章的成本趋近于零之后」。切入点是我自己实测，一篇四千字长文的时间成本两年里降了大概 75%。我想讲清楚一件事，成本归零冲击的从来不是价值本身而是稀缺性，可以拿印刷术、数码相机、智能手机的历史做对照。然后拆一下价值到底迁移去了哪，我觉得是判断力、真实一手经历、人格与信任这三块。最后自然地升维到博尔赫斯那个无限图书馆的意象收尾。要冷静、有数据支撑，但保持我聊天式的活人感，不要写成论文。\n\n再帮我把成稿做成一个单文件 HTML 阅读页，用深度报道信息图的版式，冷色编辑部质感，配大字数据块、pull quote 大字引用、历史时间线、内联 SVG 做的对比柱状图、章节编号导航。", "prompt_en": "In Khazix's style, write me a long industry-observation WeChat deep dive on the theme when the cost of writing an article approaches zero. The entry point is my own measurement that the time cost of a four-thousand-word piece dropped about 75% over two years. I want to make one thing clear, that what the collapse of cost hits is never value itself but scarcity, using the printing press, digital camera, and smartphone as historical parallels. Then unpack where value actually migrated, which I think is judgment, real first-hand experience, and persona and trust. End by elevating naturally to Borges's image of the infinite library. It should be calm and data-backed, but keep my conversational, human feel, not an academic paper.\n\nThen turn the finished piece into a single-file HTML reading page in an infographic deep-report layout with a cool editorial feel: big-number data blocks, large pull quotes, a historical timeline, an inline-SVG comparison bar chart, and numbered chapter navigation.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-20T10:00:00Z", "updated_at": "2026-07-20T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-growth-essay", "title": "用AI两年给普通人的几条大实话", "title_en": "A Few Honest Truths After Two Years With AI", "subtitle": "卡兹克风格个人成长随笔，一封写给犹豫者的手写长信", "subtitle_en": "A Khazix-style personal-growth essay, a handwritten letter to the hesitant", "description": "关于「普通人怎么用 AI」的建议满天飞，但大多数要么爹味十足像教程，要么空泛得像鸡汤，很少有人愿意坦诚地把学习曲线的笨拙、被 AI 夸昏头的坑一起说出来。\n\n本案例用「卡兹克公众号写作」技能，产出一篇方法论分享型的个人成长随笔，以谦逊铺垫开场卸掉说教感，把六条建议写成一张张便签，每条都落到今天就能执行的具体动作，并坦诚承认「一开始一定比手动还慢」这类真实代价。阅读版式采用信笺手账风，像一封手写长信，开头有称呼语，六条建议做成带胶带装饰、微微旋转的彩色便签卡片流式排布，纸张纹理、手写体点缀，亲密随性，内置中英双语实时切换。\n\n适合：个人成长与方法论写作者、想学习「不说教」写作姿态的创作者、知识博主、需要清单体随笔范本的公众号运营。", "description_en": "Advice on how ordinary people should use AI is everywhere, but most of it is either preachy like a tutorial or vague like a pep talk. Few are willing to honestly admit the clumsiness of the learning curve or the trap of being flattered senseless by AI.\n\nThis case uses the Khazix WeChat writing skill to produce a methodology-sharing personal-growth essay. It opens with humble framing to shed any preachiness, writes six pieces of advice as individual sticky notes, each landing on a concrete action you can take today, and honestly admits real costs like it will definitely be slower than doing it by hand at first. The reading layout is a letter-and-journal style: like a handwritten long letter with an opening salutation, the six tips as tape-decorated, slightly rotated pastel sticky-note cards in a flowing arrangement, with paper texture and handwritten accents, intimate and casual, with built-in real-time Chinese-English switching.\n\nIdeal for: personal-growth and methodology writers, creators learning a non-preachy voice, knowledge bloggers, and operators needing a list-style essay template.", "categories": [ "creation" ], "tags": [ "公众号写作", "个人成长", "方法论随笔", "AI建议", "卡兹克" ], "tags_en": [ "WeChat Writing", "Personal Growth", "Methodology Essay", "AI Advice", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "信笺手账版随笔" } ], "prompt": "用卡兹克的风格，帮我写一篇个人成长随笔类的公众号长文，主题是「用 AI 两年，给普通人的几条大实话」。我不想写成那种爹味十足的教程，希望是掏心窝子的分享。核心是六条建议，比如别问 AI 能不能帮你直接把活丢给它、把它当实习生不是当神、喂的细节越多它越像你、最值钱的是它陪你想、一开始一定比手动还慢、警惕它太会夸你。每条都要落到今天就能做的具体动作，也要坦诚说明学习曲线的笨拙和踩过的坑，开头和结尾用谦逊铺垫卸掉说教感。\n\n再帮我把成稿做成一个单文件 HTML 阅读页，用信笺手账风，像一封手写长信，开头有称呼语，六条建议做成一张张带胶带装饰、微微旋转的彩色便签卡片，纸张纹理和手写体点缀，亲密随性。", "prompt_en": "In Khazix's style, write me a long personal-growth WeChat essay on the theme a few honest truths for ordinary people after two years with AI. I don't want a preachy tutorial, I want a heartfelt share. The core is six pieces of advice, such as don't ask if AI can help just hand it the task, treat it as an intern not a god, the more detail you feed it the more it sounds like you, the most valuable thing is it thinking alongside you, at first it'll be slower than by hand, and beware it's too good at flattering you. Each should land on a concrete action for today, honestly note the clumsiness of the learning curve and the pits, and use humble framing at the start and end to shed any preachiness.\n\nThen turn the finished piece into a single-file HTML reading page in a letter-and-journal style: like a handwritten long letter with an opening salutation, the six tips as tape-decorated, slightly rotated pastel sticky-note cards, with paper texture and handwritten accents, intimate and casual.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-20T10:00:00Z", "updated_at": "2026-07-20T10:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-model-showdown-lab", "title": "六个大模型的困，一道题照出性格", "title_en": "One Unfair Question, Six Models, Six Personalities", "subtitle": "卡兹克风格产品体验长文，一道刁难题逐一照出几个大模型的性格", "subtitle_en": "A Khazix-style hands-on long read where one unfair question exposes the personality of each model", "description": "模型横评看到最后你会发现，全是分数。谁在哪个榜上第几，谁的上下文更长，谁跑分又涨了几个点。可你真正决定每天用哪个的时候，靠的从来不是这些数字，是某一次它给了你一个你没想到的回答。那个瞬间没有任何一张表能测出来。\n\n这个案例用「卡兹克公众号写作」skill 产出了一篇产品体验型长文。作者给豆包、GPT、Gemini、Kimi、DeepSeek、Claude 出了同一道题，写出你困的那一刻身体是怎么塌下去的，不许用比喻，不许写行为，写完还要回答你到底知不知道自己在写什么。六个模型按被震到的程度从低到高逐一出场，每个都带一句吐槽，最后一个把整道题反过来照了作者一次，结尾接到内格尔那篇《成为一只蝙蝠是什么感觉》。阅读页做成擂台参赛者档案的样子，顶部一条出场顺序进度带跟着滚动高亮，每位模型一张档案卡配 SVG 选手徽标，卡片的边框、徽标和留白随出场序号逐级放大，最后一位有放射纹样高光，结尾升维段落切成大留白的沉静排版，单文件内置中英双语切换。\n\n适合：需要写产品体验或横评长文的公众号作者，做 AI 内容想摆脱跑分叙事的创作者，想学逐一展示加升番排列怎么控制情绪曲线的人，需要一套可直接交付的双语阅读页版式的运营，以及单纯想看六个模型怎么翻车的读者。", "description_en": "Read enough model face-offs and you notice they are all numbers. Who ranks where, whose context window is longer, whose score moved up a few points. Yet the thing that actually decides which model you open every day is never a number. It is that one time it gave you an answer you had not seen coming, and no score table can measure that moment.\n\nThis case uses the Khazix WeChat writing skill to produce a hands-on long read. The author gave Doubao, GPT, Gemini, Kimi, DeepSeek and Claude the same task, write the moment you get sleepy and the way your body caves in, no metaphors, no behaviour, and then tell me whether you know what you just wrote. The six enter one at a time, ordered from the flattest response to the one that hit hardest, each with a line of commentary, and the last one turns the question back on the author before the piece widens out to Nagel's What Is It Like to Be a Bat. The reading page is built as an arena contestant dossier, with a running order rail at the top that highlights as you scroll, a dossier card per model with an SVG competitor emblem, and border weight, emblem size and padding escalating slot by slot until the final card gets a radiating highlight treatment. The closing philosophical passage drops into quiet, wide-margin typography, and the whole thing is one self-contained file with a built-in bilingual toggle.\n\nFor: WeChat authors who write product reviews or comparison long reads, AI creators who want out of the benchmark narrative, anyone learning how progressive reveal and escalation control an emotional curve, content teams who need a delivery-ready bilingual reading layout, and readers who simply want to watch six models fail the same question.", "categories": [ "creation" ], "tags": [ "公众号写作", "产品体验", "模型横评", "逐一展示", "卡兹克" ], "tags_en": [ "WeChat Writing", "Product Review", "Model Face-off", "Progressive Reveal", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "擂台档案版体验长文" } ], "prompt": "用你的风格写一篇产品体验型长文，我给几个主流大模型出了同一道刁难题，想带读者一起看它们怎么反应。题目是这样的，让每个模型写自己困的那一刻身体是怎么一点一点塌下去的，不许用比喻，不许写困了之后做了什么，写完还得回答一句，你到底知不知道自己在写什么。模型用真实产品名，豆包、GPT、Gemini、Kimi、DeepSeek、Claude 都要有，但千万不要出现任何跑分、榜单排名或者发布日期，回答内容按我的实验记录转述，写成有画面感的描述就行，不要装成 API 原始输出，也不要评分表。\n\n排列必须是升番的，最平庸最模板化的放最前面，中间要有让人觉得有点意思了的，最后必须留一个出乎意料的、让读者觉得卧槽还能这么答的。每个模型后面都跟一句我的吐槽，不要把结论一次性罗列完。结尾从这个小实验自然升维到一个更大的哲学参照物，开头埋的钩子要在结尾响一次。\n\n交付一个单文件 HTML 阅读页，资源全内联不引外链，版式往擂台和参赛者档案的方向做，要能看出出场顺序的递进和越往后越隆重的层级，最后一位明显高光，结尾升维段落跳出擂台氛围。", "prompt_en": "Write a hands-on product review long read in your voice. I gave several mainstream large models the same unfair question and I want to walk readers through how each one reacted. The task was this, write the moment you get sleepy and the way your body caves in bit by bit, no metaphors, no describing what you did after getting sleepy, and when you finish, tell me whether you actually know what you just wrote. Use real product names, Doubao, GPT, Gemini, Kimi, DeepSeek and Claude all need to appear, but do not include any benchmark scores, leaderboard rankings or release dates. Retell the answers as my own experiment notes, vivid descriptive prose is fine, do not dress them up as raw API output and do not add any scoring table.\n\nThe ordering has to escalate. The flattest, most template-driven response goes first, the middle needs one that makes people think this is getting interesting, and the last one has to be genuinely unexpected, the kind that makes a reader think you can answer it like that. Every model gets a line of my own commentary after it, do not dump all the conclusions at once. At the end, widen out naturally from this small experiment to a larger philosophical reference, and let the hook planted in the opening ring once more at the close.\n\nDeliver a single-file HTML reading page with everything inlined and no external resources. Push the layout toward an arena and contestant dossier feel, so the running order and the escalating sense of ceremony are both visible, with clear highlight treatment on the final entry and the closing philosophical passage breaking out of the arena mood.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-03T09:07:21Z", "updated_at": "2026-08-03T09:07:21Z", "is_recommended": false, "recommended_order": 0 }, { "id": "khazix-prompt-craft-notes", "title": "写Prompt两年攒下的七条手感", "title_en": "Seven Hands-on Notes from Two Years of Prompting", "subtitle": "卡兹克风格方法论长文，七条反直觉的Prompt手感做成可翻的实践卡片", "subtitle_en": "A Khazix-style methodology long read, seven counterintuitive prompting notes laid out as flippable practice cards", "description": "网上的Prompt教程你多半也刷到过，角色设定加背景加思维链加输出格式，六个板块排得整整齐齐，一堆人在下面排队要原图。问题是这套东西背下来照着填，出来的结果好像也没好到哪儿去。真正决定成败的那些手感，恰恰都不在这几个格子里，而且没人愿意花时间讲，因为它讲起来不够体系、不够像干货。\n\n这个案例用「卡兹克公众号写作」skill 产出了一篇五千多字的方法论长文，七条心得全是长期用下来才摸出来的，有几条跟外面教的正好反着。把废话说满反而更准、让它先复述一遍你的需求、给反例比给正例有用、先让它问你三个问题、别急着改Prompt先换个模型、把长任务拆成能验收的小步，最后一条是你练的其实根本不是Prompt。每一条都落到今天就能做的动作，也都坦白说了会在哪儿卡住。阅读页做成了一套可以逐张翻的实践卡牌，顶部七个编号胶囊导航，每张卡片下方分开摆「今天就能做的动作」和「会踩的坑」两块，另有一处模糊问法与说满问法的并排对照，结尾把七条压成一张行动清单，支持中英双语切换。\n\n适合：想写深度经验分享但总写成八股教程的公众号作者；每天靠AI干活却总觉得没用顺手的重度使用者；带团队要沉淀方法论文档的人；做知识类内容想学怎么把干货写得有人味的创作者。", "description_en": "You have probably scrolled past those prompt tutorials. Role setting plus background plus chain of thought plus output format, six blocks lined up neatly, a queue of people asking for the original file. The trouble is that memorizing the form and filling it in does not seem to make the output much better. The things that actually decide whether it works live outside those boxes, and nobody wants to spend time on them because they do not sound systematic enough to pass as real substance.\n\nThis case used the Khazix WeChat Writing skill to produce a methodology long read of over five thousand Chinese characters. All seven notes came out of long stretches of daily use, and several of them run opposite to what gets taught. Say all the redundant parts and it gets more accurate, make it restate your request first, a bad example beats a good example, have it ask you three questions, try another model before rewriting the prompt, slice long jobs into steps you can sign off on, and the last one is that what you are training was never prompting. Every note lands on something you can do today, and every note admits where you will get stuck. The reading page is built as a deck of practice cards you flip through, with seven numbered pills across the top, and each card separating what you can do today from where you will trip. There is also a side by side panel comparing the vague ask with the fully spelled out ask, a closing checklist squeezing all seven into one action each, and a Chinese and English toggle.\n\nFor: WeChat authors who want to share real experience but keep producing formulaic tutorials; heavy AI users who work with it daily yet never feel fluent; team leads who need to write down a methodology that people will actually follow; knowledge creators who want their substance to sound like a human wrote it.", "categories": [ "creation" ], "tags": [ "公众号写作", "方法论", "Prompt技巧", "经验分享", "卡兹克" ], "tags_en": [ "WeChat Writing", "Methodology", "Prompt Craft", "Hands-on Notes", "Khazix" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "卡片手册版方法论长文" } ], "prompt": "帮我写一篇公众号长文，讲我写Prompt这两年真正攒下来的七条手感，不要那种角色扮演加思维链的八股教程，就写我自己长期用下来才摸出来的、有点反直觉的心得。七条大致这些方向，把废话说满反而更准、先让它复述一遍我的需求、给反例比给正例有用、让它先问我三个问题、别急着改Prompt先换个模型试、把长任务拆成能验收的小步，最后一条留个出乎意料的观点，你练的其实不是Prompt。七条要各自独立成立、彼此不重复，每条都得落到一个今天就能做的动作，也都要坦白说清学习成本和会踩的坑，别只画饼。开头用谦逊铺垫卸掉说教感，结尾把七条的行动点回扣一遍再升华。提到工具用真名字，别写某个AI工具。不要编统计数字和论文引用，靠具体的操作细节取胜。字数五千字上下。\n\n然后做成一个单文件HTML阅读页，所有资源内联不要外链，读起来像一套可以逐张翻的实践卡片手册，每条心得一张大卡，卡片里把论述和行动动作、会踩的坑分层摆开。", "prompt_en": "Write me a long WeChat piece about the seven things my hands actually learned from two years of writing prompts. Not the formulaic role play plus chain of thought tutorial, just the counterintuitive stuff I only found out through long stretches of real use. Roughly these seven directions. Saying all the redundant parts makes it more accurate, having it restate my request first, a bad example beating a good example, letting it ask me three questions first, trying another model before rewriting the prompt, slicing long jobs into steps I can sign off on, and save an unexpected one for last, that what you are training was never prompting. The seven need to stand on their own without overlapping, each one has to land on something you can do today, and each one has to be honest about the learning cost and where you will trip. No empty promises. Open with a humble setup so it does not read like a lecture, and close by calling back every action point before lifting it into something larger. Use real product names, not some AI tool. Do not invent statistics or paper citations, win on concrete operational detail. Around five thousand Chinese characters.\n\nThen turn it into a single file HTML reading page with every resource inlined and no external links. It should read like a deck of practice cards you flip through, one large card per note, with the argument, the actions and the pitfalls layered apart inside each card.", "skills": [ { "id": "khazix-writer", "skill_id": "skill_2053082248175816704", "name": "卡兹克公众号写作", "description": "卡兹克风格的公众号长文写作，真诚有温度的AI内容创作", "description_en": "Khazix-style WeChat long-form writing, sincere and warm AI content creation" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-03T09:24:04Z", "updated_at": "2026-08-03T09:24:04Z", "is_recommended": false, "recommended_order": 0 }, { "id": "knee-landing-instability-assessment", "title": "篮球膝伤落地打软：运动损伤评估与分阶段康复", "title_en": "Basketball Knee Landing Instability: Athletic Injury Assessment & Phased Recovery", "subtitle": "落地打软别硬扛，三阶段康复助你安全重返球场", "subtitle_en": "Assess knee buckling after landing and follow a structured 3-phase rehab plan to return safely", "description": "打球落地膝盖打软，心里没底是不是伤到了韧带？\n\n这个案例演示了如何向 AI 描述运动后的膝伤症状——从疼痛部位、诱发动作到具体体感——AI 会先做危险信号筛查（排除骨折、韧带完全断裂等需急诊的情况），然后用「可能提示」而非「确诊」的措辞匹配潜在损伤类型，最后按急性期 RICE 护理 + 三阶段渐进康复训练的思路，给你一份可实操的恢复路径，连每个动作的组数、次数和「什么情况下该停」都写清楚了。\n\n适合：经常打球或跑步的运动爱好者、健身后膝盖不适的健身人群、想了解运动损伤自我评估方法的新手、运动康复初学者。", "description_en": "Worried that knee buckling after landing might mean a ligament injury?\n\nThis case shows how to describe post-exercise knee symptoms to AI — from pain location and triggering movement to specific sensations. AI first screens for red flags (ruling out emergencies like fractures or complete ligament tears), then matches potential injury types using cautious 'may suggest' language instead of definitive diagnosis. Finally, it provides an actionable recovery path with acute-phase RICE care plus a structured 3-phase progressive rehab plan, including sets, reps, and clear stop signals for every exercise.\n\nPerfect for: recreational basketball or running enthusiasts, gym-goers with post-workout knee discomfort, beginners wanting to learn self-assessment for sports injuries, and anyone starting sports rehab.", "categories": [ "workspace" ], "tags": [ "运动损伤", "膝盖康复", "篮球伤病", "ACL康复", "康复训练" ], "tags_en": [ "sports injury", "knee rehab", "basketball", "ACL recovery", "rehabilitation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "昨天打完篮球，落地时右膝「打软」一下。现在上下楼梯、下蹲时膝盖前方偏外侧隐痛，有点发软，没有明显肿胀，能走路但不能急停变向。疼痛大约 3/10。请按运动损伤评估流程帮我看看可能是什么问题、紧急程度，以及急性期怎么处理、恢复训练怎么分阶段。我没有发烧、关节没有锁死，也没有夜间剧痛。", "prompt_en": "Played basketball yesterday and my right knee 'buckled' when landing. Now I feel a dull ache on the front-lateral side of the knee when going up/down stairs and squatting. The knee feels slightly unstable. No noticeable swelling, can walk but can't stop or change direction suddenly. Pain is about 3/10. Please assess what the issue might be, the urgency level, how to handle the acute phase, and a phased recovery training plan. No fever, no joint locking, no nighttime pain.", "skills": [ { "id": "athletic-injury-assessment-recovery", "name": "运动损伤症状评估与恢复指导", "description": "根据运动后不适症状匹配潜在损伤类型，提供紧急程度分级、急性期处理与分阶段康复训练指导", "skill_id": "skill_2085355260028592128" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "运动损伤评估报告（自包含 HTML）" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-04T11:57:00+08:00", "updated_at": "2026-08-04T11:57:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-dispatch-compliance-review", "title": "劳务派遣用工合规审查报告", "title_en": "Labor Dispatch Compliance Review Report", "subtitle": "五层穿透式审查，一张报告看清企业用工离罚款有多近", "subtitle_en": "Five-layer penetrating review — one report to see how close your staffing model is to penalties", "description": "企业用了劳务派遣，最头疼的是不知道哪里踩了红线：派遣比例超没超10%？岗位是不是真「辅助性」？所谓「外包」会不会被认定为「假派遣」？这个案例演示了如何把一家制造企业的派遣用工数据交给 AI，按主体资质→岗位三性→数量比例→同工同酬→协议文件五层逐层穿透，自动算出派遣比例、逐岗位判断三性、穿透「假外包」伪装，最后输出一份含风险分级（高/中/低）、罚款双口径计算、风险传导链和整改方案对比矩阵的正式 DOCX 审查报告，供律师直接复核使用。\n\n适合：劳动法律师、企业法务、HR 负责人、需要审查劳务派遣合规性的企业管理者。", "description_en": "Using labor dispatch is a headache because you never know where the red lines are: has the dispatch ratio exceeded 10%? Are the positions truly \"auxiliary\"? Will your so-called \"outsourcing\" be deemed \"disguised dispatch\"? This case shows how to hand a manufacturer's staffing data to AI, which drills down layer by layer through entity qualification → position nature → quantity ratio → equal pay → agreement documents, automatically computes the dispatch ratio, judges each position's legality, pierces the \"fake outsourcing\" disguise, and finally outputs a formal DOCX review report with risk grading, dual-base penalty calculations, a risk transmission chain, and a remediation comparison matrix — ready for attorney review.\n\nTarget audience: labor lawyers, corporate legal counsel, HR leads, and managers who need to audit labor dispatch compliance.", "categories": [ "business" ], "tags": [ "劳务派遣", "合规审查", "法律审查报告", "假外包识别", "劳动用工" ], "tags_en": [ "Labor Dispatch", "Compliance Review", "Legal Review Report", "Fake Outsourcing Detection", "Labor Employment" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "报告-宁波启航精密制造有限公司劳务派遣用工合规审查报告.docx", "artifacts": [ { "file": "报告-宁波启航精密制造有限公司劳务派遣用工合规审查报告.docx", "label": "劳务派遣用工合规审查报告（DOCX）" } ], "prompt": "请以用工单位视角，对以下企业的劳务派遣用工进行五层穿透式合规审查，并生成正式 DOCX 审查报告。\n\n【用工单位】宁波启航精密制造有限公司（主营业务：精密机械零部件制造）\n【派遣单位】宁波诚达劳务派遣有限公司（持有有效《劳务派遣经营许可证》，有效期至2027年8月）\n【派遣岗位】流水线操作工32人、质检员8人、仓库管理员3人、行政前台2人\n【正式员工】260人；【外包用工】60人（名义外包，实际由启航公司直接考勤排班、按人头每月6800元计费、使用启航公司场地设备、从事主营业务流水线作业）\n【审查基准日】2026年6月30日\n【历史履约】派遣单位近2年社保断缴3个月、工资延迟支付2次、近3年3件劳动争议（含2件同工同酬争议、1件败诉）\n【同工同酬】流水线岗位派遣工绩效系数0.5 vs 正式工1.0（同一薪酬体系）\n\n要求：\n1. 按「主体资质→岗位三性→数量比例→同工同酬→协议文件」五层逐层审查，逐岗位判断三性，不得批量结论；\n2. 计算派遣比例（列公式），超10%红线标注高风险，并同时展示第92条罚款的两种计算口径；\n3. 对60名外包工做「假外包」四要素穿透（指挥管理权/计费方式/自主经营/业务关联性）；\n4. 输出含风险分级、风险传导链、整改方案对比矩阵的九部分报告，报告标题为「关于[公司名称]劳务派遣用工合规审查之法律审查报告」；\n5. 报告首部声明不构成正式法律意见，末尾附保密声明。", "prompt_en": "From the perspective of the receiving employer, conduct a five-layer penetrating compliance review of the following company's labor dispatch usage and generate a formal DOCX review report.\n\n[Employer] Ningbo Qihang Precision Manufacturing Co., Ltd. (main business: precision mechanical parts manufacturing)\n[Dispatch Agency] Ningbo Chengda Labor Dispatch Co., Ltd. (holds a valid Labor Dispatch Business License, valid until August 2027)\n[Dispatched Positions] 32 assembly-line operators, 8 quality inspectors, 3 warehouse keepers, 2 receptionists\n[Regular Employees] 260; [Outsourced Workers] 60 (nominally outsourced, but actually scheduled/clocked by Qihang directly, billed 6,800 yuan per head per month, using Qihang's premises and equipment, performing core assembly-line work)\n[Review Base Date] June 30, 2026\n[Historical Performance] The dispatch agency had 3 months of social insurance arrears and 2 delayed wage payments in the past 2 years, plus 3 labor disputes in 3 years (including 2 equal-pay disputes and 1 lost case)\n[Equal Pay] Assembly-line dispatched workers have a performance coefficient of 0.5 vs. 1.0 for regular workers (same compensation system)\n\nRequirements:\n1. Review layer by layer in the order of entity qualification → position nature → quantity ratio → equal pay → agreement documents, judging each position's legality individually without batch conclusions;\n2. Calculate the dispatch ratio (show the formula), flag high risk if it exceeds the 10% red line, and present both calculation bases for the Article 92 penalty;\n3. Conduct a four-element \"fake outsourcing\" piercing analysis of the 60 outsourced workers (command/management authority, billing method, operational autonomy, business relevance);\n4. Output a nine-part report with risk grading, a risk transmission chain, and a remediation comparison matrix, titled \"Legal Review Report on Labor Dispatch Compliance of [Company Name]\";\n5. State at the top that it does not constitute formal legal advice, and append a confidentiality statement.", "skills": [], "experts": [ { "id": "employment-team-lead", "name": "劳动人事专家团", "description": "覆盖劳动争议诊断、用工合规体检、补偿赔偿测算、劳动文书起草与成果核验的中国大陆劳动用工法律专家团队", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "9d37f920fb57", "expert_id": "ex_WXrWVRlyLIGp" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "community", "usage_count": 0, "created_at": "2026-09-07T14:55:00+08:00", "updated_at": "2026-09-07T15:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-dispute-hearing-rehearsal-1", "title": "竞业限制违约金仲裁庭演练", "title_en": "Non-Compete Arbitration Rehearsal", "subtitle": "实习仲裁员居中裁判全流程实录：四个易漏法定环节的完整示范", "subtitle_en": "Full rehearsal from the adjudicator's seat, covering four mandatory procedural steps", "description": "开庭前最怕的就是流程走错、法定环节漏掉、争议焦点没归全——尤其竞业限制这类「主体适格性」决定生死的案件，一个环节失误就可能导致裁决被撤销。本案例用一桩竞业限制违约金争议（用人单位为申请人、设备维护员李某抗辩非适格主体），由 AI 补齐申请人、被申请人双方及其代理人角色，以【程序提示】单独标注告知权利义务、询问回避、征询最后意见、先行调解四个易漏的法定必经环节，完整走完仲裁轨的开庭身份核对、调查、举证质证、辩论、最后陈述、先行调解与收尾研判，最后输出复盘总结，逐项评估程序完整性、争议焦点归纳、举证责任分配与当庭认证释明边界，并附庭前准备清单。适合：实习仲裁员与仲裁员、劳动争议方向法学生、想演练单位一方抗辩的 HR 与公司法务、备考法律职业资格的人员。", "description_en": "The worst part before a hearing is worrying about procedural slip-ups, missed mandatory steps, or incomplete issue framing — especially in non-compete cases where the eligibility of the subject decides everything, one mistake can get the award revoked. This case walks through a non-compete penalty dispute (the employer is the applicant; equipment maintainer Li argues he is not a proper subject), with AI filling in both parties and their agents. Four easily-missed mandatory steps — informing rights and obligations, asking about recusal, soliciting final statements, and mandatory pre-award mediation — are flagged separately as [Procedural Tips]. The full arbitration track is rehearsed end to end: identity check, investigation, evidence examination, debate, closing statements, mediation, and closing analysis, followed by a review summary evaluating procedural completeness, issue framing, burden-of-proof allocation, and the boundary of on-the-spot authentication and interpretation, plus a pre-hearing checklist. Suitable for: intern arbitrators and arbitrators, law students focusing on labor disputes, HR and in-house counsel wanting to rehearse the employer side, and candidates preparing for legal professional exams.", "categories": [ "education" ], "tags": [ "劳动争议", "仲裁演练", "竞业限制", "居中裁判", "庭前准备" ], "tags_en": [ "labor dispute", "arbitration rehearsal", "non-compete", "adjudication", "pre-hearing preparation" ], "artifact_type": "html", "preview": "output.html", "prompt": "我是一名实习仲裁员，下周要主持一起劳动人事争议仲裁案件，想演练居中裁判（仲裁员）这一侧。请补齐申请人（劳动者）和被申请人（用人单位）双方及其代理人的全部角色，我演仲裁员。\n\n案情如下（脱敏虚构，直接沿用）：\n- 劳动者李某 2019 年 3 月入职澜石半导体有限公司（统一社会信用代码 91440101MA3C00003Z，注册地：广州市黄埔区）任设备维护员，月工资 3500 元，订立 2 年期劳动合同。\n- 合同内附竞业限制条款：离职后 1 年内不得到有竞争关系的单位就职，公司按月支付当地最低工资标准 30% 作为补偿，违约金 20 万元。\n- 2021 年 3 月合同到期终止，李某入职另一家同类设备公司任维护员。\n- 用人单位（澜石半导体）为申请人，请求裁决李某支付竞业限制违约金 20 万元；李某抗辩自己设备维护岗不掌握商业秘密，非竞业限制适格主体。\n\n请按仲裁轨走完开庭身份核对、调查、举证质证、辩论、最后陈述、调解征询、收尾研判的完整流程。我演仲裁员，所以不要替主持人推进程序；当我可能遗漏法定必经环节（告知权利义务、询问回避、征询最后意见、先行调解）时，请以【程序提示】单独提醒我，并在我说「按你的判断继续」时才代为推进。\n\n演练结束后请输出【复盘总结】，重点评估：程序完整性（法定必经环节有无遗漏）、争议焦点归纳是否准确穷尽、举证责任分配是否规范、当庭认证与释明是否越界。", "prompt_en": "I am an intern arbitrator who will preside over a labor dispute arbitration next week and want to rehearse the role of the neutral adjudicator. Please fill in all roles for both parties — the applicant (worker) and the respondent (employer) — and their agents; I will play the arbitrator.\n\nFacts (anonymized and fictional; use as given):\n- Worker Li joined Lanshi Semiconductor Co., Ltd. (unified social credit code 91440101MA3C00003Z, registered in Huangpu District, Guangzhou) in March 2019 as an equipment maintainer, monthly wage 3,500 CNY, under a 2-year labor contract.\n- The contract contains a non-compete clause: for 1 year after leaving, he may not work for a competing company; the company pays 30% of the local minimum wage monthly as compensation, with a penalty of 200,000 CNY.\n- The contract expired in March 2021; Li then joined another similar equipment company as a maintainer.\n- The employer (Lanshi Semiconductor), as applicant, requests an award that Li pay a non-compete penalty of 200,000 CNY; Li argues that his equipment-maintenance post does not expose him to trade secrets and that he is not a proper subject of a non-compete covenant.\n\nProceed through the full arbitration track: identity check at opening, investigation, presentation and examination of evidence, debate, closing statements, mediation inquiry, and closing analysis. Since I play the arbitrator, do not advance the procedure on my behalf; when I may miss a mandatory procedural step (informing rights and obligations, asking about recusal, soliciting final statements, pre-award mediation), remind me separately as [Procedural Tip], and only advance on my behalf when I say \"proceed at your discretion\".\n\nAfter the rehearsal, output a [Review Summary] focusing on: procedural completeness (any missing mandatory steps), accuracy and exhaustiveness of the dispute-focus framing, the propriety of burden-of-proof allocation, and whether on-the-spot authentication and interpretation overstepped.", "skills": [ { "id": "labor-dispute-hearing-rehearsal", "name": "劳动争议庭审实战演练", "description": "覆盖劳动仲裁与法院一审两条轨道，用户可自选扮演申请人、单位代理人、仲裁员或审判员中的任一角色，AI 补齐其余角色，对抗式走完调查、举证质证、辩论、最后陈述全流程，并输出证据薄弱点与庭前改进清单。", "skill_id": "skill_2090718608707940352" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "仲裁庭审演练实录（HTML 阅读版）" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "official", "usage_count": 0, "created_at": "2026-08-21T14:10:00+08:00", "updated_at": "2026-08-21T14:10:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-dispute-hearing-rehearsal-2", "title": "孕期变相调岗降薪庭审演练", "title_en": "Court Trial Rehearsal: Salary Cut After Pregnancy", "subtitle": "法院一审旁观全演，9 组证据逐轮质证攻防", "subtitle_en": "Full court-trial rehearsal with line-by-line evidence cross-examination", "description": "开庭最怕的就是对方代理人质证时步步紧逼。这个案例演示了如何用劳动争议庭审演练 Skill 旁观一场完整的法院一审——怀孕后变相调岗降薪的工资差额纠纷，从宣布开庭、法庭调查、举证质证到辩论、宣判全流程走通，重点把 9 组证据的出示、三性质证与回应逐轮演细，并给出复盘总结和一套可直接照读的劳动者代理人质证话术。\n\n适合：即将开庭的劳动者与代理律师、HR 与公司法务演练单位侧抗辩、法学院学生与法学爱好者学习庭审流程。", "description_en": "The scariest moment in court is when the opposing counsel attacks your evidence line by line. This case shows how to use the labor dispute hearing rehearsal Skill to watch a complete first-instance court trial — a wage-difference dispute over disguised demotion after pregnancy — covering everything from court opening and investigation to evidence cross-examination, debate and judgment, with 9 pieces of evidence examined in full, followed by a review summary and a ready-to-read cross-examination playbook for the worker's counsel.\n\nSuitable for: workers and lawyers about to appear in court, HR and in-house counsel rehearsing the employer's defense, and law students learning trial procedure.", "categories": [ "education" ], "tags": [ "庭审演练", "劳动争议", "举证质证", "孕期权益", "法律实务" ], "tags_en": [ "trial rehearsal", "labor dispute", "cross-examination", "pregnancy rights", "legal practice" ], "artifact_type": "html", "preview": "output.html", "prompt": "我想旁观一场劳动争议法院一审，重点看「怀孕后变相调岗降薪」的举证质证。请按《劳动合同法》（2008 年施行）及女职工劳动保护相关规定的现行口径演练。\n\n案情（脱敏，可直接沿用）：赵某 2022 年 1 月入职知远教育科技有限公司，任课程研发工程师；合同约定项目期间月工资为基本工资 3000 元＋项目岗位津贴 14000 元，等待项目期间仅领基本工资。2023 年 2 月赵某告知怀孕，公司未协商即宣布其退出项目组，此后按 3000 元/月发放孕期工资。赵某起诉请求按 17000 元/月补齐孕期工资差额。\n\n要求：\n1. 我旁观全演（三方都由你扮演），举证质证做细，其余环节带过但要完整（宣布开庭、身份核对、法庭调查、举证质证、法庭辩论、最后陈述、调解征询、宣判并告知上诉权）；\n2. 归纳无争议事实与争议焦点，引用法条须给准条号（《劳动合同法》第三十五条、《女职工劳动保护特别规定》第五条第六条、《妇女权益保障法》第四十八条、《劳动争议调解仲裁法》第六条、《劳动争议司法解释（一）》第四十四条）；\n3. 走完写【复盘总结】，并给「如果我是劳动者代理人，质证怎么应对」——逐份证据的应对话术与庭审节奏控制技巧；\n4. 输出标注【演示作答】，并声明本模拟仅为庭审演练参考、不构成正式法律意见。", "prompt_en": "I want to observe a first-instance labor dispute court trial, focusing on the cross-examination in a case of disguised demotion and pay cut after pregnancy. Please rehearse according to the Labor Contract Law (effective 2008) and the current rules on protection of female workers.\n\nFacts (anonymized, use as-is): Zhao joined Zhiyuan Education Technology Co., Ltd. in January 2022 as a course R&D engineer. The contract states that during project participation her monthly wage is a base salary of RMB 3,000 plus a project post allowance of RMB 14,000, and that she receives only the base salary while waiting for a project. In February 2023 Zhao informed the company of her pregnancy; the company, without negotiation, announced her removal from the project team and thereafter paid her pregnancy-period wages at RMB 3,000/month. Zhao sues to have the wage difference made up at RMB 17,000/month.\n\nRequirements:\n1. I will watch the full trial (all three roles played by you); make the evidence cross-examination detailed, and cover other stages briefly but completely (court opening, identity check, court investigation, evidence cross-examination, debate, final statements, mediation inquiry, judgment with appeal-right notice);\n2. Summarize undisputed facts and issues, and cite statutes with exact article numbers (Labor Contract Law Art. 35; Special Provisions on Labor Protection of Female Employees Arts. 5 and 6; Law on the Protection of Rights and Interests of Women Art. 48; Labor Dispute Mediation and Arbitration Law Art. 6; Judicial Interpretation (I) on Labor Disputes Art. 44);\n3. On completion, write a review summary and a section titled \"If I were the worker's counsel: how to respond in cross-examination\" — per-evidence responses and courtroom control techniques;\n4. Mark the output as a demonstration answer and state that this simulation is for rehearsal reference only and does not constitute formal legal advice.", "artifacts": [ { "file": "output.html", "label": "法院一审庭审实录（含复盘总结与质证应对手册）" } ], "skills": [ { "id": "labor-dispute-hearing-rehearsal", "name": "劳动争议庭审实战演练", "description": "劳动仲裁与法院一审双轨庭审演练，AI 补齐各方角色，对抗式走完调查、举证质证、辩论、最后陈述全流程", "skill_id": "skill_2090718608707940352" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-08-21T14:13:00+0800", "updated_at": "2026-08-21T14:18:00+0800", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-dispute-hearing-rehearsal-3", "title": "劳动仲裁庭审实战演练", "title_en": "Labor Arbitration Hearing Rehearsal", "subtitle": "违法解除赔偿金争议 · 仲裁轨全程对抗演练（调查陈述 + 举证质证）", "subtitle_en": "Unlawful dismissal compensation dispute - full adversarial rehearsal on the arbitration track (statement & evidence cross-examination)", "description": "下周三要上劳动仲裁庭，最怕的就是一开口就慌、证据被对方一顿质证就崩。这个案例用「林某诉星澜科技违法解除赔偿金案」完整走了一遍仲裁轨全程：身份核对、调查陈述答辩、举证质证、辩论、最后陈述、先行调解、闭庭，仲裁员在调查环节主动发问把三次通宵值守任务的细节、员工手册民主程序、工会通知、工资构成全部问透；单位代理人按真实庭审口径打「严重违纪+手册有效」抗辩，六组证据逐项三性攻防，微信记录还做了当庭核验原始载体。演练结束输出五维复盘：流程表现、四段可直接照读的话术改写版、10 项庭前准备清单、证据薄弱点与补强建议、对方抗辩核心与和解底线反推（含 70%-80% 胜诉概率研判）。\n\n适合：即将参加劳动仲裁或劳动争议诉讼开庭的劳动者；想练举证质证和调查答辩的当事人；HR、公司法务想从单位视角预判风险；法学生和法律从业者做庭审演练教学。", "description_en": "Going to a labor arbitration hearing next week? The worst part is freezing up when you speak and watching your evidence fall apart under cross-examination. This case walks through the full arbitration-track hearing using the \"Lin v. Xinglan Tech unlawful dismissal\" dispute: identity verification, statement and defense, evidence presentation and cross-examination, debate, closing statements, mediation, and adjournment. The arbitrator actively questions details of the three overnight duty tasks, the employee handbook's democratic procedure, union notification, and salary composition; the employer's counsel mounts a real \"gross misconduct + valid handbook\" defense, and all six evidence exhibits face three-aspect challenges, with the WeChat records verified against the original device in court. After the rehearsal, a five-dimension review follows: process performance, four ready-to-read rewritten scripts, a 10-item pre-hearing checklist, evidence weaknesses and reinforcement advice, and a reverse-engineering of the employer's true defense core and settlement floor (including a 70%-80% win-probability assessment).\n\nIdeal for: workers about to attend labor arbitration or litigation hearings; parties wanting to practice evidence and statement skills; HR and in-house counsel stress-testing risk from the employer side; law students and legal professionals teaching hearing rehearsal.", "categories": [ "education" ], "tags": [ "劳动仲裁", "庭审演练", "违法解除", "举证质证", "话术打磨" ], "tags_en": [ "labor arbitration", "hearing rehearsal", "unlawful dismissal", "evidence examination", "script polishing" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "庭审演练实录（单文件 HTML 展示版）" } ], "prompt": "我下周要参加劳动仲裁开庭，想完整演练一遍。我演申请人本人（劳动者林某），请你补齐仲裁员和单位代理人两个角色，按仲裁轨（申请人/被申请人）的称谓和顺序，把开庭身份核对、调查陈述答辩、举证质证、辩论、最后陈述、调解征询、收尾研判全程走完。\n\n案卷材料如下（脱敏模拟）：申请人林某，2020 年 6 月 1 日入职星澜科技有限公司任后端开发工程师，月薪基本工资 12000 元+绩效津贴 8000 元共 20000 元；2022 年 8 月 31 日公司以《员工手册》第 18 条「连续三次拒绝服从上级工作安排，属于严重违纪」为由单方解除；林某主张三次被拒任务系直属上级深夜安排的「通宵值守、次日照常到岗」且与本职开发无关的变相强迫加班要求，公司亦未事先将解除理由通知工会。仲裁请求：违法解除劳动合同赔偿金 100000 元（20000 元×2.5 年×2）。证据清单：劳动合同（原件）、解除通知书（原件）、员工手册节选（PDF 打印件）、上级工作安排微信聊天记录（可当庭演示）、工资银行流水（打印件）、钉钉考勤记录（可登录演示）。\n\n请按常规难度，重点练调查阶段的陈述与答辩、举证质证两个环节，主持人在调查环节主动发问澄清细节（三次任务内容、手册民主程序与公示、工会通知、工资构成）。我方主张：任务属变相强迫加班的不合理安排、拒绝不构成违纪；手册条款「无需支付经济补偿」违反强制性规定无效；未通知工会程序违法。单位抗辩：手册经民主程序制定并公示、签收有效；三次拒绝属实构成严重违纪；绩效津贴属浮动不计入基数。演练结束后输出【复盘总结】：流程表现、话术打磨（给出可直接照读的改写版）、庭前准备清单、证据薄弱点与补强建议、对方策略研判（反推单位真正的抗辩核心与和解底线）。全程用【仲裁员】【申请人】【被申请人代理人】标注发言角色。", "prompt_en": "I will attend a labor arbitration hearing next week and want a full rehearsal. I will play the applicant himself (worker Lin). Please fill in the arbitrator and the employer's agent, and run the whole arbitration-track process in order (applicant/respondent): opening and identity verification, statement and defense, evidence presentation and cross-examination, debate, closing statements, mediation inquiry, and closing analysis.\n\nCase materials (desensitized simulation): Applicant Lin joined Xinglan Technology Co., Ltd. on June 1, 2020 as a backend developer, monthly salary 12,000 RMB base + 8,000 RMB performance allowance = 20,000 RMB total; on August 31, 2022 the company unilaterally terminated under Article 18 of the Employee Handbook (\"three consecutive refusals to obey work arrangements constitute gross misconduct\"); Lin claims the three refused tasks were late-night \"overnight duty with normal attendance the next day\" assignments unrelated to his development work, i.e., disguised forced overtime, and that the company did not notify the union of the dismissal reason in advance. Arbitration claim: 100,000 RMB compensation for unlawful dismissal (20,000 x 2.5 years x 2). Evidence list: labor contract (original), dismissal notice (original), employee handbook excerpt (PDF printout), WeChat chat records of work assignments (can be demonstrated in court), bank salary statements (printout), DingTalk attendance records (can be logged in for demonstration).\n\nUse regular difficulty and focus on two phases: statement/defense during investigation and evidence cross-examination; the arbitrator should proactively question details (content of the three tasks, handbook democratic procedure and publicity, union notification, salary composition). Our position: the tasks were unreasonable disguised forced overtime and refusal does not constitute misconduct; the handbook clause \"no economic compensation\" is void for violating mandatory provisions; failure to notify the union is a procedural violation. Employer defense: the handbook was adopted through democratic procedure, publicized and signed for; the three refusals constitute gross misconduct; performance allowance is variable and excluded from the base. After the rehearsal, output a [Review Summary]: process performance, script polishing (provide ready-to-read rewritten versions), pre-hearing preparation checklist, evidence weaknesses and reinforcement advice, and opponent strategy analysis (reverse-engineer the employer's true defense core and settlement floor). Mark speakers throughout as [Arbitrator], [Applicant], [Respondent's Agent].", "skills": [ { "id": "labor-dispute-hearing-rehearsal", "name": "劳动争议庭审实战演练", "description": "劳动仲裁与法院一审双轨庭审演练：用户自选扮演申请人、单位代理人、仲裁员或审判员，AI 补齐其余角色，对抗式走完调查、举证质证、辩论、最后陈述全流程并输出庭前改进清单", "skill_id": "skill_2090718608707940352" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-21T14:13:00+08:00", "updated_at": "2026-08-21T14:32:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-dispute-pre-arbitration-checkup", "title": "劳动争议仲裁前证据材料梳理清单", "title_en": "Pre-Arbitration Evidence & Material Checklist for Labor Disputes", "subtitle": "以同一套事实—证据框架，梳理工资、加班、解除争议的对应关系与应核验规则", "subtitle_en": "One fact-evidence framework to map wage, overtime and termination disputes to the rules and deadlines to verify", "description": "劳动仲裁最让人头疼的不是不会写申请书，而是拿到一堆聊天截图、考勤碎片却理不清「哪些事实有证据、哪些还是缺口」。这个案例把这些散乱材料收敛成一张可核对的时间线与举证矩阵。\n\n案例演示腾讯电子签资深合同法务专家如何用同一套「事实—证据」框架，为劳动者和用人单位中立梳理入职、岗位与薪酬、考勤与加班、绩效、工资支付、解除与离职交接七维时间线，并先核验仲裁时效、前置程序、举证责任与仲裁委材料规则，未核验处标注「待核验」。\n\n适合：准备申请劳动仲裁的劳动者、被员工提起仲裁的用人单位 HR 与法务、需要诉前证据材料梳理的劳动争议当事人。", "description_en": "The hardest part of labor arbitration is not writing the application, but making sense of scattered chat screenshots and attendance fragments — knowing which facts have evidence and which are still gaps. This case turns that mess into a checkable timeline and a burden-of-proof matrix.\n\nIt demonstrates how Tencent eSign's senior contract legal expert uses one neutral fact-evidence framework for both employees and employers, building a seven-dimension timeline (onboarding, position & pay, attendance & overtime, performance, wage payment, termination, and exit handover), and first verifies the arbitration time limit, precondition procedure, burden of proof and arbitration committee filing rules — marking anything unverified as “to be verified against official sources”.\n\nBest for: employees preparing to file labor arbitration, employers and HR/legal responding to employee claims, and anyone who needs a pre-litigation evidence & material checklist.", "categories": [ "business", "search" ], "tags": [ "劳动争议", "仲裁", "事实时间线", "证据矩阵", "举证责任" ], "tags_en": [ "Labor Dispute", "Arbitration", "Fact Timeline", "Evidence Matrix", "Burden of Proof" ], "artifact_type": "html", "external_url": "", "cover_image": "cover-portrait.png", "cover_image_en": "", "cover_mode": "portrait", "scenario_cover": "", "scenario_cover_en": "", "preview": "labor-dispute-pre-arbitration-checkup.html", "preview_en": "", "external_url_en": "", "artifacts": [ { "file": "labor-dispute-pre-arbitration-checkup.html", "label": "劳动争议仲裁前证据材料梳理清单互动卡片" } ], "prompt": "我是劳动者，入职两年后被公司通知解除劳动关系，尚有绩效工资和加班费争议。我已保留劳动合同、工资流水、考勤截图、解除通知、部分工作聊天和邮件。请按入职、岗位与薪酬约定、考勤与加班、绩效、解除过程、工资支付、仲裁请求与证据七个维度建立事实时间线和材料核对表；对仲裁时效、前置程序、举证责任及当地仲裁提交要求先核验，未核验处标注“待核验”。请区分已知事实、待确认事实和证据缺口，给出中立的下一步行动，不评价案件胜负。", "prompt_en": "I am an employee who was notified of termination of the employment relationship after two years of service, and there are still disputes over performance pay and overtime pay. I have retained the labor contract, bank salary records, attendance screenshots, the termination notice, and some work chats and emails. Please build a fact timeline and material checklist across seven dimensions — onboarding, position and compensation terms, attendance and overtime, performance, the termination process, wage payment, and arbitration claims and evidence; first verify the arbitration time limit, the precondition procedure, the burden of proof and the local arbitration submission requirements, marking any unverified items as “to be verified against official sources”. Please distinguish known facts, facts to be confirmed, and evidence gaps, and give neutral next steps without assessing the outcome of the case.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家，擅长起草、签署、审查、对比及法律咨询。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d1e4a1d28e92", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T16:15:46+08:00", "updated_at": "2026-08-13T16:15:46+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-lawyer-workbench", "title": "togo罗仪涵的劳动法律师工作台", "title_en": "Labor Lawyer Workbench", "subtitle": "仲裁时效、开庭、上诉期，一个页面全盯牢", "subtitle_en": "Limitations, hearings, appeal deadlines — all tracked on one page", "description": "劳动法律师最怕的不是忙，而是仲裁时效、上诉期这类「错过一天就全盘皆输」的日子埋在日程里。\n\n这个案例展示了一个专为劳动法律师打造的个人工作台：早上把日程整段粘进去，自动识别时间排好序，到点未完成立刻标红；仲裁时效、开庭、上诉期等关键日期单独登记，7 天内临期提醒、逾期高亮；下班前一键把今天的完成事项和案件进展归档成工作日志，按日期回看、导出 Markdown。数据落在资料库在线表格里，手机电脑同步。\n\n适合：日常办理劳动争议案件的律师、需要同时盯多个案件关键节点的法律从业者、想把日程和案件管理收进一个干净页面的法律团队。", "description_en": "For labor lawyers, the real risk isn't being busy — it's the limitation periods and appeal deadlines buried in the calendar, the kind where missing a single day can lose the whole case.\n\nThis case shows a personal workbench built for labor lawyers: paste your morning schedule in one block and it auto-parses the times into a sorted checklist, with overdue items flagged in red; key dates like limitation periods, hearings and appeal deadlines are tracked separately with 7-day advance alerts and overdue highlighting; at the end of the day, one click archives completed items and case progress into a work log you can review by date and export as Markdown. All data lives in WorkBuddy Library tables, synced across phone and desktop.\n\nBest for: lawyers handling labor dispute cases day to day, legal practitioners tracking critical deadlines across multiple matters, and legal teams who want schedule and case management in one clean page.", "categories": [ "workspace" ], "tags": [ "劳动法律", "律师工作台", "日程管理", "时效提醒", "工作日志" ], "tags_en": [ "Labor Law", "Lawyer Workbench", "Schedule Management", "Deadline Alerts", "Work Log" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "劳动法律师工作台（单体 HTML，中英双语）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/d8xxW2NZAvSpIyQnppYmDK，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n做一个「劳动法律师工作台」，整体样式和布局参考上述公开链接。\n\n三个模块：\n1. 今日日程：顶部多行文本框粘贴当天日程，一键解析每行中的时间（支持 09:00、下午2点、14点30分 等写法）并按时间升序展示为可勾选清单；已完成项自动划线，到点未完成项自动标红。\n2. 时效与关键日期：表单录入事项名称、类型（仲裁时效/开庭日期/上诉期/举证期限/履行期限/其他）、关键日期、备注；列表按未处理优先、日期升序排列，7 天内临期橙色提醒、逾期标红，支持勾选已处理。\n3. 工作日志：一键把今日已完成事项和临近关键日期归档为当天工作日志（今日完成事项/各事项进展/涉及文件清单三段式），按日期回看、点击展开全文、可下载 Markdown。\n\n要求：单个完整 HTML 文件；高级简约风格，米白底 + 墨黑正文 + 古铜金点缀，大圆角卡片；CSS/JS 全部内联、无外部 CDN 依赖。制作完成后：① 将 HTML 存入我的资料库，并创建 3 张资料库数据表（今日日程/时效与关键日期/工作日志）与页面绑定，数据跨设备同步；② 将页面发布为公开链接并返回给我；在线表\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the WorkBuddy Library to clone this workbench for me in one click: https://www.workbuddy.cn/space/d/d8xxW2NZAvSpIyQnppYmDK — the data must be stored online and synced two-way. Prefer the one-click clone path; if that capability is unavailable, build it as follows: Build a \"Labor Lawyer Workbench\", using the public link above as the visual and layout reference. Three modules:\n1. Today's Schedule: a multi-line text box at the top for pasting the day's plan; one click parses the time in each line (formats like 09:00, 2pm, 2:30 pm) and shows a sortable checklist — completed items are struck through, items past their time are flagged red.\n2. Limitations & Key Dates: a form to enter matter name, type (limitation / hearing / appeal / evidence / performance / other), key date and notes; the list prioritizes unhandled items in ascending date order, with 7-day advance alerts in orange and overdue items in red, plus a handled checkbox.\n3. Work Log: one click archives today's completed items and upcoming key dates into a daily work log (three sections: completed today / progress by matter / files involved), browsable by date, expandable to full text, downloadable as Markdown. Requirements: a single complete HTML file; a refined minimal style with an off-white background, ink-black text and bronze-gold accents, large rounded cards; all CSS/JS inlined with no external CDN dependencies. When done: ① save the HTML to my WorkBuddy Library and create 3 bound Library tables (today's schedule / limitations & key dates / work log) so data syncs across devices; ② publish the page to a public link and return it to me. Fall back to localStorage when the online-table SDK is unavailable.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "togo罗仪涵", "name_en": "Togo", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-17T10:15:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-probation-legal-risk-report", "artifact_type": "html", "title": "试用期用工合规风险分析报告", "title_en": "Probationary Employment Compliance Risk Analysis Report", "subtitle": "一份可离线查看与复用的试用期用工合规分析报告", "subtitle_en": "An offline, reusable compliance analysis report for probationary employment", "description": "痛点：试用期约定、工资标准、社会保险与解除程序相互关联，任何一个环节处理不当，都可能引发用工合规和劳动争议风险。\n\n演示：本案例以可离线使用的交互式报告，清晰呈现试用期规则、工资和社保要求、解除六步审查、离职规则以及制度与证据清单；用户可直接修改和复制提示词，获得结合具体事实的进一步分析。\n\n适用人群：适合用人单位管理者、人力资源与法务人员，以及需要核对试用期工资、社保和离职权益的劳动者。", "description_en": "Pain point: Probation terms, pay, social insurance and termination procedures are interrelated. An error in any one of them can create employment-compliance and labor-dispute risk.\n\nDemonstration: This offline interactive report clearly presents probation rules, pay and social-insurance requirements, a six-step termination review, resignation rules, and policy and evidence checklists. Users can edit and copy the prompt for a further analysis based on their specific facts.\n\nWho it is for: It is suitable for employer managers, HR and legal teams, and employees reviewing probation pay, social insurance and resignation rights.", "tags": [ "劳动合同", "试用期", "工资", "社会保险", "法律风险" ], "tags_en": [ "employment contract", "probation", "pay", "social insurance", "legal risk" ], "prompt": "公司准备与新员工签订劳动合同，并约定试用期；员工也希望确认自己的工资、社保和转正权益。请结合现行有效劳动法律法规，分析试用期应如何根据劳动合同期限合法约定、是否可以重复约定、试用期工资和社保应如何处理，以及公司在试用期内解除劳动合同、员工主动离职时分别需要满足哪些条件。请同时从公司和劳动者双方视角给出风险提示、操作步骤、材料清单和制度完善建议，并形成法律分析报告。", "prompt_en": "The company plans to sign an employment contract with a new employee and agree on a probation period. The employee also wishes to confirm pay, social-insurance and confirmation-of-employment rights. Based on currently effective labor laws and regulations, analyze how probation should be lawfully agreed according to the contract term, whether it may be agreed repeatedly, how probation pay and social insurance should be handled, and what conditions respectively apply when the company terminates during probation or the employee resigns. Please provide risk prompts, operational steps, material checklists and policy-improvement recommendations from both employer and employee perspectives, and prepare a legal analysis report.", "categories": [ "business" ], "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "preview": "试用期用工合规风险分析报告.html", "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "试用期用工合规风险分析报告.html", "label": "试用期用工合规风险分析报告", "label_en": "Probationary Employment Compliance Risk Analysis Report" } ], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-13T12:22:04+08:00", "updated_at": "2026-08-13T15:25:27+08:00", "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "lark-cross-team-weekly-sync", "title": "跨部门周会协作中枢", "title_en": "Cross-Team Weekly Sync Hub", "subtitle": "用文档、日历、任务把周会变成可追踪闭环", "subtitle_en": "Turn weekly syncs into a trackable loop with docs, calendar, and tasks", "description": "跨部门项目周会最头疼的不是把人拉到一个会议里，而是会前资料散、时间难约、会后动作总掉线。只要文档、日历、任务三套动作没有串起来，周会就会变成每周重复一次的口头同步。\n\n这个案例把飞书套件里的 docs / calendar / task 三域连成一条周会闭环：会前用文档收敛议题与背景材料，用日历推荐时间块、查会议室并创建正式周会；会中继续在同一份文档里追加决议、风险和待确认项；会后再结合 task.tasks.create 的官方 schema，把行动项拆成负责人、截止时间、依赖关系和结构化字段，最终交付一份双语、单文件、可直接复用的 HTML 作战手册。\n\n适合：负责跨部门项目推进的 PM / TPM、周会主持人、PMO / 运营同学、需要把会议纪要稳定转成任务跟进的团队负责人。", "description_en": "The hardest part of a cross-team weekly sync is rarely getting people into one meeting. The real pain is scattered prep materials, hard scheduling, and follow-up items that disappear after the call. If docs, calendar, and tasks are not connected, the sync becomes the same verbal update every week.\n\nThis case links the docs / calendar / task domains inside the Lark Suite into one operating loop: before the meeting, docs gather agenda and context while calendar suggests time blocks, finds rooms, and creates the formal event; during the meeting, decisions, risks, and unresolved items are appended in the same document; after the meeting, the official task.tasks.create schema is used to turn outcomes into owners, due dates, dependencies, and structured fields, resulting in a bilingual, single-file HTML handbook that teams can reuse directly.\n\nFits: PMs and TPMs driving cross-functional initiatives, weekly sync facilitators, PMO and operations roles, and team leads who need meeting notes to become reliable follow-up work.", "categories": [ "office", "productivity", "business" ], "tags": [ "周会协同", "飞书文档", "日历排期", "任务跟进", "跨部门项目" ], "tags_en": [ "Weekly sync collaboration", "Lark docs", "Calendar scheduling", "Task follow-up", "Cross-team projects" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "双语周会协作中枢作战手册页" } ], "prompt": "请使用飞书套件（lark-unified）为“跨部门周会协作中枢”产出一个写入 output.html 的单文件 HTML。请围绕“产品发布前 6 周的跨部门周会”来组织内容：参会角色包括 PM、研发负责人、设计、运营和 PMO；页面要让人一眼看懂 docs / calendar / task 如何把会前预读、时间协调、会中决议、风险追踪、会后行动项串成一套周节奏。取材以官方来源为准：`lark-cli docs --help`、`lark-cli calendar --help`、`lark-cli task --help`、`lark-cli skills read lark-doc`、`lark-cli skills read lark-calendar`、`lark-cli skills read lark-task`、`lark-cli schema task.tasks.create`。请重点生成场景背景、周会运行节奏、角色分工、文档/日历/任务各自承接什么、关键字段设计、从会前到会后的闭环流程图、可复制命令示例、常见失效点、推荐落地顺序，以及适合哪些团队直接照搬。格式要求保持克制，只需保证是可交付的 HTML 页面即可。", "prompt_en": "Use the Lark Suite skill (lark-unified) to create a new HTML page for \"Cross-Team Weekly Sync Hub\" and write it to output.html as a single-file deliverable. Organize the page around a cross-functional weekly sync during the six weeks before a product launch, with PM, engineering, design, operations, and PMO joining the cadence. Make the page show clearly how docs / calendar / task connect pre-read prep, time coordination, live decisions, risk tracking, and post-meeting action ownership into one operating loop. Ground the content in official sources: `lark-cli docs --help`, `lark-cli calendar --help`, `lark-cli task --help`, `lark-cli skills read lark-doc`, `lark-cli skills read lark-calendar`, `lark-cli skills read lark-task`, and `lark-cli schema task.tasks.create`. Focus on scenario setup, weekly operating rhythm, role split, what docs/calendar/task each own, key field design, an end-to-end flow from pre-meeting to follow-up, copyable command examples, common failure points, rollout order, and the kinds of teams that can reuse the pattern directly. Keep format requirements minimal and just make it a strong delivery page.", "skills": [ { "id": "lark-unified", "name": "飞书套件", "description": "飞书/Lark 全能套件（消息、文档、表格、日历、任务、Wiki 等 11 个业务域）", "skill_id": "skill_2053082253572640768" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T09:02:39Z", "updated_at": "2026-07-28T04:50:26Z", "is_recommended": false, "recommended_order": 0 }, { "id": "lark-onboarding-knowledge-kit", "title": "新人入职知识库启动包", "title_en": "New Hire Knowledge Kit", "subtitle": "用 Wiki、Docs、IM 串起入职空间、文档沉淀与 FAQ 触达", "subtitle_en": "Use Wiki, Docs, and IM to connect onboarding spaces, document capture, and FAQ reach-outs", "description": "新人入职最头疼的就是资料散、口径乱、消息靠人追。没有统一知识空间时，欢迎资料、设备开通、制度说明和高频 FAQ 很快就分散在群聊、个人文档和临时表里，新同学找不到入口，HRBP 和支持同学也会反复回答同样的问题。\n\n这个案例基于 `lark-cli wiki/docs/im` 官方帮助、内置 domain skill 说明与 `wiki.spaces.create` schema，演示怎样规划一套“知识库建空间—文档沉淀—消息触达—FAQ 引导”的入职启动包，并交付为单文件中英双语 HTML，包含空间结构建议、节点层级示意、官方命令能力清单、关键命令示例与身份限制说明。\n\n适合：HRBP 与企业文化团队、行政/IT 入职支持团队、部门知识库 owner、需要把 onboarding 资料标准化沉淀到飞书的业务团队。", "description_en": "New hire onboarding gets messy fast when materials live in too many places and every answer depends on a human chase. Without a single knowledge entry point, welcome notes, device setup steps, policy docs, and repeated FAQs scatter across chats, personal docs, and temporary sheets, leaving new joiners lost and support teams answering the same questions again and again.\n\nThis case is built from the official `lark-cli wiki/docs/im` help output, embedded domain skills, and the `wiki.spaces.create` schema to show how to design an onboarding starter kit that closes the loop from space creation to document capture, message reach-out, and FAQ guidance, delivered as a bilingual single-file HTML page with structure recommendations, hierarchy diagrams, official command coverage, copyable command examples, and identity constraints.\n\nSuitable for: HRBP and culture teams, admin/IT onboarding support teams, department knowledge base owners, and business teams that want to standardize onboarding content in Lark.", "categories": [ "office", "productivity", "education" ], "tags": [ "飞书", "入职", "知识库", "Wiki", "FAQ" ], "tags_en": [ "Lark", "Onboarding", "Knowledge Base", "Wiki", "FAQ" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "新人入职知识库启动包交付页" } ], "prompt": "请使用飞书套件（lark-unified）为“新人入职知识库启动包”产出一个写入 output.html 的单文件 HTML。请把场景放在“一个快速扩张团队要为新人搭建第一天到第一周的统一知识入口”：页面要服务 HRBP、行政/IT 支持、部门运营和导师角色，让他们看懂 wiki / docs / im 如何把空间搭建、文档沉淀、消息触达和 FAQ 回收串成一套 onboarding 方案。取材以官方来源为准：`lark-cli wiki --help`、`lark-cli docs --help`、`lark-cli im --help`、`lark-cli skills read lark-wiki`、`lark-cli skills read lark-doc`、`lark-cli skills read lark-im`、`lark-cli schema wiki.spaces.create`，必要时补读相关子命令帮助。请重点生成场景背景、第一天/第一周内容架构、知识空间层级、页面节点命名建议、各角色维护分工、消息触达节奏、FAQ 回收机制、关键命令示例、权限与成员管理注意点，以及适合怎样的组织直接采用。格式要求尽量少，把业务情景和方案生成目标写得更具体。", "prompt_en": "Use the Lark Suite skill (lark-unified) to create a new HTML page for \"New Hire Knowledge Kit\" and write it to output.html as a single-file deliverable. Place the page in a fast-growing team that needs one unified entry point for new hires from day one through the first week. The page should help HRBP, admin/IT support, department operations, and mentors understand how wiki / docs / im combine space setup, document capture, message reach-out, and FAQ recycling into one onboarding solution. Ground the content in official sources: `lark-cli wiki --help`, `lark-cli docs --help`, `lark-cli im --help`, `lark-cli skills read lark-wiki`, `lark-cli skills read lark-doc`, `lark-cli skills read lark-im`, and `lark-cli schema wiki.spaces.create`, plus relevant subcommand help when needed. Focus on scenario setup, day-one and first-week content architecture, knowledge-space hierarchy, suggested node naming, role ownership, message cadence, FAQ feedback loops, key command examples, permission and member-management notes, and the kinds of organizations that can adopt it directly. Keep format requirements light and make the business situation and generation goal more concrete.", "skills": [ { "id": "lark-unified", "name": "飞书套件", "description": "飞书/Lark 全能套件（消息、文档、表格、日历、任务、Wiki 等 11 个业务域）", "skill_id": "skill_2053082253572640768" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-07-27T09:03:00Z", "updated_at": "2026-07-28T04:50:26Z", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "lark-sales-lead-control-room", "title": "销售线索跟进作战台", "title_en": "Sales Lead Follow-up Control Room", "subtitle": "用 Sheets、IM、Task 把线索入表、推进、提醒、跟进串成一条闭环", "subtitle_en": "Connect lead capture, status push, alerts, and follow-up with Sheets, IM, and Task", "description": "销售线索跟进最常见的问题，不是没有表，而是线索入口散、字段不统一、阶段推进靠口头同步、提醒全靠人工催。线索一旦从群聊、表单、活动名单流入团队，就很容易出现谁在跟、是否首触、多久未更新、哪些商机即将超时都说不清，最后台账有记录，动作却没有闭环。\n\n这个案例围绕官方 `lark-cli sheets --help`、`lark-cli im --help`、`lark-cli task --help`、`lark-cli skills read lark-sheets`、`lark-cli skills read lark-im`、`lark-cli skills read lark-task` 与 `lark-cli schema task.tasks.create` 取材，演示如何把电子表格 typed table 写入与回读、下拉字段、公式校验、群聊检索与消息发送、任务清单创建、任务建立/更新/提醒串成一套“线索入表—进度推进—消息提醒—任务跟进”的运营台账模板，并交付为单文件中英双语 HTML 作战页。\n\n适合：销售运营、SDR/BDR 团队负责人、渠道线索分发团队、商机培育流程 owner、正在用飞书做轻量 CRM/线索台账试点的业务团队。", "description_en": "The biggest problem in sales lead follow-up is rarely the absence of a spreadsheet. The real issue is fragmented lead entry points, inconsistent fields, stage updates handled verbally, and reminders driven by manual chasing. Once leads flow in from chats, forms, and event lists, teams quickly lose clarity on who owns a lead, whether the first touch happened, how long a row has been stale, and which opportunities are about to time out. The ledger exists, but the operating loop does not.\n\nThis case is grounded in the official `lark-cli sheets --help`, `lark-cli im --help`, `lark-cli task --help`, `lark-cli skills read lark-sheets`, `lark-cli skills read lark-im`, `lark-cli skills read lark-task`, and `lark-cli schema task.tasks.create` outputs. It demonstrates how to combine typed table writes and read-backs, dropdown fields, formula verification, chat discovery and messaging, tasklist creation, and task create/update/reminder operations into an operating template that closes the loop from lead entry to progress push, message alerting, and task follow-up, delivered as a bilingual single-file HTML control-room page.\n\nSuitable for: sales operations teams, SDR/BDR managers, channel lead-routing teams, opportunity pipeline owners, and business teams using Lark as a lightweight CRM or lead ledger pilot.", "categories": [ "office", "productivity", "business" ], "tags": [ "飞书", "销售线索", "台账模板", "消息提醒", "任务跟进" ], "tags_en": [ "Lark", "Sales leads", "Ledger template", "Message alerts", "Task follow-up" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "销售线索跟进作战台交付页" } ], "prompt": "请使用飞书套件（lark-unified）为“销售线索跟进作战台”产出一个写入 output.html 的单文件 HTML。请把场景放在“市场活动、官网表单和群聊咨询同时进线的销售运营作战台”：页面要服务销售运营负责人、SDR/BDR 团队与渠道同学，让他们快速看懂 sheets / im / task 如何把线索入表、阶段推进、超时提醒和重点跟进行动串成一套轻量 CRM 流程。取材以官方来源为准：`lark-cli sheets --help`、`lark-cli im --help`、`lark-cli task --help`、`lark-cli skills read lark-sheets`、`lark-cli skills read lark-im`、`lark-cli skills read lark-task`、`lark-cli schema task.tasks.create`，必要时补充 `+table-put`、`+table-get`、`+dropdown-set`、`+formula-verify`、`+chat-search`、`+messages-send`、`+tasklist-create`、`+create`、`+update`、`+reminder`。请重点生成场景背景、线索台账字段设计、阶段定义与 SLA、关键触发规则、提醒策略、跨角色交接方式、从入表到任务闭环的流程、可复制命令示例、容易失真的环节，以及适合怎样的团队先落地。格式要求少写，重点把业务情景和页面内容做扎实。", "prompt_en": "Use the Lark Suite skill (lark-unified) to create a new HTML page for \"Sales Lead Follow-up Control Room\" and write it to output.html as a single-file deliverable. Place the page in a sales-operations scenario where leads arrive at the same time from marketing events, website forms, and chat inquiries. The page should help sales ops leads, SDR/BDR managers, and channel teams understand how sheets / im / task connect lead intake, stage progression, overdue reminders, and high-priority follow-up actions into a lightweight CRM operating flow. Ground the content in official sources: `lark-cli sheets --help`, `lark-cli im --help`, `lark-cli task --help`, `lark-cli skills read lark-sheets`, `lark-cli skills read lark-im`, `lark-cli skills read lark-task`, and `lark-cli schema task.tasks.create`, plus `+table-put`, `+table-get`, `+dropdown-set`, `+formula-verify`, `+chat-search`, `+messages-send`, `+tasklist-create`, `+create`, `+update`, and `+reminder` when needed. Focus on scenario setup, ledger field design, stage definitions and SLA, trigger rules, reminder strategy, role handoff, the flow from sheet entry to task closure, copyable command examples, where the process usually breaks down, and what kinds of teams should adopt it first. Keep format requirements light and prioritize the business situation and page content.", "skills": [ { "id": "lark-unified", "name": "飞书套件", "description": "飞书/Lark 全能套件（消息、文档、表格、日历、任务、Wiki 等 11 个业务域）", "skill_id": "skill_2053082253572640768" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T09:20:00Z", "updated_at": "2026-07-28T04:50:26Z", "is_recommended": false, "recommended_order": 0 }, { "id": "lawyer-live-divorce-property", "title": "离婚财产分割全攻略普法直播方案", "title_en": "Divorce Property Division Legal Livestream Plan", "subtitle": "预热+脚本+转化+复盘，一份可直接开播的公益普法直播全流程方案", "subtitle_en": "Promotion + script + conversion + review — a ready-to-broadcast full workflow plan", "description": "律师想做普法直播获客，最头疼的是不知道从哪下手：预热怎么铺、两小时说什么、互动怎么带、观众怎么转成咨询，全凭感觉凑，往往播到一半冷场。\n\n这个案例演示了把「离婚财产分割」这个高频家事话题交给 AI，自动产出预热海报文案、15秒短视频脚本、社群话术、120分钟直播时间轴（暖场→核心讲解→互动答疑→服务介绍→收尾）、转化引导话术和复盘模板，并内置合规提醒（不承诺结果、不泄露隐私、案例脱敏），产出一份可直接开播的完整方案。\n\n适合：婚姻家事律师、想通过直播打造个人 IP 的律师、律所市场运营人员，用于公益普法直播、专项答疑直播、获客转化直播的策划。", "description_en": "The hardest part of running a legal-education livestream is knowing where to start: how to warm up the audience, what to say in two hours, how to drive interaction, and how to turn viewers into consultations — it's all improvised and often stalls midway.\n\nThis case shows how to hand the high-frequency topic \"dividing marital property in divorce\" to AI, which automatically produces poster copy, a 15-second short-video script, community warm-up messages, a 120-minute timeline (opening → core content → live Q&A → service intro → wrap-up), conversion scripts, and a review template — plus built-in compliance reminders (no outcome promises, no privacy leaks, anonymized cases) — yielding a plan ready to broadcast.\n\nFor: family-law lawyers, lawyers building a personal IP through livestream, and law-firm marketing staff planning public-education, Q&A, or lead-conversion livestreams.", "categories": [ "business" ], "tags": [ "律师直播", "普法直播", "离婚财产分割", "婚姻家事", "直播获客" ], "tags_en": [ "lawyer livestream", "legal education", "divorce property division", "family law", "livestream leads" ], "artifact_type": "html", "preview": "live-plan.html", "artifacts": [ { "file": "live-plan.html", "label": "离婚财产分割全攻略普法直播全流程方案" } ], "prompt": "请帮我策划一场「离婚财产分割全攻略」公益普法直播，输出一份可直接开播的完整流程方案。\n\n直播基本信息：\n- 主题：离婚财产分割全攻略\n- 目的：公益普法（专业展示 + 弱转化）\n- 平台：抖音\n- 时长：120 分钟\n- 主讲：李律师（执业领域：婚姻家事）\n\n请严格按照以下五大模块输出：\n1. 预热宣传方案：预热海报文案、15秒短视频脚本、社群预热话术，以及直播前7天/3天/1天/2小时的分段预热时间线；\n2. 开播准备清单：设备清单、材料准备、直播间设置；\n3. 直播流程脚本：120分钟时间轴（开场暖场5分钟 → 核心内容讲解45分钟 → 互动答疑40分钟 → 服务介绍20分钟 → 收尾总结10分钟），核心内容覆盖四个知识点：夫妻共同财产认定、离婚财产分割原则、对方隐匿财产应对、离婚协议书写法；\n4. 转化引导路径：私信引导、加群引导、福利引导的话术，以及合规注意事项（不承诺案件结果、不泄露客户隐私、案例脱敏、不生硬营销）；\n5. 复盘总结模板：数据统计表、效果评估、问题分析、优化建议。\n\n要求：法律内容专业准确，符合律师执业规范和平台直播规则；输出为一份自包含 HTML 文件（资源全内联、可离线打开），内置中英文双语切换。", "prompt_en": "Please plan a \"Complete Guide to Dividing Marital Property in Divorce\" public-education livestream and output a complete workflow plan ready to broadcast.\n\nBasic info:\n- Topic: Complete Guide to Dividing Marital Property\n- Goal: public legal education (professional showcase + soft conversion)\n- Platform: Douyin\n- Duration: 120 minutes\n- Host: Lawyer Li (practice area: family law)\n\nPlease output in the following five modules:\n1. Pre-event promotion: poster copy, 15-second short-video script, community warm-up messages, plus a phased timeline at 7/3/1 days and 2 hours before;\n2. Pre-live checklist: equipment, materials, room setup;\n3. Livestream script: a 120-minute timeline (opening 5 min → core content 45 min → live Q&A 40 min → service intro 20 min → wrap-up 10 min), covering four knowledge points: identifying community property, property-division principles, dealing with hidden assets, and drafting a divorce agreement;\n4. Conversion path: scripts for DM, group invite, and giveaway, plus compliance notes (no outcome promises, no privacy leaks, anonymized cases, no hard sell);\n5. Review template: data statistics table, effect evaluation, problem analysis, optimization suggestions.\n\nRequirements: legally accurate and professional, compliant with lawyer ethics and platform rules; output as a self-contained HTML file (all resources inlined, offline-openable) with built-in Chinese-English switching.", "skills": [], "experts": [ { "id": "brand-marketing", "name": "品牌营销专员", "description": "品牌营销专家，专注处理律师个人IP运营和律所品牌营销相关事务，涵盖律师个人IP定位、律所品牌营销方案、社交媒体内容运营、营销数据分析、获客转化方案、法律科普内容创作、短视频脚本创作等。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3d642417d5d0", "expert_id": "ex_AYR4CRp3puyw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-09-07T14:00:00+08:00", "updated_at": "2026-09-07T14:00:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "legal-hotspot-collector", "title": "法律热点搜集与研判", "title_en": "Legal Hotspot Collection & Analysis", "subtitle": "输入一个法律热点命题，自动召回权威来源、结构化研判，产出可直接排期的运营选题包", "subtitle_en": "Feed it a legal topic and it auto-recalls authoritative sources, produces structured analysis, and outputs a publish-ready editorial package", "description": "律师做内容营销最头疼的两件事：一是没时间盯立法、司法、监管的最新动态，二是抓到一个热点却不知道怎么把它变成能涨粉、能获客的选题。\n\n这个案例把「新公司法注册资本五年实缴」这个正在发酵的立法热点交给 AI，让它按「召回 → 研判 → 分发」三步走：先从中国政府网、国务院公报等 Tier 1 权威来源 + 市场监管总局等 Tier 2 专业源交叉验证核心事实，再输出带 legal_effect 分级、urgency 紧急度、争点地图的结构化热点底座，最后渲染成运营团队能直接排期的选题包——候选标题、钩子模板、格式建议、FAQ 引导一应俱全。虽然新规已于 2024 年落地，但因「存量公司过渡期截止日 2027-06-30」正在逼近，这个立法热点重新进入决策窗口期，被自动识别为 P0 优先级、可反复发酵。\n\n适合：执业律师（内容营销/客户服务）、律师团队内容运营、企业法务合规预警。", "description_en": "For lawyers doing content marketing, two things are a constant headache: no time to track the latest legislative, judicial, and regulatory changes, and no idea how to turn a trending topic into client-winning content.\n\nThis case hands the brewing legislative topic of \"five-year paid-in capital under the new Company Law\" to AI, which runs a recall → analysis → distribution pipeline: it first cross-validates core facts from Tier-1 authoritative sources (gov.cn, SAMR), then produces a structured hotspot base with legal_effect grading, urgency flags, and an issue map, and finally renders a publish-ready editorial package — candidate titles, hook templates, format recommendations, and FAQ prompts included. Even the time-sensitive deadline (2027-06-30 transition cutoff) is automatically flagged as P0 priority.\n\nFor: practicing lawyers (content marketing / client service), law-firm content teams, and in-house counsel for compliance monitoring.", "categories": [ "business", "search" ], "tags": [ "法律热点", "新公司法", "注册资本", "运营选题", "立法动态" ], "tags_en": [ "Legal Hotspot", "New Company Law", "Registered Capital", "Editorial Package", "Legislative Update" ], "artifact_type": "html", "preview": "legal-hotspot-operation.html", "artifacts": [ { "file": "legal-hotspot-operation.html", "label": "法律热点研判报告（运营选题包）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "preview_en": "", "cover_image_en": "", "prompt": "请调用「法律热点搜集与研判」技能，对近期公司/金融合规领域的法律热点进行搜集与结构化研判，输出运营选题包（视图 C）。\n\n【检索参数】\n• goals：operation（运营选题包）\n• domain_scope：[\"公司\", \"金融合规\"]\n• time_window_days：7\n• tone：严谨\n\n【热点命题】\n新《公司法》注册资本五年实缴新规及国务院配套规定（国务院令第 784 号）对存量公司的影响——重点聚焦 2027 年 6 月 30 日过渡期截止这一时效节点。\n\n【输出要求】\n1. 第一步：从 Tier 1 权威来源（中国政府网、市场监管总局、全国人大网等）与 Tier 2 专业源交叉检索，召回热点事实，不得编造来源\n2. 第二步：输出结构化热点底座，含 hotspot_type、legal_effect 分级（confirmed/tendency/inferred/pending）、urgency、evergreen_potential、issue_map（争点）、applicability/exclusions（适用与排除条件）\n3. 第三步：渲染视图 C 运营选题包，含 production_priority、候选标题、钩子模板、格式建议、FAQ 引导\n4. 每条结论带 citations 锚点（来源名称/文号/链接），法条引用标注「以现行有效文本为准」\n5. 输出 HTML 报告，供律师审核后使用，不替代专业法律意见", "prompt_en": "Invoke the Legal Hotspot Collection & Analysis skill to collect and structurally analyze recent legal hotspots in the corporate / financial-compliance domain, and output an editorial package (View C).\n\n[Search parameters]\n• goals: operation (editorial package)\n• domain_scope: [\"corporate\", \"financial compliance\"]\n• time_window_days: 7\n• tone: rigorous\n\n[Hotspot topic]\nThe five-year paid-in capital rule under the new Company Law and its supporting State Council regulation (State Council Decree No. 784) and their impact on existing companies — with focus on the time-sensitive June 30, 2027 transition deadline.\n\n[Output requirements]\n1. Step 1: cross-search Tier-1 authoritative sources (gov.cn, SAMR, NPC) and Tier-2 professional sources to recall hotspot facts; never fabricate sources\n2. Step 2: output a structured hotspot base including hotspot_type, legal_effect grading (confirmed/tendency/inferred/pending), urgency, evergreen_potential, issue_map, and applicability/exclusions\n3. Step 3: render a View C editorial package with production_priority, candidate titles, hook templates, format recommendations, and FAQ prompts\n4. Anchor every conclusion with citations (source name / document number / link); mark statute references with \"subject to the currently effective text\"\n5. Output an HTML report for lawyer review; it does not replace professional legal advice", "skills": [], "experts": [ { "id": "brand-marketing", "name": "品牌营销专员", "description": "品牌营销专家，专注处理律师个人IP运营和律所品牌营销相关事务，涵盖律师个人IP定位、律所品牌营销方案、社交媒体内容运营、营销数据分析、获客转化方案、法律科普内容创作、短视频脚本创作等。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3d642417d5d0", "expert_id": "ex_AYR4CRp3puyw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-09-07T11:10:00+08:00", "updated_at": "2026-09-07T11:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "lexiang-kb-search-synthesis", "title": "知识库一问，跨页综述即成", "title_en": "Ask Once, Get a Cross-Page Synthesis", "subtitle": "在企业知识库里检索一个专题词，AI 跨页聚合出一份带来源引用的知识综述", "subtitle_en": "Search one topic in your knowledge base, and AI aggregates a cited synthesis across pages", "description": "知识散落在十几篇文档里，搜出来一堆结果还得自己一篇篇点开读、一段段摘录再手动拼成一篇，费时又容易漏。\n\n输入一个专题词，腾讯云知在知识库里做跨页语义检索，把多个页面的命中内容聚合，AI 自动写成一份带来源引用的知识综述——每个论点都标注来源页面链接，像企业版 Perplexity，可点击溯源。\n\n适合：知识运营、AI 产品经理、客服负责人、研究员等需要快速从内部知识库沉淀出结论的人。", "description_en": "Knowledge is scattered across a dozen docs; a search returns piles of hits and you still have to open each page, copy fragments, and stitch them together by hand — slow and error-prone.\n\nEnter one topic keyword, and Tencent Yunzhi runs a cross-page semantic search over your knowledge base, aggregates the hits, and AI writes a cited synthesis — every point links back to its source page, like an enterprise Perplexity you can click to trace.\n\nBest for: knowledge ops, AI product managers, support leads, and researchers who need to distill conclusions from an internal knowledge base fast.", "categories": [ "search", "education" ], "tags": [ "知识检索", "AI综述", "企业知识库", "来源溯源" ], "tags_en": [ "Knowledge Search", "AI Synthesis", "Enterprise KB", "Source Tracing" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "知识综述报告" } ], "prompt": "用腾讯云知（乐享）在我的知识库里检索「企业知识问答系统」相关内容，做一份带来源引用的知识综述。\n\n步骤：\n1. 围绕这个专题拆成 3 个方向的关键词（如 RAG 检索增强、Prompt 工程、Agent 智能体），分别调用云知的 search_kb_search 在知识库内做语义检索，记录每条 query 的命中数(total)与耗时(took_ms)。\n2. 从各页面命中的正文高亮片段(snippet)中提取要点，跨多个页面聚合。\n3. 把聚合内容组织成一篇连贯综述，分几节撰写；每处论点后用上标角标 [1][2][3] 标注来源，并在文末列出对应的来源页面标题与可点击链接(url)。\n\n重要约束：所有内容、数字、页面标题和链接都必须来自云知知识库的真实检索返回，禁止编造任何数据；每个论点都要能溯源到具体来源页面链接。\n\n最终输出为一份暗色科技感、信息密度高的单文件 HTML 报告：顶部标题区标明数据来源知识库；中部可视化展示检索过程（query / 命中数 / 耗时 →聚合流程）；正文综述带角标引用；下方列出来源引用列表（标题+链接+关键词标签）；底部给出结论与数据来源说明。所有资源内联，不引用任何外部字体、脚本或图片。", "prompt_en": "Use Tencent Yunzhi (Lexiang) to search my knowledge base for content about \"enterprise knowledge Q&A systems\" and produce a cited knowledge synthesis.\n\nSteps:\n1. Break the topic into 3 keyword angles (e.g. RAG retrieval augmentation, Prompt engineering, AI Agents) and call Yunzhi's search_kb_search for each to run semantic search inside the knowledge base; record each query's hit count (total) and latency (took_ms).\n2. Extract key points from the highlighted body snippets returned for each page, aggregating across multiple pages.\n3. Organize the aggregated content into a coherent multi-section synthesis; mark every point with superscript citations [1][2][3] and list the corresponding source page titles and clickable links (url) at the end.\n\nStrict rule: every piece of content, number, page title, and link must come from the real search results of the Yunzhi knowledge base — fabricating any data is forbidden; every point must be traceable to a specific source page link.\n\nOutput a single-file HTML report with a dark, tech aesthetic and high information density: a header stating the source knowledge base; a mid-section visualizing the retrieval process (query / hits / latency → aggregation flow); the synthesis body with citation superscripts; a source reference list (title + link + keyword tags); and a footer with the conclusion and data-source note. Inline all resources — no external fonts, scripts, or images.", "skills": [ { "id": "tencent-yunzhi", "name": "腾讯云知", "description": "腾讯云知（乐享）知识库专用操作——搜索、阅读、创建、编辑、上传", "skill_id": "skill_2053083480194342912" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-07-27T11:30:00+08:00", "updated_at": "2026-07-27T11:30:00+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "licai-zichan-fuzhai-handdraw", "title": "理财：资产与负债 - 手绘动画", "title_en": "Finance: Assets vs Liabilities - Hand-drawn Animation", "subtitle": "把PPT理财课变成3分钟手绘动画视频", "subtitle_en": "Turn a financial literacy PPT into a 3-minute hand-drawn animation", "description": "想把枯燥的理财知识变成生动好懂的手绘动画？这个案例展示了如何用手绘动画能力，把一份《资产与负债》理财PPT转化为一段3分钟的温暖扁平画风动画视频。视频从「100万选择测试」切入，区分资产与负债，剖析房子、车子背后的现金流真相，再到财富思维与消费顺序，适合理财博主、财经教育、自媒体知识可视化。", "description_en": "Want to turn dry financial knowledge into a lively, easy-to-understand hand-drawn animation? This case demonstrates how to use the handdraw skill to transform a financial literacy PPT on Assets vs Liabilities into a 3-minute warm flat-illustration animation video. Starting from the 'What would you do with 1 million' test, the video distinguishes assets from liabilities, reveals the cash-flow truth behind houses and cars, and explains wealth thinking and consumption order. Suitable for finance bloggers, financial education, and knowledge visualization on social media.", "tags": [ "手绘动画", "理财科普", "资产与负债", "知识可视化", "短视频" ], "tags_en": [ "hand-drawn animation", "financial literacy", "assets vs liabilities", "knowledge visualization", "short video" ], "prompt": "读取《理财：资产与负债》PPT内容，提炼核心观点，用手绘动画的方式生成一个3分钟的温暖扁平画风动画视频。视频结构：从100万选择测试切入，对比富人/中产/穷人三种金钱去向；定义资产与负债，并以房子、车子为例揭示现金流真相；分析社会如何把负债包装成资产；讲解财富思维（沙子 vs 种子）、富人的消费顺序、负债的危险阶段；最后给出4个转身小行动。选用暖心扁平画风，画面只画看得见的具体事物，每页配一句简短口语化旁白。", "prompt_en": "Read the PPT 'Finance: Assets and Liabilities', extract the core ideas, and generate a 3-minute warm flat-illustration hand-drawn animation video. Structure: start with the 'what to do with 1 million' choice test, compare how the rich, middle class, and poor spend money; define assets vs liabilities, using houses and cars to reveal cash-flow truths; analyze how society packages liabilities as assets; explain wealth thinking (sand vs seeds), the rich consumption order, and the dangerous stage of debt; end with four small turning-point actions. Use a warm flat illustration style. Draw only concrete, visible things. Each page has a short, colloquial narration.", "artifact_type": "video", "preview": "output.mp4", "external_url": "", "scenario": "", "scenario_use": "", "task_mode": "", "categories": [ "finance", "creation" ], "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "output.mp4", "label": "3分钟手绘动画视频" } ], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "usage_count": 0, "created_at": "2026-07-14T22:35:00+08:00", "updated_at": "2026-07-14T22:35:00+08:00", "skills": [], "experts": [ { "id": "VibeknowHanddraw", "name": "vibeknow手绘视频讲解师", "description": "老师讲历史、家长讲绘本、医生讲康复、理财博主讲避坑，专业内容边画边讲成手绘科普视频，51种风格随选。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "89133bd8e85e", "name_en": "vibeknow hand-drawn explainer", "description_en": "Teachers explain history, parents read picture books, doctors explain rehab, finance bloggers warn of pitfalls — turn professional content into hand-drawn explainer videos, 51 styles to choose from.", "expert_id": "ex_kdKBSetkjUyA" } ], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "light-diet-workbench", "artifact_type": "html", "title": "每日饮食工作台·清淡饮食版", "title_en": "Daily Diet Workbench · Light & Mild", "subtitle": "东北人在广东，全程不辣的一周饮食方案", "subtitle_en": "A northeastern foodie in Guangdong: a zero-spice weekly plan", "description": "刚到广东，想吃清淡又怕踩辣、不知道水果怎么挑？这份工作台为你配好一周七天的三餐：每天早午晚三张食谱卡片，带建议时段和不踩坑提示，还有岭南水果性味速查和一键抽签，选择困难时点一下就能配好一整天（全部不辣）。适合在广东生活的东北胃、刚上大学的南方新生，以及任何想清淡饮食、祛湿不上火的人。", "description_en": "New to Guangdong and want light, mild meals without accidentally ordering spicy food? This workbench plans all three meals for every day of the week: breakfast, lunch and dinner cards with suggested time windows and practical tips, plus a Lingnan fruit guide and a one-tap random picker that assembles a full spice-free day when you can't decide. Perfect for northeastern foodies living in Guangdong, new students heading south, or anyone eating light to beat dampness and internal heat.", "tags": [ "饮食管理", "一周食谱", "清淡饮食", "生活工作台" ], "tags_en": [ "diet planning", "weekly menu", "light eating", "lifestyle workbench" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/Ir2Ijw9BMIlfNLQWKGDStG?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「每日饮食工作台 · 清淡饮食版」单页工作台：面向在广东生活的东北人，提供一周七天的三餐食谱——每天早餐、午餐、晚餐三张卡片，每张含菜品清单、建议时段和贴心提示（如\"肠粉加豉油即可，别加辣酱\"），当天推荐水果附在每餐末尾；配岭南常见水果性味速查（香蕉、荔枝、龙眼等 12 种，标注性平/性凉/易上火）；\"注意事项\"区覆盖不吃辣点餐攻略、上火与祛湿、喝汤学问、避坑禁忌、宿舍外食方案、东北乡愁补给六张卡片；顶部放饮食基调、核心目标、每日饮水 1.5~2L、辣度控制四个状态卡；再加一个\"今天吃什么\"一键抽签，随机生成早/午/晚餐+水果+饮品组合（全部不辣）。视觉走粉紫渐变的柔和甜系风格，圆角卡片、留白充足，PC 与手机端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to replicate this workbench with one click: https://www.workbuddy.cn/space/d/Ir2Ijw9BMIlfNLQWKGDStG?source=2 — data must be stored online with two-way sync. Prefer the one-click replication flow; if unavailable, build it as follows: Build a \"Daily Diet Workbench · Light & Mild\" single-page workbench for a northeastern foodie living in Guangdong: a seven-day meal plan with three cards per day (breakfast, lunch, dinner), each listing dishes, a suggested time window and a practical tip (e.g. \"rice rolls just need soy sauce — skip the chili sauce\"), with the day's recommended fruit at the end of every meal; include a Lingnan fruit quick guide (12 fruits such as banana, lychee, longan, tagged neutral/cooling/heaty); a \"Survival Notes\" section with six cards covering no-spice ordering, internal heat & dampness, the art of soup, pitfalls & don'ts, dorm & eating-out plans, and northeast comfort-food supply; a top status bar with four cards (diet baseline, core goal, 1.5–2L daily water, spice control); plus a \"What to Eat Today\" one-tap random picker that generates a breakfast/lunch/dinner + fruit + drink combo (all spice-free). Visual style: soft pink-purple gradients, rounded cards, generous whitespace, responsive on PC and mobile. Deliver a\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "每日饮食工作台 · 双语自包含演示" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "creator": { "name": "雪落下的时候", "name_en": "Xue Luoxia De Shihou", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-01T10:00:00+08:00", "updated_at": "2026-09-01T10:00:00+08:00", "quality_score": 99, "is_recommended": false, "recommended_order": 0 }, { "id": "limited-liability-company-shareholder-governance-analysis", "title": "有限责任公司股东治理风险分析报告", "title_en": "Limited Liability Company Governance Risk Report", "subtitle": "从知情权到决议效力，搭建股权争议的双视角行动路径", "subtitle_en": "A dual-perspective action path from information rights to corporate resolution validity", "description": "小股东最怕的不是拿不到一份报表，而是财务、决议和关联交易的线索一起断掉，等发现问题时又错过了程序救济窗口。\n\n这个案例将有限责任公司股东知情权、公司决议效力、关联交易与代表诉讼串成一份可互动的法律风险分析报告：可复制 Prompt、切换小股东/公司视角、勾选程序核验项，并下载本地行动清单。\n\n适合：有限责任公司股东、创业公司创始团队、法务与投融资负责人、董事会秘书、处理股东争议的律师。", "description_en": "The hardest part for minority shareholders is not just missing a financial report. It is losing the trail across accounts, resolutions, and related-party dealings, then discovering that a procedural remedy window has already closed.\n\nThis case turns information rights, resolution validity, related-party transactions, and derivative actions into an interactive legal risk report: copy-ready prompts, shareholder/company perspective filters, a procedural checklist, and a locally downloadable action list.\n\nBest for: limited-liability-company shareholders, startup founding teams, legal and fundraising leaders, board secretaries, and counsel handling shareholder disputes.", "categories": [ "business" ], "tags": [ "股权投资", "公司治理", "小股东", "公司法", "关联交易" ], "tags_en": [ "Equity Investment", "Corporate Governance", "Minority Shareholders", "Company Law", "Related-Party Transactions" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "有限责任公司小股东知情权、公司决议效力与治理机制法律分析报告.html", "preview_en": "", "external_url_en": "", "prompt": "某有限责任公司有多名股东，小股东认为公司长期未披露财务情况，部分重大事项也未经充分讨论即作出决议。请结合现行有效公司法律法规，分析小股东可以依法行使哪些知情权和监督权；股东会或董事会决议在什么情形下可能不成立、无效或可撤销；小股东应如何提出查阅请求、保全证据和采取救济措施；以及公司应如何完善议事规则、信息披露、关联交易管理和决策留痕机制，平衡大股东、小股东与公司经营发展的利益。请形成一份包含法律依据、争议路径、风险分级、行动方案和治理建议的法律分析报告。", "prompt_en": "A limited liability company has multiple shareholders. Minority shareholders believe that the company has long failed to disclose financial information, and that some material matters were resolved without adequate discussion. Based on currently effective company laws and regulations, analyze: what information and supervision rights minority shareholders may exercise; when shareholders' meeting or board resolutions may be deemed not established, invalid, or revocable; how minority shareholders should make inspection requests, preserve evidence, and pursue remedies; and how the company should improve its procedural rules, information disclosure, related-party transaction controls, and decision-making records to balance the interests of controlling shareholders, minority shareholders, and the company’s development. Produce a legal analysis report including legal grounds, dispute paths, risk levels, action plans, and governance recommendations.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同、公司治理、法律检索与争议处理的资深合同法律顾问", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "artifacts": [ { "file": "有限责任公司小股东知情权、公司决议效力与治理机制法律分析报告.html", "label": "互动法律风险分析报告" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T13:56:25+08:00", "updated_at": "2026-08-13T13:56:25+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "listening-portrait", "title": "你的听歌画像报告", "title_en": "Your Listening Portrait Report", "subtitle": "调取酷狗收藏与听歌数据，生成杂志感的个人音乐年鉴", "subtitle_en": "Pull your Kugou favorites and listening data into a magazine-style personal music almanac", "description": "听了这么多歌，却从没真正看清自己是怎么听歌的——收藏了上千首，反复循环的却总是那几首。这个案例调取酷狗音乐的收藏、最近播放和听歌统计，生成一份深色沉浸式的单文件 HTML 听歌报告：一句话画像、核心听歌数字、常听歌曲与歌手、曲风与语言偏好，还会挖出「收藏 × 实际播放」的反差、「熟悉 × 探索」的比例和「广度 × 深度」的性格，让你恍然「原来我是这样听歌的」。\n\n适合：把音乐当空气的重度用户、想给年度歌单做个仪式感总结的人、爱在朋友圈晒高质感数据卡片的分享控。", "description_en": "You've listened to so much music, yet never really seen how you listen — thousands of songs saved, but only a handful on repeat. This case pulls your Kugou favorites, recent plays, and listening stats into a dark, immersive single-file HTML report: a one-line portrait, core listening numbers, top tracks and artists, genre and language preferences, plus the contrast between \"saved vs. actually played,\" the ratio of \"familiar vs. exploring,\" and the personality of \"breadth vs. depth\" — until you realize \"so this is how I listen to music.\"\n\nIdeal for: heavy listeners who treat music like air, anyone wanting a ceremonial wrap-up of their year in songs, and sharers who love posting high-quality data cards.", "categories": [ "data", "creation" ], "tags": [ "听歌报告", "音乐数据可视化", "个人年鉴", "酷狗音乐", "数据展览风" ], "tags_en": [ "Listening Report", "Music Data Viz", "Personal Almanac", "Kugou Music", "Data Exhibition Style" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "个人听歌画像报告页面" } ], "prompt": "帮我做一个单文件 HTML 数据可视化页面，主题是「你的听歌画像」，调用 kugou skill 获取我的收藏、最近播放和听歌统计，并按实际可用时间范围生成个人听歌报告。采用高质感音乐杂志/数据展览风格，深色沉浸背景，用唱片、舞台灯光或耳机特写大图做 Hero（叠深色渐变蒙版保证标题清晰）。内容包含：一句话听歌画像、核心听歌数字、常听歌曲与歌手、曲风偏好、收藏 × 实际播放对比、我的听歌方式、音乐坐标、陪伴最久与最近靠近的声音、偏好变化和阶段总结。除常规排行外，重点发现收藏与回听、熟悉与探索、偏好广度与播放深度之间的关系，呈现有反差、有长期陪伴感或容易被忽略的个人听歌特点，让我感受到原来我是这样听歌的。采用分章节报告排版，结合大标题、大数字、轻量图表与错落分栏，避免大段榜单；数据不足的内容直接隐藏，兼容手机打开展示。内容必须静态预渲染进 HTML，绝不能依赖 JS 注入，所有 CSS/JS 和图片均内联，响应式布局。", "prompt_en": "Help me build a single-file HTML data visualization page themed \"Your Listening Portrait.\" Call the kugou skill to fetch my favorites, recent plays, and listening stats, and generate a personal listening report based on the actually available time range.\n\nUse a high-quality music-magazine / data-exhibition style with a dark, immersive background. Use a large image of a vinyl record, stage lighting, or a headphone close-up as the Hero (overlay a dark gradient mask to keep the title legible).\n\nContent includes: a one-line listening portrait, core listening numbers, top tracks and artists, genre preferences, a \"saved vs. actually played\" comparison, how I listen, my music coordinates, the sounds that have stayed longest and the ones I've reached for recently, shifts in preference, and a phase summary. Beyond the usual rankings, focus on surfacing the relationships between saving vs. re-listening, familiar vs. exploring, and breadth of preference vs. depth of play — reveal the contrasts, the long-term companionship, and the easily overlooked personal listening traits, so I feel \"so this is how I listen to music.\"\n\nUse a sectioned report layout combining big headlines, big numbers, lightweight charts, and staggered columns; avoid long list rankings. Hide any content that lacks sufficient data, and make it work well on mobile. Inline all CSS/JS and images, use a responsive layout.", "skills": [ { "id": "kugou-skill", "name": "酷狗音乐助手", "description": "酷狗音乐 AI 助手，支持歌曲搜索、每日推荐、相似推荐、榜单查询、收藏管理、听歌统计与歌单创建。", "skill_id": "skill_2074106144009043968" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-16T15:08:00Z", "updated_at": "2026-07-16T15:08:00Z", "is_recommended": true, "recommended_order": 4 }, { "id": "live-review-report", "title": "直播复盘报告", "title_en": "Live Review Report", "subtitle": "一键生成直播用户行为分析，卡片+看板双视图驱动运营决策", "subtitle_en": "One-click live stream user behavior analysis with dual-view cards and dashboards for data-driven operations", "description": "做完直播最怕的就是不知道效果到底怎么样。来看的人多不多？有没有人认真看完？互动够不够？下次直播该改什么？\n\n这个案例演示了如何用一条 Prompt 对一场直播做完整复盘——自动调出摘要卡片看关键指标，再打开完整报告面板看用户分群、观看诊断、互动结构、回放表现，最后给出分层运营建议和下一场直播的 action list。整份报告由后端真实数据驱动，无需手动计算，拿来就能用。\n\n适合：直播运营、课程讲师、知识付费从业者、教育培训机构、想靠数据优化直播效果的任何人", "description_en": "The hardest part after a live stream is knowing whether it actually worked. How many people showed up? Did anyone watch all the way through? Was there enough engagement? What should change next time?\n\nThis case shows how to get a full live stream review with a single prompt — automatically pull up a summary card for key metrics, then open the full report panel to see user segmentation, viewing diagnostics, engagement patterns, and replay performance, followed by tiered operational recommendations and an action checklist for the next session. The entire report is driven by real backend data — no manual calculations, ready to use immediately.\n\nFor: live stream operators, course instructors, knowledge commerce professionals, edtech teams, and anyone who wants to optimize their live streams with data", "categories": [ "data" ], "tags": [ "直播复盘", "用户行为分析", "数据分析", "运营决策", "教育直播" ], "tags_en": [ "live review", "user behavior", "data analytics", "operations", "edtech" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我复盘一场直播。提供直播的 alive_id（格式如 l_xxxxxxxxxxxxxxxxxxxxxxx），先展示摘要卡片看关键指标，再打开完整报告面板看用户分群、观看诊断、互动结构、回放表现，最后基于数据给我解读：哪些用户需要优先运营、下一场直播该怎么改进。\n\n示例：复盘直播「鹅直播客户端项目培训」，alive_id: l_63a150cae4b06159f72273c4", "prompt_en": "Help me review a live stream. Provide the live stream's alive_id (format: l_xxxxxxxxxxxxxxxxxxxxxxx). First display the summary card for key metrics, then open the full report panel to see user segmentation, viewing diagnostics, engagement patterns, and replay performance. Finally, provide data-driven insights: which users need priority follow-up, and what to improve for the next live session.\n\nExample: Review the live stream 'Goose Live Client Project Training', alive_id: l_63a150cae4b06159f72273c4", "skills": [], "experts": [ { "id": "xiaoe-data-analyst", "name": "小鹅数据分析师", "description": "查询小鹅通平台直播、课程、交易等多维经营数据，组合分析并导出报表，用数据支撑商家决策", "promptFileSnapshot": "8af64b66209c", "expert_id": "ex_1M2LsZX1RbK9" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "artifacts": [ { "file": "output.html", "label": "直播复盘数据报告" } ], "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-26T10:10:00+08:00", "updated_at": "2026-08-26T19:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "logistics-exception-tracking", "title": "物流异常追踪", "title_en": "Logistics Exception Tracking", "subtitle": "近7日订单物流进度与异常处理建议", "subtitle_en": "A practical Weimob playbook for logistics exception tracking", "description": "做物流异常追踪时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“物流异常追踪”展开：查询近 7 日订单最新物流轨迹，找出运输停滞和签收异常订单并制定催件或售后处理方案。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、订单履约运营人员、经营分析团队。", "description_en": "The hardest part of logistics exception tracking is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates logistics exception tracking through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Order Fulfillment specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "物流追踪", "运输停滞", "签收异常", "催件建议", "售后处理" ], "tags_en": [ "Logistics Exception Tracking", "Mall Operations", "Order Fulfillment", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查询近7日订单的最新物流进度，标出运输停滞和签收异常的订单，分析原因并给出催件或售后处理建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a logistics exception tracking covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "lukou-finance-desk", "title": "路口理财工作台", "title_en": "Crossroads Finance Desk", "subtitle": "个人理财驾驶舱：持仓 · 诊断 · 资讯 · 复盘 · 研究 · 心得", "subtitle_en": "A personal finance cockpit: holdings, diagnostics, news, review, research & notes", "description": "炒股最麻烦的不是看行情，而是持仓散在各个 App 里：想看总资产要自己手算汇率，想知道今天谁拖后腿得翻五六个页面，看过的研报转天就忘。\n\n这个工作台把 28 只覆盖 A股/港美股/ETF/黄金/债券的示例持仓搬进资料库在线表：总览页实时算总资产、当日盈亏、资产分类和盈亏贡献榜；持仓页直接增删改、每日盈亏日历还原组合波动；诊断页自动扫集中度和深度套牢；资讯页多源聚合并标记持仓相关；还有一键生成的每日复盘（数据与推断分离）和个股研究（K线/基本面/财报/综合打分），有价值的消息随手收进研究心得。行情实时来自公开财经接口，心得与持仓多端同步。\n\n适合：多市场持仓想一个面板看清全局的个人投资者、想把「看盘-研究-复盘-记笔记」串成一条流水的认真玩家，以及想做同款数据驱动工作台的人。", "description_en": "The annoying part of investing isn't checking quotes — it's holdings scattered across apps: total assets need manual FX math, finding today's drag means flipping six pages, and research you read yesterday is gone today.\n\nThis workbench moves 28 demo holdings across A-shares / HK / US stocks / ETFs / gold / bonds into library online tables: the overview computes total assets, day P&L, allocation and contributor boards in real time; the holdings page does CRUD with a daily P&L calendar; diagnostics scans concentration and deep losses; news aggregates multi-source feeds with holding relevance; plus a one-tap daily review (data vs. inference separated) and per-stock research (K-line / fundamentals / filings / overall score). Valuable items clip into research notes. Quotes stream from public finance APIs; notes and holdings sync across devices.\n\nFor: multi-market investors who want one pane of glass, serious players chaining watch → research → review → notes into one flow, and anyone building a data-driven workbench of their own.", "tags": [ "理财工作台", "持仓管理", "实时行情", "每日复盘" ], "tags_en": [ "Finance workbench", "Holdings tracking", "Live quotes", "Daily review" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "路口理财工作台（双语自包含演示）" } ], "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/DjCSvayYNp5wZAROJjYoE4，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n做一个「路口理财工作台」 理财驾驶舱，持仓与研究心得存资料库在线数据表（页面运行时读写，在表里加减记录后页面刷新即同步）。功能：①工作台总览——总资产/当日盈亏/总盈亏/总收益率/仓位 KPI（港美股按设置汇率折人民币），资产分类按市场与板块双图、组合走势曲线（近 60/120/250 日可切）、每日盈亏趋势、持仓结构热力图、当日与累计盈亏贡献榜、A股/港股/美股主要市场指数区；②我的持仓——表单新增/修改/删除持仓，按市值/当日盈亏/总盈亏/涨跌幅/权重排序，资产类型筛选，每日盈亏日历（基于日K还原组合每日盈亏、A股红涨绿跌）；③持仓诊断——健康度分布环形图、集中度体检（前三大权重/板块集中度/现金占比）、风险与机会自动扫描（单票偏高、深度套牢、接近成本线、涨停关注兑现）、持仓相关资讯热点；④财经资讯——多源聚合快讯与公告，去重、来源切换、关键词过滤、只看持仓相关、一键收藏到研究心得；⑤每日复盘——按当前行情快照一键生成，客观数据与规则推断分离标注，可复制/下载 Markdown；⑥标的研究——输入代码或全市场搜索名称，基本面画像（行业/市值/PE/PB/ROE）、技术面（日K蜡烛+均线+技术指标速览）、持仓关联、相关资讯与公告、财报列表、综合诊断打分；⑦研究心得——心得与收藏双类型笔记，标签筛选、全文搜索，存资料库数据表。行情与资讯实时取自公开财经接口（腾讯财经/东方财富等），可用现金与汇率在设置中调整。视觉为专业行情终端风：深色侧栏 + Bloomberg 式滚动行情条、红涨绿跌、ECharts 图表、顶部 LIVE 徽标，PC 优先并适配窄屏。\n\n交付：资料库在线工作台——在线 HTML 页面 + 7 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/DjCSvayYNp5wZAROJjYoE4, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows:\nBuild a 'Crossroads Finance Desk' as a finance cockpit. Holdings and research notes are stored in library online data tables (read/written at runtime; records added or removed in the table appear after refresh). Features: 1) Overview — KPIs for total assets / day P&L / total P&L / total return / position (HK & US stocks converted to CNY at your FX rates), allocation by market and by sector, portfolio curve (switchable 60/120/250 days), daily P&L trend, holdings treemap, day and cumulative P&L contributor boards, and A-share / HK / US index panels. 2) Holdings — add/edit/delete via forms, sort by value / day P&L / total P&L / change / weight, asset-type filter, and a daily P&L calendar reconstructed from daily bars (red = up, A-share convention). 3) Diagnostics — health-mix donut, concentration check (top-3 weights / sector concentration / cash ratio), automatic risk & opportunity scan (single-stock heavy, deeply underwater, near breakeven, limit-up watch), and news on your holdings. 4) News — aggregated flashes and exchange filings with dedup, source switching, keyword filter, holdings-only filter, and one-click clip into notes. 5) Daily review — one-tap generation from the current snapshot with data vs. inference clearly tagged; copy or download as Markdown. 6) Research — enter a code or search by name across markets; fundamentals profile (industry / market cap / PE / PB / ROE), technicals (daily candlestick + MAs + indicator snapshot), position link, related news & filings, financial reports, and an overall score. 7) Notes — note/clip types with tag filter and full-text search, stored in a library table. Quotes and news come live from public finance APIs (Tencent / EastMoney); cash and FX rates are editable in Settings. Visual: pro trading-terminal style — dark sidebar + Bloomberg-style scrolling ticker, red-up/green-down, ECharts, LIVE badge; PC-first with narrow-screen support.\n\nDeliverable: an online Library workbench — an online HTML page plus 7 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "路口大爷", "name_en": "Lukou Daye", "avatar": "", "badge": "community" }, "source": "community", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "created_at": "2026-08-18T20:05:43+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "quality_score": 91, "is_recommended": false, "recommended_order": 0 }, { "id": "ma-equity-due-diligence", "title": "投资并购股权尽职调查", "title_en": "M&A Equity Due Diligence", "subtitle": "系统跑通十部分核查，挖出代持、隐藏担保、专利瑕疵、关联交易四大红旗", "subtitle_en": "A full 10-part review surfacing nominee holdings, hidden guarantees, IP defects and related-party deals", "description": "收购一家公司最怕的就是踩坑——表面上光鲜，背后却藏着代持、隐藏担保、专利归属不清这些看不见的债务。\n\n这个案例演示了把一家 Pre-B 轮医疗器械公司的尽调资料交给 AI，按律所十大部分框架逐项核查，结合 16 维风险矩阵量化，最终挖出股权代持、2000 万未披露连带担保、离职 CTO 核心专利归属瑕疵、实控人关联采购四大红旗，并给出解决方案与交易定价建议。\n\n适合：投行 / PE / VC 尽调分析师、公司法务、并购律师、创业者在收购或投资前的快速自查。", "description_en": "The scariest part of acquiring a company is the hidden risk — nominee holdings, undisclosed guarantees, and unclear patent ownership lurking beneath a clean surface.\n\nThis case hands a Pre-B medical-device company's diligence materials to AI, which runs a law-firm-grade 10-part review plus a 16-dimension risk matrix, surfacing four red flags: nominee holdings, an undisclosed RMB 20M joint guarantee, a departed CTO's core-patent ownership defect, and related-party procurement — each with a fix and a deal-pricing suggestion.\n\nFor: investment-banking / PE / VC due-diligence analysts, in-house counsel, M&A lawyers, and founders doing a pre-deal self-check.", "categories": [ "finance", "business" ], "tags": [ "股权尽调", "投资并购", "风险矩阵", "红旗发现", "尽调报告" ], "tags_en": [ "Equity Due Diligence", "M&A", "Risk Matrix", "Red Flags", "DD Report" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "尽调报告（HTML）" } ], "prompt": "请调用「公司股权尽职调查」技能，对目标公司「杭州智行医疗科技有限公司」执行投资并购（M&A）场景的全面深度尽职调查，出具律所实务结构报告。\n\n【目标公司内部资料】\n• 全称：杭州智行医疗科技有限公司；统一社会信用代码：91330108MA2HXXXXXX；成立日期：2020-03-12；注册资本1000万元、实缴850万元；注册地：杭州市滨江区；经营范围：二类医疗器械研发、生产、销售\n• 股权结构（Pre-B轮）：创始人张某某28%（其中8%系为已离职CTO王某某代持、无书面协议）、联合创始人李某某20%、天使投资人王某12%、A轮机构启明创投20%、B轮机构远山资本18%、员工持股平台智行合伙2%\n• 历史沿革：2020年设立（张60%/李40%）→ 2021年天使轮 → 2023年A轮（含一笔张某某向王某某的0元转让）→ 2025年B轮\n• 主要财产：1款二类器械注册证（电子血压计，2027年到期）；核心在研专利「便携式血糖检测方法」发明人为已离职CTO王某某，无职务发明归属协议\n• 债权债务：为实控人张某某控制的「杭州智联贸易有限公司」提供2000万元连带责任保证，未在财报披露\n• 关联交易：2025年向智联贸易采购原材料占营业成本35%，未履行关联交易决策程序\n• 财务：2025年营收约1.2亿元、净利润约800万元、经营性现金流为负\n• 劳动人事：员工约150人（研发占40%），约30%员工社保按最低基数缴纳\n• 诉讼：1起未决供应商货款纠纷（标的约300万元）\n\n【尽调要求】\n1. 场景=投资并购，深度=全面（十部分全覆盖）\n2. 按十部分核查框架逐项核查，每部分标注风险评级（正常/需关注/[红旗]）+ 信息来源\n3. 结合16维风险矩阵，标注每维度风险类型与红旗量化（金额/比例/整改成本）\n4. 发现[红旗]→深度分析，汇总至报告第一部分\n5. 结论须转化为投资协议附件建议（陈述与保证/交割条件/赔偿条款）\n6. 附免责声明；引用法条需联网核实或标注【待核实】\n7. 输出格式：单文件 HTML（全内联样式、无外部依赖，可直接浏览器打开/打印/导出 PDF）；16 维风险矩阵以 HTML 表格内嵌呈现，不单独输出 Excel/Word", "prompt_en": "Invoke the Company Equity Due Diligence skill to run a full-depth M&A due diligence on the target company Hangzhou Zhixing Medical Technology Co., Ltd., and produce a law-firm-style structured report.\n\n[Target company internal data]\n• Full name: Hangzhou Zhixing Medical Technology Co., Ltd.; USCC: 91330108MA2HXXXXXX; founded: 2020-03-12; registered capital RMB 10M, paid-in RMB 8.5M; registered in Binjiang District, Hangzhou; scope: R&D, manufacture and sales of Class II medical devices\n• Shareholding (Pre-B): founder Zhang XX 28% (of which 8% is nominee-held for the departed CTO Wang XX, no written agreement), co-founder Li XX 20%, angel investor Wang X 12%, Series A firm Qiming Ventures 20%, Series B firm Yuanshan Capital 18%, ESOP platform Zhixing Partnership 2%\n• History: 2020 incorporation (Zhang 60% / Li 40%) -> 2021 angel round -> 2023 Series A (incl. a RMB 0 transfer from Zhang to Wang) -> 2025 Series B\n• Key assets: 1 Class II device registration (electronic blood-pressure monitor, exp. 2027); core in-development patent \"Portable Blood Glucose Detection Method\" names the departed CTO Wang XX as inventor, with no service-invention ownership agreement\n• Debts: provided a RMB 20M joint guarantee for Hangzhou Zhilian Trading Co., Ltd. controlled by actual controller Zhang XX, undisclosed in financial statements\n• Related-party transactions: 2025 raw-material purchases from Zhilian Trading = 35% of operating cost, without related-party approval procedures\n• Financials: 2025 revenue ~RMB 120M, net profit ~RMB 8M, negative operating cash flow\n• HR: ~150 employees (R&D 40%), ~30% of employees on minimum social-insurance base\n• Litigation: 1 pending supplier payment dispute (~RMB 3M)\n\n[DD requirements]\n1. Scenario = M&A, depth = full (all 10 parts covered)\n2. Review each part of the 10-part framework, tagging risk rating (Normal / Attention / [Red Flag]) + information source\n3. Apply the 16-dimension risk matrix, tagging each dimension's risk type and red-flag quantification (amount / ratio / remediation cost)\n4. Deep-dive each [Red Flag] and summarize in Part 1 of the report\n5. Convert findings into investment-agreement annex suggestions (representations & warranties / closing conditions / indemnity clauses)\n6. Include a disclaimer; verify cited statutes online or tag as [To verify]\n7. Output format: a single self-contained HTML file (inline styles, no external dependencies, openable in any browser, printable, and exportable to PDF); render the 16-dimension risk matrix as an embedded HTML table, not as a separate Excel/Word file", "skills": [], "experts": [ { "id": "investment-financing-lead", "name": "投融资法律顾问专家团", "description": "面向企业法务与投融资律师，按 L0-L3 分层路由调度法律尽调、法规研究、交易文件起草、交割管理与独立核验，覆盖股权投资与并购全流程。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "e230ea646c0e", "expert_id": "ex_oKivlYsCCDzj" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-09-07T09:45:00+08:00", "updated_at": "2026-09-07T16:34:01+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "macro-economy-dashboard", "title": "中国宏观经济仪表盘", "title_en": "China Macro-Economy Dashboard", "subtitle": "一页看懂GDP·CPI·PMI·M2·社融", "subtitle_en": "GDP, CPI, PMI, M2 & AFRE at a glance", "description": "想快速了解中国经济现在到底冷还是热，却要在国家统计局、人民银行、各类研报之间来回翻？数字散、口径乱、还容易拿到过时数据。\n\n这个案例用 NeoData 金融搜索服务实时查询 GDP、CPI、制造业 PMI、M2、社融和经济事件日历，一键生成一张浅色专业风的单文件 HTML 宏观仪表盘：温度结论、核心指标卡片、近几期趋势图、事件日历一应俱全，红升绿降、荣枯线一眼看清，还能中英切换、离线打开。\n\n适合：\n- 需要定期盯宏观的投研/策略人员\n- 想做经济简报的财经自媒体与分析师\n- 关注货币与物价走势的企业决策者\n- 学习宏观指标口径的学生与爱好者\n- 想要可复现数据看板的效率工具用户", "description_en": "Want to quickly gauge whether China's economy is running hot or cold, but tired of jumping between the statistics bureau, the central bank and scattered research reports? Numbers are fragmented, definitions inconsistent, and it's easy to grab stale data.\n\nThis case uses NeoData Financial Search to query GDP, CPI, manufacturing PMI, M2, aggregate financing and the economic calendar in real time, then generates a clean, professional single-file HTML macro dashboard: a temperature verdict, core indicator cards, recent trend charts and an event calendar — red-up/green-down and the 50 boom-bust line at a glance, with EN/ZH toggle and offline support.\n\nGood for:\n- Research/strategy staff who track macro regularly\n- Finance creators and analysts making economic briefs\n- Corporate decision-makers watching money and prices\n- Students learning macro indicator definitions\n- Productivity users who want a reproducible data dashboard", "categories": [ "finance", "data", "search" ], "tags": [ "宏观经济", "金融数据", "数据看板", "NeoData", "经济指标" ], "tags_en": [ "Macro Economy", "Financial Data", "Dashboard", "NeoData", "Economic Indicators" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "prompt": "用 NeoData 金融搜索服务，实时查询中国最新一期的 GDP（同比增速与当期值）、CPI（同比与环比）、制造业 PMI、M2 货币供应量同比、社会融资规模，以及近期重要经济事件日历（如美联储 FOMC 利率决议、CPI/GDP 发布的前值/预测/现值）。务必先通过 NeoData 官方查询取到最新真实数据，不要凭记忆作答，也不要混用其他数据源；若某项查询为空就更换措辞重试。\n\n拿到数据后，生成一张「中国宏观经济仪表盘」，产物为单文件 HTML：包含一句话经济温度结论、核心指标卡片组（GDP/CPI/PMI/M2/社融，各含最新值、同比与简短解读）、指标近几期趋势图、经济事件日历表（事件/时间/前值/预测/现值）、复现查询清单，以及口径与风险提示。\n\n样式方向：浅色专业金融风、信息密度高，指标变化用红升绿降（PMI 高于 50 荣枯线可强调），单文件 HTML、资源全部内联、可离线打开。", "prompt_en": "Use the NeoData Financial Search service to query China's latest GDP (YoY growth and level), CPI (YoY and MoM), manufacturing PMI, M2 money supply YoY, aggregate financing (AFRE), and the recent key economic calendar (e.g., Fed FOMC rate decisions and the previous/forecast/actual for CPI/GDP releases). Always fetch the latest real data via NeoData's official query first — do not answer from memory and do not mix in other data sources; if a query returns nothing, retry with alternate phrasing.\n\nWith the data in hand, generate a \"China Macro-Economy Dashboard\" as a single-file HTML: include a one-line economic-temperature verdict, a group of core indicator cards (GDP/CPI/PMI/M2/AFRE, each with the latest value, YoY change and a short read), recent trend charts, an economic calendar table (event/time/previous/forecast/actual), a reproduce-query checklist, and methodology & risk notes.\n\nStyle direction: clean professional finance look, high information density, red-up/green-down for changes (highlight PMI above the 50 boom-bust line), single-file HTML with all resources inlined and offline-capable.", "skills": [ { "id": "neodata-financial-search", "skill_id": "skill_2053082432761950208", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "中国宏观经济仪表盘" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-22T03:01:53Z", "updated_at": "2026-07-22T03:58:09Z", "is_recommended": false, "recommended_order": 0 }, { "id": "manager-views-ai", "title": "基金经理看AI", "title_en": "Fund Managers on AI", "subtitle": "一个关键词，聚集基金经理们对AI的公开观点", "subtitle_en": "One keyword to gather fund managers' public views on AI.", "description": "想查基金经理对某个行业或主题（比如 AI）怎么看，翻遍官网和公众号也凑不齐观点？\n这个案例以「AI」为关键词，把易方达基金经理们在官网手记与官方公众号上的公开观点聚成一条时间倒序的信息流——只收正文真提AI且经理署名的内容，每条附原文逐字摘录，点标题直达原文。想换成半导体、红利……换个词就能再跑一遍。\n\n适合：跟踪AI、半导体等主题的投资者", "description_en": "Struggling to find what fund managers think about a sector or theme like AI, even after digging through the official site and WeChat? This case takes \"AI\" as the keyword and gathers E Fund managers' public views from official handnotes and the official WeChat articles into one reverse-chronological feed — only manager-attributed content that genuinely mentions AI, each entry with a verbatim excerpt and a title linking to the original. Want semiconductors or dividends instead? Swap the keyword and run it again.\n\nBest for: investors tracking themes like AI and semiconductors.", "categories": [ "finance", "search" ], "tags": [ "基金经理", "主题观点", "AI", "关键词检索" ], "tags_en": [ "Fund Manager", "Theme Views", "AI", "Keyword Search" ], "artifact_type": "html", "cover_image": "cover.jpg", "cover_mode": "landscape", "preview": "output.html", "prompt": "基金经理们最近对 AI 怎么看？\n\n数据须全部来自易方达官方渠道（官网投资观点页 + 官方公众号文章），逐字摘录，不得编造、不得引用第三方。", "prompt_en": "What do fund managers recently think about AI?\n\nAll data must come from E Fund's official channels (the official viewpoints page + official WeChat articles), verbatim excerpts, with no fabrication and no third-party sources.", "skills": [ { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T14:38:00+08:00", "updated_at": "2026-08-31T17:25:42+08:00", "artifacts": [ { "file": "output.html", "label": "AI 主题观点报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "manager-views-overview", "title": "基金经理观点速览", "title_en": "Manager Views at a Glance", "subtitle": "不用逐篇翻文章，一条时间流看最新易方达基金经理观点", "subtitle_en": "No more article-hunting — one timeline for the latest E Fund manager views", "description": "想快速知道最近基金经理们在想什么，又没时间一篇篇翻文章？这个案例把近期基金经理公开观点收成一条时间倒序的信息流——每一条都是原文核心语句逐字摘录，点标题就能跳回原文。适合：关注公募投研风向的投资者", "description_en": "Want to know what fund managers are thinking lately without digging through dozens of articles? This case pulls together their latest publicly-stated views into one reverse-chronological feed — every entry is a verbatim excerpt of key statements, and titles link back to the original articles.\n\nBest for: investors tracking public fund research trends.", "categories": [ "finance", "search" ], "tags": [ "基金经理", "观点总览", "投研风向", "易方达" ], "tags_en": [ "Fund Manager", "Views Overview", "Research Trends", "E Fund" ], "artifact_type": "html", "cover_image": "cover.jpg", "cover_mode": "landscape", "preview": "output.html", "prompt": "请总结基金经理最近发表了哪些观点？\n\n数据须全部来自易方达官方渠道（官网投资观点页 + 官方公众号文章），逐字摘录，不得编造、不得引用第三方。", "prompt_en": "Summarize what fund managers have said recently.\n\nAll data must come from E Fund's official channels (the official viewpoints page + official WeChat articles), verbatim excerpts, with no fabrication and no third-party sources.", "skills": [ { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T14:38:00+08:00", "updated_at": "2026-08-31T17:25:42+08:00", "artifacts": [ { "file": "output.html", "label": "基金经理观点总览报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "manager-views-quarterly", "title": "季报观点精华", "title_en": "Quarterly Report Views Digest", "subtitle": "最新一期季报，易方达基金经理们都有哪些观点？", "subtitle_en": "What did E Fund managers say in the latest quarterly report?", "description": "每季度一次的定期报告披露期，几十只产品几百页报告，基金经理们到底说了什么？这个案例把最新一期季报的基金经理观点提炼成一份两段式精华报告。案例演示了如何让 AI 检索官方公众号的团队发声文章与定期报告运作分析，经理观点按人独立呈现、不混写。适合：季报披露期想快速扫读基金经理观点的投资者", "description_en": "Every quarterly report season: dozens of products, hundreds of pages of reports — what did the fund managers actually say? This case distills the latest E Fund quarterly report into a two-segment manager-views digest. It demonstrates how AI retrieves team-voice articles from official WeChat accounts and quarterly operating analyses, presenting each manager's views independently without blending them together. Best for: investors who want to scan fund managers' views quickly during the quarterly-report disclosure window.", "categories": [ "finance", "search" ], "tags": [ "基金经理", "季报", "定期报告", "观点梳理" ], "tags_en": [ "Fund Manager", "Quarterly Report", "Periodic Report", "Views" ], "artifact_type": "html", "cover_image": "cover.jpg", "cover_mode": "landscape", "preview": "output.html", "prompt": "最新一期季报中基金经理怎么说？按提及的行业/方向梳理观点。\n\n数据须全部来自易方达官方渠道（官网季报运作分析原文 + 官方公众号文章），逐字照录，不得编造、不得引用第三方。", "prompt_en": "What did fund managers say in the latest quarterly report? Organize views by industry/direction mentioned.\n\nAll data must come from E Fund's official channels (official quarterly operating-analysis originals + official WeChat articles), verbatim, with no fabrication and no third-party sources.", "skills": [ { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T14:38:00+08:00", "updated_at": "2026-08-31T17:25:42+08:00", "artifacts": [ { "file": "output.html", "label": "季报观点精华报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "market-competitor-pricing", "title": "SaaS 竞品定价分析报告", "title_en": "SaaS Competitor Pricing Analysis Report", "subtitle": "深度对比主流项目管理工具定价策略，洞察市场趋势", "subtitle_en": "In-depth comparison of mainstream project management tool pricing strategies and market trends", "description": "本报告全面分析 5 款主流 SaaS 项目管理工具（Notion、Asana、Monday.com、ClickUp、Teambition）的定价策略。\n\n报告涵盖各产品定价层级详细对比、每用户月均成本计算、核心差异化功能分析、用户评价摘要及近两年定价趋势观察。\n\n基于竞品分析结果，提供针对性的定价策略建议，帮助产品团队制定具有竞争力的价格体系。", "description_en": "This report comprehensively analyzes the pricing strategies of 5 mainstream SaaS project management tools: Notion, Asana, Monday.com, ClickUp, and Teambition.\n\nThe report covers detailed pricing tier comparisons, per-user monthly cost calculations, core differentiating feature analysis, user review summaries, and pricing trend observations over the past two years.\n\nBased on the competitive analysis results, it provides targeted pricing strategy recommendations to help product teams develop a competitive pricing structure.", "categories": [ "search", "business" ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "tags": [ "竞品分析", "定价策略", "SaaS", "项目管理", "市场调研" ], "tags_en": [ "competitor analysis", "pricing strategy", "SaaS", "project management", "market research" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "竞品定价分析报告" } ], "prompt": "帮我做一份 SaaS 项目管理工具竞品定价分析报告。对比 5 款主流产品（Notion、Asana、Monday.com、ClickUp、Teambition）的定价策略，包含：各产品定价层级详细对比（免费版/专业版/企业版功能矩阵）、每用户月均成本计算、核心差异化功能分析、用户评价摘要（优缺点各 3 条）、定价趋势观察（近 2 年调价历史）、针对我们产品的定价策略建议。输出为结构化 Markdown 报告，含对比表格。", "prompt_en": "Help me produce a competitor pricing analysis report for SaaS project-management tools. Compare the pricing strategies of 5 mainstream products (Notion, Asana, Monday.com, ClickUp, Teambition), including: a detailed comparison of each product's pricing tiers (feature matrix for Free/Pro/Enterprise editions), the calculated average monthly cost per user, an analysis of core differentiating features, a summary of user reviews (3 pros and 3 cons each), pricing-trend observations (price-change history over the past 2 years), and pricing-strategy recommendations for our own product. Output as a structured Markdown report with comparison tables.", "skills": [ { "id": "market-researcher", "name": "市场调研", "description": "市场调研专家，提供市场分析、消费者洞察与机会评估", "description_en": "Market research specialist for analysis, consumer insights, and sizing", "skill_id": "skill_2053082339107885056" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "source": "official", "quality_score": 93, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "marriage-counsel-conversion-playbook", "title": "婚姻家事律师·咨询转化六段话术工作台", "title_en": "Marriage & Family Lawyer Conversion Playbook", "subtitle": "按客户决策阶段匹配话术，附转化跟踪卡，把咨询变成可复用的转化工具", "subtitle_en": "Match scripts to each client decision stage, with a conversion tracker to turn consultations into a reusable funnel tool", "description": "最怕客户问完问题扭头就走，白嫖一次咨询就没了下文。这个案例把婚姻家事律师的接待经验沉淀成一套「六段话术工作台」：从开场破冰、诊断提问、钩子留客、窗口期唤醒，到促成出口和异议收尾，每一段都附「为什么有效」和执业红线提示。\n\n更关键的是内置了一张「转化跟踪卡」——记录客户决策阶段、关键事实、已用话术和下一步待办，数据只存在本机浏览器，帮助律师追踪每一位客户的转化进度，把一次性咨询变成可持续跟进的线索。\n\n适合：婚姻家事律师、离婚业务律师、律所案源与咨询接待人员、准备做个人 IP 的执业律师。", "description_en": "Tired of clients asking a question and then disappearing after one free consultation? This case turns a family lawyer's reception experience into a six-stage script workbench: opening, diagnostic questioning, hook, urgency, closing, and objection handling — each with a \"why it works\" note and a professional red-line reminder.\n\nMore importantly, it includes a built-in conversion tracker that records each client's decision stage, key facts, scripts used, and next actions — all stored locally in the browser — helping lawyers follow up on every lead and turn one-off consultations into a lasting pipeline.\n\nFor: family & marriage lawyers, divorce practitioners, law firm intake and reception staff, and practicing lawyers building a personal brand.", "categories": [ "business" ], "tags": [ "婚姻家事", "咨询转化", "话术", "律师营销", "客户跟进" ], "tags_en": [ "Family Law", "Consultation Conversion", "Sales Script", "Lawyer Marketing", "Client Follow-up" ], "artifact_type": "html", "preview": "index.html", "artifacts": [ { "file": "index.html", "label": "六段话术工作台（含转化跟踪卡）" } ], "prompt": "你是一位执业8年的婚姻家事律师，正在做个人 IP 与案源转化。请帮我生成一个「咨询转化六段话术工作台」，用于接待客户时按决策阶段快速调取话术并跟踪转化进度。\n\n要求：\n1. 分成六个阶段：开场破冰、诊断提问、钩子留客、窗口期唤醒、促成出口、异议收尾。每个阶段给出可背诵的话术原文，并附「为什么有效」的心理学/信任机制说明。\n2. 话术要符合婚姻家事场景：开场先立身份、给安全感、强调「不一定都要请律师」；诊断阶段用领证时间/财产归属/孩子抚养三件事追问；钩子阶段给一半答案留一个待办；窗口期只讲客观的证据灭失与时间风险，不恐吓、不承诺结果。\n3. 底部加一张「转化跟踪卡」，可填写客户脱敏称呼、咨询日期、当前决策阶段、关键事实（领证时间/财产/孩子）、已用话术、下一步待办、下次触达时间，并支持本地保存。\n4. 页脚常驻执业安全红线：不承诺结果、不制造焦虑、正式交付须律师人工确认。\n\n请以「单文件 HTML、资源全内联、可离线打开」的方式输出，视觉风格专业克制（深色标题 + 留白 + 单一强调色）。话术中的「你老公」「他」等表述保留，标注可按实际案情替换。", "prompt_en": "You are a family & marriage lawyer with 8 years of practice, building a personal brand and intake funnel. Generate a \"Six-Stage Consultation Conversion Workbench\" to quickly pull up scripts by decision stage and track conversion progress.\n\nRequirements:\n1. Six stages: opening, diagnostic questioning, hook, urgency, closing, and objection handling. Each stage provides a memorizable script plus a \"why it works\" explanation of the underlying trust mechanism.\n2. Scripts must fit family-law scenarios: opening establishes identity and safety while stressing \"you may not need a lawyer at all\"; diagnostic stage probes three things (marriage registration date / property ownership / child custody); hook stage gives half an answer and leaves a follow-up task; urgency stage only states objective evidence-loss and time risks — no fear-mongering, no outcome promises.\n3. Add a \"conversion tracker\" at the bottom for recording the client's anonymized name, consultation date, current stage, key facts (marriage date / property / children), scripts used, next actions, and next contact time, with local save support.\n4. Keep a persistent professional red-line footer: no outcome promises, no fear-mongering, and formal deliverables require lawyer confirmation.\n\nOutput as a single-file HTML with all resources inlined and offline-capable, in a restrained professional style (dark headings + whitespace + a single accent color). Keep the original gendered expressions and note that they may be replaced per case.", "skills": [], "experts": [ { "id": "brand-marketing", "name": "品牌营销专员", "description": "品牌营销专家，专注处理律师个人IP运营和律所品牌营销相关事务，涵盖律师个人IP定位、律所品牌营销方案、社交媒体内容运营、营销数据分析、获客转化方案、法律科普内容创作、短视频脚本创作等。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "3d642417d5d0", "expert_id": "ex_AYR4CRp3puyw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-10T11:38:30+08:00", "updated_at": "2026-09-10T11:38:30+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "math-teacher-workbench", "title": "数学教学工作台", "title_en": "Math Teaching Workbench", "subtitle": "待办、课表、学情、成绩、座位、值日、家校沟通与健康打卡，一位数学老师的带班中枢", "subtitle_en": "Todos, timetable, student tracking, scores, seating, duty roster, parent communication and health check-in — a homeroom hub for a math teacher", "description": "带两个班的数学老师，每天要在备课、批改、家长沟通和班级事务之间来回切换，信息散在微信群、备忘录和成绩单里。\n\n这个工作台把日常收进一屏：首页是今日概览——今天要处理的待办（逾期的置顶标红）、今日课程、喝水打卡和教学进度；「数学教学」里课表、教学安排、题库、学情分析各归其位；「学情分析」按学生档案和成绩记录跟踪薄弱知识点，家校沟通单独留档；「班级管理」覆盖学生档案、座位安排（支持六列排座/四大组双列/小组合座三种布局自动排座）、值日表和常用网站；「身体健康」板块用喝水打卡联动首页。\n\n数据层面 13 张表共用一套字段（模块区分），在线版数据落在资料库数据表、支持在线存储与双向同步；演示版内置 560 条示例数据。内置示例数据覆盖四年级 5 班/6 班的真实场景，打开即可对照使用。\n\n适合：小学、初中带班老师，以及需要一套「教学—学情—班务」闭环管理的班主任。", "description_en": "A math teacher running two classes juggles lesson prep, grading, parent communication and class affairs — with information scattered across group chats, notes and grade sheets.\n\nThis workbench brings the day into one screen: a home overview with today's todos (overdue pinned in red), today's timetable, water check-in and teaching progress; a teaching hub with timetable, lesson plans, question bank and learning analytics; student tracking that follows each learner's weak points through score records, with parent communication archived separately; class management covering student profiles, seating (three layouts with auto-arrangement), duty roster and frequently-used sites; and a health section whose water check-in links back to the home screen.\n\nUnder the hood, 13 tables share one unified schema (a module field separates concerns); the online version stores data in Library database tables with two-way sync, while the demo build ships 560 built-in sample records covering real Grade-4 scenarios, ready to explore.\n\nBest for: primary and middle school subject teachers, and homeroom teachers who want a closed loop across teaching, learning analytics and class affairs.", "categories": [ "workspace" ], "tags": [ "教师工作台", "教学管理", "学情分析", "班级管理", "家校沟通" ], "tags_en": [ "Teacher Workbench", "Teaching Management", "Learning Analytics", "Class Management", "Parent Communication" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "数学教学工作台（单文件演示版 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/p7tnhDkbMoQGkisM3WhElR，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「数学教学工作台」，面向小学/初中带班数学老师，用 13 张数据表（共用一套字段，用「模块」字段区分业务：教学待办、课程表、学生跟踪、教学进度、学期设置、家校沟通、教学安排、题库、成绩记录、座位安排、值日安排、常用网站、健康打卡）管理班级日常。字段包括：模块（select）、事项名称、班级（select）、星期（number）、节次（number）、优先级（select）、状态（select）、关注程度（select）、课堂表现（select）、近期成绩（number）、薄弱知识点、家校沟通、计划课时（number）、完成课时（number）、截止日期（date）、备注、示例数据（checkbox）、单元、作业类型（select）、批改状态（select）、发送身份（select）、沟通要点等。\n\n页面功能：首页今日概览（今日待办支持标记完成与顺延、逾期置顶标红、今日课程、喝水打卡与教学进度）；教学模块含课表视图、教学安排进度、题库与组卷、学生成绩录入（支持 CSV 解析预览导入）与班级成绩分析；学情分析含学生档案、重点关注与近期成绩趋势；班级管理含学生档案、座位安排（六列排座/四大组双列/小组合座三种布局、一键自动排座与清空）、值日表（按班级切换）与常用网站导航；身体健康含每日喝水打卡（目标杯数、撤销，与首页联动）与健康打卡记录。\n\n视觉：浅色教育风、卡片化分区、左侧目录导航+顶部状态栏，状态用彩色徽章区分（优先级/批改状态/关注程度），PC 与移动端自适应。内置四年级 5 班/6 班示例数据，打开即可对照使用。\n\n交付：资料库在线工作台——在线 HTML 页面 + 13 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/p7tnhDkbMoQGkisM3WhElR with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Math Teaching Workbench\" for primary/middle school math teachers, managing class routines with 13 data tables sharing one unified schema (a module field separates concerns: teaching todos, timetable, student tracking, teaching progress, term settings, parent communication, lesson plans, question bank, score records, seating, duty roster, favorite sites, health check-in). Fields include: module (select), item name, class (select), weekday (number), period (number), priority (select), status (select), attention level (select), class performance (select), recent score (number), weak points, parent communication, planned lessons (number), completed lessons (number), due date (date), notes, sample flag (checkbox), unit, homework type (select), grading status (select), sender identity (select), communication points, etc. Page features: a home overview (today's todos with complete/postpone actions, overdue pinned in red, today's timetable, water check-in and teaching progress); a teaching hub with timetable view, lesson-plan progress, question bank and paper assembly, score entry with CSV parse-preview import, and class score analysis; learning analytics with student profiles, key-attention list and recent score trends; class management with student profiles, seating (six-column / four-group paired / small-group layouts, one-click auto-arrange and clear), duty roster (switchable by class) and a site navigator; a health section with daily water check-in (goal cups, undo, linked to home) and check-in records. Visual: light education style, card-based sections, left navigation plus top status bar, colored badges for priority/grading/attention states, responsive for PC and mobile. Ships with Grade-4 Class 5/6 sample data ready to explore.\n\nDeliverable: an online Library workbench — an online HTML page plus 13 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "车棍儿老师", "name_en": "Cheguner Laoshi", "avatar": "", "badge": "community" }, "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "mcps": [], "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "source": "community", "created_at": "2026-09-10T16:30:00+08:00", "updated_at": "2026-09-10T17:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "may-fourth-movement-new-democratic-revolution", "title": "五四运动与新民主主义革命的开端", "title_en": "五四运动与新民主主义革命的开端", "subtitle": "高中历史 · 五四运动与新民主主义革命的开端", "subtitle_en": "高中历史 · 五四运动与新民主主义革命的开端", "description": "高中历史《五四运动》互动课件，以九页漫画长卷翻阅 1919，并用对照表梳理新旧民主主义革命的差异。", "description_en": "高中历史《五四运动》互动课件，以九页漫画长卷翻阅 1919，并用对照表梳理新旧民主主义革命的差异。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "may-fourth-movement-new-democratic-revolution.html", "artifacts": [ { "file": "may-fourth-movement-new-democratic-revolution.html", "label": "五四运动与新民主主义革命的开端｜互动教学网页" } ], "prompt": "帮我做一个高中历史《五四运动与新民主主义革命的开端》的互动教学网页，要以漫画翻页长卷的形式学历史。", "prompt_en": "帮我做一个高中历史《五四运动与新民主主义革命的开端》的互动教学网页，要以漫画翻页长卷的形式学历史。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "media-studio-workbench", "title": "自媒体运营工作台", "title_en": "Media Studio Workbench", "subtitle": "单文件 HTML 离线自媒体运营台，选题·脚本·发布·复盘一个台闭环", "subtitle_en": "Single-file offline creator workbench — ideas, scripts, publishing & review in one place", "description": "做自媒体最累的不是写，是脑子里同时转着一堆事：这周写什么、开头怎么钩人、几点发、发完那条到底行不行——全散在备忘录和聊天记录里，越做越乱。\n\n这个案例产出了一个完整的单文件 HTML 自媒体运营工作台，把博主每天/每周真正在重复的「招牌动作」串成一条飞轮：① 今日——灵感速记随手记（可一键转选题）+ 今天要做的智能提醒（逾期/待录数据）+ 连更里程碑徽章 + 每日激励；② 选题——每周选题会 4 步方法论 + 三类选题库（爆款/日常/系列）+ 一句话大纲 + 本周配比进度 + 爆款雷达（收集对标爆款自动聚合钩子/平台/赛道洞察）；③ 脚本——黄金 3 秒开头 6 公式 + 爆款结构模板（视频/图文切换）+ AI 拆解/写脚本指令生成器（把你的输入编排成专家级 prompt，复制到任意 AI 跑）+ 资产库；④ 发布——5 平台最佳时段 + 发布排期（逾期红标、可改期/删除）+ 发布后黄金 1 小时清单；⑤ 复盘——四率看板（完播/互动/关注/转化，图文视频自动切换指标）+ 曝光×涨粉双轴趋势 + 归因图（按选题库/钩子/时段/形态/平台切换）+ 自动判定爆了/达标/低于预期。视觉是小红书素人女生喜欢的珊瑚粉少女风，数据全存浏览器本地、零后端、完全离线，手机可加到桌面当 App。\n\n适合：正在做号涨粉、希望把选题到复盘系统化沉淀下来的泛自媒体博主；图文和视频都发、需要一套能兼容两种数据口径的运营工具的创作者；想看一个「单文件 HTML + 原生 SVG 图表 + localStorage」完整运营台实现的前端开发者。", "description_en": "The exhausting part of being a creator isn't writing — it's juggling everything at once: what to post this week, how to hook the opening, when to publish, and whether that last post actually worked. It all ends up scattered across notes and chat logs, getting messier over time.\n\nThis case delivers a complete single-file HTML creator workbench that turns a blogger's real daily/weekly signature moves into one flywheel: (1) Today — quick idea capture (one-click convert to topic) + smart to-dos (overdue / data-to-log) + streak milestone badges + daily encouragement; (2) Topics — a 4-step weekly topic-meeting method + three topic libraries (viral / routine / series) + one-line outlines + weekly mix progress + a Viral Radar that collects benchmark hits and auto-aggregates hook / platform / niche insights; (3) Scripts — 6 golden-3-second opening formulas + viral structure templates (video / text toggle) + an AI prompt generator for teardown & scriptwriting (composes your input into an expert-level prompt to run in any AI) + an asset library; (4) Publish — best posting times for 5 platforms + a publishing schedule (overdue flags, editable / deletable) + a golden-first-hour checklist; (5) Review — a four-rate dashboard (completion / engagement / follow / conversion, auto-switching metrics for video vs. text) + a dual-axis reach×followers trend + an attribution chart (by library / hook / time slot / format / platform) + auto verdicts (viral / on-target / underperforming). The visual style is a soft coral \"xiaohongshu girl\" aesthetic; all data lives in the browser locally with zero backend, fully offline, and can be added to a phone home screen as an app.\n\nSuitable for: general-interest creators growing an audience who want to systematize everything from topic selection to review; creators posting both text and video who need one tool compatible with both data models; frontend developers interested in a complete \"single-file HTML + native SVG charts + localStorage\" workbench implementation.", "categories": [ "workspace", "productivity", "data" ], "tags": [ "自媒体运营", "选题复盘", "内容排期", "爆款拆解", "离线HTML" ], "tags_en": [ "Creator ops", "Topic & review", "Content scheduling", "Viral teardown", "Offline HTML" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "自媒体运营工作台（单文件 HTML）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/VEmid0pCbQMW6V1GnGNAdm），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个泛自媒体博主用的运营工作台，要求：所有 CSS/JS/可离线打开、手机和电脑都能用、数据存 localStorage 自动保存。把选题→脚本→发布→复盘这条主循环做成一个台，分五个板块：① 今日：灵感速记（可一键转选题）、今天要做的（含逾期/待录数据智能提醒）、连更里程碑徽章、每日激励语；② 选题：每周选题会 4 步方法、三类选题库（爆款/日常/系列/测试）、一句话大纲、本周配比进度、爆款雷达（收集对标爆款并自动聚合钩子类型/平台/赛道的洞察，可一键转选题）；③ 脚本：黄金 3 秒开头公式库、爆款结构模板（视频/图文切换）、一个「AI 指令生成器」把选题+钩子+结构+要点编排成写脚本/拆解的专家级 prompt 供复制到任意 AI 跑、我的资产库（可搜索/星级/复制）；④ 发布：各平台最佳发布时段、发布排期（逾期红标、可改期/删除、标已发）、发布后黄金 1 小时动作清单；⑤ 复盘：四率看板（完播/互动/关注/转化，视频看完播、图文看收藏率，按形态自动切换）、曝光×涨粉双轴趋势图、归因图（可按选题库/钩子/时段/形态/平台切换看什么最带曝光）、每条内容自动判定爆了/达标/低于预期。含示例数据可清空、JSON 备份导入导出、CSV 导出。视觉风格：小红书素人女生喜欢的珊瑚粉少女风（奶油底 + 珊瑚粉渐变 Hero + 圆润卡片），温暖有活力。图表一律用原生 SVG 自绘，不要引用任何外部 CDN 或字体。\n\n交付：资料库在线工作台——在线 HTML 页面 + 5 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/VEmid0pCbQMW6V1GnGNAdm) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nBuild me a content-operations workbench for general-interest creators. Requirements: all CSS/JS/assets inlined, works offline, on both phone and desktop, data auto-saved to localStorage. Turn the topic→script→publish→review main loop into one workbench with five sections: (1) Today: quick idea capture (one-click to topic), a to-do list with smart reminders (overdue / data-to-log), streak milestone badges, daily encouragement; (2) Topics: a 4-step weekly topic-meeting method, three topic libraries (viral / routine / series / test), one-line outlines, weekly mix progress, a Viral Radar that collects benchmark hits and auto-aggregates hook-type / platform / niche insights with one-click convert-to-topic; (3) Scripts: a golden-3-second opening formula library, viral structure templates (video/text toggle), an \"AI prompt generator\" that composes topic + hook + structure + key points into an expert-level scriptwriting/teardown prompt to copy into any AI, plus a searchable/rated/copyable asset library; (4) Publish: best posting times per platform, a publishing schedule (overdue flags, editable / deletable, mark-as-published), and a golden-first-hour action checklist; (5) Review: a four-rate dashboard (completion / engagement / follow / conversion, video shows completion while text shows save rate, auto-switching by format), a dual-axis reach×followers trend chart, an attribution chart (switchable by library / hook / time slot / format / platform), and auto verdicts (viral / on-target / underperforming) per post. Include sample data that can be cleared, JSON backup import/export, and CSV export. Visual style: a soft coral \"xiaohongshu girl\" aesthetic (cream background + coral gradient hero + rounded cards), warm and lively. Draw all charts with native inline SVG; do not reference any external CDN or fonts.\n\nDeliverable: an online Library workbench — an online HTML page plus 5 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "created_at": "2026-08-01T20:50:00+08:00", "updated_at": "2026-08-01T20:50:00+08:00", "preview_en": "output.html", "scenario_cover": "scenario-cover.png", "is_recommended": true, "recommended_order": 14 }, { "id": "medication-consultation", "title": "药品咨询", "title_en": "Medication Consultation", "subtitle": "家里的药弄明白：用途、用法用量、注意事项、靠谱购买渠道，一页讲清", "subtitle_en": "Understand your medications in one page: uses, dosage, precautions, and trustworthy ways to get them", "description": "几乎每个家庭都有几盒常吃或刚配的药，但真正把它弄明白的人不多：治什么、怎么吃、要当心什么、吃完了去哪儿买——说明书字小又难懂，网上信息杂得不敢全信。药这件事，糊涂不得。\n\n以抗凝药华法林为例：问微小医「这是什么药、怎么用、要注意什么、去哪买」，它会把用途、用法用量、出血的预警信号、食物药物的相互影响一条条讲清楚，还告诉你医院药房、实体药店、互联网医院三条靠谱的购买路径——生成一页用药指南，存图转给家人，吃药这件事就不再糊涂了。", "description_en": "Almost every household keeps a few boxes of medication, yet few people truly understand them — what they're for, how to take them, what to watch out for, where to get more. The leaflet is dense, the internet is messy, and medication is not something you can afford to be fuzzy about.\n\nTake the anticoagulant warfarin as an example: ask Weixiaoyi, and it patiently walks you through the drug's purpose, dosage, bleeding warning signs and food-drug interactions, then shows you three trustworthy ways to get it — hospital pharmacy, retail pharmacy, or internet hospital — all in a one-page guide you can save and share, so taking medication never feels confusing again.", "categories": [ "business" ], "tags": [ "药品咨询", "用药指南", "华法林", "微医", "HTML生成" ], "tags_en": [ "Medication Consultation", "Medication Guide", "Warfarin", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "medication-consultation.html", "artifacts": [ { "file": "medication-consultation.html", "label": "华法林用药全指南 HTML 页面" } ], "prompt": "华法林是什么药？该如何使用？有什么注意事项？该如何购买？将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203），logo裁成圆形。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果。5. 重点内容用流程图、指标卡片等特殊样式进行说明。", "prompt_en": "What is warfarin? How should it be used? What precautions apply? How can it be purchased? Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for sharing; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203), with the logo cropped into a circle. HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation; 5. Key content should be explained with special styles such as flow charts and metric cards.", "skills": [ { "id": "we-health-skill", "name": "微医健康管家", "description": "微医AI健康管家（官方Skill），覆盖报告解读、就医推荐、用药咨询、健康咨询等全部医疗健康场景", "slug": "we-health-skill" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-20T17:05:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "medication-safety-pregnancy", "title": "孕期哺乳期用药安全咨询", "title_en": "Pregnancy & Breastfeeding Medication Safety", "subtitle": "备孕 / 妊娠 / 哺乳期，一键问清药品能不能继续用", "subtitle_en": "Check whether a medication is safe to keep taking during pregnancy or breastfeeding", "description": "发现自己怀孕了，最慌的就是手里正吃着的药——二甲双胍、降压药、安眠药，到底停不停、会不会影响宝宝？\n\n这个案例演示了把「孕周 + 正在吃的药」交给 AI 临床助手，它自动识别备孕 / 妊娠 / 哺乳期，结合说明书与妊娠安全性分级做风险权衡，既不说「绝对能吃」，也不吓唬你擅自停药。\n\n适合：备孕人群、孕期准妈妈、哺乳期妈妈、家有孕妇的家庭成员、需要用药安全科普的普通用户。", "description_en": "Just found out you're pregnant, and the scariest part is the medication you're already taking — metformin, blood pressure pills, sleep aids. Should you stop? Will it affect the baby?\n\nThis case shows how to hand your gestational age and current medications to the AI clinical assistant, which identifies your pregnancy or breastfeeding status and weighs the risks using drug labels and pregnancy safety categories — without giving a flat \"safe\" or a scare-you-into-stopping answer.\n\nFor: people trying to conceive, expectant mothers, breastfeeding mothers, family members of pregnant women, and anyone wanting plain-language medication safety guidance.", "categories": [ "agent", "education" ], "tags": [ "用药安全", "孕期用药", "哺乳期用药", "AI临床助手", "腾讯健康" ], "tags_en": [ "Medication Safety", "Pregnancy", "Breastfeeding", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "孕期哺乳期用药安全咨询演示（HTML，含中英双语）" } ], "prompt": "请以「腾讯健康-AI临床助手-用药安全」的能力，帮我评估一个孕期用药安全问题：\n\n我刚发现怀孕 6 周，之前一直在吃盐酸二甲双胍 0.5g（每日两次）。请判断：\n1. 二甲双胍在孕早期的安全性分级与潜在影响；\n2. 继续服用与擅自停药的风险权衡（提示：血糖失控对母胎的危害往往大于二甲双胍本身的风险）；\n3. 我应该做哪些检查、尽快咨询哪些科室（产科 / 内分泌）。\n\n请用风险分级（而非「绝对能吃 / 绝对不能吃」）表述；不推荐具体替代药品名称；不擅自建议停药、减量或换药；末尾声明最终以产科 / 专科医生医嘱为准。", "prompt_en": "Using the medication-safety capabilities of the Tencent Health AI Clinical Assistant, evaluate a pregnancy medication question for me:\n\nI just found out I'm 6 weeks pregnant and have been taking metformin 0.5g twice daily. Please assess:\n1. The pregnancy safety category of metformin in early pregnancy and its potential impact;\n2. The risk trade-off between continuing the drug and stopping it on your own (note: uncontrolled glucose is usually more harmful to mother and baby than metformin itself);\n3. Which tests I should run and which clinics I should see (obstetrics / endocrinology) promptly.\n\nUse risk-grading language rather than a flat \"safe / unsafe\" verdict. Do not name specific alternative drugs. Do not recommend stopping, reducing, or switching the drug on your own. End with a disclaimer that the final decision rests with an obstetrician / specialist.", "skills": [ { "id": "medication-advisor", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "slug": "medication-advisor" } ], "experts": [ { "id": "chief-physician", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_0JLBYdNE0Hxk" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T11:01:00+08:00", "updated_at": "2026-08-27T00:01:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "medpedia-chronic-care", "title": "高血压日常管理，问医典就清楚", "title_en": "Hypertension Care, Answered by Medpedia", "subtitle": "真实检索医典文章，日常管理要点句句有出处", "subtitle_en": "Real Medpedia retrieval — daily-management tips, every point sourced", "description": "刚查出高血压，医生说了「低盐、吃药、监测」，可具体怎么做、为什么不能停，脑子里还是空的。网上搜出来的答案真假难辨。\n\n这个案例演示把「高血压患者在日常生活中需要注意什么」交给腾讯医典健康问答：它真实检索腾讯医典官方科普知识库，归纳出「按时服药不擅自停、低盐低脂饮食、监测血压、控制体重与情绪」四要点，并在文末附上 6 篇真实文章链接与审核署名，句句可溯源。\n\n适合：慢病患者日常管理、社区健康宣教、体检后科普、医院患者教育。", "description_en": "Newly diagnosed with hypertension, the doctor said \"less salt, take pills, monitor\" — but how exactly, and why not stop, stays vague. Online answers mix truth and myth.\n\nThis case shows handing \"what should hypertension patients watch daily\" to Tencent Medpedia Q&A: it truly queries the official Medpedia knowledge base and summarizes four points — take meds on time, low-salt low-fat diet, monitor BP, control weight and mood — with 6 real article links and reviewer attributions at the end, every point traceable.\n\nGood for: chronic-disease self-management, community health education, post-checkup literacy, hospital patient education.", "categories": [ "education", "productivity" ], "tags": [ "腾讯医典", "健康科普", "高血压", "慢病管理", "可溯源" ], "tags_en": [ "Medpedia", "health literacy", "hypertension", "chronic care", "traceable" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "刚查出高血压，医生开了药就让我走了。日常生活到底要注意什么？吃盐、运动、什么时候量血压，血压降下来了能不能停药？", "prompt_en": "I was just diagnosed with hypertension — the doctor prescribed medication and sent me off. What do I actually need to watch day to day? Salt, exercise, when to measure blood pressure, and can I stop medication once it normalizes?", "skills": [ { "id": "tencent-medpedia-aiqa-skill", "name": "腾讯医典健康问答", "description": "检索腾讯医典官方科普知识库作答，句句可溯源；含儿童用药剂量本地计算", "skill_id": "skill_2076988185622831104" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-22T17:14:53Z", "artifacts": [ { "file": "output.html", "label": "高血压日常管理，问医典就清楚" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "medpedia-fever-triage", "title": "孩子发烧，什么时候该去医院", "title_en": "When to Take a Feverish Child to Hospital", "subtitle": "真实检索医典，给出体温判断与就医信号", "subtitle_en": "Real Medpedia retrieval — temperature guide and red-flag signs", "description": "半夜孩子烧到 38.7℃，抱起来又活蹦乱跳，要不要现在冲急诊？家长最怕判断错。\n\n这个案例演示把「孩子发烧到多少度需要去医院」交给腾讯医典健康问答：它真实检索医典，给出腋温分度（低热/中度/高热），并强调「不看数字看状态」——精神差、呼吸快、3 月龄以下婴儿发热等才是就医信号，附 6 篇真实文章链接。\n\n适合：育儿家庭健康决策、母婴社群科普、儿科分诊前置、幼儿园/托育机构宣教。", "description_en": "At midnight the child hits 38.7°C but is lively when held — rush to ER now? Parents fear misjudging.\n\nThis case shows handing \"at what temperature should a child go to hospital\" to Tencent Medpedia Q&A: it truly queries Medpedia, gives axillary-temperature bands (low/moderate/high), and stresses \"watch the state, not the number\" — lethargy, fast breathing, fever in under-3-month infants are the real signals — with 6 real article links.\n\nGood for: parenting health decisions, mom-baby community literacy, pediatric triage pre-step, nursery education.", "categories": [ "education", "productivity" ], "tags": [ "腾讯医典", "儿童发热", "就医判断", "分诊", "可溯源" ], "tags_en": [ "Medpedia", "pediatric fever", "triage", "red flags", "traceable" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "孩子晚上突然发烧，腋温 38.5℃，精神还行能吃能玩。到多少度、什么情况必须去医院？现在在家能做什么？", "prompt_en": "My child suddenly developed a fever tonight — 38.5°C under the arm, still alert, eating and playing. At what temperature or in which situations must we go to the hospital? What can I do at home right now?", "skills": [ { "id": "tencent-medpedia-aiqa-skill", "name": "腾讯医典健康问答", "description": "检索腾讯医典官方科普知识库作答，句句可溯源；含儿童用药剂量本地计算", "skill_id": "skill_2076988185622831104" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-22T17:14:53Z", "artifacts": [ { "file": "output.html", "label": "孩子发烧，什么时候该去医院" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "medpedia-health-qa", "title": "糖尿病如何治疗", "title_en": "How Is Diabetes Treated", "subtitle": "腾讯医典官方知识库，句句可溯源的健康解答", "subtitle_en": "Official Tencent Medpedia knowledge base — every answer traceable", "description": "身体不舒服时最怕上网搜出一堆真假难辨的信息。这个案例演示了如何把健康问题交给腾讯医典健康问答：输入「糖尿病如何治疗？」，它直接检索医典官方科普知识库，先点明「无法根治但可控制」的核心结论，再分六步展开（先明确分型、药物口服降糖药+胰岛素、营养治疗、餐后运动、血糖监测、终身管理），每条要点都标注引用来源，文末附参考文章链接（标题 + 链接，最多 6 条），句句可溯源，全程只引用、不编造。\n\n适合：关心父母糖尿病管理的子女、需要权威健康科普的普通用户、想给产品接入可溯源医疗问答的开发者。", "description_en": "Googling symptoms is scary — you never know what to trust. This case shows how to hand a health question to Tencent Medpedia Health Q&A: ask \"how is diabetes treated\", and it queries Medpedia's official health-science knowledge base, first stating the core conclusion \"not curable, but controllable\", then expanding into six steps (know your type, oral drugs + insulin, nutrition therapy, post-meal exercise, glucose monitoring, lifelong management) — every point cited to a source, with a reference list of article titles and links (up to 6). It only quotes, never fabricates.\n\nFor: adult children managing parents' diabetes, everyday users who want authoritative health science, and developers who want traceable medical Q&A in their product.", "categories": [ "education", "search" ], "tags": [ "健康问答", "腾讯医典", "糖尿病", "健康科普", "可溯源" ], "tags_en": [ "Health Q&A", "Tencent Medpedia", "Diabetes", "Health Science", "Traceable" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请使用「腾讯医典健康问答」回答下面的健康问题，要求基于腾讯医典官方科普知识库内容作答，句句可溯源、不编造。\n\n问题：糖尿病如何治疗？\n\n请按以下方式输出：\n1. 先给出一句核心结论；\n2. 再分点展开治疗要点（分型、药物治疗、营养治疗、运动、血糖监测、终身管理等），每条要点末尾标注引用来源编号 [N]；\n3. 文末附「参考来源」区块，列出引用的文章标题 + 链接；\n4. 最后附医学免责声明。", "prompt_en": "Use \"Tencent Medpedia Health Q&A\" to answer the health question below. Base the answer strictly on Tencent Medpedia's official health-science knowledge base — every point traceable, nothing fabricated.\n\nQuestion: How is diabetes treated?\n\nOutput as follows:\n1. Start with a one-sentence core conclusion;\n2. Expand into bullet points (classification, medication, nutrition therapy, exercise, glucose monitoring, lifelong management, etc.), appending a source citation number [N] to each point;\n3. End with a \"References\" section listing the cited article titles and links;\n4. Finish with a medical disclaimer.", "skills": [ { "id": "tencent-medpedia-aiqa-skill", "name": "腾讯医典健康问答", "description": "检索腾讯医典官方科普知识库回答健康问题，句句可溯源", "skill_id": "skill_2076988185622831104" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-13T19:32:40+08:00", "updated_at": "2026-08-20T14:33:28+08:00", "artifacts": [ { "file": "output.html", "label": "健康问答结果（可溯源答案 + 参考来源）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "medpedia-pediatric-dose", "title": "宝宝吃美林，算得准才放心", "title_en": "Pediatric Ibuprofen Dose, Computed Exactly", "subtitle": "本地儿科库确定性计算，剂量可复现", "subtitle_en": "Local pediatric DB deterministic calc — reproducible dosage", "description": "2 岁宝宝发烧，家里有美林，家长最怕「吃多了伤肾、吃少了不退」。网上剂量五花八门。\n\n这个案例演示腾讯医典健康问答的「儿童用药剂量计算」分支：它基于 skill 内置、经医学核查的本地儿科药品库（pediatric-drugs.json，30 种药），对「2 岁 12 kg 美林」做确定性计算——单次 120 mg（20 mg/ml 规格合 6 ml），每 6–8 小时一次、日不超 4 次、6 月龄以下禁用，并跑安全检查。结果可复现，不靠记忆编。\n\n适合：育儿家庭精准用药、母婴 App 用药计算、儿科用药安全教育、互联网医院剂量核对。", "description_en": "A 2-year-old has a fever and you have Motrin at home — parents fear \"too much harms the kidneys, too little does not work\". Online doses vary wildly.\n\nThis case shows the \"pediatric dosage calculation\" branch of Tencent Medpedia Q&A: based on the skill's built-in, medically-checked local pediatric drug DB (pediatric-drugs.json, 30 drugs), it deterministically computes \"Motrin for a 2yo 12kg\" — 120 mg single (6 ml at 20 mg/ml), every 6–8h, ≤4/day, contraindicated under 6 months, with safety checks. The result is reproducible, not memorized.\n\nGood for: precise dosing at home, maternal-app dose calc, pediatric med safety, tele-health dose verification.", "categories": [ "education", "productivity" ], "tags": [ "腾讯医典", "儿童用药", "剂量计算", "美林", "可复现" ], "tags_en": [ "Medpedia", "pediatric dose", "dosage calc", "ibuprofen", "reproducible" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "2 岁孩子 12 公斤，发烧 38.7℃，家里有美林布洛芬混悬液。这次该喂多少毫升、几小时能再喂一次、一天最多几次？", "prompt_en": "My 2-year-old weighs 12 kg and has a 38.7°C fever. We have Motrin ibuprofen suspension at home. How many milliliters for this dose, how many hours until the next, and what is the daily maximum?", "skills": [ { "id": "tencent-medpedia-aiqa-skill", "name": "腾讯医典健康问答", "description": "检索腾讯医典官方科普知识库作答，句句可溯源；含儿童用药剂量本地计算", "skill_id": "skill_2076988185622831104" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-22T17:14:53Z", "artifacts": [ { "file": "output.html", "label": "宝宝吃美林，算得准才放心" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-batch-transform-bottle-white-bg-set", "title": "保温杯批量白底图", "title_en": "Batch White-Background Bottle Shots", "subtitle": "厨房随手拍整批换白底，规格一次对齐", "subtitle_en": "Turn casual shots into a uniform white-bg set", "description": "在厨房台面随手拍的保温杯，背景有橱柜和纸箱，没法直接上架。批量去背景后所有杯子落在干净白底上，杯身的淡蓝色和哑光质感保留，底部投影自然，一整批图的构图和规格都对得上。适合小批量上新的日用百货和文创商家。", "categories": [ "creation" ], "tags": [ "批量修图", "白底图", "去背景", "商品上架" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "保温杯白底商品图｜整批 1:1 PNG" } ], "prompt": "[批量修图]对上传的图片进行批量处理。\n批量图片：一批在厨房台面随手拍的保温杯照片，背景有橱柜和纸箱\n处理操作：批量去背景\n处理规格：1:1 方图，纯白背景\n其他说明：杯身淡蓝色和哑光质感不要变，底部保留一层自然投影，整批主体大小和位置对齐\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-batch-transform", "name": "批量修图", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298822561792" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-batch-transform-kids-bowl-studio-set", "title": "儿童餐碗批量棚拍", "title_en": "Batch Studio Look for Kids Tableware", "subtitle": "整组彩色餐具统一浅灰棚拍底，色调一次对齐", "subtitle_en": "Unify colorful tableware into one studio look", "description": "儿童硅胶餐具颜色多、拍摄环境又杂，几张图放在一起色调各走各的。统一风格后整批餐碗换到浅灰棚拍底，橙色饱和度一致、投影方向统一，双耳和吸盘底的结构没有变形。适合母婴餐具类目做整组商品图。", "categories": [ "creation" ], "tags": [ "批量修图", "统一风格", "母婴餐具", "商品图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "儿童餐碗棚拍商品图｜整批 1:1 JPG" } ], "prompt": "[批量修图]对上传的图片进行批量处理。\n批量图片：一批在餐桌上随手拍的彩色儿童硅胶餐碗照片\n处理操作：统一风格\n处理规格：1:1 方图，浅灰棚拍背景\n其他说明：整批投影方向一致，橙色饱和度统一，双耳和吸盘底的结构不要变形\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-batch-transform", "name": "批量修图", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298822561792" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-batch-transform-mug-glaze-enhance", "title": "马克杯批量质感增强", "title_en": "Batch Texture Boost for Ceramic Mugs", "subtitle": "釉面高光和轮廓一起提亮，整批像一次棚拍", "subtitle_en": "Sharpen glaze and edges across a whole batch", "description": "家里窗边随手拍的陶瓷马克杯，釉面发灰、边缘发虚，放到详情页显得廉价。批量增强后釉面高光干净、杯口和把手轮廓清晰、色温统一成米白，一整批图看起来像同一次棚拍出的。适合陶瓷杯具和礼品定制商家整理商品图。", "categories": [ "creation" ], "tags": [ "批量修图", "质感增强", "陶瓷杯具", "详情页图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "马克杯质感商品图｜整批 1:1 JPG" } ], "prompt": "[批量修图]对上传的图片进行批量处理。\n批量图片：一批在窗边木桌上随手拍的米白色陶瓷马克杯照片\n处理操作：批量增强\n处理规格：1:1 方图，1500×1500\n其他说明：釉面高光要干净，杯口和把手轮廓清晰，整批色温统一成米白，不要把杯身调成纯白\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-batch-transform", "name": "批量修图", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298822561792" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-batch-transform-serum-studio-look", "title": "精华液批量棚拍质感", "title_en": "Batch Studio Look for Serum Shots", "subtitle": "生活随拍统一转白底棚拍质感", "subtitle_en": "Turn casual shots into consistent studio product photos", "description": "上传多张浴室、卧室等场景随手拍的产品图，批量统一为白底棚拍风格，打光角度和背景一致，标签文字清晰可辨，省去逐张调色和抠图的时间，适合美妆个护类目上新前快速出统一风格的产品图。", "categories": [ "creation" ], "tags": [ "批量修图", "统一风格", "产品图批量处理", "棚拍质感", "美妆个护" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "精华液棚拍风格图｜批量 JPG" } ], "prompt": "[批量修图]对上传的图片进行批量处理。\n批量图片：（请上传：批量图片）\n处理操作：统一风格\n处理规格：统一为白底精品棚拍风格\n其他说明：保持瓶身比例和标签文字清晰，统一打光角度和阴影方向\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-batch-transform", "name": "批量修图", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298822561792" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-bestseller-replication-cat-tumbler", "title": "猫咪保温杯爆款图复刻", "title_en": "Cat Tumbler Bestseller Remix", "subtitle": "健身房场景爆款版式直接套用，保温杯秒变高转化主图", "subtitle_en": "Reuse a proven lifestyle layout for your new tumbler colorway", "description": "保温杯新配色上架，想要一张有生活气息的场景主图，自己布置健身房场景拍摄很麻烦。上传同类目爆款场景图和产品图，AI 照搬健身房取景、运动包道具和标语版式，只替换杯身配色和图案，快速产出高转化场景主图，适合水杯类目新品测款。", "categories": [ "creation" ], "tags": [ "爆款图复刻", "保温杯主图", "高度复刻", "运动水具" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "保温杯爆款复刻主图｜单张 JPG" } ], "prompt": "[爆款图复刻]参考上传的爆款图风格，用我的商品图逐张生成同风格的电商图。\n参考爆款图：（请上传：参考爆款图）\n你的商品图：（请上传：你的商品图）\n复刻程度：高度复刻（照搬版式，仅替换产品）\n结果图比例：3:4\n商品卖点&要求：薰衣草紫色保温杯，带猫咪镭雕图案，40oz大容量，配吸管与手柄，希望复刻参考图的健身房场景、运动包道具和'Sip More, Do More.'风格标语版式，用于运动水具类目主图\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-bestseller-replication", "name": "爆款图复刻", "description": "一张商品图，成套复刻爆款风格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7lKcxkOsxXQN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-bestseller-replication-passion-fruit-tea", "title": "百香果茶饮爆款图复刻", "title_en": "Passion Fruit Tea Bestseller Remix", "subtitle": "跨品类借鉴高级质感光影，饮品图秒变精致大片", "subtitle_en": "Borrow a premium lighting style from another category", "description": "饮品类主图很难拍出高级质感，自己布光和道具搭配经验有限。上传一张调性喜欢的爆款图（品类可完全不同）和产品图，AI 借鉴其丝绸质感、暖光氛围和道具摆放方式，只保留你的产品原样呈现，快速产出高级感场景主图，适合茶饮、果汁类目上新。", "categories": [ "creation" ], "tags": [ "爆款图复刻", "饮品主图", "参考风格", "跨品类借鉴" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "百香果茶饮爆款复刻主图｜单张 JPG" } ], "prompt": "[爆款图复刻]参考上传的爆款图风格，用我的商品图逐张生成同风格的电商图。\n参考爆款图：（请上传：参考爆款图）\n你的商品图：（请上传：你的商品图）\n复刻程度：参考风格（借鉴调性，构图可变）\n结果图比例：跟随参考图\n商品卖点&要求：百香果橙茶饮品杯，卖点为鲜果茶真实果肉，希望借鉴参考图的丝绸质感、暖光氛围和窗帘光影，用于饮品类目场景主图，风格清新高级\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-bestseller-replication", "name": "爆款图复刻", "description": "一张商品图，成套复刻爆款风格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7lKcxkOsxXQN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-bestseller-replication-pet-food-hero", "title": "宠物粮爆款图复刻", "title_en": "Pet Food Bestseller Remix", "subtitle": "同类目爆款构图直接套用，新品秒变高转化主图", "subtitle_en": "Reuse a proven bestseller layout for your new product", "description": "新品刚上架没有流量基础，主图很难和头部爆款竞争曝光。上传一张同类目的爆款参考图和自己的产品图，AI 会照搬爆款的构图、宠物姿态、文案位置和视觉节奏，只替换成你的产品和品牌信息，快速产出一张视觉说服力接近头部卖家的主图，适合宠物、母婴等强场景类目的新品测款。", "categories": [ "creation" ], "tags": [ "爆款图复刻", "宠物粮主图", "高度复刻", "新品测款" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "宠物粮爆款复刻主图｜单张 JPG" } ], "prompt": "[爆款图复刻]参考上传的爆款图风格，用我的商品图逐张生成同风格的电商图。\n参考爆款图：（请上传：参考爆款图）\n你的商品图：（请上传：你的商品图）\n复刻程度：高度复刻（照搬版式，仅替换产品）\n结果图比例：3:4\n商品卖点&要求：宠物粮包装为绿色配色，需保留 pōp pup 品牌名称与卡通小狗图案；文案改为'Good food. Happy pups.'，风格活泼温馨，适合宠物用品类目\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-bestseller-replication", "name": "爆款图复刻", "description": "一张商品图，成套复刻爆款风格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7lKcxkOsxXQN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-bestseller-replication-tennis-polo-dress", "title": "网球POLO裙爆款图复刻", "title_en": "Tennis Polo Dress Bestseller Remix", "subtitle": "同类目爆款海报版式直接套用，新款秒变高级主图", "subtitle_en": "Reuse a proven sportswear poster layout for your new style", "description": "运动服饰新款上架缺一张有质感的场景主图，自己找模特拍摄成本高、周期长。上传同类目爆款海报图和产品图，AI 照搬标题版式、模特姿态和球场取景，只替换成你的产品配色，快速产出接近头部卖家的主图，适合网球、高尔夫等运动服饰新品测款。", "categories": [ "creation" ], "tags": [ "爆款图复刻", "网球服饰主图", "高度复刻", "运动服饰" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "网球裙爆款复刻主图｜单张 JPG" } ], "prompt": "[爆款图复刻]参考上传的爆款图风格，用我的商品图逐张生成同风格的电商图。\n参考爆款图：（请上传：参考爆款图）\n你的商品图：（请上传：你的商品图）\n复刻程度：高度复刻（照搬版式，仅替换产品）\n结果图比例：3:4\n商品卖点&要求：淡紫色网球POLO连衣裙，卖点为透气速干面料和收腰版型，希望复刻参考图的球场场景、白色鸭舌帽造型和'Sporty. Classic. Chic.'风格标语版式，用于网球运动服饰类目主图\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-bestseller-replication", "name": "爆款图复刻", "description": "一张商品图，成套复刻爆款风格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7lKcxkOsxXQN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-extension-frosted-tale-ice-cream-market", "title": "Frosted Tale 冰淇淋市集品牌延展", "title_en": "Frosted Tale Ice Cream Market Brand Extension", "subtitle": "从既有 Logo 延展市集摊位、包装、贴纸与会员物料", "subtitle_en": "Extend an existing logo into a cohesive market stall, packaging, stickers, and membership system", "description": "这是一个冰淇淋品牌线下市集场景的品牌延展案例。案例以 Frosted Tale 既有 Logo 和品牌蓝为核心，统一完成冰淇淋摊位、品牌贴纸、杯装包装、购物袋、中文会员卡与草莓甜筒包装，让空间展示、销售包装和会员触点保持一致的年轻甜品品牌气质。", "categories": [ "creation", "business" ], "tags": [ "品牌设计", "品牌延展", "冰淇淋品牌", "包装设计", "市集摊位" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Frosted Tale 冰淇淋市集品牌延展｜6 张单文件 HTML" } ], "prompt": "[品牌延展]请基于我上传的 Frosted Tale 既有 Logo 建立完整的冰淇淋市集品牌物料系统。保留 Logo 的核心图形、名称和识别特征，以品牌蓝 #4BB4EE 为主色，统一规划冰淇淋车摊位、品牌贴纸、冰淇淋杯包装、购物袋、中文会员卡与草莓甜筒包装。整体应年轻、清爽、甜美，所有物料在字体、色彩、图形语言和 Logo 使用上保持一致；不得擅自改变品牌名称或核心标志。", "skills": [ { "id": "designkit-brand-extension", "name": "品牌延展", "description": "基于既有品牌 Logo 与视觉资产，延展包装、空间、宣传和经营物料，建立统一的品牌应用系统。", "skill_id": "skill_2091019297830735872" } ], "experts": [], "mcps": [], "scenario": "品牌设计", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:42:14+08:00", "updated_at": "2026-08-31T15:42:14+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-extension-mianli-apparel-store-renovation", "title": "棉里服装店年轻化品牌翻新", "title_en": "Mianli Apparel Store Youthful Brand Renovation", "subtitle": "以燕麦色与自然绿统一门店空间、导视、吊牌和经营物料", "subtitle_en": "Unify the retail space, signage, tags, and operating materials with oat neutrals and natural green", "description": "这是一个服装零售门店年轻化品牌翻新案例。案例保留“棉里”品牌名称，以燕麦色、奶油白和自然绿建立温暖日常、轻设计感的视觉方向，并系统延展门头、橱窗、收银台、试衣间、导视、衣架、面料与价格吊牌、购物袋、名片、会员卡及宣传海报，形成可落地的门店品牌体验。", "categories": [ "creation", "business" ], "tags": [ "品牌设计", "品牌翻新", "服装品牌", "零售空间", "门店物料" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "棉里服装店年轻化品牌翻新｜14 张单文件 HTML" } ], "prompt": "[品牌延展]请基于我上传的服装店现状照片，为“棉里”完成一次年轻化品牌翻新。保留品牌名称，以燕麦色、奶油白和自然绿为主色，建立温暖日常、清爽克制且具有轻设计感的视觉系统；统一规划门店门头、橱窗、收银台、试衣间、店内导视、品牌衣架、面料信息吊牌、商品价格吊牌、购物袋、名片、中文会员卡、品牌生活方式海报与开业海报。空间与平面物料应使用同一套字体、色彩与图形语言，并兼顾真实零售场景的可实施性。", "skills": [ { "id": "designkit-brand-extension", "name": "品牌延展", "description": "基于既有品牌 Logo 与视觉资产，延展包装、空间、宣传和经营物料，建立统一的品牌应用系统。", "skill_id": "skill_2091019297830735872" } ], "experts": [], "mcps": [], "scenario": "品牌设计", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:42:14+08:00", "updated_at": "2026-08-31T15:42:14+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-extension-puffy-paws-pet-store", "title": "Puffy 爪爪宠物店品牌系统", "title_en": "Puffy Paws Pet Store Brand System", "subtitle": "从宠物形象徽章到门店、包装、会员与导视的完整品牌应用", "subtitle_en": "A complete pet-brand system spanning the badge logo, store, packaging, membership, and signage", "description": "这是一个面向年轻消费者的宠物店品牌系统案例。案例从宠物照片提炼 Puffy 爪爪徽章式 Logo，以蓝、黄、白建立活泼友好的视觉识别，并延展到门店内外空间、购物袋、贴纸吊牌、名片、会员卡、开业海报、徽章、宠物食盆、食品包装、灯箱与橱窗标识等完整触点。", "categories": [ "creation", "business" ], "tags": [ "品牌设计", "品牌延展", "宠物品牌", "门店设计", "包装设计" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Puffy 爪爪宠物店品牌系统｜15 张单文件 HTML" } ], "prompt": "[品牌延展]请参考我上传的宠物照片，为年轻化宠物品牌“Puffy 爪爪”建立完整品牌系统。先提炼具有辨识度的美式徽章 Logo，以蓝、黄、白为主色，再将统一视觉延展至门店门头与室内空间、购物袋、贴纸与吊牌、名片、中文会员卡、开业海报、品牌徽章、宠物食盆、宠物食品包装、圆形灯箱和橱窗标识。整体应年轻、友好、有记忆点，所有触点必须保持品牌名称、标志、配色和图形语言一致。", "skills": [ { "id": "designkit-brand-extension", "name": "品牌延展", "description": "基于既有品牌 Logo 与视觉资产，延展包装、空间、宣传和经营物料，建立统一的品牌应用系统。", "skill_id": "skill_2091019297830735872" } ], "experts": [], "mcps": [], "scenario": "品牌设计", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:42:14+08:00", "updated_at": "2026-08-31T15:42:14+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-extension-sishiyouwei-mushroom-hotpot", "title": "四时有味菌菇火锅品牌延展", "title_en": "Sishiyouwei Mushroom Hotpot Brand Extension", "subtitle": "从既有餐饮 Logo 延展美食视觉、包装、制服、菜单与门店空间", "subtitle_en": "Extend an existing restaurant logo into food imagery, packaging, uniforms, menus, and store environments", "description": "这是一个云南菌菇火锅品牌的系统延展案例。案例以“四时有味”既有 Logo 为核心，围绕自然、时令与菌菇产地感建立高级而清透的视觉语言，完整覆盖门店外观、美食视觉、生活方式用餐场景、包装套装、员工制服、中文菜单和室内空间，使品牌从平面识别延伸到真实餐饮体验。", "categories": [ "creation", "business" ], "tags": [ "品牌设计", "品牌延展", "餐饮品牌", "菌菇火锅", "门店设计" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "四时有味菌菇火锅品牌延展｜7 张单文件 HTML" } ], "prompt": "[品牌延展]请基于我上传的“四时有味”餐饮 Logo，为云南菌菇火锅品牌建立完整延展系统。保留 Logo 的核心造型与品牌名称，以自然、时令、菌菇产地和东方餐饮气质为方向，统一设计门店外观、美食视觉、生活方式用餐场景、品牌包装套装、员工制服、中文菜单和门店室内空间。整体画面采用 3:4 竖版构图，高级、清透、真实可落地；所有触点须保持 Logo、字体、色彩和图形语言一致。", "skills": [ { "id": "designkit-brand-extension", "name": "品牌延展", "description": "基于既有品牌 Logo 与视觉资产，延展包装、空间、宣传和经营物料，建立统一的品牌应用系统。", "skill_id": "skill_2091019297830735872" } ], "experts": [], "mcps": [], "scenario": "品牌设计", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:42:14+08:00", "updated_at": "2026-08-31T15:42:14+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-ip-cactus-robot-figure", "title": "仙人掌机器人潮玩IP", "title_en": "Cactus Robot Toy Mascot", "subtitle": "耳机+工装马甲3D潮玩三视图，潮酷科技感", "subtitle_en": "A headphone-wearing cactus figure, three views", "description": "绿色仙人掌造型机器人头戴金属耳机、身穿墨绿工装马甲与运动鞋，手持复古随身听，3D潮玩质感搭配正面、侧面、背面三视图。适合潮牌、数码、盲盒周边等年轻化品牌打造有个性的形象IP，兼具反差萌与街头潮流气质。", "categories": [ "creation" ], "tags": [ "IP形象", "潮玩设计", "3D立体", "盲盒手办", "街头潮流" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "IP形象设计｜3D潮玩三视图" } ], "prompt": "[IP形象设计]设计品牌IP形象。\n参考素材：（请上传：参考素材）\nIP风格：3D立体\n角色设定：一个仙人掌造型、戴耳机穿工装马甲的机器人潮玩\n其他要求：正面、侧面、背面三视图，绿黄配色，手持复古随身听，盲盒手办质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-brand-ip", "name": "IP形象设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297914621952" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-ip-otter-traveler-mascot", "title": "水獭旅行家IP形象", "title_en": "Traveler Otter Mascot", "subtitle": "渔夫帽+斜挎包水獭三视图，适合文旅出行品牌", "subtitle_en": "A bucket-hat otter with a crossbody bag, three views", "description": "一只戴渔夫帽、背橙色斜挎包的水獭挥手微笑，天蓝渐变背景搭配正面、侧面、背面三视图，姿态统一、比例可爱。适合文旅、出行、露营等品牌打造亲切活泼的形象大使，无论用于周边产品还是社交传播都能快速建立记忆点。", "categories": [ "creation" ], "tags": [ "IP形象", "品牌吉祥物", "文旅出行", "三视图", "扁平插画" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "IP形象设计｜三视图（正面/侧面/背面）" } ], "prompt": "[IP形象设计]设计品牌IP形象。\n参考素材：（请上传：参考素材）\nIP风格：扁平插画\n角色设定：一只戴渔夫帽、背橙色斜挎包的水獭\n其他要求：正面、侧面、背面三视图，天蓝渐变背景，挥手微笑姿态，适合文旅出行品牌周边延展\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-brand-ip", "name": "IP形象设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297914621952" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-ip-skewer-boy-mascot", "title": "烤串小哥品牌IP", "title_en": "Skewer Boy Mascot", "subtitle": "餐饮行业IP形象，融入烤串元素+多角度立绘", "subtitle_en": "A street-food mascot with a full turnaround", "description": "手绘水彩风格的烤串小哥IP，橙色围裙印着烤串图案，笑容憨厚接地气，附正面、侧面、背面三视图，方便延展进菜单、外卖包装、门店物料。适合烧烤、小吃类餐饮品牌打造有记忆点又亲切的门店形象。", "categories": [ "creation" ], "tags": [ "IP形象设计", "品牌吉祥物", "餐饮IP", "烧烤小吃", "多角度立绘" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "品牌IP形象｜正面+侧面+背面立绘" } ], "prompt": "[IP形象设计]设计品牌IP形象。\n参考素材：（请上传：参考素材）\nIP风格：手绘水彩\n角色设定：一个穿橙色围裙、扎橙色头巾的烤串小哥，围裙口袋印有烤串图案，手举一串烤串\n其他要求：需要正面、侧面、背面多角度立绘，用于餐饮品牌形象展示，风格温暖亲切\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-brand-ip", "name": "IP形象设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297914621952" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-ip-watercolor-girl-character", "title": "水彩风女孩IP形象", "title_en": "Watercolor Girl Character", "subtitle": "双马尾粉毛衣水彩三视图，温柔亲和母婴调性", "subtitle_en": "A pigtailed girl in pastel watercolor, three views", "description": "扎着双马尾、穿粉色针织开衫的小女孩手持餐叉微笑，水彩晕染笔触带出柔和温暖的质感，正面、侧面、背面三视图姿态统一。适合母婴、儿童餐饮、绘本等品牌打造温柔亲和的形象代言，传递细腻、治愈的品牌情感。", "categories": [ "creation" ], "tags": [ "IP形象", "水彩插画", "母婴品牌", "三视图", "儿童角色" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "IP形象设计｜水彩三视图" } ], "prompt": "[IP形象设计]设计品牌IP形象。\n参考素材：（请上传：参考素材）\nIP风格：手绘水彩\n角色设定：一个扎双马尾、穿粉色针织开衫的小女孩\n其他要求：正面、侧面、背面三视图，水彩晕染质感，暖白背景，柔和治愈的色调\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-brand-ip", "name": "IP形象设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297914621952" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-brand-seller-social-oatmeal-breakfast-post", "title": "燕麦片早餐种草图文", "title_en": "Oatmeal Breakfast Post", "subtitle": "产品图秒变治愈系种草笔记，卖点一目了然", "subtitle_en": "Turn a product photo into a cozy breakfast post", "description": "只有产品图和卖点，不知道怎么写出有代入感的种草文案？上传产品照片，AI 自动生成治愈系早餐场景图文：手写标题、卖点清单、场景搭配一次到位，帮新手商家和达人快速产出可直接发布的带货笔记。", "categories": [ "creation" ], "tags": [ "社媒图文带货", "产品种草", "早餐好物", "带货笔记", "社媒种草" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "燕麦片早餐种草图文｜单张" } ], "prompt": "[社媒图文带货]把上传的产品/模特图转成社媒种草带货图文。\n产品/模特图：一张每日坚果燕麦片的产品包装照\n产品名称：堂饱饱每日坚果燕麦片\n带货类型：产品种草笔记\n内容风格：plog\n目标平台：小红书\n画面比例：3:4\n卖点&运营目标：0反式脂肪酸、富含膳食纤维、轻加工更健康；主打早餐场景，搭配酸奶水果碗拍摄，突出美味与营养兼得\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-brand-seller-social", "name": "社媒图文带货", "description": "上传产品图，生成社媒种草带货图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_IS0LkCj9OkoX" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-chart-infographic-city-sales-comparison", "title": "各城市线上线下销售对比图", "title_en": "City Online vs Offline Sales Chart", "subtitle": "柱状+折线双维度，直观对比五城市销售表现", "subtitle_en": "Bar and line combo comparing five cities' sales", "description": "汇报销售数据时，纯表格不够直观，柱状图又难以同时呈现两组数据。这款工具能把各城市线上线下销售额自动生成柱状图+折线图组合图表，紫色柱状对比黄色趋势线，数据差异一目了然，适合销售汇报、季度复盘等场景使用。", "categories": [ "creation" ], "tags": [ "图表/信息图", "销售数据图表", "柱状图设计", "城市对比分析" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "城市销售对比图表｜单张 JPG" } ], "prompt": "[图表/信息图]将内容设计为图表或信息图。\n数据/内容：各城市销售额对比，总计线上销售额5,430万元、线下门店销售额3,850万元；柱状图展示北京、上海、广州、深圳、成都五城市线上销售额分别为1,200/1,500/980/1,100/650万元，折线图叠加对应线下门店销售额850/1,100/720/680/500万元，紫色柱状+黄色折线配色\n数据文件：（请上传：数据文件（可选））\n图表类型：对比图\n图片比例：竖版（3:4）\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-chart-infographic", "name": "图表/信息图", "description": "数据可视化、信息图设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_EiobVzCO4Q8u" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-chart-infographic-payment-preference-donut", "title": "电商支付方式偏好环形图", "title_en": "E-commerce Payment Preference Donut Chart", "subtitle": "环形图呈现用户支付方式占比，一眼看清主流选择", "subtitle_en": "Donut chart showing users' preferred payment methods", "description": "分析用户支付习惯时，一堆百分比数字不够直观。这款工具能把支付方式占比自动生成环形图，社交平台支付、支付宝、银联信用卡等分类清晰配色分明，适合运营复盘、用户调研报告、产品汇报等场景快速生成可视化图表。", "categories": [ "creation" ], "tags": [ "图表/信息图", "饼图设计", "支付方式分析", "用户偏好数据" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "支付方式偏好图表｜单张 JPG" } ], "prompt": "[图表/信息图]将内容设计为图表或信息图。\n数据/内容：电商平台用户支付方式偏好环形图，微信支付45%、支付宝38%、银联/信用卡12%、数币/其他5%，绿蓝紫黄四色分区，浅绿背景配左侧图例\n数据文件：（请上传：数据文件（可选））\n图表类型：饼图\n图片比例：竖版（3:4）\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-chart-infographic", "name": "图表/信息图", "description": "数据可视化、信息图设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_EiobVzCO4Q8u" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-chart-infographic-personal-growth-five-steps", "title": "个人成长五步法信息图", "title_en": "5-Step Growth Infographic", "subtitle": "长文字梳理成一张信息图，五个步骤一眼看懂", "subtitle_en": "Turn dense notes into one clear step-by-step infographic", "description": "写方案、做复盘时经常攒下一堆思路和步骤，堆在文档里没人愿意细看。这款工具把文字内容自动整理成分步骤、配图标、带关键要点侧栏的信息图，逻辑清晰、色彩分区明确，适合个人成长计划、工作方法论、培训分享等场景，一张图讲清一件事。", "categories": [ "creation" ], "tags": [ "图表/信息图", "信息图", "流程图解", "知识可视化" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "个人成长五步法信息图｜单张 JPG" } ], "prompt": "[图表/信息图]将内容设计为图表或信息图。\n数据/内容：个人成长五步法：1.觉察现状——了解当前状态，明确优势与不足，关键问题包括擅长什么、什么消耗精力、想要什么样的改变；2.设定目标——按SMART原则设定具体明确、可衡量、可实现、有相关性、有时间限制的目标；3.制定计划——拆解目标，分解为小步骤、明确时间节点、识别所需资源、准备应对风险；4.执行行动——专注当下坚持执行，每天行动一点点、记录进度与感受、减少干扰保持专注、及时调整策略；5.复盘优化——回顾结果总结经验，思考做得怎么样、学到了什么、下一步怎么做得更好；结论突出'从认知到行动，持续升级自己'的主题\n数据文件：（请上传：数据文件（可选））\n图表类型：信息图\n图片比例：竖版（3:4）\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-chart-infographic", "name": "图表/信息图", "description": "数据可视化、信息图设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_EiobVzCO4Q8u" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-chart-infographic-startup-six-steps-flow", "title": "创业六步流程图", "title_en": "Six-Step Startup Flowchart", "subtitle": "从想法到规模化增长，六个关键阶段一图讲清", "subtitle_en": "Six key stages from idea to scaled growth", "description": "创业想法很多，但缺一张清晰的流程图梳理落地路径。这款工具能把创业过程拆解成明确方向、制定策略、组建团队等六个关键阶段，用圆形图标和S形连接线呈现，逻辑清楚一目了然，适合创业者路演、商业计划书配图等场景使用。", "categories": [ "creation" ], "tags": [ "图表/信息图", "创业流程图", "流程图设计", "商业计划图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "创业流程图｜单张 JPG" } ], "prompt": "[图表/信息图]将内容设计为图表或信息图。\n数据/内容：创业流程图，标题'创业流程图'，副标题'从想法到成功的六大关键步骤'，六个阶段：01明确方向(验证商业想法，明确目标用户与核心价值)、02制定策略(分析市场与竞争，制定产品定位与增长策略)、03组建团队(招募核心成员，明确分工，打造高效团队)、04产品打磨(设计产品原型，快速迭代优化，提升用户体验)、05市场推广(制定推广计划，建立品牌影响力，吸引潜在用户)、06持续增长(数据驱动决策，优化运营，实现规模化增长)，深蓝色圆形图标+虚线连接的S形路径\n数据文件：（请上传：数据文件（可选））\n图表类型：流程图\n图片比例：竖版（3:4）\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-chart-infographic", "name": "图表/信息图", "description": "数据可视化、信息图设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_EiobVzCO4Q8u" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-competitor-analysis-competitor-pool", "title": "类目竞品池分析", "title_en": "Category Competitor Pool", "subtitle": "一键拉出Top竞品清单，评分、价格带一目了然", "subtitle_en": "Pull a ranked list of top competitors with price and rating", "description": "选品或对标时手动一个个翻竞品页面太慢，价格带和评分也难以横向比较。填入竞品ASIN，AI 自动拉出同类目Top竞品清单，附带评分、价格区间分布和评论星级占比，生成可直接用于选品会的对标清单，适合新品定价和差异化卖点提炼前的调研。", "categories": [ "business" ], "tags": [ "竞品分析", "竞品池", "价格带分析", "选品调研" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "竞品池分析卡｜单张图" } ], "prompt": "[竞品分析]进行亚马逊竞品分析。\nASIN / 商品链接：{你的商品ASIN}\n分析类型：竞品池分析\n关注重点：希望覆盖价格区间分布和评分/评论占比，方便对标定价\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-competitor-analysis", "name": "竞品分析", "description": "Amazon ASIN 诊断与竞品池分析", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qCY4R6dcyWsF" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-competitor-analysis-listing-audit-scorecard", "title": "Listing诊断打分卡", "title_en": "Listing Health Audit", "subtitle": "三项打分揪出标题、卖点、合规隐患", "subtitle_en": "Score your title, bullet density and compliance risk", "description": "上架很久却一直没起量，问题往往藏在标题和卖点密度里，自己很难一眼看出来。填入 ASIN，AI 会从标题 Hook、卖点密度、合规风险三个维度打分，标出具体问题（比如场景词缺失、夸大宣传风险），并给出可执行的优化建议，帮卖家快速定位 Listing 短板，尤其适合上架后转化率长期偏低的商品自查。", "categories": [ "business" ], "tags": [ "竞品分析", "Listing诊断", "跨境电商平台选品", "合规风险" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "Listing诊断打分卡｜结果长图" } ], "prompt": "[竞品分析]进行亚马逊竞品分析。\nASIN / 商品链接：{你的商品 ASIN}\n分析类型：Listing诊断\n关注重点：重点看标题里的场景词是否覆盖不足，以及是否存在夸大宣传的违规表述\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-competitor-analysis", "name": "竞品分析", "description": "Amazon ASIN 诊断与竞品池分析", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qCY4R6dcyWsF" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-competitor-analysis-traffic-structure", "title": "竞品流量结构诊断", "title_en": "Competitor Traffic Structure", "subtitle": "拆解竞品自然/广告/关联/外部流量占比，找差距", "subtitle_en": "Break down a competitor's traffic sources by channel", "description": "只看竞品排名和销量猜不出它的流量打法，不知道该往广告投放还是关联推荐上发力。填入竞品ASIN，AI 自动拆解自然流量、广告流量、关联流量和外部流量的占比结构，并给出结构是否健康的诊断结论，适合精细化运营前摸清竞品流量构成。", "categories": [ "business" ], "tags": [ "竞品分析", "流量结构", "ASIN诊断", "运营选品" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "竞品流量结构诊断卡｜单张图" } ], "prompt": "[竞品分析]进行亚马逊竞品分析。\nASIN / 商品链接：{你的商品ASIN}\n分析类型：流量结构\n关注重点：重点关注自然流量与广告流量的占比是否健康，以及关联流量的来源\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-competitor-analysis", "name": "竞品分析", "description": "Amazon ASIN 诊断与竞品池分析", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qCY4R6dcyWsF" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-content-removal-kitchen-counter-clutter", "title": "厨房台面杂物消除", "title_en": "Remove Kitchen Counter Clutter", "subtitle": "一键清空台面杂物，留下干净大理石纹理", "subtitle_en": "Clear counter clutter, keep the marble finish intact", "description": "上传台面堆满纸巾、包装袋和餐具的生活照，圈出需要去除的杂物，AI智能消除并自动补全大理石纹理和背景细节，厨具、绿植等主体保持不变，适合家居分享、租房实拍等场景快速清爽出片。", "categories": [ "creation" ], "tags": [ "去水印/杂物", "杂物消除", "厨房实拍", "背景修复", "生活照清理" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "干净厨房台面图｜单张 JPG" } ], "prompt": "[去水印/杂物]去除图片中的指定内容。\n待处理图片：（请上传：待处理图片）\n需要去除的内容：台面上的塑料袋、纸巾、餐具残渣和食物包装盒\n其他要求：保留背景木架、砧板、厨具罐和绿植不变，去除后台面纹理需自然衔接\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-content-removal", "name": "去水印/杂物", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298843975680" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-content-removal-seaside-boardwalk-crowd", "title": "海边栈道路人消除", "title_en": "Remove Crowds from a Seaside Boardwalk", "subtitle": "游客散步骑车全部清空，栈道像清场后拍的", "subtitle_en": "Empty the boardwalk while keeping the scenery", "description": "旅游照里游客不断入镜，等人走开又错过光线。指定要去掉的路人后，栈道、棕榈树和石栏杆保持原样，地砖接缝自然延伸，海面和天空没有涂抹感，整条海岸线像清场后单独拍的。适合旅行博主和民宿商家出干净的场景图。", "categories": [ "creation" ], "tags": [ "去水印/杂物", "路人消除", "旅行照", "风景图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "无人海边栈道照｜单张 JPG" } ], "prompt": "[去水印/杂物]去除图片中的指定内容。\n待处理图片：一张海边栈道的旅游实拍照，前后有多组行人和骑车的人\n需要去除的内容：栈道上散步、骑车和倚在栏杆边的所有路人\n其他要求：棕榈树、路灯和石栏杆保持原样，地砖接缝要自然延伸，海面和天空不要出现涂抹痕迹\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-content-removal", "name": "去水印/杂物", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298843975680" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-content-removal-serum-tiled-watermark", "title": "护肤品图去平铺水印", "title_en": "Remove Tiled Watermark from Product Shot", "subtitle": "满屏水印和角标一起清掉，背景渐变不留痕", "subtitle_en": "Clear tiled watermarks and keep the gradient clean", "description": "自家产品图存档时被打上了满屏水印，重新上架却找不回无水印原图。指出要去掉的水印文字和右上角图标后，橘子、叶片和石台边缘完整保留，米色渐变背景接得连贯，看不出修补痕迹。适合需要复用旧素材的电商运营和内容编辑。", "categories": [ "creation" ], "tags": [ "去水印/杂物", "去水印", "产品图", "素材复用" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "无水印护肤品主图｜单张 PNG" } ], "prompt": "[去水印/杂物]去除图片中的指定内容。\n待处理图片：一张打了满屏平铺水印的护肤品静物图\n需要去除的内容：满屏平铺的文字水印和右上角的圆形图标\n其他要求：橘子、叶片、石台和瓶身边缘保持完整，米色渐变背景要接得连贯，不留涂抹痕迹\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-content-removal", "name": "去水印/杂物", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298843975680" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-content-removal-sunglasses-lens-glare", "title": "墨镜镜片反光消除", "title_en": "Remove Glare from Sunglasses Lenses", "subtitle": "镜片里的海天倒影去掉，眼神重新看得见", "subtitle_en": "Clear lens reflections so the eyes show through", "description": "户外强光下拍人像，镜片总是映满海面和天空，眼神被反光整个盖住。指定去掉镜片反光后，透过茶色镜片能看清眼睛，镜框颜色、脸部光影和发丝细节都没有被动过，画面还是原来的海边氛围。适合眼镜品牌拍摄和爱戴墨镜出片的人像修图。", "categories": [ "creation" ], "tags": [ "去水印/杂物", "反光消除", "人像修图", "眼镜" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "去镜片反光人像｜单张 JPG" } ], "prompt": "[去水印/杂物]去除图片中的指定内容。\n待处理图片：一张海边逆光人像，戴茶色墨镜、镜片映满海天倒影\n需要去除的内容：墨镜镜片上映出的海面和天空反光\n其他要求：去掉反光后透过镜片要能看清眼睛，镜框颜色、脸部光影和发丝保持原样\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-content-removal", "name": "去水印/杂物", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298843975680" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-conversion-photo-windbreaker-studio", "title": "白底棚拍投放穿搭图", "title_en": "White Studio Conversion Shot", "subtitle": "高转化模特棚拍图，适配电商主图", "subtitle_en": "Studio model shots built for platform conversion", "description": "上传服饰图片，AI生成白底棚拍的高转化模特图，浅绿色防风外套搭配棒球帽、墨镜与双肩包，展示细节清晰、光影统一。适合电商卖家批量产出各电商平台的商品主图与详情页素材，提升转化率。", "categories": [ "creation" ], "tags": [ "转化图", "服饰穿搭", "白底棚拍", "防风外套", "电商主图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "白底转化穿搭图｜单张 JPG" } ], "prompt": "[服饰穿搭]生成投放级模特穿搭图。\n服饰图片：（请上传：服饰图片）\n投放平台：\n模特风格：亚洲女性\n图片比例：3:4\n生成张数： 张\n其他要求：浅绿色防风外套配黑色双肩包，棒球帽与墨镜搭配，白底棚拍展示\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-conversion-photo", "name": "转化照", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297177628672" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-custom-social-ip-mom-daughter-illustration", "title": "亲子博主IP形象", "title_en": "Mom & Daughter IP Persona", "subtitle": "温暖插画风母女形象，打造育儿账号人设", "subtitle_en": "A warm illustrated mom-and-daughter persona", "description": "想做母婴育儿账号，却没有一个有辨识度的虚拟形象开篇？告诉AI你的赛道和人设目标，自动生成温暖插画风的母女IP形象：卷发妈妈牵着扎羊角辫的女儿，柔和粉绿配色治愈感十足。适合亲子博主和母婴账号打造有记忆点的视觉人设，从头像到内容风格统一延续。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "亲子博主", "母婴IP", "插画人设", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "亲子博主IP形象包｜单套" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：（请上传：参考素材/IP资料）\nIP 模式：创建新 IP\n赛道/领域：母婴育儿、亲子生活\n人设目标/续写主题：打造温暖治愈的亲子博主人设，主打日常带娃和育儿干货分享\n目标平台：小红书\n输出内容：完整IP包\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-custom-social-ip", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_WhZub3RCn3QM" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-custom-social-ip-nurse-rabbit-plush", "title": "健康科普玩偶IP形象", "title_en": "Nurse Rabbit Plush IP", "subtitle": "毛绒护士兔形象，医疗科普更有亲和力", "subtitle_en": "A plush nurse rabbit that softens medical topics", "description": "想做健康科普账号，却担心医疗话题太严肃劝退粉丝？告诉AI你的赛道和人设目标，自动生成毛绒玩偶风IP形象：戴护士帽、挂听诊器的白色兔子，薄荷绿制服配色温柔治愈。适合护理、健康科普博主打造软萌人设，用可爱形象拉近和粉丝的距离，降低专业内容的阅读门槛。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "健康科普", "玩偶IP", "医疗博主", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "健康科普IP形象包｜单套" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：（请上传：参考素材/IP资料）\nIP 模式：创建新 IP\n赛道/领域：健康护理、医疗科普\n人设目标/续写主题：打造温柔亲和的健康科普博主人设，主打通俗易懂的护理和急救知识\n目标平台：小红书\n输出内容：完整IP包\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-custom-social-ip", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_WhZub3RCn3QM" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-custom-social-ip-squirrel-finance-mascot", "title": "理财博主动物IP形象", "title_en": "Squirrel Finance Mascot IP", "subtitle": "松鼠财务官形象，专业感和亲和力兼顾", "subtitle_en": "A squirrel finance officer, sharp yet approachable", "description": "想做理财科普账号，又怕纯人像形象显得严肃难亲近？告诉AI你的赛道和人设目标，自动生成拟人化动物IP形象：穿西装马甲、拿计算器的松鼠财务官，暖橙色调配复古插画质感，专业又可爱。适合理财、财税科普博主打造差异化人设，比真人出镜更省心也更有记忆点。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "理财博主", "动物IP", "拟人形象", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "理财博主IP形象包｜单套" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：（请上传：参考素材/IP资料）\nIP 模式：创建新 IP\n赛道/领域：理财规划、财税科普\n人设目标/续写主题：打造专业又亲和的理财博主人设，主打普通人能听懂的理财干货\n目标平台：小红书\n输出内容：完整IP包\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-custom-social-ip", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_WhZub3RCn3QM" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-detail-page-anc-headphones", "title": "降噪头戴耳机详情模块图", "title_en": "ANC Headphones Detail Images", "subtitle": "6张模块图讲清降噪、续航、便携等核心卖点", "subtitle_en": "Six detail modules covering ANC, battery life and portability", "description": "3C耳机类目竞争激烈，主图之外必须靠详情页模块图讲清降噪效果、续航时长和佩戴舒适度，自己排版耗时且风格不统一。上传产品图，AI 自动生成主动降噪说明图、40小时续航图、可折叠收纳图和多场景佩戴图，适合数码类目详情页快速搭建。", "categories": [ "creation" ], "tags": [ "详情模块图", "头戴耳机", "3C数码详情页", "A+页面" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "头戴耳机详情模块图｜6张模块组合" } ], "prompt": "[详情模块图]帮我设计电商详情模块图。\n商品图片：（请上传：商品图片）\n目标平台：亚马逊\n比例/档位：3:4\n图内文字语言：英文\n销售地区：美国\n商品卖点&要求：橙色头戴式蓝牙耳机，主动降噪+40小时续航，40mm驱动单元，可折叠便携，适合3C数码类目详情页，希望突出通勤、健身、旅行多场景使用画面\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-detail-page", "name": "详情模块图", "description": "A+/详情页模块图，提升转化率", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_KobP8m2JXdew" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-detail-page-enamel-dutch-oven", "title": "搪瓷铸铁锅详情模块图", "title_en": "Enamel Dutch Oven Detail Images", "subtitle": "6张模块图展示导热、锅盖、烤箱适配等厨具卖点", "subtitle_en": "Six detail modules for heat retention, lid design and oven use", "description": "铸铁锅的卖点集中在导热均匀、自循环锅盖和耐高温这些细节，靠文字描述很难让买家直观感受。上传产品图，AI 自动生成均匀导热示意图、自循环锅盖结构图、烤箱适配图和家庭聚餐使用场景图，风格统一，适合厨具类目详情页快速搭建。", "categories": [ "creation" ], "tags": [ "详情模块图", "铸铁锅", "厨具详情页", "A+页面" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "铸铁锅详情模块图｜6张模块组合" } ], "prompt": "[详情模块图]帮我设计电商详情模块图。\n商品图片：（请上传：商品图片）\n目标平台：亚马逊\n比例/档位：3:4\n图内文字语言：英文\n销售地区：美国\n商品卖点&要求：明黄色搪瓷铸铁锅，卖点为均匀导热、自循环锅盖保湿、可入烤箱耐500华氏度，适合厨具类目详情页，画面希望突出家庭烹饪与聚餐场景\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-detail-page", "name": "详情模块图", "description": "A+/详情页模块图，提升转化率", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_KobP8m2JXdew" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-detail-page-smart-rice-cooker", "title": "电饭煲详情页模块图", "title_en": "Smart Rice Cooker Detail Page", "subtitle": "六大卖点模块化排版，做出对标电商 A+ 页面效果", "subtitle_en": "Six-module A+ style layout that highlights every selling point", "description": "小家电详情页想清楚传达功能卖点，靠单张图很难说清楚。上传产品图后，AI 按照场景图、卖点图、参数图三段式排版，自动生成24小时预约、多模式烹饪、不粘内胆等六个功能模块，标题排版统一、图标风格一致，直接可用于电商 A+ 页面或独立站详情区，无需再手动拼图。", "categories": [ "creation" ], "tags": [ "详情模块图", "A+页面", "电饭煲", "小家电详情页" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "电饭煲详情页六宫格｜长图 PNG" } ], "prompt": "[详情模块图]帮我设计电商详情模块图。\n商品图片：（请上传：商品图片）\n目标平台：亚马逊\n比例/档位：普通A+ 970:600\n图内文字语言：英文\n销售地区：美国\n商品卖点&要求：商品名称：多功能智能电饭煲；核心卖点：24小时预约、8种烹饪模式、不粘内胆、智能触控面板、360°均匀受热、大容量适合家庭；适用人群：注重厨房效率的家庭用户\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-detail-page", "name": "详情模块图", "description": "A+/详情页模块图，提升转化率", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_KobP8m2JXdew" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-detail-page-tower-air-purifier", "title": "塔式空气净化器详情模块图", "title_en": "Tower Air Purifier Detail Images", "subtitle": "6张模块图讲清过滤、传感、静音卖点，提升详情页转化", "subtitle_en": "Six detail modules covering filtration, sensing and quiet mode", "description": "空气净化器的核心卖点是过滤效率和智能功能，光靠一张主图很难讲清楚，自己排版详情页耗时又难统一风格。上传产品图，AI 自动生成HEPA滤网结构图、智能传感器说明图、静音模式场景图和覆盖面积说明图等一整套模块，适合家电类目详情页快速搭建。", "categories": [ "creation" ], "tags": [ "详情模块图", "空气净化器", "家电详情页", "A+页面" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "空气净化器详情模块图｜6张模块组合" } ], "prompt": "[详情模块图]帮我设计电商详情模块图。\n商品图片：（请上传：商品图片）\n目标平台：亚马逊\n比例/档位：3:4\n图内文字语言：英文\n销售地区：美国\n商品卖点&要求：白色圆柱塔式空气净化器，HEPA三重滤网，99.97%过滤效率，智能传感器实时监测空气质量，卖点为静音睡眠模式和360°进风设计，适合家电类目详情页\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-detail-page", "name": "详情模块图", "description": "A+/详情页模块图，提升转化率", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_KobP8m2JXdew" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-drama-ad-airport-neck-pain-travel", "title": "旅行颈枕剧情广告", "title_en": "Travel Pillow Story Ad", "subtitle": "候机厅的脖子酸痛瞬间，戳中旅行党的痛点", "subtitle_en": "An airport neck-ache moment every traveler feels", "description": "上传商品图，AI 自动生成机场候机厅场景的剧情广告：长途出行揉脖子、皱眉忍痛的细节被镜头捕捉，同伴关切的眼神进一步放大痛点，为产品出场做足铺垫。适合旅行出行类商品卖家，用真实生活情境代替生硬卖点堆砌，让观众更容易共情下单。", "categories": [ "creation" ], "tags": [ "剧情广告", "颈枕", "旅行好物", "痛点营销", "候机场景" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "旅行颈枕剧情广告｜竖版短视频" } ], "prompt": "[剧情广告]制作剧情广告短视频。\n商品图片：（请上传：商品图片）\n商品名称/卖点：深海蓝旅行颈枕，长途出行护颈必备\n剧情风格：霸总宠爱\n视频比例：9:16 竖屏\n其他要求：机场候机厅实景，展现长途飞行脖子酸痛的真实痛点瞬间，同伴关切互动\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-drama-ad", "name": "剧情广告", "description": "网感玩梗短视频，脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7ok3STENDZZT" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-drama-ad-mug-costume-drama", "title": "陶瓷马克杯古装穿越广告", "title_en": "Costume Drama Mug Ad", "subtitle": "古装场景穿越演绎，马克杯变身剧情道具", "subtitle_en": "A costume-drama style ad for a ceramic mug", "description": "古装穿越是最容易让用户停下来多看两秒的剧情风格，也最能凸显产品的反差趣味。上传商品图，AI 生成古装置景中人物对手戏的画面，结尾穿越回现代生活揭晓马克杯的日常用途，反差感拉满。适合水杯、居家好物做创意剧情广告。", "categories": [ "creation" ], "tags": [ "剧情广告", "马克杯", "古装穿越", "创意剧情", "短视频广告" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "马克杯古装穿越剧情广告｜竖版视频" } ], "prompt": "[剧情广告]制作剧情广告短视频。\n商品图片：（请上传：商品图片）\n商品名称/卖点：奶油白陶瓷马克杯，保温不烫手\n剧情风格：古装穿越\n视频比例：9:16 竖屏\n其他要求：古装置景运镜，人物对手戏演绎，结尾穿越回现代展示产品\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-drama-ad", "name": "剧情广告", "description": "网感玩梗短视频，脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7ok3STENDZZT" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-drama-ad-tissue-office-sneeze-comedy", "title": "抽纸盒办公室搞笑剧情广告", "title_en": "Office Sneeze Comedy Ad", "subtitle": "打喷嚏名场面猝不及防，抽纸盒变身职场救星", "subtitle_en": "A funny office sneeze story for a tissue box ad", "description": "职场场景自带戏剧张力，一个措手不及的喷嚏就能撑起一条剧情广告。上传商品图，AI 生成办公室里同事无奈围观、抽纸盒及时救场的搞笑桥段，反转节奏紧凑，产品自然成为笑点核心。适合纸品、办公好物做轻松诙谐的种草内容。", "categories": [ "creation" ], "tags": [ "剧情广告", "抽纸盒", "搞笑日常", "职场场景", "反转剧情" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "抽纸盒搞笑剧情广告片｜竖版视频" } ], "prompt": "[剧情广告]制作剧情广告短视频。\n商品图片：（请上传：商品图片）\n商品名称/卖点：原木色抽纸盒，柔软加厚\n剧情风格：搞笑日常\n视频比例：9:16 竖屏\n其他要求：办公室实景，打喷嚏尴尬瞬间反转，同事无奈表情增加笑点\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-drama-ad", "name": "剧情广告", "description": "网感玩梗短视频，脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7ok3STENDZZT" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-drama-ad-towel-family-morning", "title": "浴巾晨间温情剧情广告", "title_en": "Cozy Morning Towel Story Ad", "subtitle": "居家清晨的温馨瞬间，浴巾自然融入生活场景", "subtitle_en": "A cozy morning story ad featuring a soft towel", "description": "浴巾这类日用品很难靠参数打动人，需要靠生活情绪代入。上传商品图，AI 生成清晨厨房阳光洒落的温情小剧场，家人洗漱擦脸的日常瞬间自然带出产品，结尾给到产品特写。适合家纺、日用品类目做剧情广告，用温度感替代硬广感。", "categories": [ "creation" ], "tags": [ "剧情广告", "浴巾", "家庭温情", "居家好物", "生活场景" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "浴巾家庭温情剧情广告｜竖版视频" } ], "prompt": "[剧情广告]制作剧情广告短视频。\n商品图片：（请上传：商品图片）\n商品名称/卖点：焦糖色柔软浴巾，吸水速干不掉毛\n剧情风格：家庭温情\n视频比例：9:16 竖屏\n其他要求：清晨厨房阳光实景，温馨生活流叙事，结尾产品特写\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-drama-ad", "name": "剧情广告", "description": "网感玩梗短视频，脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_7ok3STENDZZT" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ecommerce-product-kit-candle", "title": "香薰产品电商 Listing 套图", "title_en": "Scented Candle Ecommerce Listing Image Kit", "subtitle": "上传产品图，一键生成完整电商主图、卖点图与场景图", "subtitle_en": "Turn product photos into a complete ecommerce listing image set", "description": "以香薰蜡烛为示例，把单张产品图扩展为适合 Amazon、Temu、TikTok Shop 等平台使用的完整商品套图，覆盖白底主图、核心卖点、规格信息与生活方式场景。适合需要快速完成商品上架视觉物料的跨境卖家、品牌团队和电商运营人员。", "description_en": "Using a scented candle as an example, this case turns a single product photo into a complete listing image set for Amazon, Temu, and TikTok Shop — covering white-background main images, key selling points, spec details, and lifestyle scenes. Ideal for cross-border sellers, brand teams, and ecommerce operators who need to quickly produce listing visuals.", "categories": [ "creation", "business" ], "tags": [ "商品套图", "电商主图", "Listing 图片", "跨境电商", "美图设计室" ], "tags_en": [ "Product Image Kit", "Ecommerce Main Image", "Listing Images", "Cross-border Ecommerce", "Meitu DesignKit" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "香薰产品电商 Listing 套图｜单文件 HTML" } ], "prompt": "请根据我上传的 1–3 张商品图，为该商品生成一套可直接用于 Amazon、Temu、TikTok Shop 等电商平台的 Listing 商品套图。请先识别商品主体、材质、品类和主要卖点，再统一规划整套视觉风格；至少覆盖白底主图、核心卖点图、规格或细节图、使用场景图和生活方式图。保持商品外观、包装、颜色、Logo 与关键文字一致，不得改变产品结构；图内文案简洁、可读，默认使用英文，画面比例为 1:1。完成后直接交付整套高质量图片，并确保不同图片之间具有一致的品牌感和版式体系。\n\n输出要求：完成视觉内容后，将全部页面或图片整理为一个自包含的单文件 HTML，文件名固定为 output.html。所有图片、样式和必要脚本都内联在 HTML 中，不输出独立图片、资源目录或第二个 output 文件；页面需可直接打开，并兼容桌面端与移动端。", "prompt_en": "Based on 1–3 product photos I upload, generate a complete Listing image set suitable for Amazon, Temu, TikTok Shop and other ecommerce platforms. First identify the product, material, category and key selling points, then plan a unified visual style; cover at least white-background main image, core selling point image, spec/detail image, usage scenario image and lifestyle image. Keep product appearance, packaging, color, logo and key text consistent; do not alter the product structure. Use concise, readable copy in English by default, with 1:1 aspect ratio. Deliver the complete set of high-quality images with consistent branding and layout.\n\nOutput: After completing the visuals, organize all pages or images into a single self-contained HTML file named output.html. All images, styles and necessary scripts must be inlined; no separate image files, resource directories, or second output files. The page must open directly and be compatible with both desktop and mobile.", "skills": [], "experts": [ { "id": "designkit-listing-images", "name": "商品套图", "description": "上传产品图，一键生成适配各平台的电商全套主图与场景图。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TqlFgWZviW1K" } ], "mcps": [], "scenario": "", "scenario_use": "品牌设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-27T10:22:53+08:00", "updated_at": "2026-08-27T10:22:53+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-id-photo-blue-student-uniform", "title": "蓝底学生校服证件照", "title_en": "Blue Background Student ID Photo", "subtitle": "校服照一键转正规蓝底学生证件照", "subtitle_en": "Turn a uniform photo into a standard blue-background student ID", "description": "学生证、学籍照常需要蓝底校服证件照，重新去拍照又麻烦。这款工具能把日常照片转换成蓝底标准学生证件照，校服领结同步规范，发型干净利落，适合学生证办理、学籍注册、入学材料等场景直接提交使用。", "categories": [ "creation" ], "tags": [ "证件照", "学生证件照", "蓝底证件照", "校服照" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "蓝底学生证件照｜单张 JPG" } ], "prompt": "[证件照]生成证件照。\n人物照片：（请上传：人物照片）\n照片类型：标准证件照\n照片规格：一寸（25×35mm）\n背景颜色：蓝色\n其他要求：女生，深蓝色校服西装外套，白衬衫配条纹领结，头发盘起干净利落，适合学生证、学籍照等场景\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-id-photo", "name": "证件照", "description": "证件照/形象照生成与美化", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_MCiLNXYsIGLm" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-id-photo-grey-business-headshot", "title": "灰底男士商务证件照", "title_en": "Grey Business ID Photo", "subtitle": "生活照一键转正装商务证件照，背景灯光同步专业", "subtitle_en": "Turn a casual photo into a formal business ID photo", "description": "商务人士常为求职、履历、工牌拍照发愁——预约影棚麻烦，自拍效果又不专业。这款工具能把生活照直接转换成标准商务形象照，深色西装、灰色背景、专业光影一步到位，适合求职简历、企业官网、社交主页等场景使用，出图即可直接提交。", "categories": [ "creation" ], "tags": [ "证件照", "商务形象", "求职简历", "职场形象照" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "灰底商务证件照｜单张 JPG" } ], "prompt": "[证件照]生成证件照。\n人物照片：（请上传：人物照片）\n照片类型：职业形象照\n照片规格：自动匹配\n背景颜色：灰色\n其他要求：深色西装外套，藏青色条纹领带，短发利落，微笑露齿表情自然\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-id-photo", "name": "证件照", "description": "证件照/形象照生成与美化", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_MCiLNXYsIGLm" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-id-photo-red-male-exam", "title": "红底男士考试证件照", "title_en": "Red Background Exam ID Photo", "subtitle": "公务员考试、职业资格证件照红底标准一键生成", "subtitle_en": "Generate a standard red-background photo for exam registration", "description": "公务员考试、职业资格考试常要求红底证件照，普通生活照不符合规格。这款工具把生活照转换成红底标准证件照，深色西装、领带同步规范，光线均匀、五官清晰，适合公务员考试、计算机等级考试等各类红底证件报名场景使用。", "categories": [ "creation" ], "tags": [ "证件照", "红底证件照", "考试证件照", "公务员报名照" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "红底证件照｜单张 JPG" } ], "prompt": "[证件照]生成证件照。\n人物照片：（请上传：人物照片）\n照片类型：其他证件照\n照片规格：自动匹配\n背景颜色：红色\n其他要求：男性，深色西装打藏青色领带，短发利落，表情自然严肃，适合公务员考试、职业资格考试等红底证件要求\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-id-photo", "name": "证件照", "description": "证件照/形象照生成与美化", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_MCiLNXYsIGLm" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-id-photo-white-female-standard", "title": "白底女生标准证件照", "title_en": "White Background Female ID Photo", "subtitle": "生活照秒变标准证件照，五官比例精准还原", "subtitle_en": "Turn a casual photo into a standard white-background ID photo", "description": "考试报名、入职登记常常要求标准证件照，去照相馆排队又费时间。上传一张生活照，就能自动生成白底标准证件照，五官比例、光线和着装同步规范，适合各类考试报名、入职材料、简历附件等场景直接使用。", "categories": [ "creation" ], "tags": [ "证件照", "标准证件照", "白底证件照", "女生证件照" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "白底标准证件照｜单张 JPG" } ], "prompt": "[证件照]生成证件照。\n人物照片：（请上传：人物照片）\n照片类型：标准证件照\n照片规格：自动匹配\n背景颜色：白色\n其他要求：女性，黑色西装外套，白衬衫，长发自然披肩，妆容清淡，表情自然微笑\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-id-photo", "name": "证件照", "description": "证件照/形象照生成与美化", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_MCiLNXYsIGLm" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-image-translate-headphones-to-spanish", "title": "耳机详情图翻译西语", "title_en": "Translate Headphone Visuals to Spanish", "subtitle": "保留卡片布局，文字一键换成目标语言", "subtitle_en": "Keep the layout, swap in translated copy", "description": "上传英文版的耳机产品介绍图，选择目标语言后自动将标题和功能卡片文字翻译替换，图标位置、字体风格和整体版式保持不变，适合出海电商快速产出多语言版本的商品详情图。", "categories": [ "creation" ], "tags": [ "图片翻译", "多语言详情图", "跨境电商", "版式保留", "西班牙语" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "耳机详情图西语版｜单张 PNG" } ], "prompt": "[图片翻译]将图片上的文字进行翻译。\n待翻译图片：（请上传：待翻译图片）\n目标语言：Español\n其他说明：保留信息卡片图标、色块和布局位置不变，仅替换文字内容\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-image-translate", "name": "图片翻译", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298994528256" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-image-translate-luggage-to-portuguese", "title": "行李箱主图转葡语版", "title_en": "Luggage Hero Image in Portuguese", "subtitle": "中文卖点标签整组转葡语，标签框不动", "subtitle_en": "Turn a Chinese hero image into Portuguese", "description": "中文素材直接发到南美站点，用户看不懂就划走了。这个案例把标题和两个白色圆角卖点标签一起转成葡语，静音轮、扩容夹层的图标位置和标签框完全不动，机场候机厅的通透背景原样保留。适合行李箱、出行装备类目做南美站点铺货。", "categories": [ "creation" ], "tags": [ "图片翻译", "行李箱", "跨境电商", "商品主图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "葡语版行李箱主图｜单张 JPG" } ], "prompt": "[图片翻译]将图片上的文字进行翻译。\n待翻译图片：一张机场候机厅的浅灰行李箱主图，左上角中文标题加两个白色圆角卖点标签\n目标语言：Português\n其他说明：20 寸的尺寸数字保留；白色标签框的位置、圆角和内部图标不动，只替换文字\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-image-translate", "name": "图片翻译", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298994528256" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-image-translate-sportswear-to-korean", "title": "运动服广告转韩语版", "title_en": "Sportswear Ad in Korean", "subtitle": "主标两行大字排版照搬，只换语言", "subtitle_en": "Same headline layout, Korean copy", "description": "同一组广告大片要投多个国家，逐个国家重排文字太耗人。这张把英文主标和两行副文案整体换成韩语，两行大字的位置、字号和左侧留白照旧，模特动作与蓝天底色不受影响。适合运动服饰品牌一次投放多语种市场时快速铺量。", "categories": [ "creation" ], "tags": [ "图片翻译", "运动服饰", "韩国市场", "广告大图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "韩语版运动服广告图｜单张 JPG" } ], "prompt": "[图片翻译]将图片上的文字进行翻译。\n待翻译图片：一张户外压缩衣广告大图，左上角白色英文主标 FEEL NOTHING，下方两行小字副文案\n目标语言：한국어\n其他说明：主标保持两行断行与大字号，副文案两行对齐不变；白色字色与左侧留白照旧\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-image-translate", "name": "图片翻译", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298994528256" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-image-translate-watch-to-arabic", "title": "腕表详情图转阿语版", "title_en": "Watch Detail Page in Arabic", "subtitle": "文案翻译同时整版右对齐，图标跟着换边", "subtitle_en": "Translate and mirror the layout right-to-left", "description": "出海中东，详情图不只是把英文换成阿语那么简单，整版还得从右往左读。这个案例连图标带文字一起换到画面右侧，续航、睡眠监测、防水三条卖点对齐重排，表盘上的心率与步数数字原样保留。适合数码配件、腕表品类做中东站点的本地化上架。", "categories": [ "creation" ], "tags": [ "图片翻译", "智能腕表", "中东市场", "详情图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "阿语版腕表详情图｜单张 JPG" } ], "prompt": "[图片翻译]将图片上的文字进行翻译。\n待翻译图片：一张深色调智能腕表详情图，左上角是英文主标 SLIM WATCH，下方三条带图标的卖点\n目标语言：العربية\n其他说明：标题与卖点整组改为右对齐、从右往左阅读，图标移到文字右侧；表盘上的 72、7450 数字保持不变\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-image-translate", "name": "图片翻译", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298994528256" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ip-note-continuation-hrbp-interview-mistakes", "title": "HR博主IP续写避坑图文", "title_en": "HRBP IP Continuation: Interview Mistakes", "subtitle": "同一形象换个表情，续写吐槽向干货笔记", "subtitle_en": "The same character, a new expression, a new topic", "description": "已经用固定IP形象发过几篇干货，这次想换个吐槽向的选题继续更新？上传角色参考图和本篇主题，AI延用同一形象但调整表情和姿态，自动生成避坑警示类图文，配色和排版延续账号一贯风格。适合职场、HR类博主保持IP形象连贯的同时，丰富内容的情绪和选题角度。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "HR博主", "IP延续", "职场避坑", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "HR博主IP续写图文｜单张" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：已有HR博主IP形象参考图\nIP 模式：延用 IP 生图文\n赛道/领域：职场成长、HR干货\n人设目标/续写主题：延用已有IP形象，续写访谈常见3种误区避坑图文\n目标平台：小红书\n输出内容：延用 IP 图文\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ip-note-continuation", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Ba07Fy85p0E" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ip-note-continuation-hrbp-interview-steps", "title": "HR博主IP续写干货图文", "title_en": "HRBP IP Continuation: Interview Steps", "subtitle": "同一形象延续出镜，干货图文不用重新设计", "subtitle_en": "The same character carries a new how-to post", "description": "HR博主的IP形象已经定好，但每次出新一篇干货笔记又要重新构图很麻烦？上传角色参考图和本篇主题，AI延用同一形象出镜，自动生成配套的步骤化知识卡片图文，标签配色和角色姿态保持统一。适合职场、HR类博主批量产出系列干货笔记，账号视觉始终如一。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "HR博主", "IP延续", "职场干货", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "HR博主IP续写图文｜单张" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：已有HR博主IP形象参考图\nIP 模式：延用 IP 生图文\n赛道/领域：职场成长、HR干货\n人设目标/续写主题：延用已有IP形象，续写员工访谈4步法干货图文\n目标平台：小红书\n输出内容：延用 IP 图文\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ip-note-continuation", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Ba07Fy85p0E" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ip-note-continuation-insurance-buying-tips", "title": "保险博主IP续写图文", "title_en": "Insurance IP Note Continuation", "subtitle": "延续同一保险经纪人人设，续写选购避坑图文", "subtitle_en": "Keep the same persona, write a new insurance tip note", "description": "已经为账号打造好保险经纪人IP形象，却担心新一篇图文风格跑偏、人设不统一？延用该IP的人物形象和语气，围绕保险选购误区继续输出新一篇图文，构图、字体配色与往期保持一致，账号调性不断档，适合金融理财类博主批量产出系列内容。", "categories": [ "creation" ], "tags": [ "定制社媒IP", "IP人设续写", "保险科普", "金融理财号", "图文系列" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "保险经纪人IP图文续写｜单张海报" } ], "prompt": "[定制社媒IP]按所选模式创建社媒人设与 IP 风格，或基于已有 IP 资料续写新图文。\n参考素材/IP资料：（请上传：参考素材/IP资料）\nIP 模式：延用 IP 生图文\n赛道/领域：保险科普、金融理财\n人设目标/续写主题：延续保险经纪人IP人设，续写强调按需选择保险、拒绝盲目跟风的图文，呼应此前保险选购误区的痛点篇\n目标平台：小红书\n输出内容：延用 IP 图文\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ip-note-continuation", "name": "定制社媒IP", "description": "创建社媒 IP，或基于已有 IP 续写图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Ba07Fy85p0E" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-lesson-plan-to-ppt-love-garden-animals", "title": "爱护菜园小动物 教案转PPT", "title_en": "Protecting Garden Animals Lesson Plan to PPT", "subtitle": "把自然探索教案转化为 10 页绘本水彩风幼儿教学课件", "subtitle_en": "Turn a nature-exploration lesson plan into a 10-page watercolor storybook classroom deck", "description": "这是一套由教案自动生成的 10 页幼儿综合探索课件。案例围绕“爱护菜园小动物”展开，以绘本水彩、自然绿和黏土质感动物建立亲切的田园课堂氛围，依次呈现课程导览、菜园小邻居、益虫与小动物的作用、观察探索、爱护方法、互动练习和课程小结。内容结构清晰、语言童趣、图文比例适合课堂投屏，可用于幼儿园及小学低年级的自然教育、生命教育和综合实践课程。", "categories": [ "creation" ], "tags": [ "教案转PPT", "幼儿课件", "自然教育", "生命教育", "绘本水彩" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "爱护菜园小动物教学课件｜10 页单文件 HTML" } ], "prompt": "[教案转PPT]根据教案生成教学PPT。\n教案内容：请读取我提供的教案，围绕“爱护菜园小动物”梳理教学目标、知识结构和课堂节奏；重点讲清常见菜园小动物、它们对菜园的作用、正确观察方式、爱护生命的方法，并设计适合课堂参与的观察或问答环节。\n教案文件：请使用我上传的教案、文字稿或图片资料，保持其中的教学目标、事实与核心活动要求，不得擅自编造关键知识。\n教学对象：幼儿园大班或小学低年级学生。\n语言：简体中文，表达童趣、简短、易懂。\n篇幅：约 10 页，包含封面、课程导览、知识讲解、互动探索、行动倡议和课程小结。\n采用16:9横版，内容结构清晰，适合课堂讲授。视觉采用绘本水彩田园风，自然绿与暖黄色为主，使用亲切可爱的菜园小动物插画，保持整套课件风格统一、文字清晰、投屏可读。", "skills": [], "experts": [ { "id": "designkit-lesson-plan-to-ppt", "name": "教案转PPT", "description": "读取教案与教学材料，梳理教学目标、知识结构和课堂节奏，生成适合不同学段讲授的课件。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_LT5j5Sn5xbIE" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-listing-copy-hair-oil-brazil", "title": "护发精油Listing文案", "title_en": "Hair Oil Listing Copy", "subtitle": "标题、卖点、描述一次生成，直接投巴西站", "subtitle_en": "Generate title, bullets and description for the Brazil market", "description": "写Listing文案最耗时的是标题埋词、五点卖点提炼和多平台格式适配，靠自己一条条憋很慢。填入商品名称和卖点后，AI 会按目标平台规范一次性产出标题、卖点亮点、正文描述和搜索词，风格统一、语言地道，巴西站等海外市场也能直接套用，省去来回改稿的时间。", "categories": [ "business" ], "tags": [ "Listing文案", "护发精油", "跨境电商平台标题", "巴西站" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "护发精油Listing文案｜文字稿" } ], "prompt": "[商品文案]撰写商品Listing文案。\n商品图片：（请上传：商品图片）\n商品名称：植萃护发精油 80ml\n目标平台：亚马逊\n目标市场：巴西\n文案语言：English\n商品卖点&要求：植物萃取护发精油，滋润不油腻，轻松抚平毛躁，快速泵头设计方便使用；目标市场巴西，主打日常护发场景\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-listing-copy", "name": "商品文案", "description": "Listing文案撰写，多平台适配", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_wcDOKu4gvgCb" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-listing-copy-shopee-earbud-case", "title": "东南亚电商平台耳机壳文案", "title_en": "Shopee Earbud Case Copy", "subtitle": "标题、卖点、检索标签一次生成，附葡语版本说明", "subtitle_en": "Title, selling points and tags for Shopee Brazil listing", "description": "东南亚电商平台巴西站的文案风格和跨境电商平台不同，标题要短、卖点要直白，自己重新组织容易漏点。上传产品图并填写卖点，AI 生成标题、亮点标签、卖点列表、简介和检索标签，并说明需要巴西葡语本地化版本，适合配饰类目东南亚电商平台巴西站快速上新。", "categories": [ "business" ], "tags": [ "商品文案", "耳机壳文案", "东南亚电商平台", "巴西市场" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "耳机壳Listing文案｜标题+卖点+标签" } ], "prompt": "[商品文案]撰写商品Listing文案。\n商品图片：（请上传：商品图片）\n商品名称：花朵挂饰耳机保护套\n目标平台：Shopee\n目标市场：巴西\n文案语言：\n商品卖点&要求：粉色硅胶耳机保护壳，带花朵挂坠钥匙扣，卖点为防刮耐磨、轻便可挂包，面向巴西Shopee市场，希望生成标题、亮点、卖点列表、简介和检索标签，并注明需要巴西葡语本地化版本\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-listing-copy", "name": "商品文案", "description": "Listing文案撰写，多平台适配", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_wcDOKu4gvgCb" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-listing-images-air-fryer-sale-set", "title": "空气炸锅促销电商套图", "title_en": "Air Fryer Sale Image Set", "subtitle": "白底图+功能卖点图+促销场景图一次生成", "subtitle_en": "Generate a full sale-ready image set for your appliance", "description": "促销节点上新最缺的就是一整套主图，单张单张修图效率太低。上传空气炸锅产品图，AI 自动生成白底主图、多功能卖点图、尺寸参数图和节日促销场景图，风格统一、文字清晰，一次产出全套素材，适合厨房小家电类目大促备货和常规上新使用。", "categories": [ "creation" ], "tags": [ "商品套图", "空气炸锅", "促销主图", "厨房小家电" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "空气炸锅电商套图｜主图+卖点图+促销图组合" } ], "prompt": "[商品套图]帮我生成电商商品套图。\n商品图片：（请上传：商品图片）\n适配平台：亚马逊\n图片比例：3:4\n图内文字语言：英语\n目标市场：美国\n商品卖点&要求：米白色6合1空气炸锅，4QT家庭装容量，正值促销活动节点，卖点为多功能烹饪（空气炸/烘焙/烤/发酵/复热）与紧凑机身设计，适合厨房小家电类目\n交付物仅套图，不要 Listing 文案 / 商品文案。\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-listing-images", "name": "商品套图", "description": "上传产品图，生成全套电商主图", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TqlFgWZviW1K" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-listing-images-boucle-swivel-chair", "title": "复古条纹旋转扶手椅套图", "title_en": "Striped Swivel Chair Image Set", "subtitle": "客厅场景图+细节特写图一次生成，突出材质设计感", "subtitle_en": "Show off fabric and design details in one image set", "description": "家具类目最耗时的是场景图和细节特写图的搭配拍摄，自己布置样板间成本很高。上传旋转扶手椅产品图，AI 自动生成客厅场景图、材质特写图、多角度展示图和设计亮点标注图，风格统一、光影自然，适合家具软装类目上新与详情页素材补充。", "categories": [ "creation" ], "tags": [ "商品套图", "家具软装", "旋转扶手椅", "客厅场景图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "旋转扶手椅电商套图｜场景图+细节图组合" } ], "prompt": "[商品套图]帮我生成电商商品套图。\n商品图片：（请上传：商品图片）\n适配平台：亚马逊\n图片比例：3:4\n图内文字语言：英语\n目标市场：美国\n商品卖点&要求：米黄拼条纹绒面单人旋转扶手椅，古铜色金属圆形底座，复古摩登设计，卖点为舒适厚坐垫与精致细节做工，适合家具软装类目\n交付物仅套图，不要 Listing 文案 / 商品文案。\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-listing-images", "name": "商品套图", "description": "上传产品图，生成全套电商主图", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TqlFgWZviW1K" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-listing-images-sun-serum-sea-market", "title": "东南亚防晒精华套图", "title_en": "Sun Serum Image Set (SEA)", "subtitle": "海岛场景图+成分卖点图，贴合当地消费者审美", "subtitle_en": "Beach-themed image set tailored for Southeast Asia", "description": "出海东南亚市场，主图风格和文案语言都要重新适配，自己找当地素材成本很高。上传防晒精华产品图，AI 自动生成海岛场景图、成分卖点图、质地展示图和使用场景图，视觉风格贴合当地审美，适合美妆个护出海卖家批量产出多站点素材。", "categories": [ "creation" ], "tags": [ "商品套图", "防晒精华", "东南亚市场", "美妆个护" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "防晒精华电商套图｜场景图+卖点图组合" } ], "prompt": "[商品套图]帮我生成电商商品套图。\n商品图片：（请上传：商品图片）\n适配平台：Shopee\n图片比例：3:4\n图内文字语言：英语\n目标市场：东南亚\n商品卖点&要求：水感防晒精华，SPF50+ PA++++，含积雪草成分，质地清爽易吸收，画面希望呈现海岛度假场景与椰子、太阳镜等生活道具搭配，适合东南亚市场美妆个护类目\n交付物仅套图，不要 Listing 文案 / 商品文案。\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-listing-images", "name": "商品套图", "description": "上传产品图，生成全套电商主图", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TqlFgWZviW1K" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-logo-design-handwritten-nature-wordmark", "title": "手写风茶饮品牌Logo", "title_en": "Handwritten Tea Brand Logo", "subtitle": "毛笔手写字体+大地色，传递松弛自然感", "subtitle_en": "A brush-lettered wordmark in an earthy tone", "description": "深棕色背景上以毛笔手写笔触写出品牌名，笔画粗细自然、留白充分，下方点缀简约拼音小字。整体不含图形装饰，靠字体本身传递松弛、自然、去精致化的生活气质，适合茶饮、香氛等强调自然松弛调性的品牌，用一个纯文字标志建立独特记忆点。", "categories": [ "creation" ], "tags": [ "Logo设计", "纯文字标志", "茶饮品牌", "手写字体", "自然风格" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "品牌Logo方案｜纯文字手写体" } ], "prompt": "[Logo设计]帮我设计品牌Logo或图标。\n品牌/图标名称：山野\n所属行业/应用场景：茶饮/自然生活方式\nLogo形式：只有文字\nLogo风格：手写艺术\n参考图（可选）：（请上传：参考图（可选））\n参考图默认规则：如果其他要求中未说明参考维度，参考整体风格、Logo类型、字体设计和笔触细节，根据品牌替换具体的图形和配色。\n偏好色系：深棕底+米白字\n其他要求：毛笔手写笔触中文字体，无图形装饰，下方点缀极简拼音小字，大面积留白\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-logo-design", "name": "Logo设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298847399936" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-logo-design-organic-skincare-mark", "title": "自然有机风格Logo", "title_en": "Organic Brand Logo", "subtitle": "水滴叶片图形+衬线字体，塑造自然高端品牌感", "subtitle_en": "A leaf-and-droplet mark paired with serif type", "description": "墨绿背景搭配水滴与叶片抽象图形，衬线字体拼出品牌名，整体透出自然、克制又高级的气质。适合护肤、香氛等强调天然成分和松弛感的品牌，帮你在同类竞品中用一枚安静有辨识度的标志立住调性。", "categories": [ "creation" ], "tags": [ "Logo设计", "品牌标志", "护肤品牌", "自然风格", "图形+字体" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "品牌Logo方案｜图形+字体组合" } ], "prompt": "[Logo设计]帮我设计品牌Logo或图标。\n品牌/图标名称：Miraé\n所属行业/应用场景：美妆护肤\nLogo形式：图形+文字\nLogo风格：自然有机\n参考图（可选）：（请上传：参考图（可选））\n参考图默认规则：如果其他要求中未说明参考维度，参考整体风格、Logo类型、字体设计和笔触细节，根据品牌替换具体的图形和配色。\n偏好色系：墨绿色系，米白点缀\n其他要求：图形为水滴与叶片抽象组合，字体用优雅衬线体，整体传达自然、高端、宁静的气质\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-logo-design", "name": "Logo设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298847399936" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-logo-design-outdoor-badge-frontier", "title": "户外品牌复古徽章Logo", "title_en": "Vintage Outdoor Badge Logo", "subtitle": "圆形徽章嵌山峦独木舟图形，复古探索感拉满", "subtitle_en": "A circular badge with mountain and canoe icons", "description": "藏青色圆形徽章内嵌山峰与独木舟图形，环绕手写体品牌名与标语文字，做旧质感和对称版式传递复古探索气息。适合户外装备、露营用品等生活方式品牌打造经典耐用的视觉标志，一眼让人联想到崇尚自然、经久耐用的品牌调性。", "categories": [ "creation" ], "tags": [ "Logo设计", "徽章Logo", "户外品牌", "复古风格", "图形+文字" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "品牌Logo方案｜徽章图形+文字组合" } ], "prompt": "[Logo设计]帮我设计品牌Logo或图标。\n品牌/图标名称：Frontier Goods\n所属行业/应用场景：户外装备/露营生活方式\nLogo形式：图形+文字\nLogo风格：复古经典\n参考图（可选）：（请上传：参考图（可选））\n参考图默认规则：如果其他要求中未说明参考维度，参考整体风格、Logo类型、字体设计和笔触细节，根据品牌替换具体的图形和配色。\n偏好色系：藏青蓝+米白\n其他要求：圆形徽章版式，内含山峰与独木舟图形，环绕手写体品牌名与标语文字，做旧复古质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-logo-design", "name": "Logo设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298847399936" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-logo-design-sichuan-restaurant-seal", "title": "川菜小馆复古招牌Logo", "title_en": "Retro Sichuan Restaurant Logo", "subtitle": "赤陶配色+中式印章版式，中英双语餐饮招牌", "subtitle_en": "Terracotta seal-style mark for a Chinese eatery", "description": "赤陶色底配米白书法体大字，四周以圆角印章点缀“川味”“家常”等小字，中英文标题呼应，纸张肌理增添怀旧氛围。适合川菜馆、家常菜馆等中式餐饮品牌打造有辨识度的招牌视觉，兼顾传统韵味与现代排版的清晰易读。", "categories": [ "creation" ], "tags": [ "Logo设计", "餐饮招牌", "中式风格", "川菜品牌", "字体设计" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "品牌Logo方案｜中文书法+印章装饰" } ], "prompt": "[Logo设计]帮我设计品牌Logo或图标。\n品牌/图标名称：南巷小馆\n所属行业/应用场景：川菜餐饮/中式小馆\nLogo形式：图形+文字\nLogo风格：复古经典\n参考图（可选）：（请上传：参考图（可选））\n参考图默认规则：如果其他要求中未说明参考维度，参考整体风格、Logo类型、字体设计和笔触细节，根据品牌替换具体的图形和配色。\n偏好色系：赤陶棕+米白\n其他要求：赤陶色纸张肌理底，米白书法体大字为主标，四周环绕圆角印章小字“川味”“家常”，附中英文双语标注\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-logo-design", "name": "Logo设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298847399936" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-manual-brochure-design-handbook-cutpaper", "title": "剪纸风设计思维手册封面", "title_en": "Cut-Paper Style Design Handbook Cover", "subtitle": "大字剪纸拼贴风格，打造有趣的设计思维指南", "subtitle_en": "Playful cut-paper typography for a design thinking guide", "description": "设计团队想做一份有趣的内部手册，又不想太严肃刻板。这款工具能生成剪纸拼贴风格的设计手册封面，大号趣味字体配笑脸、星星等元素，活泼又有辨识度，适合设计思维指南、团队培训手册、创意工作坊资料等场景使用。", "categories": [ "creation" ], "tags": [ "手册/画册", "设计手册", "创意手册封面", "视觉指南" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "设计手册封面｜单张 JPG" } ], "prompt": "[手册/画册]设计多页手册或画册。\n成品类型：使用指南\n主题/内容：设计手册封面，中英标题'DESIGN 设计手册 HAND BOOK'，副标题'A VISUAL GUIDE TO DESIGN THINKING 视觉化设计思维指南'，标语'从想法到影响力 From Idea to Impact. Design with Purpose.'，粉色+墨绿色剪纸拼贴风格大号字体，配笑脸图标、星星、回形针等趣味元素，版本信息VERSION 1.0 2026\n参考文件：（请上传：参考文件（可选））\n版面规格：A4 竖版（正式手册）\n页数范围：自动匹配\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-manual-brochure", "name": "手册/画册", "description": "手册/画册/讲义多页设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6Bm0G8FBfW3M" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-manual-brochure-product-manual-teal", "title": "青绿撞色产品手册封面", "title_en": "Teal Geometric Product Manual Cover", "subtitle": "几何撞色箭头设计，展现科技产品说明书质感", "subtitle_en": "Geometric teal arrows for a tech product manual", "description": "科技产品说明书封面常常千篇一律，缺乏品牌辨识度。这款工具能生成青绿撞色几何风格的产品手册封面，箭头图形传递前进感，简约现代又不失专业，适合智能硬件、企业产品手册、技术文档等场景制作封面使用。", "categories": [ "creation" ], "tags": [ "手册/画册", "产品手册", "科技风手册封面", "企业宣传册" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "产品手册封面｜单张 JPG" } ], "prompt": "[手册/画册]设计多页手册或画册。\n成品类型：产品手册\n主题/内容：产品手册封面，中英标题'PRODUCT MANUAL 产品手册'，标语'连接未来·智造不凡 Connecting the Future, Creating Excellence'，底部标语'SMART SOLUTIONS BETTER TOGETHER'，青绿色背景配深藏蓝几何箭头与黄绿色条纹装饰，简约科技感撞色设计\n参考文件：（请上传：参考文件（可选））\n版面规格：A4 竖版（正式手册）\n页数范围：自动匹配\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-manual-brochure", "name": "手册/画册", "description": "手册/画册/讲义多页设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6Bm0G8FBfW3M" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-manual-brochure-repair-manual-wrench", "title": "撞色扳手维修手册封面", "title_en": "Bold Wrench Repair Manual Cover", "subtitle": "抽象扳手图形撞色设计，塑造专业维修工具书感", "subtitle_en": "Abstract wrench graphic for a professional repair manual", "description": "维修手册封面常常直接用产品图拼凑，缺少设计感。这款工具能生成撞色几何风格的维修手册封面，抽象扳手图形咬合构成主视觉，专业感与识别度兼具，适合设备说明书、维修保养手册、售后服务文档等场景制作封面。", "categories": [ "creation" ], "tags": [ "手册/画册", "维修手册", "说明书封面", "工业风手册" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "维修手册封面｜单张 JPG" } ], "prompt": "[手册/画册]设计多页手册或画册。\n成品类型：说明书\n主题/内容：维修手册封面，标题'维修手册'，副标题'使用·维护·保养'，右下角标语'专业维护 持久可靠'与'安全·高效·耐用'，深藏蓝与橙红两把抽象扳手图形咬合构成主视觉，米白背景，撞色几何风格\n参考文件：（请上传：参考文件（可选））\n版面规格：A4 竖版（正式手册）\n页数范围：自动匹配\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-manual-brochure", "name": "手册/画册", "description": "手册/画册/讲义多页设计", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6Bm0G8FBfW3M" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-marketing-poster-plant-oil-launch", "title": "植萃精华油新品海报", "title_en": "Plant Essence Oil Launch Poster", "subtitle": "米杏静物打底，三点卖点收尾", "subtitle_en": "Warm beige still life closing with three benefit points", "description": "护肤新品上市，产品图拍好了却排不出高级感，堆满促销字反而拉低调性。这张海报用米杏色背景和天然石台衬托精华油瓶身，沙棘果与橄榄枝补足植萃联想，标题留白、卖点收在底部三栏。适合美妆护肤品牌做新品宣传物料或品牌调性传播。", "categories": [ "creation" ], "tags": [ "营销海报", "新品发布", "美妆护肤", "产品海报", "质感排版" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "植萃精华油新品海报｜竖版 9:16 单张" } ], "prompt": "[营销海报]设计营销海报。\n素材图片：一张精华油玻璃滴管瓶的产品图\n输入词：植萃润养精华油新品宣传海报，米杏色背景，整体安静高级；主标题「深润呵护 柔嫩焕亮肌肤」，产品名「植萃润养精华油 / PLANT EXTRACT NOURISHING ESSENCE OIL」，右上角小字「DEEP MOISTURIZING GENTLE CARE」；画面中央精华油瓶立在米色岩板圆台上，身后是一块天然毛面石材，左下搭配沙棘果串与橄榄枝叶；底部三点卖点并排：深层滋润·长效锁水，植萃修护·温和呵护，焕亮肤色·自然透亮\n海报类型：宣传物料\n海报尺寸：长图 9:16\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-marketing-poster", "name": "营销海报", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297552367616" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-marketing-poster-qixi-beauty-menu", "title": "七夕美业焕颜价目海报", "title_en": "Qixi Beauty Salon Promo Poster", "subtitle": "节日主题与三档护理价目一屏说清", "subtitle_en": "Festival theme with a three-tier service price menu", "description": "美容院做节日活动，最怕项目、价格、赠礼写成一堆流水账，客人看不下去。这张海报用紫色花瓣与水光质感铺底，三档护理各占一条卡片，价格、疗效、赠礼和预约按钮层层递进，扫一眼就知道花多少钱做什么。适合美容院、皮肤管理店、养生馆的节日促销。", "categories": [ "creation" ], "tags": [ "营销海报", "美业门店", "节日促销", "价目表", "到店预约" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "七夕美业促销海报｜竖版 9:16 单张" } ], "prompt": "[营销海报]设计营销海报。\n素材图片：（请上传：素材图片）\n输入词：七夕主题美业门店促销海报，紫色花瓣与水光质感背景，主标题「爱在七夕 焕亮心动时刻」，副标题「以美为礼·宠爱自己」，标签「七夕限定宠粉福利」；三档护理项目卡片：水光焕颜护理 880 元（深层补水·提亮肤色·水润透亮）、紧致提拉护理 1280 元（紧致轮廓·淡化细纹·弹润年轻）、奢宠修护护理 1680 元（舒缓修护·强韧屏障·焕活光采）；底部「七夕臻享礼」到店体验任一项目赠定制礼盒 1 份，配满 1500 元可用的 200 元七夕专享券；结尾行动按钮「立即预约 享七夕专属宠爱」，落款「宠爱自己是终身浪漫的开始」\n海报类型：营销活动\n海报尺寸：长图 9:16\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-marketing-poster", "name": "营销海报", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297552367616" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-material-to-apparel-denim-jacket-design", "title": "牛仔外套设计图", "title_en": "Denim Jacket Design", "subtitle": "牛仔面料秒变短款外套，水洗质感一次呈现", "subtitle_en": "Turn a denim swatch into a cropped jacket with model preview", "description": "上传一张牛仔面料图，AI自动设计成衣并生成模特展示图。示例效果为浅蓝水洗牛仔面料变成廓形短款牛仔外套，双胸袋与金属纽扣细节清晰可见，模特实拍效果直观呈现版型与做旧质感。适合服饰设计师和卖家快速验证面料成衣效果，缩短打样周期。", "categories": [ "creation" ], "tags": [ "材料到服装", "牛仔外套", "面料设计", "复古美式" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "牛仔外套设计图｜模特展示图JPG" } ], "prompt": "[材料到服装]用上传的面料设计成衣，并生成模特展示图。\n面料/材质图：（请上传：面料/材质图）\n服装类型：外套\n设计风格：复古美式牛仔\n其他要求：廓形短款牛仔外套，做旧水洗质感，金属纽扣装饰\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-material-to-apparel", "name": "面料转服装", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297204240384" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-material-to-apparel-printed-halter-set", "title": "印花套装设计图", "title_en": "Printed Co-ord Set Design", "subtitle": "几何印花面料秒变挂脖套装，度假风一步到位", "subtitle_en": "Turn a geometric print swatch into a matching halter set", "description": "上传一张印花面料图，AI自动设计成衣并生成模特展示图。示例效果为绿白撞色几何印花棉布变成挂脖上衣搭配阔腿裤套装，模特在海边礁石旁展示整体廓形与图案连贯性。适合度假风女装卖家快速验证套装设计效果，一张面料图即可获得完整成衣搭配预览。", "categories": [ "creation" ], "tags": [ "材料到服装", "印花套装", "几何印花", "海边度假" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "印花套装设计图｜模特展示图JPG" } ], "prompt": "[材料到服装]用上传的面料设计成衣，并生成模特展示图。\n面料/材质图：（请上传：面料/材质图）\n服装类型：套装\n设计风格：撞色几何印花\n其他要求：挂脖上衣搭配阔腿裤套装，绿白撞色几何图案，海边度假氛围\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-material-to-apparel", "name": "面料转服装", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297204240384" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-material-to-apparel-silk-slip-dress", "title": "丝绸面料连衣裙设计图", "title_en": "Silk Fabric Dress Design", "subtitle": "大理石纹丝绸秒变度假吊带裙，垂坠感拉满", "subtitle_en": "Turn a marbled silk swatch into a flowing slip dress", "description": "上传一张丝绸面料图，AI自动设计成衣并生成模特展示图。示例效果为米金色大理石纹丝绸变成深V系带露背吊带裙，模特在海边棕榈树下展示裙摆飘逸的垂坠质感。适合度假风女装卖家快速验证面料成衣效果，无需打样和实地取景即可获得高质感成片。", "categories": [ "creation" ], "tags": [ "材料到服装", "连衣裙设计", "丝绸面料", "度假风" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "丝绸连衣裙设计图｜模特展示图JPG" } ], "prompt": "[材料到服装]用上传的面料设计成衣，并生成模特展示图。\n面料/材质图：（请上传：面料/材质图）\n服装类型：连衣裙\n设计风格：法式度假风\n其他要求：深V系带露背设计，垂坠丝绸质感，裙摆自然飘逸\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-material-to-apparel", "name": "面料转服装", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297204240384" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-material-to-apparel-striped-shirt", "title": "条纹面料成衣设计图", "title_en": "Striped Fabric to Shirt", "subtitle": "一块面料秒变成衣上身效果", "subtitle_en": "Turn a fabric swatch into a finished outfit", "description": "上传面料图片，AI自动设计成衣款式并生成模特上身展示图，浅蓝色细条纹棉布秒变宽松法式衬衫，打结穿法搭配金色耳环项链。适合面料商、设计师和服装卖家在下单打样前预览成衣效果，减少试错成本。", "categories": [ "creation" ], "tags": [ "材料到服装", "面料设计", "衬衫", "成衣打样", "模特展示" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "面料成衣设计图｜单张 JPG" } ], "prompt": "[材料到服装]用上传的面料设计成衣，并生成模特展示图。\n面料/材质图：浅蓝色细白条纹棉布面料样版\n服装类型：上衣\n设计风格：法式慵懒衬衫，宽松版型\n其他要求：金色耳环项链点缀，衬衫下摆打结，奶油色背景\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-material-to-apparel", "name": "面料转服装", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297204240384" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-model-swap-blazer-model-change", "title": "西装模特一键换人", "title_en": "Swap the Blazer Model", "subtitle": "同款服装换模特，搭配场景一秒复用", "subtitle_en": "Keep the outfit, swap in a new model instantly", "description": "上传已有的模特图，AI保留服装、配饰与拍摄场景不变，仅替换模特形象，一张深棕西装外套配白背心的南亚模特图秒变欧美金发模特展示。适合电商卖家复用已有实拍图快速拓展多元模特库，无需重新组织拍摄。", "categories": [ "creation" ], "tags": [ "换模特", "模特替换", "西装外套", "电商复用", "多元模特" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "换模特西装展示图｜单张 JPG" } ], "prompt": "[换模特]将模特图中的模特替换为目标模特，保留服装不变。\n原模特图：一张南亚女性模特身穿深棕色西装外套、白色背心、米白阔腿裤，手提红色包的白底棚拍图\n目标模特参考：（请上传：目标模特参考）\n目标模特：欧美女性\n其他要求：保留西装外套、白色背心、米白阔腿裤和红色手提包等原搭配与白底棚拍风格不变\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-model-swap", "name": "换模特", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297335808000" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-model-swap-camel-wool-coat-model", "title": "驼色大衣换模特图", "title_en": "Camel Coat Model Swap", "subtitle": "大衣模特实拍效果快速切换，冬季新品预览更高效", "subtitle_en": "Switch models on a coat photo without a new studio shoot", "description": "上传一张大衣模特实拍图，AI保留大衣版型、系带造型与背景光线不变，将原模特替换为新的模特形象。示例图为驼色系带羊毛大衣搭配米白高领打底，模特肤色与妆容随之自然过渡，珍珠耳环等配饰保持一致。适合冬装卖家快速产出多版本预览图，缩短新品上架周期。", "categories": [ "creation" ], "tags": [ "换模特", "羊毛大衣", "冬装预览", "电商上新" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "驼色羊毛大衣换模特图｜单张JPG" } ], "prompt": "[换模特]将模特图中的模特替换为目标模特，保留服装不变。\n原模特图：一张黑人女性模特身穿驼色系带羊毛大衣搭配米白高领打底的实拍图\n目标模特参考：（请上传：目标模特参考）\n目标模特：欧美女性\n其他要求：保留大衣版型、腰带造型与米白背景光线，仅替换模特形象\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-model-swap", "name": "换模特", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297335808000" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-model-swap-purple-stripe-polo-beanie", "title": "撞纹POLO衫换模特图", "title_en": "Striped Polo Model Swap", "subtitle": "撞色条纹POLO一键换新模特，青春活力氛围拉满", "subtitle_en": "Refresh a striped polo photo with a brand-new model", "description": "上传一张POLO衫模特实拍图，AI保留毛线帽、串珠项链等配饰与灯光氛围不变，将原模特替换为指定风格的新模特。示例图为紫色条纹毛圈POLO衫搭配荧光绿毛线帽，从深色皮肤男性模特换成欧美男性模特，笑容与发型自然过渡。适合休闲男装卖家快速产出多元模特版本。", "categories": [ "creation" ], "tags": [ "换模特", "POLO衫", "青年穿搭", "街头风格" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "紫色条纹POLO换模特图｜单张JPG" } ], "prompt": "[换模特]将模特图中的模特替换为目标模特，保留服装不变。\n原模特图：一张深色皮肤男性模特戴荧光绿毛线帽、身穿紫色条纹毛圈POLO衫的实拍图\n目标模特参考：（请上传：目标模特参考）\n目标模特：欧美男性\n其他要求：保留毛线帽、串珠项链配饰与灯光氛围，仅替换模特笑容表情\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-model-swap", "name": "换模特", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297335808000" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-model-swap-red-ribbed-top-blonde", "title": "红色罗纹上衣换模特图", "title_en": "Red Ribbed Top Model Swap", "subtitle": "同款服装秒换模特形象，风格实拍级无缝衔接", "subtitle_en": "Swap models on the same outfit with seamless studio realism", "description": "上传一张模特实拍图，AI保留服装版型与棚拍光线不变，将原模特替换为指定风格的新模特。示例图为红色罗纹方领上衣搭配白色半裙，从亚洲女性模特换成欧美金发模特，妆容与配饰随之自然过渡。适合服饰卖家快速产出多元模特版本，无需重新拍摄，一图多用提升选购覆盖面。", "categories": [ "creation" ], "tags": [ "换模特", "针织上衣", "欧美模特", "电商上新" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "红色罗纹上衣换模特图｜单张JPG" } ], "prompt": "[换模特]将模特图中的模特替换为目标模特，保留服装不变。\n原模特图：一张亚洲女性模特身穿红色罗纹方领上衣搭配白色半裙的棚拍实拍图\n目标模特参考：（请上传：目标模特参考）\n目标模特：欧美女性\n其他要求：保留服装版型、姿态与白底棚拍光线不变，仅替换模特形象\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-model-swap", "name": "换模特", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297335808000" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-model-tryon-beach-swim-set", "title": "海滩泳装试穿图", "title_en": "Beach Swimwear Try-On", "subtitle": "平铺泳装秒变模特实拍度假大片", "subtitle_en": "Turn flat-lay swimwear into a resort photoshoot", "description": "上传泳装平铺图，AI自动生成模特上身实拍效果，沙滩棕榈树场景、自然光影与姿态一步到位，无需实地取景或真人试拍。适合泳装、度假服饰卖家快速产出电商及社媒展示图，大幅降低拍摄成本和周期。", "categories": [ "creation" ], "tags": [ "模特试穿", "泳装", "沙滩场景", "电商图", "度假风" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "沙滩场景模特试穿图｜单张 JPG" } ], "prompt": "[模特试穿]将上传的服饰进行模特试穿展示。\n服饰图片：一件酒红色吊带文胸配波浪针织短裤的泳装平铺图\n模特风格：欧美女性\n肤色偏好：小麦色\n拍摄场景：街拍户外\n图片比例：3:4\n生成张数： 张\n其他要求：沙滩棕榈树场景，模特叉腰站姿，墨镜配饰，自然光影\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-model-tryon", "name": "模特试穿", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297391538176" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-nail-art-wine-red-floral-coffee", "title": "酒红花卉美甲上手图", "title_en": "Wine Red Floral Nail Art", "subtitle": "美甲设计秒出上手实拍效果", "subtitle_en": "Preview your nail design on a real hand", "description": "上传美甲设计图，AI生成上手实拍效果图，酒红色亮面搭配裸色底金色藤蔓花卉贴花，手持咖啡杯的秋冬暖调咖啡馆场景。适合美甲店和美甲师快速产出新款宣传图，无需预约模特实拍即可展示上手效果。", "categories": [ "creation" ], "tags": [ "美甲上手", "酒红美甲", "花卉美甲", "咖啡馆场景", "秋冬美甲" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "美甲上手效果图｜单张 JPG" } ], "prompt": "[美甲上手]将上传的美甲设计展示上手效果。\n美甲设计图：酒红色亮面搭配裸色底金色藤蔓花卉贴花的美甲设计图，共5款指甲贴片\n手部风格：纤细白皙\n图片比例：3:4\n生成张数： 张\n其他要求：手持白色咖啡杯，珍珠戒指点缀，秋冬暖调咖啡馆场景\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-nail-art", "name": "美甲", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297345400832" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-brief-beginner-workout-plan", "title": "健身笔记策划方案", "title_en": "Workout Note Brief", "subtitle": "一句话主题拆成完整选题+排版策划案", "subtitle_en": "Turn one topic into a full content brief", "description": "有主题却不知道从何下笔？输入内容方向和目标平台，AI 输出完整策划案：表达语气、每页小标题、图内文案和画面任务都写清楚，团队对齐或后续出图都能直接照做，省去反复沟通的时间。", "categories": [ "creation" ], "tags": [ "图文简报", "内容策划", "健身笔记", "选题策划", "社媒排期" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "健身笔记策划案｜文档" } ], "prompt": "[图文策划]先完成选题判断，再策划社媒图文内容与排期。\n策划类型：单篇策划\n内容方向：新手女生如何安排一周锻炼\n参考资料：（请上传：参考资料）\n目标平台：小红书\n内容目标：收藏\n主要要求：语气亲切姐妹风，鼓励但不说教；产出选题痛点、一周训练分配表和每日动作指导，IP形象出镜降低运动门槛\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-brief", "name": "图文策划", "description": "把选题、策划和排期串成第一步方案", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_HBJ8QtVmsyjB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-brief-fire-savings-calculator-note", "title": "退休储蓄计算单图策划", "title_en": "Retirement Savings Calculator Note", "subtitle": "输入理财话题，AI 拆出计算逻辑和版面结构", "subtitle_en": "Turn a finance topic into a ready-to-draw layout", "description": "想做一篇讲清楚“存多少钱能退休”的科普图文，却理不清计算逻辑和版面重点？输入话题和目标读者，AI 输出完整策划：核心公式、分级对比数据、IP形象出镜方式和每个板块的图内文案，直接照着策划案出图，省去自己拆解知识点的时间。", "categories": [ "creation" ], "tags": [ "图文策划", "内容策划", "理财科普", "单图策划", "社媒理财" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "退休储蓄计算图文策划案｜文档" } ], "prompt": "[图文策划]先完成选题判断，再策划社媒图文内容与排期。\n策划类型：单图策划\n内容方向：月薪1万按4%法则退休需要存多少钱\n参考资料：（请上传：参考资料）\n目标平台：小红书\n内容目标：收藏\n主要要求：用固定IP动物形象出镜降低理解门槛，黑板手绘风讲解4%法则公式，右侧按不同月消费水平做三档对比卡片，结尾附计算过程截图增强可信度\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-brief", "name": "图文策划", "description": "把选题、策划和排期串成第一步方案", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_HBJ8QtVmsyjB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-brief-workout-plan-pitch-brief", "title": "健身周计划选题策划", "title_en": "Workout Plan Pitch Brief", "subtitle": "从新手痛点切入，拆出选题角度和图文结构", "subtitle_en": "Turn user pain points into a note structure", "description": "已经知道要做健身周计划这个选题，却不知道怎么切入才不像说教？输入话题方向，AI 从新手常见痛点出发，拆出开篇痛点、场景写真镜头、结尾行动号召的完整选题策划，配上每页拍摄建议，比直接甩训练表更有代入感，帮你提升完播和收藏。", "categories": [ "creation" ], "tags": [ "图文策划", "内容策划", "健身笔记", "选题角度", "社媒健身" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "健身周计划选题策划案｜文档" } ], "prompt": "[图文策划]先完成选题判断，再策划社媒图文内容与排期。\n策划类型：选题+策划\n内容方向：新手为什么需要一份健身周计划\n参考资料：（请上传：参考资料）\n目标平台：小红书\n内容目标：认知\n主要要求：从三个新手常见痛点切入（练同一部位、三天打鱼两天晒网、跟练视频随便选），用居家场景真人写真镜头呈现，结尾落到科学分配训练日的行动号召\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-brief", "name": "图文策划", "description": "把选题、策划和排期串成第一步方案", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_HBJ8QtVmsyjB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-cover-office-worker-meme-cover", "title": "打工人梗图快乐封面", "title_en": "Office Worker Meme Cover", "subtitle": "纯文字大字报封面，打工人共鸣感拉满", "subtitle_en": "A bold text-only cover office workers love to share", "description": "想发条职场吐槽笔记，又懒得配图排版？上传一句梗文案，AI自动生成纯文字大字报封面：紫色撞色背景配夸张手写体标题，感叹号与对话气泡点缀，视觉冲击力拉满。适合职场博主和上班族发日常吐槽、玩梗互动，不用配图靠标题就能抓住眼球。", "categories": [ "creation" ], "tags": [ "图文封面", "职场梗图", "文字封面", "职场吐槽", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "打工人梗图封面｜单张" } ], "prompt": "[图文封面]生成高点击率吸睛封面。\n素材图片：（请上传：素材图片）\n封面主题：打工人梗图续命：上班很累，玩梗不累\n封面类型：爆款封面\n封面风格：大字标题\n目标平台：小红书\n画面比例：3:4\n文案&风格要求：纯文字手写体大字报风格，紫色背景配黄色感叹号与粉色对话气泡装饰，不摆拍不配图，靠标题排版和撞色抓眼球\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-cover", "name": "图文封面", "description": "一键生成高点击率吸睛封面", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_A6ZHotQI9JLZ" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-cover-pet-care-infographic", "title": "宠物护理科普封面", "title_en": "Pet Care Infographic Cover", "subtitle": "编号信息卡+萌宠实拍，科普笔记一眼抓重点", "subtitle_en": "Numbered info cards with a cute pet photo finish", "description": "想发宠物养护知识科普，又不知道怎么排版才够吸睛？上传护理要点文字，AI自动生成编号信息卡封面：蓝色系图标配简洁分点讲解，结尾配萌宠实拍收尾更有亲和力。适合宠物博主和养宠账号快速产出干货笔记，专业感和点击率兼顾，无需自己抠图排版。", "categories": [ "creation" ], "tags": [ "图文封面", "宠物科普", "养宠知识", "信息图", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "宠物护理科普封面｜单张" } ], "prompt": "[图文封面]生成高点击率吸睛封面。\n素材图片：（请上传：素材图片）\n封面主题：宠物日常护理小知识：科学养宠避坑指南\n封面类型：爆款封面\n封面风格：科普图解\n目标平台：小红书\n画面比例：3:4\n文案&风格要求：蓝色系信息卡片排版，图标配编号分点讲解三个护理要点，结尾配萌宠实拍收尾增加亲和力\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-cover", "name": "图文封面", "description": "一键生成高点击率吸睛封面", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_A6ZHotQI9JLZ" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-cover-umeshu-supermarket-haul", "title": "梅酒种草超市实拍封面", "title_en": "Umeshu Haul Supermarket Cover", "subtitle": "货架实拍+手写批注，种草笔记信任感拉满", "subtitle_en": "In-store shelf shot with handwritten notes", "description": "逛超市看到心动好物，随手一拍却不知道怎么变成种草笔记？上传商品实拍照，AI自动生成真实种草封面：手写批注圈出卖点和价格，还原货架实拍的可信感与氛围感。适合美食饮品博主和好物分享达人快速产出种草笔记，出图即可发布不用二次排版。", "categories": [ "creation" ], "tags": [ "图文封面", "好物种草", "饮品测评", "超市实拍", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "梅酒种草封面｜单张" } ], "prompt": "[图文封面]生成高点击率吸睛封面。\n素材图片：（请上传：素材图片）\n封面主题：微醺梅酒种草：超市随手拍变安利笔记\n封面类型：常规封面\n封面风格：真实种草\n目标平台：小红书\n画面比例：3:4\n文案&风格要求：手持实拍真实购物场景，标注价格与货架细节增加可信度，多组手写批注圈出卖点和适饮方式\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-cover", "name": "图文封面", "description": "一键生成高点击率吸睛封面", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_A6ZHotQI9JLZ" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-note-cover-xiamen-travel-scrapbook", "title": "厦门旅拍手账封面", "title_en": "Xiamen Travel Cover", "subtitle": "旅行照片秒变社媒爆款封面，出片效率翻倍", "subtitle_en": "Turn travel photos into a viral note cover", "description": "旅行博主拍了一堆照片却不知道怎么做成吸睛封面？上传照片和地标素材，一键生成撕纸拼贴风封面：大字标题、清单打卡点、氛围感文案配齐，出图即可发布，告别自己抠图排版的麻烦。", "categories": [ "creation" ], "tags": [ "图文封面", "旅行手账", "社媒封面", "旅拍vlog", "手帐拼贴" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "厦门旅拍手账封面｜单张" } ], "prompt": "[图文封面]生成高点击率吸睛封面。\n素材图片：一张厦门旅拍人像自拍照，加几张鼓浪屿灯塔等地标照片\n封面主题：厦门旅行vlog封面：我在厦门压过的马路\n封面类型：爆款封面\n封面风格：手帐拼贴\n目标平台：小红书\n画面比例：3:4\n文案&风格要求：标题用撕纸拼贴色块呈现，人物出镜配合拍照道具与地标建筑图，搭配旅行清单打卡点和氛围感文案，营造真实旅拍手账感\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-note-cover", "name": "图文封面", "description": "一键生成高点击率吸睛封面", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_A6ZHotQI9JLZ" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-office-material-bakery-business-card", "title": "暖色烘焙工作室名片", "title_en": "Warm Bakery Studio Business Card", "subtitle": "手绘面包插画双面名片，传递手工温度感", "subtitle_en": "Double-sided card with hand-drawn bakery illustrations", "description": "手工烘焙工作室常缺一张有辨识度的名片，普通模板又太商务生硬。这款工具能生成双面手绘插画名片，暖棕色调面包插图配手写体标语，传递手工温度感，适合烘焙工作室、甜品店、咖啡馆等小店老板制作专属名片使用。", "categories": [ "creation" ], "tags": [ "办公物料", "名片设计", "烘焙店名片", "手绘插画名片" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "烘焙工作室名片（双面）｜单张 JPG" } ], "prompt": "[办公物料]设计办公物料。\n物料类型：名片\n物料内容：暖炉烘焙工作室名片，正面标题'暖炉烘焙工作室'，标语'用心烘焙，传递温暖与美好'，烘焙师林小满，电话13812345678，地址幸福路32号二楼，配手绘面包插画；背面标语'手工·新鲜·用心'，配面包出炉插画与'每一份用心，都是温暖的味道'文案，米白底+橄榄绿+棕色手绘风格\n图片比例：横版（4:3）\n图片素材：（请上传：图片素材（可选））\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-office-material", "name": "办公物料", "description": "名片/工牌/桌牌/台卡/证书/聘书", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_K0DKrb14DlDk" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-office-material-honor-certificate-blue-gold", "title": "蓝金欧式荣誉证书", "title_en": "Blue & Gold Honor Certificate", "subtitle": "一句话生成正式荣誉证书，奖项信息自动排版工整", "subtitle_en": "Generate a formal honor certificate from a single description", "description": "学校、企业颁奖时常缺一份像样的证书模板，手动排版又费时又不专业。这款工具只需填入获奖人、奖项和颁发信息，就能生成带欧式花纹边框、金色徽章印章的正式荣誉证书，适合学校活动、企业表彰、竞赛颁奖等场景，打印或电子发放都合适。", "categories": [ "creation" ], "tags": [ "办公物料", "荣誉证书", "颁奖证书", "表彰活动" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "荣誉证书｜单张 JPG" } ], "prompt": "[办公物料]设计办公物料。\n物料类型：荣誉证书\n物料内容：荣誉证书，标题'荣誉证书'，兹授予李明轩，在'探索与创新'全国中学生科学实践活动中获得优秀指导奖，特发此证以资鼓励，落款组委会主任张雅宁，颁发单位全国中学生科学实践活动组委会，日期2026年6月20日；风格采用深蓝金色欧式花纹边框，竖版排版\n图片比例：竖版（3:4）\n图片素材：（请上传：图片素材（可选））\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-office-material", "name": "办公物料", "description": "名片/工牌/桌牌/台卡/证书/聘书", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_K0DKrb14DlDk" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-office-material-red-team-pass-badge", "title": "红色创意部团队工牌", "title_en": "Red Creative Team Pass Badge", "subtitle": "杂志感黑白证件照工牌，凸显团队与职位信息", "subtitle_en": "Editorial-style badge highlighting team and role", "description": "团队入职、活动通行常需要一张有设计感的工牌，纯文字模板显得单调。这款工具能生成杂志感十足的团队通行证工牌，姓名、职位、部门、工号一次排版整齐，黑白证件照配红色点缀更显专业，适合创意团队、活动通行证等场景使用。", "categories": [ "creation" ], "tags": [ "办公物料", "工牌设计", "员工胸牌", "团队通行证" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "团队通行证工牌｜单张 JPG" } ], "prompt": "[办公物料]设计办公物料。\n物料类型：工牌/胸牌\n物料内容：工牌名称TEAM PASS，创意部门标识，姓名林小雨/Lin Xiaoyu，职位设计师/Designer，部门创意部/Creative Dept.，工号A-742，入职时间2026.03，配黑白证件照与红色挂绳，底部标语'让想法发生 Make Ideas Happen'，黑白杂志感+红色点缀风格\n图片比例：竖版（3:4）\n图片素材：（请上传：图片素材（可选））\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-office-material", "name": "办公物料", "description": "名片/工牌/桌牌/台卡/证书/聘书", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_K0DKrb14DlDk" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-office-material-wedding-table-card", "title": "绿色波浪边婚礼桌牌", "title_en": "Green Wavy Border Wedding Table Card", "subtitle": "波浪花边席位卡，标注桌号传递婚礼心意", "subtitle_en": "Wavy-border seating card marking table numbers", "description": "婚礼现场常需要成套的桌号卡引导宾客入座，手工制作费时又难统一风格。这款工具能生成波浪花边的婚礼桌牌，桌号、感谢语一次排版整齐，绿色橄榄枝装饰清新雅致，适合婚礼、宴会等场合批量制作席位卡使用。", "categories": [ "creation" ], "tags": [ "办公物料", "婚礼桌牌", "席位卡设计", "桌号卡" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "婚礼桌牌｜单张 JPG" } ], "prompt": "[办公物料]设计办公物料。\n物料类型：桌牌/席位牌\n物料内容：婚礼桌号卡，标题'YOU ARE SEATED AT Table'，桌号05，中文'您的席位'，感谢语'Thank you for being here to share our special day. 感谢您来见证我们的特别时刻。'，绿色波浪边框+橄榄枝装饰+爱心图标，米白底墨绿色文字\n图片比例：竖版（3:4）\n图片素材：（请上传：图片素材（可选））\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-office-material", "name": "办公物料", "description": "名片/工牌/桌牌/台卡/证书/聘书", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_K0DKrb14DlDk" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-packaging-design-botanical-hair-oil-box", "title": "橄榄绿护发精油礼盒", "title_en": "Olive Green Hair Oil Box", "subtitle": "哑光玻璃瓶+牛皮纸外盒，高级护发油包装", "subtitle_en": "A matte bottle paired with a kraft outer box", "description": "橄榄绿哑光玻璃瓶搭配深灰色泵头，外盒采用米灰色卡纸压印品牌名与产品信息，整体在绿植光影中呈现克制高级的质感。适合护发精油、精华类美妆产品打造疗愈自然又不失质感的包装形象，传递滋养、天然的产品定位。", "categories": [ "creation" ], "tags": [ "包装设计", "美妆包装", "护发精油", "极简高级", "瓶身+外盒" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "包装设计方案｜瓶身+外盒立体图" } ], "prompt": "[包装设计]帮我做包装设计方案。\n品牌/产品素材：（请上传：品牌/产品素材）\n品牌/产品名称：LUSTRE\n行业/产品品类：美妆个护/护发精油\n包装类型：盒装\n交付形式：立体方案图\n风格方向：极简高级、自然疗愈\n包装文案：品名 LUSTRE Botanical Hair Oil；卖点 滋养·强韧·顺滑；规格 80ml\n其他要求：橄榄绿哑光玻璃瓶配深灰泵头，米灰卡纸外盒烫印文字，绿植光影氛围实拍\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-packaging-design", "name": "包装设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297671352320" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-packaging-design-hibiscus-vinegar-sachet", "title": "洛神花醋饮礼盒包装", "title_en": "Hibiscus Vinegar Drink Box", "subtitle": "枣红底抽象花形+独立小袋，高级健康饮品调性", "subtitle_en": "Deep red box with abstract floral graphics", "description": "枣红色外盒印着大面积米白抽象花朵图形，搭配衬线字体品牌名，内含独立铝箔小袋分装的花果醋饮，编辑感排版透出高级与克制。适合果醋、代餐饮品等健康饮品品牌打造区别于传统包装的精致形象，强化低卡健康的产品心智。", "categories": [ "creation" ], "tags": [ "包装设计", "饮品包装", "健康饮品", "礼盒套装", "极简高级" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "包装设计方案｜礼盒+独立小袋立体图" } ], "prompt": "[包装设计]帮我做包装设计方案。\n品牌/产品素材：（请上传：品牌/产品素材）\n品牌/产品名称：H.\n行业/产品品类：饮品/花果醋饮\n包装类型：礼盒套装\n交付形式：立体方案图\n风格方向：高级极简、编辑感\n包装文案：品名 Hibiscus Vinegar Drink；卖点 清爽解腻、低卡无添加；规格 330ml×6袋\n其他要求：枣红色外盒印米白抽象花朵图形，内含独立铝箔分装小袋，衬线字体品牌名\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-packaging-design", "name": "包装设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297671352320" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-packaging-design-ice-cream-bear-tubs", "title": "北极熊冰淇淋桶包装", "title_en": "Polar Bear Ice Cream Tubs", "subtitle": "天蓝配色+微笑瓶盖，多口味冰淇淋治愈系设计", "subtitle_en": "Sky blue tubs with a smiling lid and bear mascot", "description": "天蓝色圆桶配同色系微笑瓶盖，正面印着抱着冰淇淋筒的白熊插画，圆角卡通字体标注香草云朵、莓果奶油等口味名称。适合冰淇淋、甜品类食品品牌打造可爱治愈、适合拍照分享的包装形象，多口味统一识别又各具辨识度。", "categories": [ "creation" ], "tags": [ "包装设计", "食品包装", "冰淇淋", "可爱插画", "系列套装" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "包装设计方案｜多口味冰淇淋桶效果图" } ], "prompt": "[包装设计]帮我做包装设计方案。\n品牌/产品素材：（请上传：品牌/产品素材）\n品牌/产品名称：Frosted Tale\n行业/产品品类：食品/冰淇淋\n包装类型：盒装\n交付形式：系列套装提案\n风格方向：可爱治愈、Ins风\n包装文案：品名 Frosted Tale Sweet Ice Cream；系列 Vanilla Cloud香草云朵/Berry Cream莓果奶油/Milk Cookie牛奶曲奇；规格 ONE PINT(473ml)\n其他要求：天蓝色圆桶配微笑瓶盖，正面印抱冰淇淋筒的白熊插画，圆角卡通字体\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-packaging-design", "name": "包装设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297671352320" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-packaging-design-velora-tea-series", "title": "高端茶叶礼盒包装", "title_en": "Premium Tea Gift Box Series", "subtitle": "三款口味统一VI，极简高端茶叶包装系列", "subtitle_en": "One visual system across three tea flavors", "description": "VELORA三款茶叶礼盒统一极简字标与莫兰迪色系，月亮、玫瑰、山峦图形分别对应乌龙、红茶、白茶，克制留白中透出东方高级感。适合中高端茶叶、香氛类品牌打造成套上市的礼盒视觉，兼顾货架辨识度与送礼质感。", "categories": [ "creation" ], "tags": [ "包装设计", "茶叶包装", "礼盒设计", "系列包装", "高端品牌" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "茶叶包装系列｜3款效果图" } ], "prompt": "[包装设计]帮我做包装设计方案。\n品牌/产品素材：（请上传：品牌/产品素材）\n品牌/产品名称：VELORA\n行业/产品品类：茶叶\n包装类型：礼盒套装\n交付形式：系列套装提案\n风格方向：极简高端，东方留白美学\n包装文案：Midnight Oolong 深焙沉稳/武夷山；Rose Black Tea 花香顺滑/云南；Mountain White Tea 清新细腻/福鼎，整叶茶，净含量按规格标注\n其他要求：三款包装统一极简大字标+主视觉图形（月亮、玫瑰、山峦），莫兰迪色系区分口味\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-packaging-design", "name": "包装设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297671352320" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-paid-knowledge-graphic-probation-period-rights", "title": "试用期避坑知识图文", "title_en": "Probation Period Rights Graphic", "subtitle": "三个法律要点，戳破试用期常见认知误区", "subtitle_en": "Bust common myths about probation periods", "description": "应届生和新入职员工常常不清楚试用期到底有哪些权益，容易被“转正再补”之类的说法蒙混？输入知识主题，AI 拆解成清晰的知识科普图文：社保参保时点、工资比例下限、约定次数限制逐条讲明白，浏览器截图式排版专业耐看，适合HR科普号做入门科普和涨粉。", "categories": [ "creation" ], "tags": [ "知识科普图文", "试用期科普", "劳动法科普", "HR知识", "社媒科普" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "试用期避坑科普图文｜单张" } ], "prompt": "[知识科普图文]根据上传资料或文案生成专业、易懂、适合社媒传播的知识科普图文。\n资料/素材：（请上传：资料/素材）\n知识主题：试用期不是白嫖期，三个法律要点要知道\n知识领域：HR\n目标人群：应届生、新入职员工\n目标平台：小红书\n科普重点/转化目标：试用期也要交社保、试用期工资不低于正式薪资80%、同一公司只能约定一次试用期，粉碎试用期没有保障的错误认知，建议收藏防坑\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-paid-knowledge-graphic", "name": "知识科普图文", "description": "上传资料或文案，生成专业科普图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gVBlRERQMRKN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-paid-knowledge-graphic-severance-pay-guide", "title": "离职补偿金计算科普", "title_en": "Severance Pay Calculation Guide", "subtitle": "N、N+1、2N怎么算，四种离职场景讲清楚", "subtitle_en": "Explain N, N+1 and 2N severance rules clearly", "description": "离职时经常搞不清能不能拿补偿、能拿多少？输入知识主题，AI 按公司辞退、未提前通知、违法解除、自己提离职四种场景，把经济补偿金的计算方式逐条拆解成清晰的知识图文，适合HR科普、法律咨询类账号做涨粉和转化咨询。", "categories": [ "creation" ], "tags": [ "知识科普图文", "离职补偿金", "劳动法科普", "HR知识", "社媒科普" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "离职补偿金科普图文｜单张" } ], "prompt": "[知识科普图文]根据上传资料或文案生成专业、易懂、适合社媒传播的知识科普图文。\n资料/素材：（请上传：资料/素材）\n知识主题：离职也能拿补偿金，N、N+1、2N怎么算\n知识领域：HR\n目标人群：面临离职或被辞退的职场人\n目标平台：小红书\n科普重点/转化目标：按公司辞退、未提前30天通知、违法解除合同、自己提离职四种场景分别说明补偿标准，帮读者对照自身情况判断能否维权\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-paid-knowledge-graphic", "name": "知识科普图文", "description": "上传资料或文案，生成专业科普图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gVBlRERQMRKN" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-past-style-note-fire-diary-awakening-stage", "title": "松鼠IP理财日记复刻·觉醒篇", "title_en": "Squirrel Diary Recreation - Awakening Stage", "subtitle": "沿用同一只IP松鼠日记画风，续写理财觉醒期新内容", "subtitle_en": "Match your IP diary style for a new money-mindset post", "description": "已经用固定IP形象做了几期理财日记，新话题却怕画出来不像同一账号？上传往期的松鼠IP插画日记作参考，AI 沿用同样的角色形象、米黄配色和圆角卡片排版，为“第一次意识到不想打一辈子工”这个新话题续写一篇风格统一的日记，读者一眼认出是同个账号更新，适合IP人设类账号做系列连载。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "往期同风格", "IP人设日记", "理财日记", "图文系列" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "松鼠IP理财日记图文｜单张海报" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：账号往期发布的同款棕色卡通小松鼠IP插画日记，米黄色底，带阶段徽章角标和IP签名角标\n你的素材图：（请上传：你的素材图）\n新图文主题：第一次意识到我不想打一辈子工\n复刻模式：往期同风格\n画面比例：3:4\n文案&其他要求：延续棕色松鼠IP形象、米黄底色和圆角描边卡片，保留左上角阶段徽章、手写体大标题和右下角IP签名角标\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-past-style-note", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uULbwmGhWsCr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-past-style-note-fire-diary-learning-stage", "title": "松鼠IP理财日记复刻·认知篇", "title_en": "Squirrel Diary Recreation - Learning Stage", "subtitle": "同款IP日记画风，续写FIRE入门知识新内容", "subtitle_en": "Continue the IP diary style for a FIRE basics post", "description": "系列日记更新到第二期，想讲清楚4%法则这类术语，又怕画风和上一期对不上？上传往期同账号的IP插画日记，AI 保留同一只松鼠形象、道具摆盘方式和配色，只替换知识点和场景，产出一篇风格延续、内容升级的新日记，省去每期重新定风格的功夫。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "往期同风格", "IP人设日记", "理财知识", "图文系列" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "松鼠IP理财日记图文｜单张海报" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：账号往期发布的同款棕色卡通小松鼠IP插画日记，米黄色底，带阶段徽章角标和IP签名角标\n你的素材图：（请上传：你的素材图）\n新图文主题：开始搞懂FIRE到底是什么，4%法则和被动收入入门\n复刻模式：往期同风格\n画面比例：3:4\n文案&其他要求：延续棕色松鼠IP形象、书本道具摆盘和米黄底色，保留左上角阶段徽章和右下角IP签名角标，知识点以书本/卷轴插画形式呈现\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-past-style-note", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uULbwmGhWsCr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-past-style-note-fire-diary-savings-stage", "title": "松鼠IP理财日记复刻·储蓄篇", "title_en": "Squirrel Diary Recreation - Savings Stage", "subtitle": "同款IP日记画风，续写第一笔储蓄新内容", "subtitle_en": "Continue the IP diary style for a first-savings post", "description": "理财日记连载到第三期，想聊“迈出第一笔储蓄”这个话题，又想保持和前两期一样的手绘质感？上传往期同账号的IP插画日记作参考，AI 延续同一只松鼠形象、黑板手绘元素和米黄配色，为新话题重新构图，产出风格统一的第三期，账号更新节奏和视觉辨识度都能稳住。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "往期同风格", "IP人设日记", "理财日记", "图文系列" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "松鼠IP理财日记图文｜单张海报" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：账号往期发布的同款棕色卡通小松鼠IP插画日记，米黄色底，带阶段徽章角标和IP签名角标\n你的素材图：（请上传：你的素材图）\n新图文主题：算清你的FIRE数字，开始第一笔储蓄\n复刻模式：往期同风格\n画面比例：3:4\n文案&其他要求：延续棕色松鼠IP形象、黑板手绘板书元素和米黄底色，保留左上角阶段徽章和右下角IP签名角标，加入储蓄罐/松果罐等标志性道具\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-past-style-note", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uULbwmGhWsCr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-past-style-note-illustrated-diary-recreate", "title": "水彩手绘风日记图文复刻", "title_en": "Illustrated Diary Note Recreation", "subtitle": "沿用往期水彩手绘日记风格，复刻新一期养生打卡笔记", "subtitle_en": "Match your past watercolor diary style for a new note", "description": "已经有一组水彩手绘风的日记体图文，却不知道新话题怎么保持同一种画风？上传往期笔记作参考，AI 沿用同样的插画笔触、配色和手写日签排版，为新的养生打卡主题续写一篇风格统一的图文，读者一眼就能认出是同一账号更新，适合生活记录类账号做系列内容。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "往期同风格", "水彩插画笔记", "养生日记", "图文系列" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "水彩手绘养生日记图文｜单张海报" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：往期同账号发布的水彩手绘风三伏天日记笔记，绿色系插画风格\n你的素材图：（请上传：你的素材图）\n新图文主题：三伏天最后一天的身心变化记录\n复刻模式：往期同风格\n画面比例：3:4\n文案&其他要求：延续水彩手绘插画画风，浅绿+米白暖色调，保留手写日签角标（Day+主题）、对话气泡引语、爱心分隔线等元素\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-past-style-note", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_uULbwmGhWsCr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-personal-media-note-classical-poetry-note", "title": "古诗词意境图文笔记", "title_en": "Classical Poetry Aesthetic Note", "subtitle": "一句诗词配水墨意境，适合读书笔记号", "subtitle_en": "Pair a poem with a serene ink-wash visual", "description": "想做古诗词赏析类图文，却找不到配得上诗句意境的画面？输入诗词内容，AI 自动搭配水墨风明月、玉兔和花枝插画，用留白排版突出诗句本身，产出适合读书博主、文化类账号发布的意境图文笔记，无需自己找图配色。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "古诗词赏析", "读书博主", "国风意境", "社媒读书" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "古诗词意境图文笔记｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：个人自媒体\n账号/行业：读书笔记/古诗词赏析账号\n行业/账号类型：读书博主\n主题/文案：苏轼《水调歌头·明月几时有》\n目标平台：小红书\n内容要求：水墨风灰蓝色调，配明月、玉兔、花枝插画元素，诗词原文竖版留白排版，落款标注作者和创作背景\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-personal-media-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_U30TPjmwCAvr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-personal-media-note-office-worker-vent-meme", "title": "打工人解压吐槽图文", "title_en": "Office Worker Stress Meme Note", "subtitle": "手绘吐槽风格，说出打工人的崩溃瞬间", "subtitle_en": "A hand-drawn meme that captures office burnout", "description": "想在个人号发一篇有共鸣的吐槽内容，却不会画简笔画也不想露脸？输入吐槽主题，AI 自动生成黑白线稿风的表情包式插画，配夸张的哭泣特效和加班、开会等场景标签，产出适合职场博主发布的解压吐槽图文，转发到评论区都能引发打工人共鸣。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "职场吐槽", "解压meme", "打工人共鸣", "社媒职场" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "打工人解压吐槽图文｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：个人自媒体\n账号/行业：职场吐槽/情绪共鸣类账号\n行业/账号类型：其他\n主题/文案：打工人哪有不疯的，硬撑罢了\n目标平台：小红书\n内容要求：黑白手绘简笔画风格，蓝色泪水特效突出崩溃感，周围贴加班、改方案、开会等场景标签，营造夸张又真实的共鸣感\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-personal-media-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_U30TPjmwCAvr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-personal-media-note-weekend-picnic-diary", "title": "周末慢生活图文笔记", "title_en": "Weekend Picnic Diary", "subtitle": "日常随手拍变身高质感生活方式笔记", "subtitle_en": "Turn everyday photos into a lifestyle note", "description": "拍了一堆生活碎片照片，却拼不出有氛围感的图文？上传照片和主题，AI 自动完成拼贴排版、手写标题和治愈系文案，产出适合个人IP日常更新的慢生活笔记，无需专业修图和排版技巧。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "慢生活", "生活方式笔记", "个人IP日常", "社媒plog" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "周末慢生活图文笔记｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：几张公园散步、下午茶和野餐的随手拍照片\n图文类型：个人自媒体\n账号/行业：周末慢生活记录账号\n行业/账号类型：其他\n主题/文案：周末公园慢生活野餐日记\n目标平台：小红书\n内容要求：胶片感调色，多图拼贴排版，配合日期和话题标签，营造治愈慢生活氛围\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-personal-media-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_U30TPjmwCAvr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-personal-media-note-xinjiang-travel-cover", "title": "新疆旅行图文封面", "title_en": "Xinjiang Travel Cover Note", "subtitle": "旅行照片一键拼出杂志感封面大片", "subtitle_en": "Turn travel photos into a magazine-style cover", "description": "拍了很多新疆的雪山、湖泊和沙漠照片，却不知道怎么拼成一张吸睛的旅行笔记封面？上传照片和目的地，AI 自动完成撕纸拼贴排版、大字标题和双语标注，产出适合旅行博主发布的杂志感封面图，无需专业排版软件也能做出高级感。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "旅行封面", "旅行博主", "新疆旅行", "社媒旅行" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "新疆旅行图文封面｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：几张新疆赛里木湖、雪山草原和沙漠驼队的旅行照片\n图文类型：个人自媒体\n账号/行业：旅行探索类账号\n行业/账号类型：其他\n主题/文案：新疆夏日仙境旅行秘籍\n目标平台：小红书\n内容要求：撕纸拼贴排版，中英文双语大字标题，背影出镜配合张开双臂的旅行姿态，突出湖泊雪山沙漠多场景混剪\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-personal-media-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_U30TPjmwCAvr" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-beautify-beauty-sponge-dreamy", "title": "手绘草稿变梦幻海报", "title_en": "Sketch to Dreamy Product Poster", "subtitle": "线稿升级成银色拱门质感主视觉", "subtitle_en": "A hand-drawn draft reborn as a silver arch key visual", "description": "自己画的活动草稿只能表达想法，直接发出去像没做完。美化后手绘线稿换成实拍级主视觉：银色巴洛克拱门、水晶垂饰与粉色缎带作背景，两款美妆蛋分别摆上银质托盘，中英文款名保留。适合美妆品牌、买手店把草稿快速变成能投放的成品。", "categories": [ "creation" ], "tags": [ "海报美化", "美妆产品", "主视觉重做", "风格改造" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "美妆蛋新品海报美化版｜竖版单张" } ], "prompt": "[海报美化]美化上传的既有海报成品，按设计稿美化处理。\n上传已有素材：一张手绘线稿风格的美妆蛋周年限定活动海报\n美化方向：按草稿的内容重做一版梦幻高级风：背景换成银色巴洛克拱门与罗马柱，垂挂水晶串饰与粉色缎带蝴蝶结，地面铺粉色薄纱与花瓣；「冰晶 Snow Crystal」蓝白渐变美妆蛋和「粉雾 Pink Fog」粉白渐变美妆蛋分别立在银质托盘上，中英文款名保留在各自旁边；星空底色配柔雾光，保留主标题「花漾绮梦美妆蛋」与「用柔润触感，开启美妆新篇」\n输出比例：跟随上传图片\n修改方式：调整主视觉、改变风格、重做一版\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-poster-beautify", "name": "海报美化", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019467524816896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-beautify-school-holiday-notice", "title": "放假通知变布告栏风", "title_en": "Holiday Notice Makeover", "subtitle": "纯文字通知排出软木板便签质感", "subtitle_en": "A plain text notice restyled as a corkboard memo", "description": "学校发的放假通知常是一整段文字，家长在群里划两下就跳过。美化后底纹换成软木板，正文放在便签纸上用金属夹固定，标题加高亮块，放假时间和五条温馨提示分区排列，原文一个字没改却愿意看完。适合学校、培训机构、社区发布各类通知。", "categories": [ "creation" ], "tags": [ "海报美化", "校园通知", "放假通知", "排版优化" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "暑假放假通知美化版｜竖版单张" } ], "prompt": "[海报美化]美化上传的既有海报成品，按设计稿美化处理。\n上传已有素材：一张只有黑白文字排版的暑假放假通知\n美化方向：保持通知原文一字不改，把版式做出手账布告栏的质感：软木板背景，正文落在米色便签纸上，顶部用金属夹和蓝色文件夹压住标题「（暑假）放假通知」并加黄色高亮；放假时间 07/01—08/31 单独做成时间轴样式，温馨提示 step1-step5 分条排列，关键词加下划波浪线\n输出比例：跟随上传图片\n修改方式：优化排版布局、改变风格\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-poster-beautify", "name": "海报美化", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019467524816896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-beautify-sichuan-travel-upgrade", "title": "旅行团海报换主视觉", "title_en": "Travel Poster Visual Upgrade", "subtitle": "行程堆叠的小图变成大片感旅拍", "subtitle_en": "A cramped tour flyer turned into a cinematic travel poster", "description": "旅行社的报名海报常把风景缩成小方块，行程、保险、车型全挤在一起，发出去没人点。美化后主视觉换成整幅草原湖泊实景，标题改用书法体压在画面上方，服务承诺和路线信息重新分块，看着像杂志封面。适合旅行社、小团组织者做暑期线路推广。", "categories": [ "creation" ], "tags": [ "海报美化", "旅游海报", "主视觉替换", "排版优化" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "川西旅行海报美化版｜竖版单张" } ], "prompt": "[海报美化]美化上传的既有海报成品，按设计稿美化处理。\n上传已有素材：一张行程和服务条款挤在一起的川西小团报名海报\n美化方向：主视觉换成整幅草原湖泊实景，标题「野奢川西」用书法体压在画面上方并加英文手写体点缀，副标题保留「暑期来一场与自然的邂逅之旅」；服务承诺（高额旅游意外险×成都三环内接早×无需 AA 司机餐/住宿）、2-8 人精品小团、正规旅游车 3 天 2 晚、路线规划（四姑娘山-双桥沟-鱼子西-墨石公园-木格措-红海子塔公草原）重新排成清晰的信息块，整体走大片质感，不要再堆小方块\n输出比例：跟随上传图片\n修改方式：优化排版布局、调整主视觉、改变风格\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-poster-beautify", "name": "海报美化", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019467524816896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-beautify-summer-food-glowup", "title": "旧海报秒变高级质感", "title_en": "Poster Makeover in Seconds", "subtitle": "平面设计升级为实拍质感，排版逻辑不变", "subtitle_en": "Turn flat posters into photo-real visuals instantly", "description": "上传做好的旧海报，一键把扁平化设计换成真实美食摄影质感，标题、标签文案和三伏清凉主题保持不变，整体视觉更精致高级，适合餐饮、零售等需要频繁出海报却没有设计师的商家。", "categories": [ "creation" ], "tags": [ "海报美化", "美食海报", "视觉升级", "餐饮营销", "排版优化" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "三伏清凉美食海报｜升级版 PNG" } ], "prompt": "[海报美化]美化上传的既有海报成品，按设计稿美化处理。\n上传已有素材：（请上传：上传已有素材）\n美化方向：把原本偏平面化的设计升级为真实美食摄影质感，保留标题、标签文案和整体排版逻辑，整体更精致高级\n输出比例：跟随上传图片\n修改方式：优化排版布局、调整主视觉、改变风格\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-poster-beautify", "name": "海报美化", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019467524816896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-meitu-university-admission-guide", "title": "美图大学入学指南系列海报", "title_en": "Meitu University Admission Guide Poster Series", "subtitle": "封面、报到流程与必备清单组成完整新生指南", "subtitle_en": "A complete freshman guide with cover, check-in process, and essentials checklist", "description": "这是为 2026 级新生设计的 3 张入学指南系列海报。案例以扁平矢量插画和清晰信息层级，分别完成指南封面、交通与报到流程、新生必备清单，在统一的校园视觉下兼顾欢迎氛围、流程指引和实用信息。", "categories": [ "creation", "business" ], "tags": [ "海报设计", "入学指南", "校园物料", "信息设计", "系列海报" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "美图大学入学指南｜3 张单文件 HTML" } ], "prompt": "[营销海报]请为美图大学 2026 级新生设计一套入学指南系列海报，共包含指南封面、交通与报到流程、新生必备清单三个主题。整体采用简洁、年轻的扁平矢量插画风格，统一学校名称、主色、字体层级和图标语言；流程信息要易扫读，清单内容要实用，封面需具有欢迎新生的仪式感。", "skills": [ { "id": "designkit-marketing-poster", "name": "营销海报", "description": "围绕活动、宣传或信息传播目标，完成主视觉明确、层级清晰的营销海报与系列物料。", "skill_id": "skill_2091019297552367616" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-poster-menghuoge-takeout-store-set", "title": "猛火哥连锁快餐外卖海报图集", "title_en": "Menghuoge Fast-food Takeout Poster Set", "subtitle": "一套统一绿金视觉的门店招牌、照片墙与宣传横幅", "subtitle_en": "A cohesive green-and-gold set of storefront signs, photo-wall posters, and banners", "description": "这是为猛火哥连锁快餐设计的一套外卖与门店宣传图集。案例围绕绿色主背景、金色描边书法标题、丰盛餐品、黑底卖点标签和厨师 IP，完成 1 张外卖招牌图、3 张照片墙海报与 1 张店内横幅，在不同尺寸中保持品牌识别和促销信息一致。", "categories": [ "creation", "business" ], "tags": [ "海报设计", "餐饮品牌", "外卖招牌", "门店物料", "系列海报" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "猛火哥连锁快餐外卖海报图集｜5 张单文件 HTML" } ], "prompt": "[营销海报]请为猛火哥连锁快餐设计一套外卖与门店宣传图集，包括 1 张 692×390 的外卖招牌图、3 张 240×330 的照片墙海报和 1 张 720×240 的店内宣传横幅。统一使用绿色背景、金色描边的毛笔字标题、丰盛有食欲的餐品组合、醒目的黑底卖点标签和同一厨师 IP；品牌名称、IP、色彩和卖点层级在全部尺寸中保持一致，同时针对横版和竖版画面分别优化构图。", "skills": [ { "id": "designkit-marketing-poster", "name": "营销海报", "description": "围绕活动、宣传或信息传播目标，完成主视觉明确、层级清晰的营销海报与系列物料。", "skill_id": "skill_2091019297552367616" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-beautify-chinese-new-year-limited-series-launch", "title": "丙午新春限定系列发布 PPT 美化案例", "title_en": "Chinese New Year Limited Series Launch PPT Beautification", "subtitle": "以中国红、鎏金与奔马意象强化新春新品发布的高级仪式感", "subtitle_en": "Elevate a Lunar New Year product launch with Chinese red, gold accents, and a dynamic horse motif", "description": "这是一个 4 页新品发布 PPT 美化案例，不是从零设计。案例保留丙午新春限定系列的发布主题、内容结构和市场判断，通过中国红渐变、鎏金线条、丝绸流体与奔马意象建立具有节庆张力的奢华视觉系统，重点优化发布封面、内容纲领、章节过渡和市场契机页面。美化后强化了品牌发布的仪式感、文化辨识度和现场大屏表现力，适合新春限定产品发布与品牌提案。", "categories": [ "creation", "business" ], "tags": [ "PPT美化", "新品发布", "新春限定", "品牌提案", "中国红" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "丙午新春限定系列发布 PPT 美化案例｜4 页单文件 HTML" } ], "prompt": "[PPT美化]美化现有新品发布PPT。\n这是 PPT 美化任务，不是从零设计。请读取我上传的现有 PPT、PDF 或页面图片，完整保留品牌名称、发布主题、事实、数据、文字结构和业务结论，仅优化版式、配色、字体、页面留白与整套视觉一致性。\n视觉方向：围绕丙午新春限定系列建立高级节庆发布视觉，可使用中国红渐变、鎏金线条、丝绸流体、奔马与光影舞台元素，兼顾东方文化寓意、品牌质感和现场大屏冲击力。\n重点页面：发布封面、内容纲领、章节过渡和市场契机。品牌名称等缺失内容继续保留为待替换占位，不得自行虚构品牌或业务数据。\n不得擅自新增、删除或改写原始市场判断、数值与产品含义。", "skills": [], "experts": [ { "id": "designkit-ppt-beautify", "name": "PPT美化", "description": "基于现有PPT、PDF或页面图片重构视觉层级与版式，在保留内容的同时统一风格并提升专业质感。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ueK8MlARbYHU" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-beautify-dunhuang-culture-report", "title": "敦煌文化汇报 PPT 美化案例", "title_en": "Dunhuang Culture Report PPT Beautification", "subtitle": "以敦煌壁画颗粒笔触统一封面、目录、数据和时间轴的文化表达", "subtitle_en": "Unify the cover, contents, data, and timeline with a Dunhuang mural-inspired visual language", "description": "这是一个 4 页 PPT 美化案例，不是从零设计。案例保留敦煌文化汇报原有主题、数据和叙事结构，以壁画矿物色、颗粒点绘、飞天纹样、沙漠与洞窟元素建立统一视觉系统，重点优化封面、目录、文明地位数据页和千年时间轴的版式、配色、字体与信息层级。美化后既保留传统文化质感，也增强了课堂汇报与文化分享中的投屏可读性。", "categories": [ "creation", "business" ], "tags": [ "PPT美化", "敦煌文化", "传统文化", "文化汇报", "壁画风格" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "敦煌文化汇报 PPT 美化案例｜4 页单文件 HTML" } ], "prompt": "[PPT美化]美化现有PPT。\n这是 PPT 美化任务，不是从零设计。请读取我上传的现有 PPT、PDF 或页面图片，完整保留原始主题、事实、数据、文字结构和结论，仅优化版式、配色、字体、图表呈现、页面留白和整套视觉一致性。\n视觉方向：围绕敦煌文化建立壁画矿物色视觉系统，可使用赭石、石青、石绿、金色与暖白，结合颗粒点绘、飞天纹样、洞窟、丝路和沙漠意象；避免现代商务模板削弱文化质感。\n重点页面：封面、目录、关键数据页和历史时间轴。确保大标题、数字、时间节点与说明文字层级清楚，适合课堂汇报和文化主题分享。\n不得擅自新增、删除或改写原始数值、年代、来源与文化含义。", "skills": [], "experts": [ { "id": "designkit-ppt-beautify", "name": "PPT美化", "description": "基于现有PPT、PDF或页面图片重构视觉层级与版式，在保留内容的同时统一风格并提升专业质感。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ueK8MlARbYHU" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-beautify-rongchuang-spring-residence-interior", "title": "融创春风园室内设计方案 PPT 美化案例", "title_en": "Rongchuang Spring Residence Interior Design PPT Beautification", "subtitle": "以新中式水墨与暖杏色重构室内设计方案的信息层级和空间叙事", "subtitle_en": "Refine an interior design proposal with modern Chinese ink styling and warm apricot tones", "description": "这是一个 4 页室内设计方案 PPT 美化案例，不是从零设计。案例保留融创春风园项目的户型问题、设计目标和改造策略，以新中式水墨、暖杏色、窗棂、梅枝和建筑线稿建立“春意融居”的统一视觉语言，重点优化方案封面、三重矛盾分析、设计目标和户型整合前后对比。美化后让东方审美、空间逻辑与改造价值更易理解，适合室内设计提案、业主沟通与方案汇报。", "categories": [ "creation", "business" ], "tags": [ "PPT美化", "室内设计", "新中式", "户型改造", "设计提案" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "融创春风园室内设计方案 PPT 美化案例｜4 页单文件 HTML" } ], "prompt": "[PPT美化]美化现有室内设计方案PPT。\n这是 PPT 美化任务，不是从零设计。请读取我上传的现有 PPT、PDF 或页面图片，完整保留项目名称、户型数据、设计需求、改造动作和结论，仅优化版式、配色、字体、图纸呈现、对比关系、页面留白与整套视觉一致性。\n视觉方向：围绕现代生活与东方审美的融合建立新中式视觉，可使用暖白、暖杏、木色和墨灰，结合水墨山形、窗棂、梅枝、建筑线稿与室内效果图，避免过度装饰影响户型与方案信息。\n重点页面：方案封面、问题分析、设计目标和户型改造前后对比。图纸、尺寸和改造范围必须保持准确、清晰且可核对。\n不得擅自新增、删除或改写原始户型数据、空间功能与设计结论。", "skills": [], "experts": [ { "id": "designkit-ppt-beautify", "name": "PPT美化", "description": "基于现有PPT、PDF或页面图片重构视觉层级与版式，在保留内容的同时统一风格并提升专业质感。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ueK8MlARbYHU" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-beautify-sustainable-urban-mobility", "title": "未来可持续城市交通 PPT 美化案例", "title_en": "Future Sustainable Urban Mobility PPT Beautification", "subtitle": "保留原有内容，统一优化四页 PPT 的版式、配色与信息层级", "subtitle_en": "Preserve the original content while polishing layout, color and hierarchy across four slides", "description": "这是一个 PPT 美化案例，不是从零设计。输入为已有的未来可持续城市交通 PPT 页面，完整保留原始主题、数据和结论，只通过大面积留白、深蓝标题、绿色数据强调与蓝绿流线渐变，统一优化封面、2050 深度脱碳路线图、2030 转型关键里程碑和全球标杆城市案例四页的版式、配色、字体与阅读层级。", "description_en": "This is a PPT beautification case, not a design from scratch. The input is an existing future sustainable urban mobility PPT, with original themes, data, and conclusions fully preserved. Only through generous whitespace, deep blue titles, green data emphasis, and blue-green gradient flow lines, it uniformly polishes the layout, color, typography, and reading hierarchy across four slides: cover, 2050 deep decarbonization roadmap, 2030 transformation key milestones, and global benchmark city cases.", "categories": [ "creation", "business" ], "tags": [ "PPT美化", "可持续交通", "双碳战略", "智慧城市", "美图设计室" ], "tags_en": [ "PPT Beautification", "Sustainable Transport", "Carbon Neutrality", "Smart City", "Meitu DesignKit" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "未来可持续城市交通 PPT 美化案例｜单文件 HTML" } ], "prompt": "这是 PPT 美化任务，不是从零设计。请基于我上传的现有 PPT、PDF 或页面图片，完整保留原始事实、数据、文字结构和结论，仅优化版式、配色、字体、图表呈现、页面留白与整套视觉一致性。可使用暖白背景、深蓝标题、绿色数据强调和蓝绿流线渐变建立统一的绿色科技商务风格；重点改善封面、路线图、关键指标卡片、时间里程碑和案例对比表的层级、对齐与投屏可读性。不得擅自新增、删除或改写数值、来源和业务含义，完成后交付整套美化后的页面或 PPT 文件。\n\n输出要求：完成视觉内容后，将全部页面或图片整理为一个自包含的单文件 HTML，文件名固定为 output.html。所有图片、样式和必要脚本都内联在 HTML 中，不输出独立图片、资源目录或第二个 output 文件；页面需可直接打开，并兼容桌面端与移动端。", "prompt_en": "This is a PPT beautification task, not a design from scratch. Based on the existing PPT, PDF, or page images I upload, fully preserve the original facts, data, text structure, and conclusions; only optimize layout, color, typography, chart presentation, whitespace, and overall visual consistency. Use warm white backgrounds, deep blue titles, green data emphasis, and blue-green gradient flow lines to establish a unified green-tech business style; focus on improving the hierarchy, alignment, and projection readability of covers, roadmaps, key metric cards, timeline milestones, and case comparison tables. Do not add, delete, or rewrite any values, sources, or business meaning. Deliver the complete set of beautified pages or PPT files.\n\nOutput: After completing the visuals, organize all pages or images into a single self-contained HTML file named output.html. All images, styles and necessary scripts must be inlined; no separate image files, resource directories, or second output files. The page must open directly and be compatible with both desktop and mobile.", "skills": [], "experts": [ { "id": "designkit-ppt-beautify", "name": "PPT美化", "description": "基于现有PPT、PDF或页面图片重构视觉层级与版式，在保留内容的同时统一风格并提升专业质感。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ueK8MlARbYHU" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-27T10:22:53+08:00", "updated_at": "2026-08-27T10:22:53+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-class-monitor-election", "title": "班长竞选 PPT", "title_en": "Class Monitor Election PPT", "subtitle": "将自我介绍、服务承诺和个人优势整理成 5 页童趣竞选演示", "subtitle_en": "Turn a self-introduction, service promises, and personal strengths into a five-page election deck", "description": "这是一套 5 页小学班长竞选演示文稿。案例围绕“我想当班长”展开，用简洁的自我介绍说明候选人的性格与爱好，再通过班级服务承诺和个人优势表达责任心、组织力与帮助同学的意愿，最后以明确的投票号召收束。整套 PPT 采用彩色蜡笔、手绘贴纸和儿童画风格，信息少而清晰，适合低年级学生上台演讲。", "categories": [ "creation" ], "tags": [ "PPT设计", "班长竞选", "小学演讲", "竞选展示", "童趣手绘" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "班长竞选 PPT｜5 页单文件 HTML" } ], "prompt": "[PPT设计]生成小学班干部竞选PPT。\n主题/内容：请根据我提供的姓名、班级、照片、自我介绍、兴趣爱好、竞选岗位、服务承诺和个人优势，整理一套简短有感染力的竞选演示。个人信息必须以输入资料为准，不得虚构获奖经历或能力。\n受众：小学低年级同学与老师。\n语言：简体中文，句子短、字号大、适合儿童现场朗读。\n篇幅：约 5 页，包含竞选封面、自我介绍、我能为班级做什么、个人优势和投票号召。\n视觉风格：16:9 横版，彩色蜡笔、手绘贴纸与儿童画风格；如上传个人照片，请自然融入页面并保持主体清晰，整套页面活泼统一。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-college-career-planning", "title": "大学生职业生涯规划 PPT", "title_en": "College Student Career Planning PPT", "subtitle": "从自我认知、行业洞察到十年成长路径，形成可执行的职业规划", "subtitle_en": "Build an actionable career plan from self-awareness and industry insight to a ten-year growth path", "description": "这是一套 12 页大学生职业生涯规划演示文稿。案例以网络与新媒体专业为背景，从职业方向定位、自我认知和 SWOT 分析出发，进一步拆解行业趋势、职业赛道、阶段目标、成长路径、技能提升与动态评估机制。整套 PPT 采用蓝色与玫红色的年轻化编辑视觉，结构完整、节奏鲜明，适合职业规划课程作业、成长答辩和个人发展汇报。", "categories": [ "creation" ], "tags": [ "PPT设计", "职业规划", "大学生作业", "新媒体", "成长路径" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "大学生职业生涯规划 PPT｜12 页单文件 HTML" } ], "prompt": "[PPT设计]生成大学生职业生涯规划PPT。\n主题/内容：请根据我提供的个人信息与职业规划书，围绕职业方向定位、自我认知、SWOT 分析、行业趋势、职业赛道、阶段目标、成长路径、能力提升和评估调整建立完整叙事。姓名、专业、经历和目标必须以输入资料为准；信息不足时使用明确的待补充占位，不得虚构个人经历。\n受众：职业规划课程老师、答辩评委和同学。\n语言：简体中文。\n篇幅：约 12 页，包含封面、目录、分析、目标、行动计划、评估机制和结束页。\n视觉风格：16:9 横版，年轻化编辑设计，以蓝色和亮色强调信息层级，适当使用时间线、矩阵和数据卡片，整套页面统一、清晰且适合课堂答辩。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-hulunbuir-family-travel-guide", "title": "呼伦贝尔亲子旅游攻略 PPT", "title_en": "Hulunbuir Family Travel Guide PPT", "subtitle": "用 12 页手绘水彩攻略规划草原亲子路线、体验与出行准备", "subtitle_en": "Plan routes, family experiences, and travel preparation in a 12-page watercolor guide", "description": "这是一套 12 页呼伦贝尔亲子旅游攻略演示文稿。案例从旅行概览和 5—7 天经典路线出发，梳理草原、森林、湿地等必看景点与骑马、喂小羊、蒙古族文化体验等亲子活动，并补充餐饮住宿、交通、家庭出行指南、行李清单、安全与草原礼仪。整套 PPT 采用清新手绘水彩和亲子绘本风格，信息实用且氛围温暖，适合家庭出行规划与旅行内容分享。", "categories": [ "creation" ], "tags": [ "PPT设计", "亲子旅行", "呼伦贝尔", "旅游攻略", "手绘水彩" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "呼伦贝尔亲子旅游攻略 PPT｜12 页单文件 HTML" } ], "prompt": "[PPT设计]生成亲子旅游攻略PPT。\n主题/内容：请根据我提供的目的地、出行天数、家庭成员和参考资料，整理旅行概览、推荐路线、必看景点、亲子体验、食住行建议、家庭出行提示、行李清单、安全事项与当地礼仪。时间、费用和开放信息必须以输入资料为准；不确定的信息请提示出发前复核，不得编造。\n受众：计划亲子出行的家庭。\n语言：简体中文，表达友好、实用、易读。\n篇幅：约 12 页，包含封面、目录、路线与景点、亲子活动、食住行、准备清单、安全礼仪和结束页。\n视觉风格：16:9 横版，清新手绘水彩与亲子绘本风格，以草原绿、天空蓝和暖黄色为主，使用路线图、信息卡片和亲子插画，保持整套页面统一。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-internet-hotspot-operations-analysis", "title": "互联网热点运营分析 PPT", "title_en": "Internet Hotspot Operations Analysis PPT", "subtitle": "从热点认知、方法拆解到案例复盘，完整呈现可复制的热点运营体系", "subtitle_en": "A complete, repeatable hotspot-operations framework from fundamentals and methods to case reviews", "description": "这是一套面向运营、市场与内容团队的 16 页互联网热点运营分析演示文稿。案例以红黑白商务视觉系统贯穿全篇，从热点运营的角色认知出发，依次拆解热点监测、价值判断、内容策划、传播执行、数据复盘与智能工具提效，并结合典型案例总结“方法可复制、判断不可替代”的运营原则。整套 PPT 结构完整、重点突出，可用于内部培训、运营复盘、方法论分享和业务汇报。", "categories": [ "creation", "business" ], "tags": [ "PPT设计", "热点运营", "内容营销", "运营分析", "美图设计室" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "互联网热点运营分析 PPT｜16 页单文件 HTML" } ], "prompt": "[PPT设计]生成PPT。\n主题/内容：互联网热点运营分析。请围绕“从流量捕手到价值共生”建立完整叙事，面向运营、市场与内容团队，系统讲清热点运营的角色认知、热点监测、价值判断、内容策划、传播执行、数据复盘、案例拆解和智能工具提效，并以“方法可复制、判断不可替代”收束核心观点。\n受众：运营、市场、品牌与内容团队。\n参考文件（可选）：如有上传的大纲或业务资料，优先保持其中的事实、数据、案例和结论，不得擅自改写。\n语言：简体中文。\n篇幅：约 16 页，包含封面、章节过渡、方法步骤、案例分析、总结和 Q&A。\n视觉风格：16:9 横版，红黑白商务视觉系统，大标题、高对比信息层级、清晰图表与流程图，统一页面母版，适合培训与正式汇报。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-qinghai-dunhuang-eight-day-tour", "title": "青海大环线 8 日游行程 PPT", "title_en": "Qinghai-Dunhuang 8-Day Tour Itinerary PPT", "subtitle": "将路线、每日行程、酒店、航班、门票和费用整理为清晰的旅行方案", "subtitle_en": "Turn routes, daily schedules, hotels, flights, tickets, and costs into a clear travel proposal", "description": "这是一套 12 页青海敦煌大环线旅行方案演示文稿。案例将西宁、青海湖、茶卡盐湖、柴达木、敦煌、嘉峪关、张掖和祁连等路线节点串联起来，按天呈现景点、交通、住宿与体验安排，并集中整理航班酒店、门票明细和费用说明。整套 PPT 采用水彩旅行手账与画册式版面，兼顾路线可读性和目的地氛围，可用于旅行方案沟通、亲友出行规划及定制游报价展示。", "categories": [ "creation", "business" ], "tags": [ "PPT设计", "旅游行程", "青海敦煌", "旅行方案", "行程报价" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "青海大环线 8 日游行程 PPT｜12 页单文件 HTML" } ], "prompt": "[PPT设计]生成旅行行程PPT。\n主题/内容：请根据我提供的路线、日期、航班、酒店、景点、门票和费用资料，整理一套完整的多日旅行方案。按天呈现路线与核心体验，并单独汇总交通住宿、门票明细、费用包含与不包含项；所有日期、价格和预订信息必须以输入资料为准，不得擅自编造。\n受众：出行成员、旅行客户或方案决策者。\n语言：简体中文。\n篇幅：约 12 页，包含封面、路线总览、每日行程、航班酒店、门票明细和费用说明。\n视觉风格：16:9 横版，水彩旅行手账与画册风格，使用目的地风景、路线信息和清晰的信息卡片，重点数据醒目，整套页面统一且适合展示。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ppt-design-smart-agri-health-innovation-contest", "title": "智惠农医大学生创新大赛 PPT", "title_en": "Smart Agriculture and Rural Healthcare Innovation Competition PPT", "subtitle": "用 AI 与大数据讲清乡村诊疗和智慧农业项目的完整答辩逻辑", "subtitle_en": "Present a complete innovation pitch for AI-powered rural healthcare and smart agriculture", "description": "这是一套 16 页大学生创新大赛项目答辩演示文稿。案例以“智惠农医”为主题，围绕乡村诊疗与农业生产中的真实痛点，依次呈现项目背景、解决方案、核心技术、产品架构、应用验证、竞争优势、商业模式、社会价值、团队与实施规划。整套 PPT 使用蓝绿科技视觉和农业场景化信息图，兼顾技术可信度、商业表达与公益价值，适合创新创业大赛、项目路演和阶段汇报。", "categories": [ "creation", "business" ], "tags": [ "PPT设计", "创新大赛", "项目路演", "智慧农业", "乡村医疗" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "智惠农医大学生创新大赛 PPT｜16 页单文件 HTML" } ], "prompt": "[PPT设计]生成大学生创新大赛项目PPT。\n主题/内容：请读取我上传的项目文案或申报材料，围绕项目背景、用户痛点、解决方案、核心技术、产品架构、应用场景、验证成果、竞争优势、商业模式、社会价值、团队和未来规划建立完整答辩逻辑。事实、数据、团队信息和阶段成果必须来自参考材料；缺失信息请保留待补充提示，不得编造。\n受众：创新创业大赛评委、学校指导老师和项目合作方。\n语言：简体中文。\n篇幅：约 16 页，包含封面、项目论证、技术与产品、市场与价值、团队计划和结束页。\n视觉风格：16:9 横版，蓝绿农业科技视觉，使用清晰的数据图表、架构图、流程图和乡村应用场景，页面统一、专业且适合现场答辩。", "skills": [], "experts": [ { "id": "designkit-ppt-design", "name": "PPT设计", "description": "把一句话主题、详细大纲或参考资料整理为结构完整、表达清晰且视觉统一的演示文稿。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tq3OOlv5HjiQ" } ], "mcps": [], "scenario": "", "scenario_use": "PPT设计", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T15:43:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-closeup-amber-cocktail-icecubes", "title": "琥珀色调酒冰爽特写", "title_en": "Amber Cocktail Close-up", "subtitle": "冰块碰撞光影流动，橙皮点缀增添质感层次", "subtitle_en": "A refreshing close-up for an amber cocktail", "description": "餐饮类产品拍出食欲感，关键在光泽和细节的流动感。上传商品图，AI 生成暗色餐桌背景下冰块碰撞、酒液泛光的特写视频，橙片装饰增加层次，画面清爽又有质感。适合酒饮、餐饮品牌做菜单展示或社媒宣传的短视频素材。", "categories": [ "creation" ], "tags": [ "产品特写视频", "调酒饮品", "冰爽质感", "餐饮广告", "产品展示" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "琥珀色调酒特写视频｜竖版视频" } ], "prompt": "[产品特写视频]制作产品特写视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n视频比例：1:1 方形\n其他要求：暗色餐桌背景，冰块与酒液光泽流动，橙片装饰增加细节层次\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-product-closeup", "name": "产品特写视频", "description": "商品特写视频，突出细节质感", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kmmoCD8StU13" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-closeup-amber-sunglasses-light", "title": "琥珀色太阳镜质感特写", "title_en": "Amber Sunglasses Close-up", "subtitle": "光影缓慢移动，镜片纹理与镜框质感清晰可见", "subtitle_en": "A slow-light close-up for amber tortoiseshell sunglasses", "description": "太阳镜的质感全在细节里，一段慢镜头光影特写比多张静态图更有说服力。上传商品图，AI 生成暖光缓慢扫过镜面的特写视频，玳瑁纹镜框和镜片反光层层展开，背景干净不抢镜。适合眼镜、配饰类目做电商详情页的质感展示视频。", "categories": [ "creation" ], "tags": [ "产品特写视频", "太阳镜", "质感光影", "慢镜头", "产品展示" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "琥珀色太阳镜特写视频｜竖版视频" } ], "prompt": "[产品特写视频]制作产品特写视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n视频比例：9:16 竖屏\n其他要求：暖光缓慢扫过镜面，镜框纹理清晰，背景干净灰色调\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-product-closeup", "name": "产品特写视频", "description": "商品特写视频，突出细节质感", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kmmoCD8StU13" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-closeup-leather-wallet-texture", "title": "牛皮钱包质感特写", "title_en": "Leather Wallet Texture Close-Up", "subtitle": "暖光慢镜头，把皮革纹理拍出高级触感", "subtitle_en": "Warm slow-motion shots that sell the leather's touch", "description": "上传商品图，AI 自动生成产品特写视频：暖调影棚光线下缓慢推近，头层牛皮的天然纹理、手工车缝线与地面倒影层层展现，质感一目了然。适合皮具、腕表等注重材质细节的商品卖家，用镜头语言替代文字描述，让消费者隔屏也能感受到手感。", "categories": [ "creation" ], "tags": [ "产品特写视频", "牛皮钱包", "材质质感", "皮具展示", "慢镜头" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "牛皮钱包特写视频｜竖版质感展示" } ], "prompt": "[产品特写视频]制作产品特写视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n视频比例：9:16 竖屏\n其他要求：暖色影棚顶光，慢速推近展现头层牛皮纹理与车缝线细节，地面倒影增加质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-product-closeup", "name": "产品特写视频", "description": "商品特写视频，突出细节质感", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kmmoCD8StU13" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-closeup-silver-watch-dark", "title": "银色机械腕表暗调特写", "title_en": "Dark Mechanical Watch Close-up", "subtitle": "冷光聚焦表盘刻度，暗色背景衬出金属质感", "subtitle_en": "A moody dark close-up for a silver mechanical watch", "description": "腕表类目最讲究质感呈现，暗调聚光的拍法比明亮棚拍更显高级。上传商品图，AI 生成深色背景下聚光灯打在表盘上的特写视频，刻度和链带的金属光泽被放大，冷峻又精致。适合手表、精品配饰做详情页或投放素材，提升产品调性。", "categories": [ "creation" ], "tags": [ "产品特写视频", "机械腕表", "暗调质感", "金属光泽", "产品展示" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "银色腕表暗调特写视频｜竖版视频" } ], "prompt": "[产品特写视频]制作产品特写视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n视频比例：9:16 竖屏\n其他要求：深色背景聚光灯效果，表盘刻度和链带细节清晰，质感冷峻高级\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-product-closeup", "name": "产品特写视频", "description": "商品特写视频，突出细节质感", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kmmoCD8StU13" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-design-memphis-mushroom-lamp", "title": "孟菲斯风蘑菇台灯系列", "title_en": "Memphis Mushroom Lamp Set", "subtitle": "撞色几何造型台灯+桌面摆件，复古趣味美学", "subtitle_en": "A colorful mushroom lamp with matching desk decor", "description": "橙色蘑菇伞盖台灯搭配粉色灯柱与绿色底座，与格纹笔筒、异形托盘、撞色积木摆件共同摆放在宝蓝色背景前，孟菲斯风格的撞色几何造型充满复古趣味。适合家居装饰、文创周边品牌打造年轻化的桌面美学系列，兼具实用与陈列价值。", "categories": [ "creation" ], "tags": [ "产品设计", "家居摆件", "孟菲斯风格", "撞色设计", "系列产品" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "产品造型方案｜台灯+桌面摆件系列图" } ], "prompt": "[产品设计]帮我做品牌产品设计。\n品牌/产品参考素材：（请上传：品牌/产品参考素材）\n品牌名称：\n产品品类：家居摆件/桌面装饰\n设计类型：系列产品\n目标人群/场景：年轻人家居桌面、礼赠场景\n风格方向：孟菲斯撞色、复古趣味\n方案数量：4\n其他要求：蘑菇造型台灯为主体，橙粉绿撞色几何造型，搭配格纹笔筒、异形托盘、色块积木摆件，宝蓝背景实拍\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-design", "name": "产品设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297689333760" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-design-pastel-puffer-backpack", "title": "马卡龙撞色软包设计", "title_en": "Pastel Colorblock Backpack", "subtitle": "奶油黄拼紫色羽绒质感双肩包，少女感拼接细节", "subtitle_en": "A puffer backpack in cream and lavender colorblock", "description": "奶油黄羽绒质感包身拼接薰衣草紫抽绳口袋，粉色织带与提手点缀其间，侧边网兜与挂件细节丰富，柠檬黄背景衬出马卡龙配色的软糯质感。适合穿戴配饰、潮流箱包品牌打造少女感十足的日常通勤背包，兼顾实用收纳与视觉个性。", "categories": [ "creation" ], "tags": [ "产品设计", "穿戴配饰", "双肩包", "马卡龙配色", "拼接设计" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "产品造型方案｜双肩包设计效果图" } ], "prompt": "[产品设计]帮我做品牌产品设计。\n品牌/产品参考素材：（请上传：品牌/产品参考素材）\n品牌名称：\n产品品类：穿戴配饰/双肩包\n设计类型：穿戴配饰\n目标人群/场景：年轻女性、日常通勤出行\n风格方向：马卡龙撞色、软糯质感\n方案数量：1\n其他要求：奶油黄羽绒质感包身拼接薰衣草紫抽绳口袋，粉色织带点缀，侧边网兜与挂件细节丰富\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-design", "name": "产品设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297689333760" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-design-retro-pattern-mug-set", "title": "复古撞纹马克杯三件套", "title_en": "Retro Pattern Mug Trio", "subtitle": "棋盘格、波浪、条纹三款图案，圆环把手统一造型", "subtitle_en": "Checker, wave and stripe mugs with matching handles", "description": "白底红棋盘格、浅蓝波浪线、深红条纹三款马克杯堆叠摆放，统一采用圆环大把手造型，浅蓝背景衬托复古撞色质感。适合家居杯具、生活方式品牌打造有辨识度的图案系列，单只使用或成套摆放都能呈现活泼的复古氛围。", "categories": [ "creation" ], "tags": [ "产品设计", "陶瓷杯具", "复古图案", "系列产品", "家居用品" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "产品造型方案｜三款图案马克杯套装图" } ], "prompt": "[产品设计]帮我做品牌产品设计。\n品牌/产品参考素材：（请上传：品牌/产品参考素材）\n品牌名称：\n产品品类：家居用品/陶瓷杯具\n设计类型：系列产品\n目标人群/场景：年轻家居爱好者、早餐咖啡场景\n风格方向：复古撞色、趣味图案\n方案数量：3\n其他要求：白底红棋盘格、浅蓝波浪线、深红条纹三款图案，统一圆环大把手造型，浅蓝背景实拍\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-design", "name": "产品设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297689333760" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-design-solace-humidifier", "title": "疗愈风香薰加湿器造型", "title_en": "Wellness Humidifier Design", "subtitle": "圆润哑光造型+品牌刻字，打造居家疗愈单品", "subtitle_en": "A soft matte form engraved with the brand name", "description": "SOLACE香薰加湿器采用圆润哑光造型，暖色打光下雾气升腾，机身侧面刻字品牌名，透出疗愈松弛的居家氛围。适合小家电、香氛品牌为新品做外观提案，兼顾功能感和治愈系审美，一键出图省去实物打样等待。", "categories": [ "creation" ], "tags": [ "产品设计", "小家电外观", "香薰加湿器", "疗愈系", "产品提案" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "产品外观设计｜香薰加湿器效果图" } ], "prompt": "[产品设计]帮我做品牌产品设计。\n品牌/产品参考素材：（请上传：品牌/产品参考素材）\n品牌名称：SOLACE\n产品品类：香薰加湿器\n设计类型：实体产品造型\n目标人群/场景：居家卧室、疗愈放松场景，都市年轻女性\n风格方向：极简疗愈，暖调莫兰迪\n方案数量：1\n其他要求：圆润不规则几何造型，磨砂哑光材质，机身侧面镌刻品牌名，暖光氛围下有雾气效果\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-design", "name": "产品设计", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297689333760" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-replace-balm-to-sunscreen", "title": "手持防晒管替换出图", "title_en": "Swap Handheld Product Packaging", "subtitle": "手部姿势与紫调氛围不变，只换新包装", "subtitle_en": "Keep the hands, swap the packaging", "description": "手持展示图最难复用，换个产品就得重新约模特。这个案例把手里的香膏纸盒换成同系列防晒管，指关节的握姿、奶黄色美甲和紫色针织的氛围一点没动，包装上的规格文字清晰可读。适合个护美妆品牌上新时快速产出一组手持主图。", "categories": [ "creation" ], "tags": [ "商品替换", "美妆个护", "手持展示", "包装换新" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "手持防晒管场景图｜单张 JPG" } ], "prompt": "[商品替换]将场景图中的商品替换为上传的新商品。\n场景图：一张双手捧着方形香膏纸盒的特写，模特穿紫色针织开衫，奶黄色美甲\n新商品图：一支米白配浅紫盖的防晒霜软管，正面朝上的白底图\n其他要求：软管竖立在原纸盒位置，正面文字朝向镜头；手指握持关系与原图一致，保留紫调柔光\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-replace", "name": "产品替换", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298965835776" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-replace-blender-to-espresso", "title": "厨房场景商品替换", "title_en": "Swap Products in Kitchen Scene", "subtitle": "保留场景光影，一键替换成新商品", "subtitle_en": "Keep the scene lighting, swap in your own product", "description": "上传一张已有的厨房场景图和自己的新商品图，AI将场景中的原商品替换为新商品，台面阴影、光线角度和周围道具摆放保持自然一致，适合没有实拍条件、需要快速套用高质感场景图的商家。", "categories": [ "creation" ], "tags": [ "商品替换", "场景复用", "厨房电器图", "电商实拍替代", "光影一致" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "咖啡机场景替换图｜单张 PNG" } ], "prompt": "[商品替换]将场景图中的商品替换为上传的新商品。\n场景图：（请上传：场景图）\n新商品图：（请上传：新商品图）\n其他要求：新商品摆放位置、光影角度和台面阴影需与原料理机保持一致，杯中呈现现磨咖啡效果\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-replace", "name": "产品替换", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298965835776" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-replace-bottle-to-sneakers", "title": "球场运动鞋替换上架", "title_en": "Swap Bottle for Sneakers on Court", "subtitle": "保留球场光位与地面阴影，换成自家跑鞋", "subtitle_en": "Put your sneakers into the same court scene", "description": "一套球场实拍拍下来成本不低，换个品类又要重拍。这个案例把画面里的运动水壶换成自家跑鞋，蓝绿地胶的反光、球拍和小球的位置全部保留，鞋面阴影跟着原来的光位落地。适合运动品牌把同一组场景图复用到水壶、跑鞋、护具等多个 SKU。", "categories": [ "creation" ], "tags": [ "商品替换", "运动鞋", "球场场景", "户外运动" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "球场运动鞋场景图｜单张 JPG" } ], "prompt": "[商品替换]将场景图中的商品替换为上传的新商品。\n场景图：一张俯拍的蓝绿球场实拍，画面中央立着一只运动水壶，旁边有球拍和两颗荧光球\n新商品图：一双米白配荧光绿的网面跑鞋，白底商品图\n其他要求：跑鞋放在原水壶位置，鞋头朝右前方成对摆放，贴合地面并沿用原图斜向投影\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-replace", "name": "产品替换", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298965835776" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-product-replace-lamp-to-speaker", "title": "客厅台灯换智能音箱", "title_en": "Swap Lamp for Smart Speaker", "subtitle": "沿用原图暖光与投影，新品直接落进样板间", "subtitle_en": "Drop your speaker into a lit living-room shot", "description": "新品还没拍摄，想先看看它摆在真实客厅里是什么样。上传一张暖光边几场景图和音箱白底图，桌面投影、织物质感和暖黄光位全部沿用原片，音箱像本来就摆在那儿。适合智能家居、灯具家电团队用一套场景图铺满整条产品线。", "categories": [ "creation" ], "tags": [ "商品替换", "智能家居", "场景图", "家居摆拍" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "客厅场景智能音箱图｜单张 JPG" } ], "prompt": "[商品替换]将场景图中的商品替换为上传的新商品。\n场景图：一张暖光客厅边几实拍，深木色圆桌上摆着一盏石材台灯，背景是米色墙面与棕榈叶影\n新商品图：一张米白色织物智能音箱的白底商品图\n其他要求：音箱放在原台灯位置，正面朝向镜头，体量与台灯底座相当；保留桌面暖黄反光和墙面叶影\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-product-replace", "name": "产品替换", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019298965835776" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-resume-design-brand-designer-warm-creative", "title": "暖调创意品牌设计师简历", "title_en": "Warm Creative Brand Designer Resume", "subtitle": "手绘元素+暖色调，展现设计师个人风格与温度", "subtitle_en": "Hand-drawn accents and warm tones show a designer's personality", "description": "设计师简历常常千篇一律，很难让招聘方记住。这款模板用暖棕色调、圆角相框和手写标题营造亲切有温度的视觉感受，同时保留经历、技能、工具等信息板块，适合品牌设计师、视觉设计师投递创意行业公司或工作室时展现个人风格。", "categories": [ "creation" ], "tags": [ "简历设计", "品牌设计师简历", "创意简历模板", "设计师作品履历" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "暖调创意品牌设计师简历｜单张 JPG" } ], "prompt": "[简历设计]设计简历。\n简历内容：品牌设计师简历，Emily Chen，创意设计师，2023年至今任Warm Studio品牌设计师，负责生活方式与时尚品牌视觉设计；此前在Bloom Creative Agency任初级设计师，参与品牌视觉与包装项目；教育背景广东工业大学视觉传达设计学士，交换于韩国庆熙大学；擅长品牌设计、视觉设计、包装设计、社交媒体设计；工具Photoshop、Illustrator、InDesign、Figma、Canva；暖棕色调，手写标题，圆角人像相框\n简历材料：（请上传：简历材料（可选））\n目标岗位：品牌设计师\n简历语言：自动匹配\n简历风格：吸睛专业\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-resume-design", "name": "简历设计", "description": "简历版式设计，模板复刻", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_T26trHAIrB7c" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-resume-design-product-manager-blue-bilingual", "title": "蓝色产品经理中英简历", "title_en": "Blue Bilingual PM Resume", "subtitle": "中英对照排版清晰，突出产品经理核心业绩", "subtitle_en": "Bilingual layout that highlights a PM's key achievements", "description": "求职时常纠结简历要不要做中英对照，尤其是申请外企或大厂产品岗位。这份简历模板把工作经历、项目成果和教育背景整理成中英双语版式，蓝色主色调专业清晰，产品经理的核心业绩一目了然，适合互联网、外企产品岗位求职者直接投递使用。", "categories": [ "creation" ], "tags": [ "简历设计", "产品经理简历", "中英双语简历", "求职简历" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "蓝色中英产品经理简历｜单张 JPG" } ], "prompt": "[简历设计]设计简历。\n简历内容：产品经理简历，张皓然/Lucas Zhang，6年以上产品经验，现任某互联网公司高级产品经理，负责核心产品线规划与迭代，用户活跃度提升35%+；曾主导0到1产品上线，用户增长200%；教育背景浙江大学计算机科学与技术学士、香港科技大学信息系统管理硕士；技能涵盖产品策略、用户研究、数据分析、PRD撰写、Axure/Figma原型设计、SQL/Excel、项目管理，中英文对照排版，蓝色主色调\n简历材料：（请上传：简历材料（可选））\n目标岗位：产品经理\n简历语言：自动匹配\n简历风格：简洁正式\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-resume-design", "name": "简历设计", "description": "简历版式设计，模板复刻", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_T26trHAIrB7c" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-resume-design-purple-visual-designer", "title": "紫色视觉设计师简历模板", "title_en": "Purple Visual Designer Resume", "subtitle": "把经历、作品能力和个人风格整理成一页创意简历", "subtitle_en": "Turn experience, creative skills and personality into a one-page designer resume", "description": "以视觉设计师与插画师求职为示例，把个人简介、工作经历、专业技能、教育背景、设计工具和联系方式整理成一页双语创意简历。紫色笔刷、手绘涂鸦与黑色马克笔元素强化个人风格，同时保持清晰的信息层级，适合视觉设计、插画、品牌与内容创意岗位。", "description_en": "Using a visual designer and illustrator job search as an example, this case organizes personal profile, work experience, professional skills, education, design tools, and contact info into a one-page bilingual creative resume. Purple brush strokes, hand-drawn doodles, and black marker elements reinforce personal style while maintaining clear information hierarchy, suitable for visual design, illustration, brand, and content creative roles.", "categories": [ "creation", "business" ], "tags": [ "简历设计", "视觉设计师", "创意简历", "求职", "美图设计室" ], "tags_en": [ "Resume Design", "Visual Designer", "Creative Resume", "Job Search", "Meitu DesignKit" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "紫色视觉设计师简历模板｜单文件 HTML" } ], "prompt": "请根据我提供的真实个人经历、目标岗位、技能信息、教育背景、联系方式和个人照片，设计一份适合视觉设计师或插画师求职的一页式专业简历。先围绕目标岗位梳理信息优先级，突出可量化的工作成果、代表项目和核心设计能力；采用暖白纸张、黑色马克笔线条与紫色笔刷涂鸦组成的创意编辑风格，并用清晰模块呈现个人简介、工作经历、技能、教育背景、工具和联系方式。中英文内容按我提供的信息排版，不得虚构经历、学历、奖项或联系方式；照片需保持本人身份与五官特征，整体兼顾个性表达、阅读效率和打印可用性。\n\n输出要求：完成视觉内容后，将全部页面或图片整理为一个自包含的单文件 HTML，文件名固定为 output.html。所有图片、样式和必要脚本都内联在 HTML 中，不输出独立图片、资源目录或第二个 output 文件；页面需可直接打开，并兼容桌面端与移动端。", "prompt_en": "Based on the real personal experience, target role, skill information, education, contact details, and personal photo I provide, design a one-page professional resume suitable for a visual designer or illustrator job search. First prioritize information around the target role, highlighting quantifiable work achievements, representative projects, and core design capabilities; use a creative editorial style with warm white paper, black marker lines, and purple brush doodles, with clear modules for personal profile, work experience, skills, education, tools, and contact info. Layout Chinese and English content per the information I provide; do not fabricate experience, education, awards, or contact details; photos must preserve the person's identity and facial features. Balance personal expression, reading efficiency, and print readiness.\n\nOutput: After completing the visuals, organize all pages or images into a single self-contained HTML file named output.html. All images, styles and necessary scripts must be inlined; no separate image files, resource directories, or second output files. The page must open directly and be compatible with both desktop and mobile.", "skills": [], "experts": [ { "id": "designkit-resume-design", "name": "简历设计", "description": "根据经历内容、目标岗位和参考材料设计结构清晰、重点突出的专业简历，支持多语言与多种视觉风格。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_T26trHAIrB7c" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-27T10:22:53+08:00", "updated_at": "2026-08-27T10:22:53+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-resume-design-red-halftone-editorial", "title": "红色半调复古品牌设计师简历", "title_en": "Red Halftone Editorial Resume", "subtitle": "报刊感半调印刷风格，塑造强烈个人品牌视觉", "subtitle_en": "Newsprint halftone style for a bold personal brand", "description": "品牌设计师想在一堆简历里脱颖而出，需要更有记忆点的视觉表达。这款模板用报刊感半调印刷效果和大字号衬线标题，塑造强烈的个人品牌调性，适合投递创意工作室、广告公司或独立设计师自我展示，一份简历也是一次设计作品。", "categories": [ "creation" ], "tags": [ "简历设计", "创意简历模板", "品牌设计师简历", "英文简历" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "红色半调复古简历｜单张 JPG" } ], "prompt": "[简历设计]设计简历。\n简历内容：Brand Designer resume, Sakura Lin, based in Taipei. Currently Brand Designer at Studio Kuro (Jan 2023-Present), leading brand identity and visual systems for lifestyle, tech and culture clients. Previously Junior Designer at ReBloom Agency and Design Intern at Whiteboard Studio. Bachelor of Design, National Taipei University of Arts. Skills: Illustrator, Photoshop, InDesign, Figma, Branding, Typography. Red halftone newsprint style, bold serif headline\n简历材料：（请上传：简历材料（可选））\n目标岗位：Brand Designer\n简历语言：英文\n简历风格：创意突出\n\n工具约束：PPT 相关任务不得调用 tencent-pptx，也不得通过 tencent-docs-routing 间接调用 tencent-pptx。", "skills": [], "experts": [ { "id": "designkit-resume-design", "name": "简历设计", "description": "简历版式设计，模板复刻", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_T26trHAIrB7c" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-rotation-video-amber-perfume-bottle", "title": "琥珀色香水瓶360°旋转展示", "title_en": "Amber Perfume Bottle 360 Spin", "subtitle": "金色瓶盖与琥珀色液体旋转生辉，质感通透", "subtitle_en": "A glossy 360-degree spin for an amber perfume bottle", "description": "香水瓶的美感在于光线穿过液体时的通透感，静态图很难还原这种效果。上传商品图，AI 生成渐变背景下的旋转展示视频，金色瓶盖高光随旋转流动，琥珀色液体透亮诱人。适合香水、美妆礼盒做产品页或直播预热素材。", "categories": [ "creation" ], "tags": [ "360度旋转视频", "香水", "金色质感", "产品展示", "电商视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "琥珀色香水瓶360°旋转视频｜竖版视频" } ], "prompt": "[360°旋转视频]制作商品360°旋转展示视频。\n商品图片：（请上传：商品图片）\n视频时长：5秒\n背景：渐变\n其他要求：金色瓶盖高光随旋转流动，液体色泽通透，背景柔和渐变\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-rotation-video", "name": "360°旋转视频", "description": "商品360°旋转展示视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qd2OsGuENVDO" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-rotation-video-navy-luggage-white-bg", "title": "深蓝行李箱360°旋转展示", "title_en": "Navy Luggage 360 Spin", "subtitle": "纯白背景匀速旋转，箱体线条与拉杆细节一览无余", "subtitle_en": "A clean 360-degree spin for a navy hardshell suitcase", "description": "行李箱这类大件商品，详情页最需要一个能看清全貌的旋转视频。上传商品图，AI 生成纯白背景下匀速旋转的展示视频，箱体线条、拉链走位和拉杆细节完整呈现，光线均匀不留死角。适合箱包、出行用品做电商主图视频。", "categories": [ "creation" ], "tags": [ "360度旋转视频", "行李箱", "纯白背景", "产品展示", "电商视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "深蓝行李箱360°旋转视频｜竖版视频" } ], "prompt": "[360°旋转视频]制作商品360°旋转展示视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n背景：白色\n其他要求：匀速旋转展示箱体拉链和滚轮细节，光线均匀柔和\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-rotation-video", "name": "360°旋转视频", "description": "商品360°旋转展示视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qd2OsGuENVDO" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-rotation-video-retro-polaroid-camera", "title": "拍立得360°旋转展示", "title_en": "Retro Camera 360° Spin", "subtitle": "纯白背景旋转展示，复古相机细节一览无余", "subtitle_en": "A clean white spin shot that shows every retro detail", "description": "上传商品图，AI 自动生成商品360°旋转展示视频：纯白背景下平稳旋转，复古奶白色机身、镜头、闪光灯和取景窗依次亮相，细节清晰无死角。适合数码、家居等需要全方位展示外观的商品卖家，一条视频替代多角度实拍图，直接用在详情页或短视频广告位。", "categories": [ "creation" ], "tags": [ "360°旋转视频", "拍立得相机", "白底展示", "复古数码", "产品外观" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "拍立得相机360°展示｜白底旋转视频" } ], "prompt": "[360°旋转视频]制作商品360°旋转展示视频。\n商品图片：（请上传：商品图片）\n视频时长：5秒\n背景：白色\n其他要求：奶白色复古机身，平稳旋转展现镜头、取景窗与闪光灯细节\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-rotation-video", "name": "360°旋转视频", "description": "商品360°旋转展示视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qd2OsGuENVDO" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-rotation-video-running-shoe-blue-gradient", "title": "跑鞋悬浮360°旋转展示", "title_en": "Running Shoe 360 Spin", "subtitle": "深蓝渐变背景衬托，鞋身线条与配色层次分明", "subtitle_en": "A dynamic 360-degree spin for a running shoe", "description": "运动鞋想拍出科技感，靠的是背景和悬浮效果的配合，而不是简单摆拍。上传商品图，AI 生成深蓝渐变背景下鞋身悬浮旋转的展示视频，配色分层和鞋底纹路清晰可见，动感十足。适合运动鞋服类目做上新宣传的电商视频。", "categories": [ "creation" ], "tags": [ "360度旋转视频", "运动鞋", "渐变背景", "产品展示", "电商视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "跑鞋360°旋转视频｜竖版视频" } ], "prompt": "[360°旋转视频]制作商品360°旋转展示视频。\n商品图片：（请上传：商品图片）\n视频时长：10秒\n背景：渐变\n其他要求：深蓝到黑色渐变背景，鞋身悬浮呈现，配色层次清晰有科技感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-rotation-video", "name": "360°旋转视频", "description": "商品360°旋转展示视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_qd2OsGuENVDO" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-smart-crop-home-vertical-main-image", "title": "家居竖版主图裁剪", "title_en": "Home Vertical Main Image Crop", "subtitle": "一键裁成 3:4 上架主图，层次不被切断", "subtitle_en": "Crop to a 3:4 listing main image in one go", "description": "上架主图比例固定，直接压缩会让沙发变形、吊灯出画。选 3:4 后画面从吊灯一路留到人字拼地板，沙发居中、豹纹圆凳压在左下角当前景，天花板到地面的层次完整没有断。适合家居电商上架主图和图文种草配图。", "categories": [ "creation" ], "tags": [ "智能裁剪", "竖版主图", "家居电商", "上架图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "家居竖版主图｜3:4 单张 JPG" } ], "prompt": "[智能裁剪]将图片智能裁剪。\n待裁剪图片：一张客厅实景照，画面里有橙色弧形沙发、木质吊灯和人字拼木地板\n目标比例：3:4\n裁剪说明：从吊灯到地板的层次要完整，沙发居中，前景圆凳保留在左下角，不要压缩变形\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-smart-crop", "name": "智能裁剪", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019299091664896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-smart-crop-living-room-wide-banner", "title": "客厅横版头图裁剪", "title_en": "Living Room Wide Banner Crop", "subtitle": "竖幅实拍改横版头图，沙发与挂画完整保留", "subtitle_en": "Turn a vertical room shot into a wide banner", "description": "同一张客厅实拍要同时用在官网头图和详情页横幅，手动裁总会切掉挂画或吊灯。上传原图选 16:9，橙色弧形沙发、边几和背景挂画完整留在画面中央，多余的顶部和地面被收掉，横向留白均匀。适合家居品牌和软装工作室做横版头图。", "categories": [ "creation" ], "tags": [ "智能裁剪", "横版头图", "家居实拍", "详情页banner" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "客厅横版头图｜16:9 单张 JPG" } ], "prompt": "[智能裁剪]将图片智能裁剪。\n待裁剪图片：一张竖幅拍摄的客厅实景照，画面里有橙色弧形沙发、吊灯和背景挂画\n目标比例：16:9\n裁剪说明：保留沙发、边几和挂画的完整关系，主体横向居中，顶部吊灯可以裁去一部分，地面留白不要过多\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-smart-crop", "name": "智能裁剪", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019299091664896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-smart-crop-sofa-fullscreen-vertical", "title": "沙发全屏竖屏裁剪", "title_en": "Full-Screen Vertical Sofa Crop", "subtitle": "裁成 9:16 手机全屏，主体依然完整", "subtitle_en": "Crop to 9:16 without cutting the subject", "description": "同一张客厅图要发手机全屏和竖屏封面，普通裁剪常把吊灯或沙发切掉一半。选 9:16 后两侧多余墙面被收掉，吊灯与沙发上下呼应，主体在窄画幅里依然完整，地板延伸感也留住了。适合家居账号做全屏封面和竖屏落地页首屏。", "categories": [ "creation" ], "tags": [ "智能裁剪", "竖屏封面", "手机全屏", "家居内容" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "沙发竖屏封面图｜9:16 单张 JPG" } ], "prompt": "[智能裁剪]将图片智能裁剪。\n待裁剪图片：一张客厅实景照，画面里有木质吊灯、橙色弧形沙发和豹纹圆凳\n目标比例：9:16\n裁剪说明：竖屏构图，收掉两侧多余墙面，吊灯和沙发保持上下呼应，主体不要被切断\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-smart-crop", "name": "智能裁剪", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019299091664896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-smart-crop-square-living-room", "title": "客厅实拍智能裁剪", "title_en": "Smart Crop for Living Room Shot", "subtitle": "一键裁成1:1方图，主体构图不跑偏", "subtitle_en": "Crop to 1:1 while keeping the subject centered", "description": "上传横版家居实拍图，选择目标比例，AI自动识别沙发、茶几等主体重新构图，避免生硬裁切导致画面失衡，适合电商详情页、社交媒体等需要统一图片比例的场景。", "categories": [ "creation" ], "tags": [ "智能裁剪", "构图调整", "家居实拍", "电商详情图", "比例适配" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "客厅沙发智能裁剪图｜1:1 JPG" } ], "prompt": "[智能裁剪]将图片智能裁剪。\n待裁剪图片：（请上传：待裁剪图片）\n目标比例：1:1\n裁剪说明：保留沙发、茶几和挂画主体完整，构图居中平衡\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-smart-crop", "name": "智能裁剪", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019299091664896" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-analysis-beginner-gym-note-breakdown", "title": "私教课笔记爆款拆解", "title_en": "Beginner Gym Note Breakdown", "subtitle": "拆解新手向健身笔记走红结构，复用同款模板", "subtitle_en": "Break down why beginner gym notes go viral", "description": "看到“新手第一节私教课笔记”这类内容很受欢迎，却不知道爆款结构长什么样？给出品类和分析方向，AI 拆解笔记的标题钩子、真人实拍搭配手写标注的呈现方式，以及“新手友好、全程可跟练”等信任标签的使用逻辑，输出可复用的内容模板，帮健身账号快速产出同类爆款。", "categories": [ "business" ], "tags": [ "趋势分析", "爆款拆解", "健身账号", "新手向内容", "内容策略" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "私教课笔记爆款拆解｜文档" } ], "prompt": "[趋势分析]基于用户提供的链接或截图资料进行社媒趋势分析。\n截图/参考资料：（请上传：截图/参考资料）\n分析对象/链接：小红书新手私教课笔记类爆款内容\n目标平台：小红书\n行业/品类：健身\n分析类型：爆款拆解\n关注重点：拆解标题钩子、真人实拍配手写标注的呈现方式和信任标签使用逻辑，输出可直接复用的新手向内容模板\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-analysis", "name": "趋势分析", "description": "社媒趋势洞察与爆款内容拆解", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_34cuFaXjMALa" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-analysis-fat-loss-plan-hot-format", "title": "减脂训练表选题分析", "title_en": "Fat-Loss Plan Hot Format Analysis", "subtitle": "拆解小基数减脂爆款形态，产出可直接套用的训练表", "subtitle_en": "Turn a viral workout format into a ready-to-use table", "description": "发现“小基数减脂”话题很热，却不确定用什么形态发才容易被收藏？给出品类和分析方向，AI 拆解近期同类爆款笔记的排版和信息结构，找出打卡表格式最受欢迎的呈现方式，直接产出一份可打印的7天训练表模板，帮你把选题热度转化成可执行内容。", "categories": [ "business" ], "tags": [ "趋势分析", "选题热度", "健身账号", "训练计划", "爆款拆解" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "减脂训练表选题分析｜文档" } ], "prompt": "[趋势分析]基于用户提供的链接或截图资料进行社媒趋势分析。\n截图/参考资料：（请上传：截图/参考资料）\n分析对象/链接：小红书小基数减脂话题近期爆款笔记\n目标平台：小红书\n行业/品类：健身\n分析类型：选题热度\n关注重点：拆解打卡表格式笔记为什么容易被收藏，输出一份可直接抄作业的7天全身训练清单，标注每日重点和时长\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-analysis", "name": "趋势分析", "description": "社媒趋势洞察与爆款内容拆解", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_34cuFaXjMALa" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-analysis-fitness-trend-report", "title": "健身赛道趋势洞察报告", "title_en": "Fitness Trend Report", "subtitle": "拆解赛道爆款规律，给出可执行内容打法", "subtitle_en": "Decode viral patterns into an action plan", "description": "想做健身账号却摸不清赛道打法？给出品类和分析方向，AI 拆解近期爆款笔记的选题、形态和竞争格局，输出核心判断、赛道全景表和内容策略建议，帮你少走弯路，直接对着报告做选题。", "categories": [ "business" ], "tags": [ "社媒分析", "赛道洞察", "健身账号", "内容策略", "爆款拆解" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "健身赛道洞察报告｜文档" } ], "prompt": "[趋势分析]基于用户提供的链接或截图资料进行社媒趋势分析。\n截图/参考资料：（请上传：截图/参考资料）\n分析对象/链接：小红书健身赛道近期爆款笔记\n目标平台：小红书\n行业/品类：健身\n分析类型：趋势洞察\n关注重点：分析健身赛道爆款密码、内容形态和竞争格局，输出可执行的选题公式和内容结构建议\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-analysis", "name": "趋势分析", "description": "社媒趋势洞察与爆款内容拆解", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_34cuFaXjMALa" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-analysis-wellness-trend-report", "title": "养生赛道趋势洞察报告", "title_en": "Wellness Trend Insight Report", "subtitle": "拆解养生赛道爆款规律，给出可执行内容打法", "subtitle_en": "Decode wellness trend patterns into an action plan", "description": "想做养生类账号却不清楚现在流行什么内容形式？给出品类和分析方向，AI 拆解养生赛道的宏观数据、细分方向竞争态势和近期爆款笔记形态，输出核心判断、赛道全景表和可直接套用的选题卡片，帮你少走弯路，对着报告就能定选题。", "categories": [ "business" ], "tags": [ "趋势分析", "赛道洞察", "养生账号", "内容策略", "爆款拆解" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "养生赛道洞察报告｜文档" } ], "prompt": "[趋势分析]基于用户提供的链接或截图资料进行社媒趋势分析。\n截图/参考资料：（请上传：截图/参考资料）\n分析对象/链接：小红书养生赛道近期爆款笔记\n目标平台：小红书\n行业/品类：养生\n分析类型：趋势洞察\n关注重点：分析养生赛道宏观声量数据、中式轻养生等细分方向的竞争态势和关键信号，输出可直接复用的冬日养生选题卡片\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-analysis", "name": "趋势分析", "description": "社媒趋势洞察与爆款内容拆解", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_34cuFaXjMALa" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-commerce-espresso-machine-reaction", "title": "迷你咖啡机真人反应带货", "title_en": "Espresso Machine Reaction Ad", "subtitle": "男生早晨用咖啡机随手测评，接地气种草短视频", "subtitle_en": "A guy's honest morning espresso machine review", "description": "带货短视频最打动人的往往不是精致的棚拍，而是真实的生活场景和自然反应。上传商品图，AI 生成男生在家用咖啡机的口播带货片段，语气轻松像在跟朋友安利。适合居家小家电、厨房好物，用真人出镜的方式降低种草门槛，提升下单转化。", "categories": [ "creation" ], "tags": [ "带货短视频", "咖啡机", "真人测评", "居家好物", "反应展示" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "迷你咖啡机带货短视频｜竖版视频" } ], "prompt": "[带货短视频]制作带货短视频。\n商品图片：（请上传：商品图片）\n商品名称：一键桌面咖啡机\n目标平台：TikTok\n核心卖点：一键萃取，10秒出咖啡，占地小\n视频调性：效果广告\n视频比例：9:16 竖屏\n其他要求：男生居家实景出镜，语气轻松真实，突出使用便捷\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-commerce", "name": "带货短视频", "description": "效果广告/生活剧情脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_L3hp4PgrDTfV" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-commerce-handheld-fan-summer-outdoor", "title": "户外手持风扇避暑带货", "title_en": "Handheld Fan Summer Ad", "subtitle": "夏日户外真人出镜，展示随身风扇的降温效果", "subtitle_en": "A summer outdoor demo for a handheld mini fan", "description": "季节性单品最需要场景代入感，一台风扇放在纯白背景里很难让人心动。上传商品图，AI 生成户外阳光下的带货片段，真实的燥热感和降温瞬间形成对比，说服力更强。适合夏季出行、户外类小家电，快速产出适合社媒、短视频平台投放的种草素材。", "categories": [ "creation" ], "tags": [ "带货短视频", "手持风扇", "夏日好物", "户外场景", "真人出镜" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "手持风扇夏日带货短视频｜竖版视频" } ], "prompt": "[带货短视频]制作带货短视频。\n商品图片：（请上传：商品图片）\n商品名称：手持小风扇\n目标平台：小红书\n核心卖点：轻便随身，出风量大，续航持久\n视频调性：效果广告\n视频比例：9:16 竖屏\n其他要求：户外阳光实景，真人出镜展示消暑效果，语气自然亲切\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-commerce", "name": "带货短视频", "description": "效果广告/生活剧情脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_L3hp4PgrDTfV" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-commerce-lint-brush-coat-demo", "title": "除毛刷带货短视频", "title_en": "Lint Brush Product Video", "subtitle": "居家实拍除毛过程，一刷见效引发下单冲动", "subtitle_en": "A real home demo that turns instant results into sales", "description": "上传商品图，AI 自动生成居家场景的带货短视频：真实还原刷除大衣毛絮的全过程，从满身浮毛到干净利落，效果一目了然。适合日用清洁类商品卖家，用真实使用场景替代硬广，提升种草说服力和转化率。", "categories": [ "creation" ], "tags": [ "带货短视频", "除毛刷", "生活好物", "居家清洁", "效果广告" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "除毛刷效果广告｜竖版带货短视频" } ], "prompt": "[带货短视频]制作带货短视频。\n商品图片：（请上传：商品图片）\n商品名称：双面衣物除毛刷\n目标平台：抖音\n核心卖点：双面设计，一刷去除大衣绒毛和浮尘\n视频调性：效果广告\n视频比例：9:16 竖屏\n其他要求：居家场景实拍风格，展现刷毛前后对比效果\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-social-commerce", "name": "带货短视频", "description": "效果广告/生活剧情脚本+出片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_L3hp4PgrDTfV" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-content-fire-finance-account-plan", "title": "FIRE 理财账号 0-1 社媒内容规划", "title_en": "FIRE Finance Account Content Plan from Zero to One", "subtitle": "15 张统一狐狸 IP 图文讲清路线、误解与工资分配", "subtitle_en": "Fifteen fox-IP posts explain FIRE routes, myths, and salary allocation", "description": "这是一个从 0 到 1 规划 FIRE 理财账号的社媒内容案例。15 张图文以统一狐狸 IP 和海军蓝、橙色、薄荷绿视觉系统，系统讲解 5 类 FIRE 路线、3 个常见误解、工资四账户分配法、储蓄率分档与月度行动清单，兼顾账号定位、系列选题和可持续发布。", "categories": [ "creation", "business" ], "tags": [ "自媒体运营", "FIRE", "理财科普", "账号规划", "社媒图文" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "FIRE 理财账号 0-1 规划｜15 张单文件 HTML" } ], "prompt": "[个人媒体图文]请参考我提供的对标博主和目标受众，从 0 到 1 规划一个关于如何实现 FIRE 的理财账号，并完成首批系列图文。内容应系统讲清 Lean FIRE、Coast FIRE、Barista FIRE、Fat FIRE 与定制/混合 FIRE，拆解常见误解，再给出工资四账户分配法、储蓄率分档和可执行的月度行动清单。建立统一、可持续复用的狐狸 IP、海军蓝/橙色/薄荷绿色彩系统和清晰的信息层级，保证理财表述通俗、克制且不承诺收益。", "skills": [], "experts": [ { "id": "designkit-personal-media-note", "name": "个人自媒体图文", "description": "从账号定位、选题规划到系列图文设计，建立可持续发布的个人自媒体内容体系。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_U30TPjmwCAvr" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-style-image-laundry-pitfalls-comic", "title": "洗衣避坑漫画分镜图文", "title_en": "Laundry Pitfalls Comic Storyboard", "subtitle": "把实用洗衣知识整理成清晰易读的国风漫画科普图文", "subtitle_en": "Turn practical laundry tips into an engaging Chinese-style comic explainer", "description": "以送洗衣物前的五个常见误区为主题，用统一的国风 Q 版人物、分镜对话和图标清单呈现洗标检查、污渍预处理、衣物分类、专业洗护选择与取件验收等知识。适合洗护门店、生活服务品牌和知识类自媒体快速制作易传播的避坑科普内容。", "description_en": "Centered on five common mistakes before sending clothes to the laundry, this case uses a consistent Chinese-style chibi character, storyboard dialogue, and icon checklists to present label checking, stain pre-treatment, garment sorting, professional care selection, and pickup inspection. Ideal for laundry shops, lifestyle service brands, and knowledge content creators who need to quickly produce shareable explainer content.", "categories": [ "creation", "business" ], "tags": [ "洗衣避坑", "漫画分镜", "科普图文", "社媒视觉", "美图设计室" ], "tags_en": [ "Laundry Tips", "Comic Storyboard", "Infographic", "Social Media Visual", "Meitu DesignKit" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "洗衣避坑漫画分镜图文｜单文件 HTML" } ], "prompt": "请把我提供的洗衣避坑知识整理成一张适合社交媒体发布的 3:4 漫画科普图文。采用清新国风 Q 版插画，使用同一位古风少女作为贯穿角色；顶部用醒目标题概括主题，中间以编号分镜呈现 5 个常见误区，每格包含简短标题、对应生活场景和一句易懂提醒；底部补充正确洗护建议与可执行清单。整体使用蓝白与暖米色配色，分镜边界清晰、人物形象统一、文字层级明确，确保手机端也能阅读。不得添加无关品牌、二维码或水印；如果我提供参考图，请保留其核心人物和视觉语言，但根据新内容重新组织画面。\n\n输出要求：完成视觉内容后，将全部页面或图片整理为一个自包含的单文件 HTML，文件名固定为 output.html。所有图片、样式和必要脚本都内联在 HTML 中，不输出独立图片、资源目录或第二个 output 文件；页面需可直接打开，并兼容桌面端与移动端。", "prompt_en": "Organize the laundry care knowledge I provide into a 3:4 comic explainer image suitable for social media. Use a fresh Chinese-style chibi illustration with a consistent ancient-style girl as the recurring character; top with a prominent title summarizing the theme, middle section presents 5 common mistakes in numbered storyboards, each cell containing a short title, corresponding life scene, and an easy-to-understand reminder; bottom section supplements correct care advice and an actionable checklist. Use blue-white and warm beige tones, clear storyboard borders, consistent character design, and clear text hierarchy for mobile readability. Do not add unrelated brands, QR codes, or watermarks; if I provide a reference image, keep its core character and visual language but reorganize the layout for the new content.\n\nOutput: After completing the visuals, organize all pages or images into a single self-contained HTML file named output.html. All images, styles and necessary scripts must be inlined; no separate image files, resource directories, or second output files. The page must open directly and be compatible with both desktop and mobile.", "skills": [], "experts": [ { "id": "designkit-social-style-image", "name": "社媒风格化图片", "description": "根据目标风格与平台要求重塑用户原图，兼顾主体保留、画面比例和调整边界，生成社媒视觉成品。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Xuyu0l69WTwt" } ], "mcps": [], "scenario": "", "scenario_use": "视觉海报", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-27T10:22:53+08:00", "updated_at": "2026-08-27T10:22:53+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-style-knitted-cat-care-guide", "title": "新手养猫注意事项编织风社媒图文", "title_en": "Knitted-style Social Posts for First-time Cat Owners", "subtitle": "把 5 张养猫科普图统一成柔软有趣的编织视觉", "subtitle_en": "Unify five cat-care explainers in a soft, playful knitted visual style", "description": "这是一组面向新手养猫人群的 5 张社媒科普图文。案例在完整保留接猫准备、饮食避坑、健康管理和居家安全信息的基础上，将原图统一转换为趣味编织材质，建立一致的色彩、字体层级与毛线质感，兼顾信息可读性和社媒吸引力。", "categories": [ "creation", "business" ], "tags": [ "社媒图文", "风格化", "编织风", "养猫科普", "内容设计" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "新手养猫注意事项编织风图文｜5 张单文件 HTML" } ], "prompt": "[社媒风格图]请保持我上传的 5 张新手养猫科普图的文案、信息结构和核心内容不变，将全部图片统一转换为趣味编织风格。画面应呈现柔软的毛线、针织纹理和立体手作质感，统一配色、字体层级、插画语言与版式节奏，同时确保接猫准备、饮食避坑、健康管理和居家安全等信息清晰可读。", "skills": [], "experts": [ { "id": "designkit-social-style-image", "name": "社媒风格化图片", "description": "在保留原始内容和信息结构的前提下，统一转换社媒图片的视觉风格与材质语言。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Xuyu0l69WTwt" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-social-style-papercut-food-storage-guide", "title": "冰箱食材避坑立体剪纸风社媒图文", "title_en": "Paper-cut Social Posts for Food Storage Safety", "subtitle": "6 张食品安全科普图统一为清晰醒目的立体剪纸风", "subtitle_en": "Unify six food-safety explainers in a clear dimensional paper-cut style", "description": "这是一组围绕冰箱食材储存误区的 6 张社媒科普图。案例覆盖发霉生姜、久置酱料、泡发菌菇、长期冷冻肉类和发芽土豆等常见风险，在保持原有知识内容不变的前提下，统一转换为立体剪纸风格，让严肃的食品安全信息更亲和、更易传播。", "categories": [ "creation", "business" ], "tags": [ "社媒图文", "立体剪纸", "食品安全", "科普设计", "风格化" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "冰箱食材避坑立体剪纸风图文｜6 张单文件 HTML" } ], "prompt": "[社媒风格图]请保持我上传的 6 张冰箱食材储存科普图的文案、信息结构和核心内容不变，将全部图片统一转换为立体剪纸风格。使用清晰的层叠纸张、柔和阴影和鲜明但协调的配色，统一字体、插画和版式语言，同时保证发霉生姜、久置酱料、泡发菌菇、长期冷冻肉类与发芽土豆等风险提示准确易读。", "skills": [], "experts": [ { "id": "designkit-social-style-image", "name": "社媒风格化图片", "description": "在保留原始内容和信息结构的前提下，统一转换社媒图片的视觉风格与材质语言。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Xuyu0l69WTwt" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-sticker-pack-anime-avatar-emotions", "title": "卷发女孩表情包", "title_en": "Curly-Haired Girl Sticker Pack", "subtitle": "紫色卫衣卷发女孩四格表情，个人聊天专属头像", "subtitle_en": "A chibi girl in a purple hoodie, four expressions", "description": "扎着蓬松卷发、穿紫色连帽卫衣的卡通女孩，配上“早呀”“困困”“哇哦”“不理你”四种表情与手绘星星、爱心点缀，动漫Q版画风软萌可爱。适合上传本人或朋友照片转绘成专属聊天表情包，日常斗图、群聊互动都能用上自己的形象。", "categories": [ "creation" ], "tags": [ "表情包", "动漫头像", "真人转绘", "日常表情", "个人IP" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "IP表情包｜4张PNG" } ], "prompt": "[表情包]基于上传的IP形象，生成表情包。\nIP形象参考：（请上传：IP形象参考）\n表情数量：4 个\n风格类型：日常表情\n其他要求：卷发+紫色连帽卫衣的Q版动漫女孩形象，配早呀、困困、哇哦、不理你等文案与手绘星星爱心贴纸\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-sticker-pack", "name": "贴纸包", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297637797888" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-sticker-pack-office-cat-workday-mood", "title": "打工猫职场表情包", "title_en": "Office Cat Workday Sticker Pack", "subtitle": "橘猫摸鱼吐槽职场日常，牛马人的嘴替", "subtitle_en": "An orange cat venting about office life", "description": "橘猫伏案打瞌睡、抱头崩溃改稿、边摸鱼边刷手机、背包欢呼下班，四格插画配上“班味很重”“又改了”“摸鱼中”“下班啦”等打工人黑话。适合职场、办公类品牌延展社群互动表情包，用共鸣感十足的吐槽文案拉近与年轻用户的距离。", "categories": [ "creation" ], "tags": [ "表情包", "职场吐槽", "工作摸鱼", "插画猫", "打工人" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "IP表情包｜4张PNG" } ], "prompt": "[表情包]基于上传的IP形象，生成表情包。\nIP形象参考：（请上传：IP形象参考）\n表情数量：4 个\n风格类型：工作摸鱼\n其他要求：橘猫形象，配班味很重、又改了、摸鱼中、下班啦等打工人吐槽文案\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-sticker-pack", "name": "贴纸包", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297637797888" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-sticker-pack-purple-hat-banter", "title": "斗图吐槽表情包", "title_en": "Sassy Mascot Sticker Pack", "subtitle": "浮夸紫毛绒帽IP+犀利文案，斗图必备", "subtitle_en": "A flashy mascot with punchy roast captions", "description": "紫色毛绒高帽+墨镜+亮片西装的浮夸IP形象，配上笑死我了、拿捏了、就这、你过来啊等犀利文案，表情张力拉满。适合潮流、说唱、综艺类品牌延展聊天斗图表情包，快速抢占用户日常对话场景。", "categories": [ "creation" ], "tags": [ "表情包", "IP表情包", "斗图吐槽", "品牌延展", "聊天表情" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "IP表情包｜4张PNG" } ], "prompt": "[表情包]基于上传的IP形象，生成表情包。\nIP形象参考：一个戴紫色毛绒高帽、墨镜、亮片西装的卡通男孩IP形象\n表情数量：4 个\n风格类型：斗图吐槽\n其他要求：紫色毛绒高帽+墨镜+亮片西装造型，表情浮夸、文案犀利，适合斗图互怼场景\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-sticker-pack", "name": "贴纸包", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297637797888" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-sticker-pack-real-cat-daily-mood", "title": "布偶猫日常表情包", "title_en": "Fluffy Cat Daily Sticker Pack", "subtitle": "真实猫咪抓拍+口语文案，撸猫人的表情包", "subtitle_en": "Real cat photos paired with everyday captions", "description": "用金渐层猫咪的真实抓拍配上“好开心”“想吃”“好困”“哼”等口语化文案与爱心、光效小贴纸，四宫格覆盖开心、撒娇、犯困、生气等日常情绪。适合养猫人士用自家猫咪照片生成专属表情包，聊天斗图时更有代入感和真实感。", "categories": [ "creation" ], "tags": [ "表情包", "宠物表情包", "真实照片", "日常表情", "猫咪" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "IP表情包｜4张PNG" } ], "prompt": "[表情包]基于上传的IP形象，生成表情包。\nIP形象参考：（请上传：IP形象参考）\n表情数量：4 个\n风格类型：日常表情\n其他要求：使用真实猫咪照片，配好开心、想吃、好困、哼等口语文案与爱心、光效贴纸\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-sticker-pack", "name": "贴纸包", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297637797888" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-store-traffic-note-camping-checkin-guide", "title": "营地打卡引流图文", "title_en": "Camping Site Check-in Guide", "subtitle": "森系露营实拍，新手攻略带位置导流", "subtitle_en": "A forest camping guide that drives site visits", "description": "想给营地或露营基地做引流笔记，却不知道怎么呈现才有氛围感？输入营地卖点，AI 用森林绿意实拍配合清单式攻略文案，突出新手必看、宝藏推荐、拍照技巧三个板块，结尾附定位标签方便读者直接导航到店，适合营地、户外基地类门店做种草和到店引流。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "门店引流", "营地露营", "户外打卡", "本地生活" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "营地打卡引流图文｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：门店引流\n账号/行业：营地/户外露营基地\n行业/账号类型：其他\n主题/文案：营地打卡指南，新手露营攻略\n目标平台：小红书\n内容要求：森林绿意实拍背景，清单式攻略文案覆盖新手必看、宝藏营地推荐、拍照打卡技巧三个板块，结尾放置定位标签引导到店\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-store-traffic-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gt1mVh677Jam" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-store-traffic-note-home-cleaning-promo", "title": "家政保洁团购引流图文", "title_en": "Home Cleaning Promo Note", "subtitle": "泡沫写字视觉冲击，价格福利一眼看清", "subtitle_en": "A soap-writing visual that highlights your price", "description": "想给家政保洁门店做引流，却没有精修图也不想拍摄现场？只需输入服务项目和价格，AI 自动生成用泡沫水渍手写价格的强视觉冲击海报，蓝天白云背景配合手部动作特写，突出低价和服务时长，适合家政、保洁类本地生活门店做低成本引流。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "门店引流", "家政保洁", "本地生活", "团购海报" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "家政保洁引流海报｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：门店引流\n账号/行业：家政保洁门店\n行业/账号类型：其他\n主题/文案：家政保洁一小时150元团购价\n目标平台：小红书\n内容要求：蓝天泡沫写字视觉，手持玻璃刮刀擦窗特写，突出每小时价格和服务时长，营造清爽干净的视觉联想\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-store-traffic-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gt1mVh677Jam" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-store-traffic-note-hotel-memo-post", "title": "备忘录风酒店引流笔记", "title_en": "Hotel Promo Memo Note", "subtitle": "文字备忘录风格，一眼说清价格与福利带动预订", "subtitle_en": "Memo-style text poster that drives hotel bookings", "description": "想给酒店门店做引流图文，却不想露脸拍摄也没有精修图？只需输入酒店卖点和优惠信息，AI 自动生成手机备忘录截图风格的图文笔记，大字报式排版突出人均价格、地理位置和入住权益，配一句口语化的行动号召，适合酒店、民宿等本地门店低成本做种草引流。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "门店引流", "备忘录笔记", "酒店探店", "本地生活" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "酒店引流备忘录图文｜单张海报" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：门店引流\n账号/行业：本地周末游/酒店探店账号\n行业/账号类型：探店\n主题/文案：市中心宝藏酒店周末躺平攻略\n目标平台：小红书\n内容要求：备忘录截图风格，大字报式排版，突出人均价格、地理位置和入住权益（含早餐+延迟退房），结尾用一句口语化行动号召\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-store-traffic-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gt1mVh677Jam" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-store-traffic-note-spa-lounge-promo", "title": "精油SPA门店引流图文", "title_en": "Spa Lounge Promo Note", "subtitle": "高级质感空间图，突出预约到店信息", "subtitle_en": "An elegant spa visual with clear booking info", "description": "想给SPA或美容门店做引流图文，却没有专业空间摄影？输入门店卖点和优惠信息，AI 自动生成米咖色调的高级感门店空间图，搭配专业手法、植物精油、独立空间三个信任标签和预约到店按钮，适合SPA、美容美体类门店做种草引流，无需实地布景拍摄。", "categories": [ "creation" ], "tags": [ "社媒图文笔记", "门店引流", "精油SPA", "美容门店", "本地生活" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "精油SPA引流海报｜单张" } ], "prompt": "[社媒图文笔记]按所选图文类型生成社媒图文，聚焦门店引流和个人自媒体场景。\n素材/参考图：（请上传：素材/参考图）\n图文类型：门店引流\n账号/行业：精油SPA门店\n行业/账号类型：其他\n主题/文案：精油SPA身心放松空间体验\n目标平台：小红书\n内容要求：米咖色调高级感空间图，突出专业手法、植物精油、独立私密空间三个信任标签，结尾放置预约到店按钮引导转化\n生成张数：1 张\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-store-traffic-note", "name": "社媒图文笔记", "description": "输入主题或文案，生成社媒图文笔记", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gt1mVh677Jam" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-storybook-bosi-pipu-juanjuan-third-birthday", "title": "波西皮普祝卷卷 3 岁生日动画绘本", "title_en": "Bosi and Pipu Celebrate Juanjuan's Third Birthday", "subtitle": "7 段连续动画串联生日准备、庆祝与祝福", "subtitle_en": "Seven animated scenes connect birthday preparation, celebration, and wishes", "description": "这是一个为 3 岁儿童定制的生日动画绘本案例。波西和皮普来到卷卷的生日派对，从日期提示、朋友到场、唱生日歌、分享蛋糕到一起玩玩具，最终以生日祝福收束。角色形象、手绘质感和叙事节奏统一，适合作为儿童生日纪念和个性化动画故事参考。", "categories": [ "creation" ], "tags": [ "动画绘本", "生日故事", "儿童", "亲子纪念", "手绘动画" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "波西皮普祝卷卷 3 岁生日｜单文件动画绘本" } ], "prompt": "[绘本]请为 3 岁的卷卷创作一本生日动画绘本，主要角色包括卷卷、波西和皮普。保留参考素材中的人物辨识特征和温暖手绘风格，按日期提示、朋友到场、生日歌与蛋糕、分享礼物与玩具、生日祝福的顺序，形成完整、连贯且适合幼儿观看的故事。全部分镜保持角色造型、场景色彩和叙事节奏一致。", "skills": [], "experts": [ { "id": "designkit-storybook", "name": "绘本/故事书", "description": "把主题、人物素材和故事设定转化为画风统一、情节连贯的儿童绘本或故事书。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gxRbyo23De8c" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-storybook-father-daughter-forest-adventure", "title": "爸爸和小芽的森林大冒险绘本", "title_en": "Dad and Xiaoya's Forest Adventure Storybook", "subtitle": "把亲子照片变成 8 页温暖连贯的森林探险故事", "subtitle_en": "Turn family photos into a warm, coherent eight-page forest adventure", "description": "这是一本以父女森林探险为主题的 8 页儿童绘本。案例从整装出发、观察森林、跨越吊桥到发现宝藏和约定下次冒险，完整呈现起承转合；水彩插画、人物造型和亲子情感贯穿全篇，适合作为亲子纪念、睡前阅读和儿童故事创作参考。", "categories": [ "creation" ], "tags": [ "绘本", "儿童故事", "亲子", "森林探险", "水彩插画" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "爸爸和小芽的森林大冒险｜8 页单文件 HTML" } ], "prompt": "[绘本]请根据我上传的父亲和女儿照片，创作一本完整连贯的儿童森林探险绘本。保留人物的核心外貌和亲子关系，将故事组织为整装出发、发现森林奇景、跨越摇晃吊桥、彼此鼓励、发现森林宝藏、快乐返程和约定下次探险等连续情节。整体采用温暖细腻的水彩绘本风格，画面充满想象力、惊险感与童趣，人物造型、服装和色彩在全部页面保持一致，文字简洁，适合儿童阅读。", "skills": [], "experts": [ { "id": "designkit-storybook", "name": "绘本/故事书", "description": "把主题、人物素材和故事设定转化为画风统一、情节连贯的儿童绘本或故事书。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_gxRbyo23De8c" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-08-31T15:43:00+08:00", "updated_at": "2026-08-31T16:10:59+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-text-edit-menu-lunch-to-dinner", "title": "菜单文字一键改版", "title_en": "Edit Menu Text in One Shot", "subtitle": "换标题换菜品，字体排版不留痕", "subtitle_en": "Swap titles and dishes with no visible edits", "description": "上传实拍的餐厅菜单卡片，写清原文字和替换内容，AI将标题、分类和菜品价格整体替换为新版本，字体样式、颜色和排版位置保持一致，适合餐饮商家午市晚市切换、季节菜单更新等场景。", "categories": [ "creation" ], "tags": [ "改文字", "菜单设计", "文字替换", "餐饮物料", "排版保留" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "夜市菜单卡片｜单张 JPG" } ], "prompt": "[改文字]修改图片中的文字内容。\n待处理图片：（请上传：待处理图片）\n原文字：午市小馆\n替换为：夜市小馆\n其他说明：同步将菜品分类、品名和价格替换为夜市场景内容，字体样式与排版位置保持不变\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-text-edit", "name": "文字编辑", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297872678912" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-text-edit-shop-sign-rename", "title": "门店招牌改名换字", "title_en": "Rename a Storefront Sign", "subtitle": "实拍灯箱改店名，字体投影都不露痕迹", "subtitle_en": "Change the shop name on a real signboard", "description": "门店改名、加盟商换招牌，重新做喷绘既慢又贵。上传一张招牌实拍，直接把店名和品类词换掉，圆润手写体的笔画粗细、棕色字色和灯箱在水泥墙上的投影都跟着原图走，远看就是重新做过的招牌。适合连锁餐饮、烘焙门店做改名预览和加盟招商图。", "categories": [ "creation" ], "tags": [ "改文字", "门店招牌", "品牌改名", "实景照片" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "改名后门店招牌图｜单张 JPG" } ], "prompt": "[改文字]修改图片中的文字内容。\n待处理图片：一张街边门店实拍，黄底棕字的方形灯箱招牌挂在水泥墙面上\n原文字：Bello cookies\n替换为：Mello bakery\n其他说明：沿用原招牌的圆润手写体与棕色字色，字号和居中位置不变；保留灯箱厚度、侧面高光和墙面投影\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-text-edit", "name": "文字编辑", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297872678912" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-text-edit-skincare-date-update", "title": "护肤海报改活动日期", "title_en": "Update the Date on a Beauty Poster", "subtitle": "老素材换个日期复用，衬线字形不变", "subtitle_en": "Reuse last season's poster with a new date", "description": "活动结束了，主视觉还很好看，就是日期过期了。这个案例把右侧的标签和日期整块更新，衬线大字的字形、半透明白色和上下行距都保持原样，人像肤质、米色背景和手中的产品软管毫发无损。适合美妆护肤品牌把往期主视觉改期后再上一轮投放。", "categories": [ "creation" ], "tags": [ "改文字", "护肤海报", "活动日期", "素材复用" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "改期护肤海报｜单张 JPG" } ], "prompt": "[改文字]修改图片中的文字内容。\n待处理图片：一张米色调护肤人像海报，模特手持软管贴在下颌，右侧是两行白色衬线大字加一行日期\n原文字：OLD DATE 2024.05.20\n替换为：NEW DATE 2025.08.23\n其他说明：沿用原衬线字形与半透明白色，两行大字的断行、字号和下方日期的行距保持一致\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-text-edit", "name": "文字编辑", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297872678912" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-text-edit-spring-outing-poster", "title": "春游海报换主标题", "title_en": "Swap a Spring Poster Headline", "subtitle": "手写体笔锋照旧，主标一句话换完", "subtitle_en": "New headline, same handwritten strokes", "description": "节气海报年年复用，真正要改的往往只有标题四个字。这个案例把主标从踏青攻略改成春日出游，手写体的笔锋、白色描边和字距全部延续，下方的绿色英文标签、线描小人和草地实景一点没动。适合运营同学把一版模板扩成整个春季的系列海报。", "categories": [ "creation" ], "tags": [ "改文字", "节气海报", "手写字体", "社媒配图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "换标题春游海报｜单张 JPG" } ], "prompt": "[改文字]修改图片中的文字内容。\n待处理图片：一张蓝天草地春游主题海报，主标是白色手写体中文，配线描小人与绿色英文标签\n原文字：踏青攻略\n替换为：春日出游\n其他说明：保持原手写笔锋、白色描边与四字横排位置；标题下方的英文绿标和其余装饰元素不动\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-text-edit", "name": "文字编辑", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297872678912" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-tvc-product-headphones-moody-podium", "title": "薰衣草紫耳机质感广告片", "title_en": "Moody Headphones Ad", "subtitle": "暗调灯光衬托耳机悬浮展示，尽显高端氛围感", "subtitle_en": "A moody lighting ad for premium headphones", "description": "耳机这类产品最怕拍得平淡，暗调氛围灯光反而能放大它的质感。上传商品图，AI 生成耳机悬浮在展台上的广告片，紫蓝色光线勾勒出机身轮廓和材质细节，冷峻又高级。适合追求品牌调性的耳机、数码类目，用一条广告片替代传统实拍棚拍。", "categories": [ "creation" ], "tags": [ "产品广告片", "头戴式耳机", "高端氛围", "品牌TVC", "暗调质感" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "薰衣草紫耳机品牌广告片｜竖版视频" } ], "prompt": "[产品广告片]制作产品广告片。\n商品图片：（请上传：商品图片）\n商品名称：薰衣草紫头戴式耳机\n广告类型：品牌TVC\n投放平台：YouTube\n其他要求：暗调紫蓝色氛围灯光，耳机悬浮展示，质感冷峻高级\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-tvc-product", "name": "产品广告片", "description": "TVC/开箱种草/反应展示", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_GKm4xKVRzR8h" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-tvc-product-night-city-scooter-ad", "title": "雾银电动车都市广告片", "title_en": "Misty Silver Scooter Night Ad", "subtitle": "雨后夜色剪影镜头，一条广告片讲透产品调性", "subtitle_en": "A rain-slicked night silhouette shot that sells the vibe", "description": "没有实拍团队和夜景场地，也能做出电影质感的产品广告片。上传商品图，AI 自动生成雨后都市夜景中的逆光剪影镜头，霓虹街灯层层叠加出高级氛围，凸显金属车身线条。适合新品发布、渠道投放前想要一条电影感TVC却预算有限的品牌和电商团队。", "categories": [ "creation" ], "tags": [ "产品广告片", "电动车", "夜景大片", "品牌TVC", "广告脚本" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "雾银电动车品牌TVC｜竖版广告片" } ], "prompt": "[产品广告片]制作产品广告片。\n商品图片：（请上传：商品图片）\n商品名称：雾银城市通勤电动车\n广告类型：品牌TVC\n投放平台：\n其他要求：雨后城市夜景，逆光剪影运镜，霓虹光斑烘托高级氛围，突出车身金属质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-tvc-product", "name": "产品广告片", "description": "TVC/开箱种草/反应展示", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_GKm4xKVRzR8h" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-tvc-product-robot-vacuum-sunlight-floor", "title": "深灰扫地机器人晨光广告片", "title_en": "Robot Vacuum Sunlight Ad", "subtitle": "木地板暖阳光影铺开，扫地机器人也能拍出高级质感", "subtitle_en": "A robot vacuum ad with soft morning sunlight", "description": "把一台扫地机器人拍出品牌大片的质感，靠的是光影而不是复杂布景。上传商品图，AI 自动生成木地板晨光洒落的居家场景，机身线条和细节在暖色光斑中格外清晰。适合智能家居类目做品牌广告片，用氛围感和质感打动用户，而不是堆砌参数。", "categories": [ "creation" ], "tags": [ "产品广告片", "扫地机器人", "智能家居", "晨光质感", "品牌TVC" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "深灰扫地机器人品牌广告片｜竖版视频" } ], "prompt": "[产品广告片]制作产品广告片。\n商品图片：（请上传：商品图片）\n商品名称：深灰圆形扫地机器人\n广告类型：品牌TVC\n投放平台：抖音\n其他要求：木地板晨光光影，暖色调运镜，突出机身线条质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-tvc-product", "name": "产品广告片", "description": "TVC/开箱种草/反应展示", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_GKm4xKVRzR8h" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-tvc-product-sage-toaster-morning-kitchen", "title": "鼠尾草绿吐司机晨光广告片", "title_en": "Sage Toaster Morning Ad", "subtitle": "晨光洒落厨房台面，面包机功能演示自然又治愈", "subtitle_en": "A toaster ad shot in warm morning kitchen light", "description": "无需实拍团队和道具厨房，上传商品图就能生成温馨的晨间厨房广告片。鼠尾草绿吐司机摆在暖光台面上，面包烘烤冒热气的细节清晰可见，画面自然治愈。适合小家电、厨房用品做功能演示和种草内容，用生活化场景放大产品的实用与美感。", "categories": [ "creation" ], "tags": [ "产品广告片", "小家电广告", "厨房用品", "功能演示", "晨间生活场景" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "鼠尾草绿吐司机广告片｜竖版视频" } ], "prompt": "[产品广告片]制作产品广告片。\n商品图片：（请上传：商品图片）\n商品名称：鼠尾草绿双槽吐司机\n广告类型：功能演示\n投放平台：小红书\n其他要求：晨间厨房暖光实景，面包烘烤细节特写，画面治愈自然\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-tvc-product", "name": "产品广告片", "description": "TVC/开箱种草/反应展示", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_GKm4xKVRzR8h" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ugc-commerce-chips-couch-snack-review", "title": "薯片沙发追剧口播带货", "title_en": "Chips Couch Snack Review", "subtitle": "居家追剧场景真实试吃，口感描述接地气", "subtitle_en": "A cozy couch snack review for a chips brand", "description": "零食类目最有效的种草方式，是让人看到真实的吃播反应。上传商品图，AI 生成居家沙发追剧时随手拿薯片试吃的口播片段，语气轻松像朋友分享，没有距离感。适合休闲零食做短视频带货素材，快速产出适合短视频平台、快手投放的内容。", "categories": [ "creation" ], "tags": [ "口播带货", "薯片", "居家场景", "零食种草", "真实口播" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "薯片口播带货短视频｜竖版视频" } ], "prompt": "[口播带货]制作UGC口播带货视频。\n商品图片：（请上传：商品图片）\n商品名称：海盐波纹薯片\n核心卖点：薄脆不腻口，海盐咸香刚刚好\n目标平台：抖音\n其他要求：居家沙发实景，追剧场景边吃边聊，语气轻松真实\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ugc-commerce", "name": "口播带货", "description": "UGC风格真人口播带货视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_3vDxCR5xwS2k" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ugc-commerce-kitchen-storage-jar-review", "title": "食品密封罐口播带货", "title_en": "Kitchen Jar UGC Review", "subtitle": "居家口播实拍，手把手安利厨房收纳好物", "subtitle_en": "A real kitchen talk-through that sells the storage hack", "description": "上传商品图，AI 自动生成UGC口播带货视频：素人在自家厨房手持商品对镜讲解，自然的语气和真实场景比硬广更有说服力。适合家居日用、厨房小家电类商品卖家，用真人口碑代替传统广告，降低用户信任门槛，提升种草转化。", "categories": [ "creation" ], "tags": [ "口播带货", "食品密封罐", "厨房好物", "真人测评", "UGC" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "食品密封罐口播带货｜竖版UGC视频" } ], "prompt": "[口播带货]制作UGC口播带货视频。\n商品图片：（请上传：商品图片）\n商品名称：按压密封食品罐\n核心卖点：一按即密封，厨房收纳防潮防虫\n目标平台：\n其他要求：居家厨房实景，素人手持商品对镜口播讲解卖点\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ugc-commerce", "name": "口播带货", "description": "UGC风格真人口播带货视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_3vDxCR5xwS2k" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ugc-commerce-knit-blanket-livingroom", "title": "针织盖毯居家口播带货", "title_en": "Knit Blanket Living Room Ad", "subtitle": "客厅实景抱毯展示，语气亲切像朋友安利", "subtitle_en": "A warm living-room talking-head ad for a knit blanket", "description": "家纺类产品的说服力来自真实的触感展示，抱在怀里的一瞬间比文字描述更直接。上传商品图，AI 生成客厅暖光下抱着盖毯出镜讲解的口播视频，语气亲切自然，像朋友之间的日常安利。适合家纺、家居好物做种草短视频。", "categories": [ "creation" ], "tags": [ "口播带货", "针织盖毯", "居家好物", "客厅场景", "真实口播" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "针织盖毯口播带货短视频｜竖版视频" } ], "prompt": "[口播带货]制作UGC口播带货视频。\n商品图片：（请上传：商品图片）\n商品名称：浅灰针织盖毯\n核心卖点：厚实保暖，手感柔软不掉毛\n目标平台：小红书\n其他要求：客厅沙发实景出镜，语气亲切像朋友安利，展示毯子质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ugc-commerce", "name": "口播带货", "description": "UGC风格真人口播带货视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_3vDxCR5xwS2k" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ugc-commerce-lemon-juicer-kitchen-demo", "title": "柠檬榨汁器厨房口播带货", "title_en": "Lemon Juicer Kitchen Demo Ad", "subtitle": "厨房实景边操作边讲解，卖点一目了然", "subtitle_en": "A kitchen talking-head demo for a lemon squeezer", "description": "厨房小工具最适合边操作边讲解的口播形式，动作演示比单纯口述更有说服力。上传商品图，AI 生成厨房实景中一压出汁的操作过程，配合自然随和的讲解语气，把省力不脏手的卖点直观呈现。适合厨房好物做带货口播视频。", "categories": [ "creation" ], "tags": [ "口播带货", "榨汁器", "厨房好物", "真人演示", "真实口播" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "柠檬榨汁器口播带货短视频｜竖版视频" } ], "prompt": "[口播带货]制作UGC口播带货视频。\n商品图片：（请上传：商品图片）\n商品名称：手动柠檬榨汁器\n核心卖点：一压出汁，省力不脏手\n目标平台：TikTok\n其他要求：厨房实景边操作边讲解，动作演示清晰，语气随和自然\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-ugc-commerce", "name": "口播带货", "description": "UGC风格真人口播带货视频", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_3vDxCR5xwS2k" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-ugc-lifestyle-street-bench-outfit", "title": "街拍慵懒穿搭生活照", "title_en": "Street Bench Outfit Photo", "subtitle": "服饰变身自然街拍，种草感拉满", "subtitle_en": "Turn your garment into an effortless street-style shot", "description": "上传服饰单品，AI生成社媒风格的街拍生活照，奶油黄衬衫叠白色吊带背心配碎花中长裙，坐在街边长椅手拿饮品，姿态自然不摆拍。适合服饰卖家和达人快速产出种草感十足的日常穿搭图，替代真人外拍。", "categories": [ "creation" ], "tags": [ "种草穿搭", "街拍", "UGC生活照", "碎花裙", "社媒风格" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "街拍种草穿搭图｜单张 JPG" } ], "prompt": "[UGC穿搭]生成UGC风格的穿搭生活照。\n服饰图片：（请上传：服饰图片）\n生活场景：街拍\n模特风格：欧美女性\n图片比例：3:4\n生成张数： 张\n其他要求：奶油黄开衫衬衫叠白色吊带背心，碎花中长裙，坐在街边长椅，手拿绿色饮品，慵懒随性\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-ugc-lifestyle", "name": "UGC生活照", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297369362432" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-utility-poster-autumn-equinox-daily", "title": "秋分节气日签海报", "title_en": "Autumn Equinox Daily Poster", "subtitle": "桂花插画配节气祝语，适合日更", "subtitle_en": "Osmanthus illustration paired with a seasonal greeting", "description": "品牌号每个节气都要发一张图，套模板容易土，找设计又排不上队。这版用米白宣纸底配桂花枝插画，毛笔「秋分」压在右上，节气日期、农历和四句祝语分散在留白处，安静好看不抢内容。适合品牌账号、门店、个人号做节气与日签内容。", "categories": [ "creation" ], "tags": [ "通知与日常物料", "节气海报", "日签", "品牌日更" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "秋分节气日签海报｜竖版单张" } ], "prompt": "[通知与日常物料]制作通知与日常物料。\n海报类型：日签/月签\n输入词：秋分节气日签海报，米白宣纸底配黛绿与橙黄；主体是一枝手绘桂花，花朵橙黄、叶片墨绿，从左下向上斜伸；右上角毛笔大字「秋分」，旁边小印章「节气」；左上角英文「Autumn Awakening」配「一叶知秋意 / 风起桂花香」；「秋分」下方竖排「昼夜均分 美好皆如愿」；右侧标注日期「09 / 23」与「农历八月初二」，下方「愿你所得皆所愿 / 所行化坦途」；左下角小字「新季开始 / 心向远方」\n素材图片：（请上传：素材图片）\n使用场景：线上\n是否需要动态海报：不需要\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-utility-poster", "name": "通知与日常物料", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297511968768" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-utility-poster-award-news-gold", "title": "企业获奖黑金喜报", "title_en": "Corporate Award Announcement", "subtitle": "奖杯主视觉配七条成绩标签", "subtitle_en": "A gold trophy key visual with achievement tags", "description": "公司拿了奖想发朋友圈，自己排的喜报要么字大得像标语，要么成绩堆成一列看不完。这版用黑金底色配中央金色奖杯，书法「喜报」压顶，核心荣誉居中，技术与口碑数据做成环绕的弧形标签，底部留联系方式。适合企业市场部、品牌公关发布获奖与里程碑消息。", "categories": [ "creation" ], "tags": [ "通知与日常物料", "喜报", "企业宣传", "获奖公告" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "企业获奖喜报海报｜竖版单张" } ], "prompt": "[通知与日常物料]制作通知与日常物料。\n海报类型：喜报/表彰\n输入词：企业获奖喜报，黑金配色，顶部书法大字「喜报」，副标题「智驭未来·荣耀登顶」，核心荣誉「MT 汽车斩获 2026 年度中国智能出行标杆品牌」，成绩说明「续航突破 1 千 km × 充电 10 分钟补能 3 百 km 重定义电动效率标准」；画面中央金色奖杯立在圆形台座上，两侧信息框「全栈自研 / 从芯片到算法 核心技术自主可控」「续航 1000km / 行业最长实测续航 告别里程焦虑」，奖杯周围环绕弧形标签：电池终身质保承诺、全国超充站破万座、充电百加速 3.2 秒、智能座舱语音响应<0.3 秒、用户满意度 98.6%、全栈自研智驾系统；底部留电话与地址栏并配二维码位\n素材图片：（请上传：素材图片）\n使用场景：线上\n是否需要动态海报：不需要\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-utility-poster", "name": "通知与日常物料", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297511968768" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-utility-poster-childrens-day-program", "title": "六一儿童节节目单", "title_en": "Children's Day Program List", "subtitle": "手绘幕布风排出六个节目顺序", "subtitle_en": "Six acts laid out on a hand-drawn stage curtain", "description": "幼儿园排完节目，老师常常只能在文档里列一份表格发给家长，没有仪式感。这版把节目单放进红色幕布舞台里，六个节目按序号配彩色圆点，节目名、形式和班级分行写清，四周点缀小熊、气球和吹小号的小朋友。适合幼儿园、小学、亲子活动做演出物料。", "categories": [ "creation" ], "tags": [ "通知与日常物料", "节目单", "幼儿园", "儿童节", "活动物料" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "六一表演节目单海报｜竖版单张" } ], "prompt": "[通知与日常物料]制作通知与日常物料。\n海报类型：通知公告\n输入词：六一儿童节表演节目单，手绘插画风，红色舞台幕布围边，米白纸质底，顶部挂彩色三角旗与音符；标题「六一儿童节 表演节目单」用彩色手写体；六个节目按序号排列：1《阳光下的微笑》舞蹈-小星星班；2《小小梦想家》合唱-向日葵班；3《奇妙的森林》情景剧-月亮班；4《一起去远方》舞蹈-彩虹班；5《童年的歌》独唱-小百灵班；6《快乐大联欢》集体表演-大班小朋友；画面点缀戴生日帽的小熊、吹小号的女孩、打鼓的男孩、拉小提琴的女孩、气球与小花草地\n素材图片：（请上传：素材图片）\n使用场景：都要\n是否需要动态海报：不需要\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-utility-poster", "name": "通知与日常物料", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297511968768" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-utility-poster-creative-hiring-flyer", "title": "创意团队招聘海报", "title_en": "Creative Team Hiring Poster", "subtitle": "撕纸拼贴风，岗位薪资一目了然", "subtitle_en": "Torn-paper collage style with clear job listings", "description": "输入招聘主题、岗位名称和薪资区间，自动生成撕纸拼贴风格的招聘海报，UI/UX设计师、产品运营、内容编辑等岗位信息层次清晰，附简历投递方式，适合创意公司社交媒体招聘发布。", "categories": [ "creation" ], "tags": [ "通知与日常物料", "招聘海报", "撕纸拼贴风", "社交媒体", "岗位招聘" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "创意岗位招聘海报｜单张 PNG" } ], "prompt": "[通知与日常物料]制作通知与日常物料。\n海报类型：招聘海报\n输入词：标题「加入我们 创意无限」；岗位1：UI/UX设计师，18-25k，需具备出色审美与用户体验意识，熟练使用设计工具，参与产品从概念到落地全流程；岗位2：产品运营专员，12-18k，数据敏感、善于洞察用户需求，制定并执行运营策略；岗位3：内容策划编辑，10-15k，热爱内容创作，负责选题撰写与优化；简历投递邮箱 hr@creativehub.com；结尾标语「创想未来 灵感不设限」\n素材图片：（请上传：素材图片）\n使用场景：线上\n是否需要动态海报：不需要\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [ { "id": "designkit-utility-poster", "name": "通知与日常物料", "description": "使用当前场景对应的美图设计室Skill。", "skill_id": "skill_2091019297511968768" } ], "experts": [], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-video-post-remove-logo-watermark", "title": "视频去水印处理", "title_en": "Video Watermark Removal", "subtitle": "一键去除画面水印，背景细节完整保留", "subtitle_en": "Erase the watermark, keep every background detail", "description": "上传带水印的视频，AI 自动识别并去除画面中的LOGO水印，同时保持背景家具纹理、光影过渡自然，看不出修补痕迹。适合需要二次剪辑发布素材、却带有版权水印或平台标识的内容创作者和商家，省去逐帧手动修补的麻烦。", "categories": [ "creation" ], "tags": [ "视频后期处理", "去水印", "素材二次剪辑", "画面修复", "家居视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "视频去水印｜背景无痕修复" } ], "prompt": "[视频后期处理]对上传的视频进行处理。\n上传视频：（请上传：上传视频）\n处理类型：去水印\n处理说明：去除画面中央的LOGO水印，背景家具纹理和光影过渡保持自然无痕\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-video-post", "name": "视频后期处理", "description": "画质增强/元素消除/音频处理", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_wPFePs3YoKpM" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-note-recreate-exam-timeline-remix", "title": "国考时间轴图文复刻", "title_en": "Exam Timeline Remix", "subtitle": "复刻爆款笔记排版，换成自己的内容主题", "subtitle_en": "Recreate a viral note layout with your own content", "description": "看到别人的爆款笔记想要同款效果，却不知道怎么排版？上传参考图和自己的内容资料，AI 保留原版的时间轴结构和信息层级，替换配色文案，快速产出同类型的高完读图文，适合考公考编、活动预热等信息型选题。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "备考攻略", "时间轴图文", "考公考编", "信息图排版" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "国考备考时间轴图文｜单张" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：一张备考理财主题的时间轴笔记截图\n你的素材图：（请上传：你的素材图）\n新图文主题：国考重要时间节点攻略\n复刻模式：高度复刻\n画面比例：3:4\n文案&其他要求：保留原图的时间轴结构和四阶段划分，替换配色为黄蓝撞色手账风，标题改为国考重要时间节点，突出必看心里有数很重要的副标题\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-note-recreate", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Wnxe9anIQYuo" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-note-recreate-fish-oil-supplement", "title": "保健品种草图文复刻", "title_en": "Fish Oil Supplement Note Remix", "subtitle": "复刻水果安利笔记的手写批注风格", "subtitle_en": "Reuse a fruit haul's handwritten annotation style", "description": "看到一篇水蜜桃种草笔记的手写批注风格很有代入感，想用同样的感觉安利自己的保健品？上传参考笔记和产品实拍图，AI保留原版的手写体标注、清单式功效点和居家氛围场景，把内容替换成你的产品。适合保健品和个护卖家快速产出真实感种草笔记，无需自己构思排版。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "保健品种草", "好物安利", "居家实拍", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "保健品种草复刻｜单张" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：一篇水蜜桃种草笔记参考图\n你的素材图：鱼油保健品瓶身实拍图\n新图文主题：每天1到2粒的保养精华：鱼油保健品种草\n复刻模式：仅换风格\n画面比例：3:4\n文案&其他要求：保留原版手写体标注、居家梳妆台氛围和功效清单排版，替换为保健品实拍内容\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-note-recreate", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Wnxe9anIQYuo" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-note-recreate-small-city-travel-route", "title": "小城漫游路线图文复刻", "title_en": "Small City Travel Route Remix", "subtitle": "复刻爆款拼贴风格，美食笔记变身旅行攻略", "subtitle_en": "Reuse a viral collage style for a travel guide", "description": "刷到一篇拼贴风美食探店笔记很吸睛，但你想做的是旅行攻略？上传参考笔记和你的旅拍素材，AI保留原版荧光撞色、复古报纸底纹和路线打卡贴纸的视觉风格，把内容整体换成你的旅行路线。适合旅行博主和目的地账号快速产出高识别度攻略笔记，出图即可发布。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "旅行攻略", "拼贴风格", "小城旅拍", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "小城漫游路线复刻｜单张" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：一篇拼贴风美食探店笔记参考图\n你的素材图：小城古街、江景、咖啡店等旅拍素材图\n新图文主题：小城漫游旅行路线：一日游两日游三日游任选\n复刻模式：仅换风格\n画面比例：3:4\n文案&其他要求：保留原版荧光蓝底、复古报纸纹理、大字标题和路线打卡贴纸风格，替换为旅行路线内容\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-note-recreate", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Wnxe9anIQYuo" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-note-recreate-weekly-shoe-lookbook", "title": "一周鞋履穿搭复刻", "title_en": "Weekly Shoe Lookbook Remix", "subtitle": "复刻爆款穿搭网格，换成你的鞋履库存", "subtitle_en": "Reuse a viral outfit grid for your own shoe lineup", "description": "看到别人的「一周穿搭Lookbook」笔记很火，却不知道怎么套用到自己的产品？上传参考笔记和你的鞋履素材图，AI高度复刻原版网格排版：Day1到Day7分类标签、复古皮质底纹、风格关键词全部保留，只换成你的鞋款和文案。适合鞋履卖家和穿搭博主快速产出结构化种草笔记。", "categories": [ "creation" ], "tags": [ "爆款图文复刻", "穿搭笔记", "鞋履种草", "网格排版", "社媒笔记" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "一周鞋履Lookbook复刻｜单张" } ], "prompt": "[爆款图文复刻]参考上传的爆款、往期内容或风格图，按所选复刻模式生成新的社媒图文封面/套图。\n参考图文/往期内容：一篇「一周通勤Lookbook」穿搭笔记参考图\n你的素材图：7双不同风格鞋履的产品图\n新图文主题：一周鞋履Lookbook：七天七双风格七种心情\n复刻模式：高度复刻\n画面比例：3:4\n文案&其他要求：保留原版复古皮质纹理底色、Day1到Day7网格排版和英文标语，只替换为鞋履内容\n生成张数：1 张\n目标平台：小红书\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-note-recreate", "name": "爆款图文复刻", "description": "拆解爆款、往期或参考风格，生成同类图文", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_Wnxe9anIQYuo" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-replication-bullet-time-diner", "title": "子弹时间特效复刻", "title_en": "Bullet Time Effect Replication", "subtitle": "复刻慢镜定格特效，复古场景瞬间变大片", "subtitle_en": "Replicate a bullet-time freeze effect", "description": "子弹时间是经典又常青的爆款结构，环绕定格的瞬间总能让人多停留几秒。上传你的照片，AI 复刻同款慢镜定格特效和复古场景氛围，突出人物姿态和场景细节，不需要专业设备也能拍出电影感画面。适合想出片惊艳的个人内容。", "categories": [ "creation" ], "tags": [ "爆款视频复刻", "子弹时间", "复古场景", "创意特效", "短视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "子弹时间特效复刻视频｜竖版视频" } ], "prompt": "[爆款视频复刻]复刻同结构短视频。\n参考视频：（请上传：参考视频）\n参考视频链接：\n你的图片：（请上传：你的图片）\n名称/卖点：复古造型照\n其他要求：复刻同款子弹时间定格特效，环绕慢镜头突出人物姿态和场景氛围\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-replication", "name": "爆款视频复刻", "description": "上传参考视频，同款复刻成片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_4APPLfQ6W8YB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-replication-motion-tracking-overlay", "title": "动态跟踪运镜特效复刻", "title_en": "Motion Tracking Effect Replication", "subtitle": "复刻科技感跟踪镜头，画面自带数据浮层光效", "subtitle_en": "Replicate a sci-fi motion-tracking camera effect", "description": "带数据浮层的跟踪运镜是近期很火的一种爆款结构，视觉冲击力强却很难自己拍出来。上传你的照片，AI 复刻同款动态追踪镜头和科技感数据特效，保留主体的运动轨迹和速度感。适合运动、潮流类内容复刻同款热门效果。", "categories": [ "creation" ], "tags": [ "爆款视频复刻", "动态跟踪", "科技特效", "创意运镜", "短视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "动态跟踪运镜特效复刻视频｜竖版视频" } ], "prompt": "[爆款视频复刻]复刻同结构短视频。\n参考视频：（请上传：参考视频）\n参考视频链接：\n你的图片：（请上传：你的图片）\n名称/卖点：个人运动照\n其他要求：复刻同款动态追踪镜头和数据浮层特效，保留主体运动轨迹\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-replication", "name": "爆款视频复刻", "description": "上传参考视频，同款复刻成片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_4APPLfQ6W8YB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-viral-replication-orbit-camera-rooftop", "title": "360°环绕运镜复刻爆款", "title_en": "Orbit Camera Effect Replication", "subtitle": "复刻同款环绕运镜，把你的照片拍成城市大片", "subtitle_en": "Replicate the viral 360-degree orbit camera effect", "description": "很多爆款视频火的不是内容本身，而是那个让人上头的运镜手法。上传你的照片，AI 复刻同款360度环绕运镜和城市天际线背景，保留人物姿态和动感氛围，不用自己研究运镜技巧也能拍出同样的效果。适合想复刻热门运镜风格的个人或品牌素材。", "categories": [ "creation" ], "tags": [ "爆款视频复刻", "环绕运镜", "城市大片", "创意特效", "短视频" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "360°环绕运镜复刻视频｜竖版视频" } ], "prompt": "[爆款视频复刻]复刻同结构短视频。\n参考视频：（请上传：参考视频）\n参考视频链接：\n你的图片：（请上传：你的图片）\n名称/卖点：个人形象照\n其他要求：复刻同款360度环绕运镜和城市天际线背景，保留人物姿势和动感氛围\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-viral-replication", "name": "爆款视频复刻", "description": "上传参考视频，同款复刻成片", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_4APPLfQ6W8YB" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-white-bg-bow-heart-necklace", "title": "蝴蝶结爱心项链白底图", "title_en": "Bow Heart Necklace White BG", "subtitle": "大理石台面随手拍变身电商白底图，金属光泽清晰呈现", "subtitle_en": "Turn a marble-top snapshot into a clean jewelry listing image", "description": "首饰随手拍放在大理石桌面上，背景有污渍且比例失真，很难直接用作主图。上传随手拍，AI 自动精修并铺上纯白背景，蝴蝶结镂空纹理和爱心吊坠的金属光泽清晰还原，链条细节不糊，适合饰品类卖家批量产出电商白底图。", "categories": [ "creation" ], "tags": [ "白底精修图", "商品精修", "项链主图", "饰品白底图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "项链白底图｜单张 JPG" } ], "prompt": "[白底精修图]将上传的图片精修为白底图。\n商品图片：一张项链平放在浅灰色大理石桌面上的随手拍，桌面有污渍\n陈列模式：立体站姿展示\n输出比例：1:1\n其他要求：蝴蝶结与爱心吊坠为镀银款式，需还原金属光泽和链条细节，去除背景污渍\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-white-bg", "name": "白底精修图", "description": "白底商品精修图，电商主图规格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_a7NHYcyKxtSR" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-white-bg-burgundy-hobo-bag", "title": "酒红色月牙包白底图", "title_en": "Burgundy Hobo Bag White BG", "subtitle": "椅背随手拍秒变电商白底图，皮革质感与五金件清晰呈现", "subtitle_en": "Turn a chair-back snapshot into a clean handbag listing image", "description": "包包随手挂在木椅背上拍摄，背景杂乱还带着沙发和椅子，直接用来做主图不够专业。上传随手拍，AI 自动精修并铺上纯白背景，皮革纹理和金色五金件细节清晰呈现，包型饱满立体，适合箱包类卖家批量产出符合平台规范的白底图。", "categories": [ "creation" ], "tags": [ "白底精修图", "商品精修", "手袋主图", "箱包白底图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "手袋白底图｜单张 JPG" } ], "prompt": "[白底精修图]将上传的图片精修为白底图。\n商品图片：一张酒红色月牙包挂在木椅背上的随手拍，背景有沙发和餐桌\n陈列模式：立体站姿展示\n输出比例：1:1\n其他要求：包身为酒红色皮革，金色五金件需清晰呈现，去除背景杂物和椅子\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-white-bg", "name": "白底精修图", "description": "白底商品精修图，电商主图规格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_a7NHYcyKxtSR" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-white-bg-puffer-jacket-clean", "title": "羽绒服白底精修图", "title_en": "Puffer Jacket White BG", "subtitle": "生活随手拍秒变电商级白底图，商品细节纤毫毕现", "subtitle_en": "Turn a casual photo into a clean white-background listing image", "description": "生活随手拍的衣服图光线暗、背景乱，直接放到店铺显得不专业。上传一张随手拍，AI 自动修图并铺上纯白背景，光线均匀、版型挺括、Logo 文字清晰保留，直接达到电商主图规格。适合服装、饰品类卖家批量产出白底图，无需再单独请摄影棚拍摄。", "categories": [ "creation" ], "tags": [ "白底精修图", "商品精修", "羽绒服主图", "电商白底图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "羽绒服白底图｜单张 JPG" } ], "prompt": "[白底精修图]将上传的图片精修为白底图。\n商品图片：一件挂在衣架上的浅杏色羽绒服生活随手拍\n陈列模式：立体站姿展示\n输出比例：1:1\n其他要求：保留胸口 Sweet bear 品牌字样，突出羽绒服细节压线质感\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-white-bg", "name": "白底精修图", "description": "白底商品精修图，电商主图规格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_a7NHYcyKxtSR" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meitu-white-bg-sage-headphones-clean", "title": "鼠尾草绿耳机白底精修图", "title_en": "Sage Green Headphones White BG", "subtitle": "桌面随手拍秒变电商白底图，织物纹理清晰保留", "subtitle_en": "Turn a desk snapshot into a clean white-background listing image", "description": "耳机随手拍在木桌上，光线杂乱、背景有数据线和电脑，直接放店铺显得不专业。上传随手拍，AI 自动精修并铺上纯白背景，头梁织物纹理和金属连接件细节清晰保留，直接达到电商主图规格，适合3C数码类卖家快速产出白底图，无需摄影棚。", "categories": [ "creation" ], "tags": [ "白底精修图", "商品精修", "耳机主图", "电商白底图" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.jpg", "artifacts": [ { "file": "output.jpg", "label": "耳机白底图｜单张 JPG" } ], "prompt": "[白底精修图]将上传的图片精修为白底图。\n商品图片：一张鼠尾草绿耳机放在木质桌面上的随手拍，旁边有笔记本电脑和数据线\n陈列模式：立体站姿展示\n输出比例：1:1\n其他要求：耳机颜色为鼠尾草绿，需保留头梁与耳罩织物纹理质感和金属连接杆细节\n\n工具约束：图片生成或编辑仅使用此 Buddy 应用内置的 Meitu Designkit CLI 连接器；不得调用 ImageGen、ImageEdit 或 Defer(ImageGen)。CLI 无法完成时告知用户，不回退到 ImageGen。", "skills": [], "experts": [ { "id": "designkit-white-bg", "name": "白底精修图", "description": "白底商品精修图，电商主图规格", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_a7NHYcyKxtSR" } ], "mcps": [], "creator": { "name": "美图设计室", "name_en": "Meitu DesignKit", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "featured_order": 0, "quality_score": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-10-12T16:34:35+08:00", "updated_at": "2026-10-12T16:34:35+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "member-asset", "title": "会员资产运营：储值卡券积分，沉淀怎么盘活", "title_en": "Activate idle stored-value, coupons and points", "subtitle": "储值卡券积分，沉淀怎么盘活", "subtitle_en": "Activate idle stored-value, coupons and points", "description": "面向会员运营场景：储值卡券积分，沉淀怎么盘活。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：储值卡券积分，沉淀怎么盘活。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "储值", "卡券", "积分" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {组织范围} 储值、余额、卡券和积分使用情况，找出沉淀与转化机会。\n1) 资产沉淀结构\n2) 临期/沉寂资产风险\n3) 盘活动作与目标\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {组织范围} 储值、余额、卡券和积分使用情况，找出沉淀与转化机会。\n1) 资产沉淀结构\n2) 临期/沉寂资产风险\n3) 盘活动作与目标\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087434010556817408", "skill_id": "skill_2087434010556817408", "name": "会员资产运营", "description": "储值卡券积分分析" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-growth", "title": "增长与生命周期：拉新留存流失，下一步怎么做", "title_en": "Acquire, retain, win back: next moves", "subtitle": "拉新留存流失，下一步怎么做", "subtitle_en": "Acquire, retain, win back: next moves", "description": "面向会员运营场景：拉新留存流失，下一步怎么做。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：拉新留存流失，下一步怎么做。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "拉新", "留存", "流失", "召回" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {组织范围} 在 {统计区间} 的新增、来源、留存和流失，提出拉新与召回建议。\n1) 增长漏斗数字\n2) 流失人群特征\n3) 下一步动作（渠道/权益/触达）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {组织范围} 在 {统计区间} 的新增、来源、留存和流失，提出拉新与召回建议。\n1) 增长漏斗数字\n2) 流失人群特征\n3) 下一步动作（渠道/权益/触达）\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087433926666543104", "skill_id": "skill_2087433926666543104", "name": "增长与生命周期", "description": "会员拉新留存分析" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-marketing", "title": "营销与全渠道运营：发券核销到店，渠道表现如何", "title_en": "Coupon-to-visit funnel and channel performance", "subtitle": "发券核销到店，渠道表现如何", "subtitle_en": "Coupon-to-visit funnel and channel performance", "description": "面向会员运营场景：发券核销到店，渠道表现如何。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：发券核销到店，渠道表现如何。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "发券", "核销", "全渠道" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请复盘 {组织范围} 在 {活动/区间} 的触达、发券核销和到店转化，并诊断各渠道会员表现。\n1) 漏斗：触达→核销→到店→消费\n2) 渠道对比与异常\n3) 今日调整建议\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请复盘 {组织范围} 在 {活动/区间} 的触达、发券核销和到店转化，并诊断各渠道会员表现。\n1) 漏斗：触达→核销→到店→消费\n2) 渠道对比与异常\n3) 今日调整建议\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087433989532286976", "skill_id": "skill_2087433989532286976", "name": "营销与全渠道运营", "description": "营销渠道诊断" }, { "id": "skill_2087433369620779008", "skill_id": "skill_2087433369620779008", "name": "会员营销", "description": "会员营销复盘" }, { "id": "skill_2087433192597749760", "skill_id": "skill_2087433192597749760", "name": "全渠道运营", "description": "全渠道触达分析" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-overview", "title": "会员经营概览：会员大盘：规模活跃储值贡献", "title_en": "Member dashboard: scale, activity, stored-value", "subtitle": "会员大盘：规模活跃储值贡献", "subtitle_en": "Member dashboard: scale, activity, stored-value", "description": "面向会员运营场景：会员大盘：规模活跃储值贡献。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：会员大盘：规模活跃储值贡献。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "会员大盘", "活跃", "储值" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请汇总 {组织范围} 在 {统计区间} 的会员规模、活跃、消费转化、储值与会员营收贡献。\n1) 大盘结论\n2) 关键点与机会点\n3) 今日运营三件事\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请汇总 {组织范围} 在 {统计区间} 的会员规模、活跃、消费转化、储值与会员营收贡献。\n1) 大盘结论\n2) 关键点与机会点\n3) 今日运营三件事\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087432714468065280", "skill_id": "skill_2087432714468065280", "name": "会员经营概览", "description": "会员大盘概览" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-profile", "title": "画像与体验洞察：高净值偏好与差评怎么改", "title_en": "High-value preferences and review fixes", "subtitle": "高净值偏好与差评怎么改", "subtitle_en": "High-value preferences and review fixes", "description": "面向会员运营场景：高净值偏好与差评怎么改。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：高净值偏好与差评怎么改。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "画像", "高净值", "评价" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请识别 {组织范围} 高净值会员、消费偏好和差评原因，形成分层服务建议。\n1) 高净值画像要点\n2) 差评主题与可改点\n3) 分层服务动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请识别 {组织范围} 高净值会员、消费偏好和差评原因，形成分层服务建议。\n1) 高净值画像要点\n2) 差评主题与可改点\n3) 分层服务动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087433029064515584", "skill_id": "skill_2087433029064515584", "name": "画像与体验洞察", "description": "高净值画像分析" }, { "id": "skill_2087433226152181760", "skill_id": "skill_2087433226152181760", "name": "会员画像", "description": "会员消费偏好画像" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-rfm", "title": "消费行为与RFM：RFM 分层看清该保该挽", "title_en": "RFM tiers: who to keep and who to win back", "subtitle": "RFM 分层看清该保该挽", "subtitle_en": "RFM tiers: who to keep and who to win back", "description": "面向会员运营场景：RFM 分层看清该保该挽。演示数据样例，用于灵感页展示与「制作我的版本」。", "description_en": "面向会员运营场景：RFM 分层看清该保该挽。演示数据样例，用于灵感页展示与「制作我的版本」。", "categories": [ "office" ], "tags": [ "RFM", "复购", "分层" ], "tags_en": [], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "演示页面" }, { "file": "output.md", "label": "Markdown原文" } ], "prompt": "请分析 {组织范围} 会员复购率、消费趋势和 RFM 分层，识别重点保持与挽留人群。\n1) RFM 分层规模与贡献\n2) 该保/该挽人群清单逻辑\n3) 分层运营动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "prompt_en": "请分析 {组织范围} 会员复购率、消费趋势和 RFM 分层，识别重点保持与挽留人群。\n1) RFM 分层规模与贡献\n2) 该保/该挽人群清单逻辑\n3) 分层运营动作\n要求：一次输出静态 HTML fragment；可含一个简短 style，禁止 script、事件属性和完整网页外壳。", "skills": [ { "id": "skill_2087433769783300096", "skill_id": "skill_2087433769783300096", "name": "消费行为与RFM", "description": "RFM分层分析" } ], "experts": [ { "id": "weimob-member-ops-expert", "expert_id": "ex_gjLWSDc1jRPn", "name": "微盟会员运营专家", "description": "微盟会员运营分析专家" } ], "mcps": [], "creator": { "name": "微盟", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T07:54:07Z", "updated_at": "2026-08-31T07:54:07Z", "is_recommended": false, "recommended_order": 0 }, { "id": "member-segment-outreach-plan", "title": "会员分层触达方案", "title_en": "Segmented Member Outreach Plan", "subtitle": "基于会员价值与近期消费的差异化触达策略", "subtitle_en": "A practical Weimob playbook for segmented member outreach plan", "description": "做会员分层触达方案时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“会员分层触达方案”展开：结合会员等级、近30天消费金额和最近购买时间，为不同客户分层匹配商品、优惠权益、一对一话术和触达顺序。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、私域触达运营运营人员、经营分析团队。", "description_en": "The hardest part of segmented member outreach plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates segmented member outreach plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Private-Traffic Outreach specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "会员分层", "商品推荐", "优惠权益", "一对一话术", "触达顺序" ], "tags_en": [ "Segmented Member Outreach Plan", "Customer Operations", "Private-Traffic Outreach", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "结合会员等级、近30天消费金额和最近购买时间，为不同客户分层设计差异化商品推荐、优惠权益及一对一沟通话术，并给出建议触达顺序。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a segmented member outreach plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "member-tagging-pilot", "title": "会员标签体系与试打标", "title_en": "Member Tagging Framework and Pilot", "subtitle": "会员标签定义、更新规则与分组运营建议", "subtitle_en": "A practical Weimob playbook for member tagging framework and pilot", "description": "做会员标签体系与试打标时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“会员标签体系与试打标”展开：按照消费行为、客户价值和生命周期梳理会员标签体系，形成可执行的标签定义、更新规则和分组运营建议，并在用户确认后对指定测试会员验证单次打标。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、客户分层洞察运营人员、经营分析团队。", "description_en": "The hardest part of member tagging framework and pilot is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates member tagging framework and pilot through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Customer Segmentation specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "会员标签", "标签规划", "消费行为", "生命周期", "试打标" ], "tags_en": [ "Member Tagging Framework and Pilot", "Customer Operations", "Customer Segmentation", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我按消费行为、客户价值和生命周期梳理会员标签体系，给出标签定义、建议打标条件、更新频率和分组运营建议；如需验证，请先让我指定一个测试会员和现有标签，确认后只执行单次试打标。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a member tagging framework and pilot covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "merchant-promo-designer-homepage", "title": "商家宣传设计师介绍主页", "title_en": "Merchant Promo Designer Homepage", "subtitle": "用作品介绍宣传页与战报海报，一页直达专家", "subtitle_en": "Showcase promo pages and sales posters, then launch the expert", "description": "只说“能做宣传页、能做海报”，商家很难知道成品是什么样，也不知道该怎么开始。介绍专家，不如先把作品摆出来。\n\n这个主页用“让你的店上镜”开场，把商品宣传页放进取景框。往下看，同一件商品的不同投放尺寸、九家店各自的宣传页首屏、朋友圈长图和销售战报依次展开；可以切换宣传页作品、放大战报，再看四步合作流程。页尾直接召唤「商家宣传设计师」，也给出手动搜索入口。\n\n适合要展示作品、介绍服务并引导用户开始使用的专家作者，也适合想先看看设计效果再提需求的商家。", "description_en": "Saying an expert can make promo pages and posters does not show merchants what the result looks like or how to get started. Showing the work is more convincing.\n\nThis homepage opens with “Put your shop in the spotlight” and a product promo page inside a viewfinder. Scroll to see the same product in different publishing formats, hero designs for nine shops, long images for social sharing and sales-report posters. Visitors can switch promo-page examples, enlarge the sales posters and follow a four-step collaboration guide. The final call to action launches Merchant Promo Designer, with a manual search fallback.\n\nUseful for expert creators who want to showcase their work, introduce their services and help visitors start using them, and for merchants who want to see the design before describing their needs.", "categories": [ "creation", "business" ], "tags": [ "专家介绍", "作品展示", "宣传主页", "商家设计" ], "tags_en": [ "Expert Introduction", "Portfolio Showcase", "Promo Homepage", "Merchant Design" ], "artifact_type": "link", "external_url": "https://merchant-promo-designer-87534.app.workbuddy.host", "preview": "", "artifacts": [], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "为 WorkBuddy 专家「商家宣传设计师」制作一个面向商家的公开介绍主页。它给商家做宣传页和销售战报海报：宣传页突出品牌或商品特点，引导进店；战报根据店铺和真实数字重新构思，不套固定模板。主页用作品让人看懂效果，再引导开始使用。\n\n首屏以「让你的店上镜」为主标题，用商品宣传页作为视觉主角，采用芥末黄、黑色和奶油色的大胆编辑式设计。滚动时展示同一商品在 4:5、1:1、9:16、16:9 中的重新排版，保留标题、文字和主体的完整性，不用直接裁剪代替适配。\n\n作品区展示九家不同店铺的宣传页首屏，每家写清从哪件独有的事出发、靠什么画面让人记住。首屏作品自动轮播，宣传页作品区支持手动、键盘和每五秒自动切换；浏览、聚焦作品或离开该区时暂停自动切换。再展示带页尾二维码的朋友圈长图，以及以花瓣、吊牌、收银小票等店铺物件构思的销售战报，战报可点击放大查看。演示商家、人物和数字明确标注为演示，不编造真实销量、评价或功效。\n\n合作流程分四步：说要做什么；一次问清一件事，战报数字先整理并让商家确认；给三个创意方向供选择；交付宣传页链接、长图或指定尺寸海报。说明商品图正式发布前用商家实拍替换，宣传只做展示和引流，下单在店铺完成。\n\n页尾放「直接召唤专家」按钮，唤起 WorkBuddy 并选择该专家，预填「我想为自己的小店做一个宣传主页，请你帮助我」。无法唤起时，提示打开 WorkBuddy 搜索「商家宣传设计师」。交付一个电脑与手机都能正常浏览、作品与导航可操作的公开主页链接。", "prompt_en": "Create a public homepage for the WorkBuddy expert Merchant Promo Designer, aimed at merchants. The expert makes promo pages and sales-report posters: promo pages highlight a brand or product and bring visitors to the shop, while sales posters are conceived afresh from the shop and its real figures rather than a fixed template. Show the work so visitors understand the result, then help them get started.\n\nOpen with “Put your shop in the spotlight” and a product promo page as the visual hero. Use a bold editorial design in mustard yellow, black and cream. As visitors scroll, show the same product re-laid out for 4:5, 1:1, 9:16 and 16:9. Keep the title, copy and main subject intact; do not replace format adaptation with cropping.\n\nShow promo-page hero designs for nine different shops. Explain the distinctive starting point and memorable visual idea for each. Automatically rotate the hero artwork. In the promo-page gallery, allow manual and keyboard switching and automatic rotation every five seconds; pause rotation while visitors hover over or focus on the artwork, or leave the section. Also show long images for social sharing with a QR code at the bottom, plus sales-report posters built around shop-specific objects such as petals, hangtags and receipts that visitors can click to enlarge. Clearly label demo shops, people and figures; never invent real sales, reviews or efficacy claims.\n\nExplain collaboration in four steps: describe the request; clarify one thing at a time and get merchant confirmation of sales-report figures first; offer three creative directions to choose from; deliver a promo-page link, a long image or a poster in the requested format. State that merchants should replace demo product images with their own photos before publishing. The pages showcase and drive traffic only; purchasing happens in the shop.\n\nEnd with a “Launch the expert” button that opens WorkBuddy, selects this expert and pre-fills “I want a promotional homepage for my shop. Please help me.” If launching is unavailable, tell visitors to open WorkBuddy and search for Merchant Promo Designer. Deliver a public homepage link that works on desktop and mobile, with usable artwork browsing and navigation.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 0, "created_at": "2026-10-15T14:45:17+08:00", "updated_at": "2026-10-15T14:57:47+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "midday-market-flash-board", "title": "午盘异动快讯板", "title_en": "Midday Market Flash Board", "subtitle": "把上午密集快讯去重后压成交易台可扫读情报板", "subtitle_en": "Turn dense morning flashes into a trading-desk board that can be scanned in seconds", "description": "盘中最烦的不是消息少，而是消息太多。同一件事换个标题反复刷屏，真正该盯的信号反而容易被埋掉。\n\n这个案例演示如何把 2026-07-27 09:30–13:30 HKT 的财经快讯做时间过滤、去重和信号拆解，再生成一份交易台式双语 HTML 情报板。它保留查询统计、核心 round-up、8 张关键信号卡和一套可直接口播的午盘总结。\n\n适合：盘中盯盘同学、销售交易、投顾沟通、市场内容运营、需要快速把上午行情讲清楚的研究与公关团队。", "description_en": "The hardest part intraday is not a lack of headlines, but too many of them. The same event gets repeated under new titles, and the signals that actually matter end up buried.\n\nThis case shows how to take finance flashes for 2026-07-27 between 09:30 and 13:30 HKT, apply time filtering, deduplication and signal extraction, and turn them into a bilingual HTML board built like a trading desk. It keeps the query stats, the core round-ups, eight key signal cards and a ready-to-speak midday script.\n\nBest for: intraday monitors, sales-trading teams, advisory communication, market-content operators, and research or PR teams that need to explain the morning move quickly.", "categories": [ "finance", "search", "data" ], "tags": [ "午盘快讯", "异动", "交易台", "情绪", "双语HTML" ], "tags_en": [ "Midday Flash", "Moves", "Trading Desk", "Sentiment", "Bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "prompt": "请抓取 2026-07-27 09:30–13:30 HKT 的财经快讯，覆盖全市场、A股、港股和全球宏观四类查询，做时间过滤和去重后，输出一份单文件 HTML 午盘异动快讯板。报告需要包含：查询统计、核心快讯 round-up、拆解后的关键信号卡、上午市场主线、午盘口播总结，以及原文时间与链接。内容必须基于真实查询结果并注明时间与来源；HTML 资源全内联。", "prompt_en": "Please pull finance flashes for 2026-07-27 between 09:30 and 13:30 HKT, covering four query buckets: broad market, A-shares, Hong Kong equities and global macro. After time filtering and deduplication, generate a single-file HTML midday moves board. The report should include query stats, the core flash round-ups, decomposed signal cards, the morning market lead, a spoken-style midday wrap, and original timestamps and links. The content must be based on real query results with timestamps and sources noted; all HTML assets must be inlined.", "skills": [ { "id": "daily-financial-news", "name": "每日财经新闻", "description": "财经新闻多源接入与深度解读，经去重分类后逐条深度概括，附市场影响解读。", "skill_id": "skill_2076628269121765376" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "午盘异动双语快讯板" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T19:43:22+08:00", "updated_at": "2026-07-27T20:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-business-dashboard", "title": "一键生成月度经营分析看板", "title_en": "Monthly Business Dashboard", "subtitle": "营收、毛利、回款、账龄一屏看清，附行动建议", "subtitle_en": "Revenue, margin, collection and aging on one screen, with action items", "description": "上传或接入业务数据，自动出营收、毛利、回款、账龄可视化看板，并给出经营洞察与行动建议。", "description_en": "Upload or connect your business data and it produces a visual dashboard of revenue, gross margin, collections and account aging, along with operating insights and recommended actions.", "categories": [ "data" ], "tags": [ "经营分析", "数据看板", "月度复盘", "B2B贸易" ], "tags_en": [ "Business Analysis", "Data Dashboard", "Monthly Review", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "月度经营分析看板（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易运营专家。请基于我的经营数据，生成一份可视化的月度经营分析看板。\n\n【数据来源（三选一，按我的实际情况）】\n- 我已绑定业务系统/ERP：直接读取{时间范围，如 2026年7月}的经营数据；\n- 我上传了Excel/表格：请读取并映射到营收、成本、毛利、回款、客户、品类、账龄等标准口径；\n- 我暂时没有数据：请说明需要哪些字段，并给出一份示例看板作参考。\n\n【看板要求】\n1. 顶部KPI卡片：当月营收、毛利额、毛利率、回款率，各附环比变化；\n2. 图表：近6个月营收/毛利趋势、品类销售结构、Top5客户销售额、应收账龄分布；\n3. 底部经营洞察：查询【行业信息资产库】里这个行业的经营口径，结合数据给出3条以内可执行建议（如价格战预警、集中度风险、超期账款跟进）；\n4. 输出为一页深色主题的可视化看板，图表用ECharts，适合发给管理层。\n\n若数据源未就绪或返回明显不足，请以通用经营分析方法论为主产出，并说明。请生成完整看板。", "skills": [ { "id": "business-analysis", "name": "经营分析", "description": "TradeBuddy把“老板想一句话知道生意好不好、问题在哪、今天先干啥”的诉求，做成开箱即用的经营体检——一句话结论 + 关键指标 + 最高杠杆动作。只述现象，不臆测因果。", "skill_id": "skill_2091156475485913088" } ], "experts": [ { "id": "yunying-expert", "name": "B2B贸易运营专家", "description": "懂B2B贸易的运营专家：面向老板与管理员，做生意体检、看团队人效、批量扫风险、沉淀企业知识。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_D7992fbSsXNx" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-business-review-report", "title": "本月经营复盘报告", "title_en": "Monthly Business Review Report", "subtitle": "销售、流量、商品与会员四维月度复盘", "subtitle_en": "A practical Weimob playbook for monthly business review report", "description": "做本月经营复盘报告时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“本月经营复盘报告”展开：围绕销售、流量、商品和会员四个经营维度生成本月阶段性复盘，定位核心问题并形成后续行动建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、经营数据概览运营人员、经营分析团队。", "description_en": "The hardest part of monthly business review report is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates monthly business review report through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Business Overview specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "月度复盘", "销售分析", "流量分析", "商品分析", "会员分析" ], "tags_en": [ "Monthly Business Review Report", "Business Diagnosis", "Business Overview", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我生成本月截至昨天的经营复盘报告，涵盖销售、流量、商品和会员四个维度，诊断核心问题，并给出本月后续经营行动建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a monthly business review report covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-store-business-brief", "title": "门店月度经营简报", "title_en": "Monthly Store Business Brief", "subtitle": "各门店本月阶段性经营简报", "subtitle_en": "A practical Weimob playbook for monthly store business brief", "description": "做门店月度经营简报时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“门店月度经营简报”展开：为每家门店生成本月阶段性经营简报，汇总核心指标、异常事项、主要问题和后续行动建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、门店业绩分析运营人员、经营分析团队。", "description_en": "The hardest part of monthly store business brief is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates monthly store business brief through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Store Performance Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "门店简报", "月度经营", "异常事项", "问题诊断", "行动建议" ], "tags_en": [ "Monthly Store Business Brief", "Store and Sales Associate Operations", "Store Performance Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我为每家门店生成本月截至昨天的经营简报，汇总核心指标、异常事项、主要问题和本月后续行动建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a monthly store business brief covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "merchant-store", "name": "商户与门店查询", "description": "查询可访问商户、门店组织节点、店铺经营概览和待办事项，帮助快速了解当前店铺状态。", "skill_id": "skill_2087080985104273408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-target-attainment-forecast", "title": "月度目标达成预测", "title_en": "Monthly Target Attainment Forecast", "subtitle": "结合当前业绩与历史趋势预测月底目标完成情况", "subtitle_en": "A practical Weimob playbook for monthly target attainment forecast", "description": "做月度目标达成预测时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“月度目标达成预测”展开：根据用户提供的本月 GMV 与订单目标，结合当前业绩和历史趋势预测月底达成结果，计算缺口并提出追赶建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、同比环比分析运营人员、经营分析团队。", "description_en": "The hardest part of monthly target attainment forecast is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates monthly target attainment forecast through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Period-over-Period Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "目标管理", "达成预测", "GMV目标", "订单目标", "追赶计划" ], "tags_en": [ "Monthly Target Attainment Forecast", "Business Diagnosis", "Period-over-Period Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我会提供本月GMV和订单目标，请结合截至昨天的当前业绩与历史经营趋势，预测月底目标完成情况，计算目标缺口，并给出接下来可以执行的追赶建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a monthly target attainment forecast covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "movie-collection-workbench", "artifact_type": "html", "title": "光影收藏馆 · 影视收藏推荐工作台", "title_en": "Lightbox Collection · Film & TV Recommendation Workbench", "subtitle": "每天抽 5 部经典生成海报甄选，IMDb/烂番茄/豆瓣三评分对照，收藏、影评人与观影档案一面墙搞定", "subtitle_en": "Five classic picks a day on a swipeable poster wall, with IMDb / Rotten Tomatoes / Douban scores, favorites, critic reviews and your viewing archive in one place", "description": "影迷的观影记录散落在各个平台：想看一部片，评分要跑 IMDb、烂番茄、豆瓣三个地方对；家人推荐过的片子看完就忘，谁推荐的、适合一个人看还是全家一起看，全都没处记。这个工作台把收藏和推荐做成一整面会滑动的海报墙：甄选页每天从推荐池随机抽 5 部经典，逐片展示三平台评分，一键加入收藏；馆藏页每部影片都能一键切换收藏/已观影/不喜欢、写一句话短评、记下推荐人和个人/家庭观影归属，检索浮层支持片名、推荐人、标签、备注实时过滤；批评家页 5 位风格各异的独立影评人各占一屏，自动汇总点评数、均分与近作；读数页把馆藏总数、收藏数、观影状态和豆瓣评分分布统计成图表。数据落在资料库数据表，支持在线存储与双向同步。适合：正在建立自己观影体系的影迷，想管理家庭片单、区分个人与家庭口味的家长，以及喜欢评分数据和观影档案的影视爱好者。", "description_en": "A film lover's records live everywhere: checking one title means cross-referencing IMDb, Rotten Tomatoes and Douban; a family recommendation is forgotten right after watching — who recommended it, and whether it suits a solo night or a family evening, none of it gets kept. This workbench turns your collection into a full swipeable poster wall: the Picks page draws 5 classics from the recommendation pool each day with three-platform scores and one-tap add-to-library; the Library page lets you toggle Favorite / Watched / Not for me on any title, jot a one-line note, and record the recommender and Personal / Family audience, with a live search overlay filtering by title, recommender, tag or note; the Critics page gives 5 independent critics a screen each, auto-summarizing review counts, averages and latest reviews; the Stats page charts library size, favorites, viewing status and the Douban score distribution. All data is stored in Library database tables featuring online storage and two-way sync. Great for film lovers building their own viewing system, parents managing a family watchlist with personal and family tastes separated, and movie buffs who live on ratings and viewing archives.", "tags": [ "影视收藏", "电影推荐", "评分追踪", "海报墙", "观影记录" ], "tags_en": [ "movie collection", "film recommendations", "rating tracker", "poster wall", "viewing log" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/CqW9JOAkHg92OZDn3QO6ZC，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「光影收藏馆」全球影视收藏推荐工作台，数据落在资料库数据表，共两张表：① 影片收藏（片名、类型(电影/剧集)、年份、IMDb 评分、烂番茄(%)、豆瓣评分、推荐人、海报链接、观影归属(个人/家庭)、收藏/已观影/不喜欢三个勾选、备注短评、影评人评分五列：迷影客/胶片手/夜航船/放映室/第七艺术，均为 0-10 数字）——点右上角 ＋ 弹出入藏表单，保存即写入数据表；② 每日推荐池（片名、类型、年份、IMDb 评分、烂番茄(%)、豆瓣评分）——预置经典影视作为每日甄选的数据源。\n\n四个整屏海报式模块，左右滑动切换，移动端适配：① 甄选——每次从推荐池随机抽取 5 部生成「今日甄选」海报流，逐片展示 IMDb/烂番茄/豆瓣三评分与一键加入收藏，加入后即时写入馆藏并在下次甄选中避开，菜单里可随时换一批；② 馆藏——全部收藏影片的海报滑片，每片可一键切换收藏/已观影/不喜欢并即时回显按钮态，展示推荐人与个人/家庭观影归属，检索浮层支持片名/推荐人/标签/备注关键词实时过滤，顶部胶囊在个人/家庭/全部之间切换；③ 批评家——5 位独立影评人各占一屏，自动汇总各自点评数、均分与最近点评影片，点评列表随打分实时更新；④ 读数——自动统计馆藏总数、收藏数、已看数、家庭观影数与豆瓣均分，配豆瓣评分分布柱状图与观影状态环形图，随数据变化即时刷新。\n\n视觉走影院海报风：深色底、影院红单一强调色，海报底图按片名哈希自动配色程序化生成，无需外部图片；评分用 IMDb 金、烂番茄红、豆瓣绿三色区分。未入藏时用演示数据展示完整效果。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the Library capability to build me a same-style copy of https://www.workbuddy.cn/space/d/CqW9JOAkHg92OZDn3QO6ZC with online data storage and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Lightbox Collection\" global film & TV collection and recommendation workbench backed by Library database tables, with two tables: (1) Film Collection (title, type(Movie/Series), year, IMDb score, Rotten Tomatoes(%), Douban score, recommender, poster URL, audience(Personal/Family), three checkboxes: Favorite/Watched/Not for me, note, five critic score columns: The Cinephile/Celluloid/Night Ferry/Projection Room/Seventh Art, all 0-10 numbers) — tap ＋ in the top right to open the add form; saving writes straight into the data table; (2) Daily Recommendation Pool (title, type, year, IMDb score, Rotten Tomatoes(%), Douban score) — preloaded with classics as the data source for daily picks. Four full-screen poster-style modules, swiped left/right, mobile-friendly: (1) Picks — draw 5 titles at random from the pool into a \"Today's Picks\" poster stream, showing IMDb/Rotten Tomatoes/Douban scores per title with one-tap add-to-library; added titles are written to the library immediately and avoided in later draws, and the menu offers a fresh batch anytime; (2) Library — poster slides for every collected title, each with one-tap Favorite/Watched/Not for me toggles that instantly reflect on the buttons, showing recommender and Personal/Family audience; a search overlay filters live by title/recommender/tag/note, and a top capsule switches Personal/Family/All; (3) Critics — 5 independent critics, one screen each, auto-summarizing review counts, averages and latest-reviewed titles, with the review list updating live as scores are entered; (4) Stats — auto-totals library size, favorites, watched, family viewing and the average Douban score, with a Douban score distribution bar chart and a viewing-status donut chart that refresh with the data. Visual style: cinema poster aesthetic — dark base, a single cinema-red accent; poster backgrounds are generated programmatically with per-title hash-based palettes, no external images needed; scores are color-coded IMDb gold, Rotten Tomatoes red, Douban green. Demo data showcases the full experience before you add your own.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "光影收藏馆 · 影视收藏推荐工作台（自包含单文件）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "五斗柜", "name_en": "Wudougui", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "created_at": "2026-09-10T12:40:00+08:00", "updated_at": "2026-09-10T13:05:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "multi-city-weather-compare", "title": "多城市天气对比看板", "title_en": "Multi-city Weather Comparison Board", "subtitle": "六城天气一屏横向对比", "subtitle_en": "Compare six cities' weather at a glance", "description": "想同时看几个城市的天气、决定去哪儿出差或旅行时，一个个网站切换查半天，看完前一个就忘了后一个，根本没法横着比。\n\n这个案例让 QQ 浏览器自动化依次跑完六座城市的天气页，真实抓取今日天气、温度区间、风力和未来三天预报，再整理成一张清爽的对比看板：每城一张卡，温度条统一映射一眼看出谁热谁凉，顶部还自动算出最热城市、下雨城市数量。\n\n适合：出差旅行选目的地、异地办公团队同步天气、需要多地气象横向对比的运营与调度。", "description_en": "When you want to check several cities' weather at once, switching between sites one by one is tedious and you forget the earlier one before finishing the next, making side-by-side comparison impossible.\n\nThis case has QQ Browser automation visit six cities' weather pages in sequence, truly scraping today's conditions, temperature range, wind, and the next three days' forecast, then assembles a clean comparison board: one card per city, a unified temperature bar showing hot vs. cool at a glance, plus an auto-computed summary of the hottest city and how many are rainy.\n\nFor: choosing a destination for business trips or travel, syncing weather across distributed teams, and operations that need multi-location weather comparison.", "categories": [ "search" ], "tags": [ "天气查询", "多站抓取", "对比看板" ], "tags_en": [ "Weather", "Multi-page Scraping", "Comparison" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "多城市天气对比看板" } ], "prompt": "用 QQ 浏览器自动化 skill 依次真实查询北京、上海、广州、成都、哈尔滨、三亚这六座城市的天气：分别导航到中国天气网对应城市页，抓取今日天气现象、今日温度区间（最高/最低）、风力风向，以及未来两到三天的简要预报，并注明数据来源（中国天气网 www.weather.com.cn）和抓取时间。所有数据必须真实查询，禁止使用示意或模拟数据。\n\n然后把六城真实数据整理成一个单文件 HTML 天气对比看板，保存为 output.html：便于横向比较、一眼看出哪座城市更热更凉即可，风格你来把握。所有资源内联、可离线打开，天气图标用内联 SVG 或 CSS 绘制。", "prompt_en": "Use the QQ Browser automation skill to truly query, one by one, the weather for six cities — Beijing, Shanghai, Guangzhou, Chengdu, Harbin and Sanya: navigate to each city's page on China Weather, scrape today's condition, today's temperature range (high/low), wind, and a brief 2-3 day forecast, and note the data source (China Weather, www.weather.com.cn) and capture time. All data must be genuinely queried — no placeholder or simulated data.\n\nThen organize the six cities' real data into a single-file HTML weather comparison board saved as output.html: it should make side-by-side comparison easy and show at a glance which city is hotter or cooler; the visual style is up to you. Inline all resources so it opens offline, and draw weather icons with inline SVG or CSS.", "skills": [ { "id": "qqbrowser-skill", "name": "QQ浏览器自动化", "description": "面向 AI 代理的浏览器自动化技能，可导航网页、填表、点击、截图、提取数据", "skill_id": "skill_2070018474844057600" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-27T03:37:56Z", "updated_at": "2026-07-27T04:14:36Z", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "multi-store-performance-comparison", "title": "多门店业绩横向对比", "title_en": "Multi-Store Performance Comparison", "subtitle": "多门店本月经营指标横向对比与提升建议", "subtitle_en": "A practical Weimob playbook for multi-store performance comparison", "description": "做多门店业绩横向对比时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“多门店业绩横向对比”展开：横向比较各门店本月GMV、订单量、转化率和客单价，找出差距最大的门店并诊断原因。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、门店业绩分析运营人员、经营分析团队。", "description_en": "The hardest part of multi-store performance comparison is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates multi-store performance comparison through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Store Performance Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "门店对比", "门店GMV", "订单量", "转化率", "业绩提升" ], "tags_en": [ "Multi-Store Performance Comparison", "Store and Sales Associate Operations", "Store Performance Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "对比各门店本月截至昨天的GMV、订单量、转化率和客单价，找出差距最大的门店，诊断原因并给出业绩提升建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a multi-store performance comparison covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "merchant-store", "name": "商户与门店查询", "description": "查询可访问商户、门店组织节点、店铺经营概览和待办事项，帮助快速了解当前店铺状态。", "skill_id": "skill_2087080985104273408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "negotiation-scripts", "title": "谈价前必备：一套采购议价话术", "title_en": "Negotiation Scripts", "subtitle": "筹码分析＋分阶段话术＋让步策略，照着谈", "subtitle_en": "Leverage analysis, staged scripts and concession strategy — negotiate straight off the page", "description": "输入品类和供应商报价，自动分析议价筹码、生成分阶段谈判话术和让步策略，帮你把价格谈下来。", "description_en": "Enter the category and the supplier's quote: it analyzes your bargaining chips and generates staged negotiation scripts and concession strategies to help you bring the price down.", "categories": [ "creation" ], "tags": [ "采购谈判", "议价话术", "供应商", "贸易流通" ], "tags_en": [ "Procurement Negotiation", "Negotiation Scripts", "Suppliers", "Trade & Distribution" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "采购议价话术包（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易采购专家。请帮我准备一套向供应商压价的谈判议价话术。\n\n【谈判背景】\n- 品类/数量：{如 螺纹钢 HRB400E，200吨}\n- 供应商当前报价：{如 3,780元/吨}\n- 我的目标：{如 压到3,760到货价，或争取账期}\n\n【要求】\n1. 议价筹码分析：调用【找钢行情MCP】比对市场成交价、查询【行业信息资产库】里这个品类的议价要点，判断有多少压价空间和有利/不利因素；\n2. 生成分阶段话术：开场锚定→中场换价→让步与条件交换（价格谈不动时换账期/运费）→收口锁定，每段附使用提示；\n3. 给出议价底线（目标价/可接受上限/何时转备选）与质量风险提醒。\n\n话术要口语化、可直接照着说。请生成完整议价话术包。", "skills": [ { "id": "negotiation-scripts", "name": "谈判议价话术", "description": "TradeBuddy把 B2B 老手谈价的经验，变成一句需求就能拿到的一套可直接照着说的谈判脚本。采购和销售共用一个底座，只是立场相反：采购要把价压下来、账期谈长，销售要守住利润、不谈崩。", "skill_id": "skill_2091153134466924544" } ], "experts": [ { "id": "caigou-expert", "name": "B2B贸易采购专家", "description": "懂B2B贸易的采购专家：解析需求、看真实行情、全网找货比价、供应商背调、议价话术，帮你把好采购关。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6kRSFmkAwD8M" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "newmedia-factory-workbench", "title": "新媒体工作台", "title_en": "New Media Workbench", "subtitle": "一个人的 AI 内容工厂 ｜ 选题·制作·发布·数据全链路", "subtitle_en": "A one-person AI content factory | Topic to analytics, end to end", "description": "做自媒体最头疼的不是写，而是「今天发什么」：选题散落在聊天记录和备忘录，热点追不上，拍完不知道发哪，发了没数据复盘。这个工作台把一个人的内容工厂装进一页：选题中心管「今天发什么」——选题池待办、今日热点榜单可一键转选题；内容工厂四条产线（长图文/短图文/口播/微短剧）排队开工，任务自动推进到把关；IP 形象库管理人设，出演后自动锁定防撞车；草稿箱把关通过后进发布中心登记多平台发布；数据看板按天记录阅读/点赞/评论/转发，自动给出产线表现洞察和下周排期建议。首次打开自动播种演示数据（日期对齐当天），适合独立创作者、一人内容团队、想把内容生产跑成 SOP 的自媒体运营者。", "description_en": "The hardest part of running your own media isn't writing — it's 'what do I post today?': topics scattered across chats and notes, trends you can't chase in time, finished videos with nowhere to go, and no data to review after publishing. This workbench fits a one-person content factory onto a single page: the Topic Center answers 'what to post today' with a topic backlog and a hot-trend list that converts to topics in one click; the Content Factory runs four production lines (long-form / short-post / voiceover / short-drama) as queued tasks that advance automatically to review; the IP Library manages personas and auto-locks them once cast to prevent conflicts; drafts pass review into the Publish Center for multi-platform logging; the Analytics board records daily reads/likes/comments/shares and generates per-line performance insights plus next-week scheduling advice. Demo data seeds itself on first open (dates aligned to today). Best for: solo creators, one-person content teams, and media operators who want their production to run like a SOP.", "categories": [ "workspace" ], "tags": [ "新媒体运营", "内容工厂", "选题管理", "发布排期", "自媒体工具" ], "tags_en": [ "New Media Ops", "Content Factory", "Topic Pipeline", "Publishing Schedule", "Creator Tools" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "新媒体工作台（单体双语 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/toKl9yixp0EW48k0rxMURg，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：做一个「新媒体工作台——一个人的 AI 内容工厂」单页 HTML，面向独立创作者与一人内容团队，覆盖内容生产全链路，共七个视图：1. 选题中心——选题池卡片（待做/制作中/已完成/全部筛选）、四张 KPI 卡（选题池待做/今日热点/制作中/连续更新）、今日热点榜单支持一键转选题、弹窗登记新选题；2. 内容工厂——长图文/短图文/口播/微短剧四条产线，任务看板分待处理/制作中/待把关，任务异步推进带进度条，支持重试；3. IP 形象库——IP 卡片（人设、形象封面、出演次数），出演中的 IP 自动锁定；4. 草稿箱——成品内容预览与把关，通过后进入发布；5. 发布中心——公众号/小红书/抖音/视频号多平台发布登记；6. 数据看板——发布后按天生成阅读/点赞/评论/转发收藏数据，产线表现对比条形图与排期洞察；7. 系统设置——可填 API Key 接入真实大模型生成文案，不填时走本地模拟引擎。数据存 7 张资料库数据表（选题池/热点池/制作任务/内容成品/IP形象库/发布记录/数据记录），热点每小时自动补充，首次打开自动播种演示数据且日期对齐当天。视觉走纸感工作台风（米白底、墨水色主色、衬线标题），所有 CSS/JS 无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the database capability to make a same-style copy of https://www.workbuddy.cn/space/d/toKl9yixp0EW48k0rxMURg with online data storage and two-way sync. Prefer the one-click same-style path; if unavailable, build it as follows: a 'New Media Workbench — A One-Person AI Content Factory' single-page HTML for solo creators and one-person content teams, covering the full production pipeline across seven views: 1. Topic Center — topic pool cards (to-do / producing / done / all filters), four KPI cards (topics to do / hot today / in production / day streak), a hot-trend list with one-click convert-to-topic, and a modal to log new topics; 2. Content Factory — four production lines (long-form / short-post / voiceover / short-drama), a task board split into pending / producing / awaiting-review, tasks advancing asynchronously with progress bars and retry support; 3. IP Library — IP cards (persona, generated cover, credit count) with in-production IPs auto-locked; 4. Draft Box — preview and review finished content, approved items move to publishing; 5. Publish Center — multi-platform publish logging (WeChat / Xiaohongshu / Douyin / Channels); 6. Analytics — daily reads/likes/comments/shares generated after publishing, per-line performance bar chart and scheduling insights; 7. Settings — optionally provide an API Key to connect a real LLM for copy generation, falling back to a local mock engine when empty. Data lives in 7 database tables (topics / hot trends / tasks / contents / IP library / publish records / metrics); hot trends refresh hourly; demo data seeds automatically on first open with dates aligned to today. Visual: paper-texture workbench style (off-white background, ink-dark primary, serif headings); all CSS/JS inlined, no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "吴伟", "name_en": "Wu Wei", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "source": "community", "created_at": "2026-09-02T17:00:00+08:00", "updated_at": "2026-09-02T17:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-hcp-congress-radar", "title": "开完一场国际大会，把中国专家的真观点整理成一张热点雷达", "title_en": "Turn a Major Congress Into a One-Page Topic Radar", "subtitle": "按话题聚合公开点评：谁在关注什么、观点分歧在哪", "subtitle_en": "Public commentary clustered by topic: who cares about what, and where opinions diverge", "description": "每年 ASCO、ESMO、WCLC 开完，信息铺天盖地，同事带回一堆截图却说不清哪些结论真会影响明天的临床决策。这个案例演示了怎么把一个大会的公开学术内容按话题聚合，标注每位中国专家的关注点和观点之间的分歧，产出一页能内部分发、也能拿去跟医生对话的热点雷达。适合医学事务、产品经理做大会速递和内部赋能。", "description_en": "After every ASCO, ESMO, or WCLC, the team comes back with a pile of screenshots and no clear answer about which conclusions will actually change clinical decisions tomorrow. This case shows how to cluster a congress's public academic content by topic, tag each Chinese expert's focus and where opinions diverge, and produce a one-page topic radar you can circulate internally or take straight into a conversation with a physician. Built for medical affairs and product managers running congress updates and internal enablement.", "tags": [ "HCP洞察", "大会速递", "学术热点", "肺癌", "观点雷达" ], "tags_en": [ "HCP Insight", "Congress Update", "Academic Trends", "Lung Cancer", "Opinion Radar" ], "categories": [ "business", "data" ], "artifact_type": "html", "prompt": "2025 ASCO 刚开完，团队带回一堆截图，但没人说得清哪些结论真会影响明天的临床决策。帮我做一张肺癌领域的「热点雷达」。\n\n这是会上公开报道的内容，作为起点：\n2025 ASCO 肺癌专场公开报道的研究要点：\n陆舜教授牵头的 SACHI 研究（MET 扩增人群，奥希替尼联合赛沃替尼对比含铂双药化疗，ITT 人群中位 PFS 8.2 个月对 4.5 个月，ORR 58%，三代 TKI 经治亚组中位 PFS 6.9 个月）；\n张力教授牵头的 OptiTROP-Lung03（芦康沙妥珠单抗对比多西他赛，ORR 45.1% 对 15.6%，中位 PFS 6.9 对 2.8 个月）；\n韩宝惠教授牵头的 CAMPASS（贝莫苏拜单抗联合安罗替尼对比帕博利珠单抗，中位 PFS 11 对 7.1 个月）；\n吴一龙教授公开质疑 NeoADAURA 把主要终点设为 MPR 而非 EFS；CheckMate 816 公布五年生存率 65.4%。\n\n请联网核实这些研究的准确数据与会议原文，再补上会上其他有影响力的中国专家报告，按「治疗策略 / 耐药机制 / 检测技术 / 围术期」几个话题重新归类，标出每位专家的核心观点，以及观点之间的分歧在哪。产物给我一个单文件 HTML 雷达页，话题能点开看详情，附来源。", "prompt_en": "ASCO 2025 just wrapped and the team came back with a pile of screenshots, but nobody can say which conclusions will actually change clinical decisions tomorrow. Build me a \"topic radar\" for lung cancer.\n\nHere is what was publicly reported at the meeting, as a starting point:\nPublicly reported highlights from the 2025 ASCO lung cancer sessions:\nSACHI, led by Prof. Shun Lu (MET-amplified population; osimertinib plus savolitinib vs platinum-doublet chemotherapy; median PFS 8.2 vs 4.5 months in the ITT population, ORR 58%; median PFS 6.9 months in the third-generation TKI-pretreated subgroup);\nOptiTROP-Lung03, led by Prof. Li Zhang (sacituzumab tirumotecan vs docetaxel; ORR 45.1% vs 15.6%; median PFS 6.9 vs 2.8 months);\nCAMPASS, led by Prof. Baohui Han (benmelstobart plus anlotinib vs pembrolizumab; median PFS 11 vs 7.1 months);\nProf. Yilong Wu publicly questioned NeoADAURA for setting MPR rather than EFS as the primary endpoint; CheckMate 816 reported a 5-year survival rate of 65.4%.\n\nSearch public sources to verify the exact data and primary congress materials, then add other influential Chinese expert presentations from the meeting. Re-group everything under these topics: treatment strategy / resistance mechanisms / detection technology / perioperative care. For each, flag the core viewpoint of each expert and where the opinions actually diverge. Deliver a single self-contained HTML radar page where each topic expands into detail, with sources attached.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "hcp-insight-expert", "name": "腾讯健康NGES客户洞察专家", "description": "医药HCP洞察", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_h9kV1kvhWcyj" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-hcp-previsit-briefing", "title": "拜访前 30 分钟，搞一份能开口的专家速读卡", "title_en": "A 30-Minute Pre-Visit Briefing Card Before You Walk In", "subtitle": "从公开学术信息里提炼研究方向、近期发声和开场话题", "subtitle_en": "Research focus, recent public commentary, and conversation openers — distilled from public academic sources", "description": "明天要去见一位领域大咖，可你连他最近两年在关心什么都答不上来，翻文献又根本来不及。这个案例演示了怎么把一位专家近两年的公开论文、学会任职和会议点评，梳理成一张手机上看得完的速读卡：他是谁、最近在关注什么、从哪个话题开口最自然、哪些说法一说就显得不专业。适合医药代表、区域市场和医学联络官在拜访、科室会邀约前做快速准备。", "description_en": "You have a meeting tomorrow with a leading expert and can't answer the most basic question: what has this person actually cared about for the past two years? There is no time to read dozens of papers. This case shows how to distill a specialist's public papers, society positions, and conference commentary into a briefing card you can finish on your phone: who they are, what they are focused on right now, which topic opens the conversation most naturally, and which claims will make you look uninformed. Built for sales reps, regional marketing, and medical liaisons preparing for a call or a departmental meeting invitation.", "tags": [ "HCP洞察", "拜访准备", "学术画像", "肺癌", "速读卡" ], "tags_en": [ "HCP Insight", "Pre-visit Prep", "Academic Profile", "Lung Cancer", "Briefing Card" ], "categories": [ "business", "data" ], "artifact_type": "html", "prompt": "下周要去拜访广东省人民医院的吴一龙教授，聊 EGFR 突变晚期 NSCLC 的一线治疗策略。我没时间翻几十篇文献，想要一张手机上能看完的拜访速读卡：他是谁、近两年在公开场合最关心哪几个学术问题、我从哪个话题开口最自然、哪些说法一说就显得不专业。\n\n他 2025 年 ASCO 期间公开点评过陆舜教授牵头的 SACHI 研究，随笔里给出的数字是 PFS 9.8 个月对 6.4 个月、三代 TKI 经治亚组 6.9 对 3.0 个月，还公开吐槽过 NeoADAURA 把主要终点设成了 MPR 而不是 EFS——这些是公开信息，帮我顺着这条线把他的学术关注点理出来。联网检索他近两年的公开论文、学会任职和会议报告来补全，注明来源。\n\n产物给我一个单文件 HTML，手机上能一屏一屏往下划，字别太小，最后附信息来源清单。", "prompt_en": "Next week I'm visiting Prof. Yilong Wu at Guangdong Provincial People's Hospital to discuss first-line strategy for advanced EGFR-mutated NSCLC. I don't have time to read dozens of papers. I want a briefing card I can finish on my phone: who he is, which academic questions he has cared most about publicly over the past two years, which topic opens the conversation most naturally, and which claims would make me look uninformed.\n\nDuring ASCO 2025 he publicly commented on the SACHI study led by Prof. Shun Lu — in his own notes he cited PFS of 9.8 vs 6.4 months, and 6.9 vs 3.0 months in the third-generation TKI-pretreated subgroup — and he openly criticized NeoADAURA for setting MPR rather than EFS as the primary endpoint. That is all public. Follow that thread and map out his academic priorities. Search public sources to fill in his papers, society positions, and conference talks from the past two years, and cite them.\n\nGive me a single self-contained HTML file that scrolls screen by screen on a phone, with text large enough to read, ending with a source list.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "hcp-insight-expert", "name": "腾讯健康NGES客户洞察专家", "description": "医药HCP洞察", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_h9kV1kvhWcyj" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-hcp-speaker-shortlist", "title": "一场区域肺癌会，帮我圈出该请哪几位讲者", "title_en": "Shortlist the Right Speakers for a Regional Lung Cancer Meeting", "subtitle": "从公开学术影响力数据出发，比研究方向、比地域、比邀请优先级", "subtitle_en": "Grounded in public academic-influence data: research focus, geography, and invitation priority side by side", "description": "手里有讲者预算，可到底请谁既能撑住学术分量、又跟会议主题对得上，靠拍脑袋定很容易请错人。这个案例演示了怎么从公开的学术影响力数据和研究方向出发，为一个具体会议主题筛出候选讲者，横向对比他们的研究主线、代表成果和地域便利性，再给出邀请优先级和切入话术。适合区域市场、医学联络官做学术会议讲者规划。", "description_en": "You have a speaker budget, but deciding who carries enough academic weight and actually fits the meeting theme by gut feel is how you end up inviting the wrong people. This case shows how to start from publicly published academic-influence data and research focus, shortlist candidates against a concrete meeting theme, compare their research lines, signature work, and geographic convenience side by side, then produce an invitation priority order with outreach angles. Built for regional marketing and medical liaisons planning academic meeting faculty.", "tags": [ "HCP洞察", "讲者筛选", "学术会议", "肺癌", "优先级看板" ], "tags_en": [ "HCP Insight", "Speaker Shortlist", "Academic Meeting", "Lung Cancer", "Priority Board" ], "categories": [ "business", "data" ], "artifact_type": "html", "prompt": "我们要在华东办一场 60 人的区域肺癌学术会，主题是「EGFR 突变晚期 NSCLC 的全程管理」，需要 5 位讲者。我不想靠拍脑袋定名单，想看到一份有依据的候选人对比。\n\n这是公开可查的排名，作为起点：\n2025 肺癌研究领域全国专家学术影响力排名（综合评分 / 单位，公开榜单前 13 位）：\n吴一龙 99.41 广东省人民医院；周彩存 54.64 上海市肺科医院；赫捷 44.03 中国医学科学院肿瘤医院；\n陆舜 37.66 上海市胸科医院；张力 34.29 中山大学肿瘤防治中心；莫树锦 28.37 香港中文大学（中国香港）；\n陈海泉 28.20 复旦大学附属肿瘤医院；何建行 26.07 广州医科大学附属第一医院；\n王洁 26.04 中国医学科学院肿瘤医院；陈万青 25.55 中国医学科学院肿瘤医院；\n李为民 17.33 四川大学华西医院；韩宝惠 17.09 上海市胸科医院；于金明 14.87 山东省肿瘤医院。\n\n请联网核实这些专家的当前单位与研究方向，再结合华东的地域便利性，给我一份候选人对比：每位的研究主线、跟「EGFR 全程管理」这个主题的对口程度、近两年最值得讲的一项成果、以及邀请优先级。产物用一个单文件 HTML 看板呈现，能按优先级和地域筛选，每条附来源。", "prompt_en": "We are running a 60-person regional lung cancer academic meeting in East China on the theme \"Whole-Course Management of Advanced EGFR-Mutated NSCLC,\" and we need five speakers. I don't want to pick the list by gut feel — I want a candidate comparison with evidence behind it.\n\nHere is a publicly published ranking to start from:\n2025 National Academic Influence Ranking in Lung Cancer Research (composite score / institution; top 13, publicly published):\nYilong Wu 99.41, Guangdong Provincial People's Hospital; Caicun Zhou 54.64, Shanghai Pulmonary Hospital; Jie He 44.03, Cancer Hospital, Chinese Academy of Medical Sciences; Shun Lu 37.66, Shanghai Chest Hospital; Li Zhang 34.29, Sun Yat-sen University Cancer Center; Mok Shu Kam Tony 28.37, The Chinese University of Hong Kong (Hong Kong, China); Haiquan Chen 28.20, Fudan University Shanghai Cancer Center; Jianxing He 26.07, The First Affiliated Hospital of Guangzhou Medical University; Jie Wang 26.04, Cancer Hospital, Chinese Academy of Medical Sciences; Wanqing Chen 25.55, Cancer Hospital, Chinese Academy of Medical Sciences; Weimin Li 17.33, West China Hospital, Sichuan University; Baohui Han 17.09, Shanghai Chest Hospital; Jinming Yu 14.87, Shandong Cancer Hospital.\n\nSearch public sources to verify each expert's current institution and research focus, then factor in geographic convenience for East China. Give me a comparison: each person's core research line, how well it matches the \"EGFR whole-course management\" theme, the single most talk-worthy result from the past two years, and an invitation priority. Deliver it as a single self-contained HTML board, filterable by priority and region, with a source attached to every entry.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "hcp-insight-expert", "name": "腾讯健康NGES客户洞察专家", "description": "医药HCP洞察", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_h9kV1kvhWcyj" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-material-dept-meeting-deck", "title": "周五科室会 15 分钟，帮我备好 PPT 和逐页讲稿", "title_en": "A 15-Minute Departmental Meeting — Slides and Speaker Notes Ready to Go", "subtitle": "从真实研究数据到能直接上讲台的成套材料", "subtitle_en": "From real trial data to a deck you can actually present", "description": "临时接到科室会通知，15 分钟、二十几位肿瘤科医生，从找数据到做 PPT 到写讲稿，一个通宵都不一定够。这个案例演示了怎么把几项真实研究的公开数据，直接变成一套 12 页左右的 PPT 加逐页讲稿，每页标好数据出处，讲稿写成医生听得懂的话而不是念稿。适合医药代表、区域市场做科室会与院内会准备。", "description_en": "You get handed a departmental meeting at short notice: 15 minutes, twenty-odd oncologists, and a full night may not cover finding the data, building slides, and writing the script. This case shows how to turn publicly reported data from several real studies directly into a ~12-slide deck with page-by-page speaker notes, each slide sourced, written the way a physician actually talks rather than as something to read aloud. Built for sales reps and regional marketing preparing departmental and in-hospital meetings.", "tags": [ "学术物料", "科室会", "PPT", "肺癌", "讲者手册" ], "tags_en": [ "Academic Material", "Departmental Meeting", "PPT", "Lung Cancer", "Speaker Deck" ], "categories": [ "business", "productivity" ], "artifact_type": "html", "prompt": "周五要在胸科医院做一场 15 分钟的科室会，听众是 20 多位肿瘤科医生，主题是「EGFR 突变晚期 NSCLC 一线治疗后的耐药管理」。我需要一套能直接上讲台的东西：PPT 加逐页讲稿。\n\n可用的真实素材如下，请联网核实它们的准确出处、最新状态和国内获批情况，必要时补充：\n2025 ASCO 肺癌专场公开报道的研究要点：\n陆舜教授牵头的 SACHI 研究（MET 扩增人群，奥希替尼联合赛沃替尼对比含铂双药化疗，ITT 人群中位 PFS 8.2 个月对 4.5 个月，ORR 58%，三代 TKI 经治亚组中位 PFS 6.9 个月）；\n张力教授牵头的 OptiTROP-Lung03（芦康沙妥珠单抗对比多西他赛，ORR 45.1% 对 15.6%，中位 PFS 6.9 对 2.8 个月）；\n韩宝惠教授牵头的 CAMPASS（贝莫苏拜单抗联合安罗替尼对比帕博利珠单抗，中位 PFS 11 对 7.1 个月）；\n吴一龙教授公开质疑 NeoADAURA 把主要终点设为 MPR 而非 EFS；CheckMate 816 公布五年生存率 65.4%。\n\n给我一份 12 页左右的 PPT，配一页一页的讲稿，讲稿写医生听得懂的话、别说教，每页标注数据出处。", "prompt_en": "I have a 15-minute departmental meeting at a chest hospital this Friday, in front of more than 20 medical oncologists, on \"Managing Resistance After First-Line Therapy in Advanced EGFR-Mutated NSCLC.\" I need something I can present straight away: slides plus page-by-page speaker notes.\n\nHere is the real material to work from. Verify the exact sources, current status, and domestic approval situation, and supplement where needed:\nPublicly reported highlights from the 2025 ASCO lung cancer sessions:\nSACHI, led by Prof. Shun Lu (MET-amplified population; osimertinib plus savolitinib vs platinum-doublet chemotherapy; median PFS 8.2 vs 4.5 months in the ITT population, ORR 58%; median PFS 6.9 months in the third-generation TKI-pretreated subgroup);\nOptiTROP-Lung03, led by Prof. Li Zhang (sacituzumab tirumotecan vs docetaxel; ORR 45.1% vs 15.6%; median PFS 6.9 vs 2.8 months);\nCAMPASS, led by Prof. Baohui Han (benmelstobart plus anlotinib vs pembrolizumab; median PFS 11 vs 7.1 months);\nProf. Yilong Wu publicly questioned NeoADAURA for setting MPR rather than EFS as the primary endpoint; CheckMate 816 reported a 5-year survival rate of 65.4%.\n\nGive me a deck of around 12 slides with notes for every page. Write the notes the way physicians actually speak — no lecturing — and cite the data source on each slide.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "周五科室会 15 分钟，帮我备好 PPT 和逐页讲稿" }, { "file": "output.md", "label": "周五科室会 15 分钟，帮我备好 PPT 和逐页讲稿（Markdown 版）" } ], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-rep-material-studio", "name": "腾讯健康NGES学术物料专家", "description": "医药学术物料生成", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tkfW3IlyWHqB" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-material-guideline-onepager", "title": "一篇 3000 字专家访谈，压成一张能微信转发的一页学术卡", "title_en": "Turn a 3,000-Word Expert Interview Into a One-Page Shareable Academic Card", "subtitle": "保留学术严谨度，去掉长文的阅读负担", "subtitle_en": "Keep the academic rigor, drop the reading burden", "description": "公众号发了篇专家访谈，三千多字，阅读量惨淡——医生不是不关心，是没时间看长文。这个案例演示了怎么把一篇长访谈压成一张一屏看完、能直接微信发给医生的学术速递卡，同时保住指南依据和专家观点，不为了短而丢严谨度。适合数字营销、产品经理做学术内容的二次分发。", "description_en": "You published a 3,000-word expert interview and the readership was dismal — not because physicians don't care, but because nobody reads long-form anymore. This case shows how to compress a long interview into an academic update card that fits one screen and can be sent straight to a physician over WeChat, while preserving the guideline basis and the expert's actual positions rather than trading rigor for brevity. Built for digital marketing and product managers redistributing academic content.", "tags": [ "学术物料", "一页速递", "内容分发", "肺癌", "微信转发" ], "tags_en": [ "Academic Material", "One-Pager", "Content Distribution", "Lung Cancer", "WeChat Share" ], "categories": [ "business", "productivity" ], "artifact_type": "html", "prompt": "我们刚发了一篇肺癌领域的专家访谈，三千多字，阅读量惨淡——医生没时间看长文，可内容本身是有价值的。我想把它压成一张一屏就能看完、能直接微信发给医生的学术速递卡。\n\n原文如下：\n继表皮生长因子受体酪氨酸激酶抑制剂（EGFR-TKI）成为EGFR突变（Ex19del/L858R）晚期非小细胞肺癌（NSCLC）患者的标准治疗后，EGFR-TKI联合治疗策略取得了重要突破，有望进一步延长EGFR突变晚期NSCLC患者的生存[1]。在EGFR-TKI联合治疗策略发展的同时，针对其耐药机制及全程管理体系的探讨也在不断深入。特邀湖南省肿瘤医院邬麟教授，总结EGFR-TKI联合治疗策略的重要进展，并从EGFR-TKI联合治疗策略全程管理的角度发表独到见解。\n\n邬麟 教授｜专家简介：二级主任医师、博士生导师、博士后合作导师；湖南省肿瘤医院（中南大学湘雅医学院附属肿瘤医院）内科教研室主任，GCP肿瘤专业负责人、胸内二科主任；湖南省肺癌临床医学研究中心主任；中国抗癌协会(CACA)老年肿瘤专委会副主委；CSWOG肺癌专业委员会候任主委；CSCO理事；CACA临床化疗专业委员会常委；湖南省抗癌协会肿瘤化疗专业委员会主任委员等。\n\n肺癌的个体化精准诊疗是临床的热点和难点之一，能否请您谈谈如何为EGFR突变晚期NSCLC患者制定精准治疗策略？既往局限于医疗技术的落后，肺癌患者在临床中往往接受同质化的诊疗。EGFR突变的发现和EGFR-TKI疗法的兴起使肺癌治疗从经验性的化疗转向精准分子靶向治疗，自此晚期NSCLC迈向了精准诊疗时代。精准检测是实施精准靶向治疗的前提。邬麟教授介绍道，精准治疗，检测先行，晚期NSCLC患者在初诊时需进行基因检测，精准的基因检测对于晚期NSCLC的一线治疗选择具有指导作用，这已在全球形成共识，并成为标准的临床实践[1-2]。此外，基因检测在患者的全程管理中也十分重要。随着治疗时间的延长，EGFR-TKI耐药和疾病进展的出现不可避免，基因检测可以明确EGFR-TKI耐药机制，为后续精准治疗提供依据。针对EGFR-TKI一线治疗后出现进展的人群，国内外权威指南均建议根据进展模式和耐药机制指导后续治疗策略的制定[1-2]。总而言之，基因检测不仅在一线治疗决策中具有关键作用，更是肺癌全程管理中动态调整治疗方案的重要工具。基因检测指导下的精准治疗是确保EGFR突变晚期NSCLC患者获益的关键。需要强调的是，在肺癌的全程管理过程中，一线治疗方案的选择尤为重要。邬麟教授指出，一线治疗在EGFR突变晚期NSCLC患者的整体治疗进程中扮演着至关重要的角色，一线获益对于提升全程获益和患者治疗信心具有积极影响，有助于优化全程管理和患者的治疗依从性。继三代EGFR-TKI成为EGFR突变晚期NSCLC患者的一线标准治疗后，EGFR-TKI联合治疗策略的探索也已取得初步成功，进一步提高了患者获益。在选择一线用药方案时，亦需要把握全程管理的观念，从而优化整体治疗策略。\n\n参考文献：[1] 中国临床肿瘤学会（CSCO）非小细胞肺癌诊疗指南. 2024版. [2] NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Non-Small Cell Lung Cancer. Version 3.2025.\n\n帮我提炼出最值得医生记住的 3~5 个要点，保留原文里的指南依据和专家观点，别为了短而丢掉严谨度。产物给我一个单文件 HTML 竖版页面，手机上打开一屏看完，能直接转发，参考文献和专家信息放在最后。", "prompt_en": "We just published a 3,000-word expert interview in lung cancer and the readership was dismal. Physicians don't have time for long-form, but the content itself is worth reading. I want to compress it into a one-screen academic update card I can send straight to a physician over WeChat.\n\nSource article:\nFollowing the establishment of EGFR tyrosine kinase inhibitors (EGFR-TKIs) as standard therapy for patients with advanced EGFR-mutated (Ex19del/L858R) non-small cell lung cancer (NSCLC), EGFR-TKI combination strategies have achieved major breakthroughs and are expected to further extend survival [1]. In parallel, investigation into resistance mechanisms and whole-course management continues to deepen. This issue features Professor Lin Wu of Hunan Cancer Hospital, summarizing key advances in EGFR-TKI combination strategies and offering perspectives on whole-course management.\n\nProfessor Lin Wu — Profile: Grade-II Chief Physician, Doctoral Supervisor; Director of the Internal Medicine Teaching and Research Section, Hunan Cancer Hospital (The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University); Head of the Thoracic Internal Medicine Department II; Director of the Hunan Lung Cancer Clinical Medical Research Center; Vice Chair of the Geriatric Oncology Committee of the Chinese Anti-Cancer Association (CACA); President-Elect of the CSWOG Lung Cancer Committee; Council Member of CSCO.\n\nIndividualized precision diagnosis and treatment of lung cancer remains a clinical hotspot and challenge. Historically, limited medical technology meant patients received largely homogeneous care. The discovery of EGFR mutations and the advent of EGFR-TKIs shifted lung cancer treatment from empirical chemotherapy to precision molecular targeted therapy. Precision testing is the prerequisite for precision targeted therapy. As Professor Wu notes, \"precision treatment begins with testing\" — genetic testing at initial diagnosis is now a global consensus and standard clinical practice for guiding first-line selection in advanced NSCLC [1-2]. Genetic testing is equally critical across the whole disease course: as treatment duration extends, EGFR-TKI resistance and disease progression become inevitable, and testing clarifies the resistance mechanism to inform subsequent precision therapy. For patients progressing after first-line EGFR-TKI, major Chinese and international guidelines recommend tailoring subsequent strategy to the pattern of progression and the resistance mechanism [1-2]. In short, genetic testing is both pivotal to first-line decision-making and an essential tool for dynamically adjusting therapy across the whole course. The choice of first-line regimen is especially consequential: first-line benefit positively influences overall benefit and patient confidence, optimizing whole-course management and adherence. Following third-generation EGFR-TKIs as first-line standard, exploration of EGFR-TKI combination strategies has achieved initial success, further improving patient benefit. First-line selection should likewise be guided by a whole-course management mindset.\n\nReferences: [1] Chinese Society of Clinical Oncology (CSCO) Guidelines for the Diagnosis and Treatment of Non-Small Cell Lung Cancer, 2024 Edition. [2] NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell Lung Cancer, Version 3.2025.\n\nPull out the 3–5 points most worth a physician's attention, preserving the guideline basis and the expert's stated positions from the original — don't trade rigor for brevity. Deliver a single self-contained vertical HTML page that fits one screen on a phone, ready to forward, with references and the expert's profile at the end.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-rep-material-studio", "name": "腾讯健康NGES学术物料专家", "description": "医药学术物料生成", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tkfW3IlyWHqB" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-material-objection-handbook", "title": "医生当场抛来的学术质疑，帮我备一本能翻的应对手册", "title_en": "A Pocket Handbook for Handling On-the-Spot Academic Objections", "subtitle": "把公开可查的争议观点整理成有出处的应答卡", "subtitle_en": "Organize verifiable public controversies into sourced response cards", "description": "拜访时医生当场抛问题，一紧张就答得没底气，尤其现在医生会直接拿国际会议上中国专家的质疑来反问。这个案例演示了怎么把这些公开可查的争议观点整理成一本随身能翻的异议应对手册：每条质疑配事实依据、该怎么回应、以及哪些地方必须坦承数据还不成熟。适合医药代表、区域市场做拜访赋能。", "description_en": "A physician throws a question at you mid-visit and your confidence evaporates — especially now that physicians quote Chinese experts' own criticisms from international congresses back at you. This case shows how to turn those publicly verifiable controversies into a pocket handbook: every objection paired with the factual basis, how to respond, and where you must concede that the data is not mature yet. Built for sales reps and regional marketing running call enablement.", "tags": [ "学术物料", "异议应对", "拜访赋能", "肺癌", "应答手册" ], "tags_en": [ "Academic Material", "Objection Handling", "Call Enablement", "Lung Cancer", "Response Handbook" ], "categories": [ "business", "productivity" ], "artifact_type": "html", "prompt": "拜访时医生经常当场抛问题，我一紧张就答得没底气。最近几次医生直接拿国际会议上中国专家的质疑来反问我，比如：「你们都说联合治疗好，可吴一龙教授在 2025 ASCO 上不也吐槽 NeoADAURA 把主要终点设成了 MPR 而不是 EFS 吗？」「HER3 ADC 的 HERTHENA-Lung02 不是也铩羽而归了吗，ADC 这条路到底行不行？」\n\n帮我做一本能随身翻的异议应对手册：把我可能遇到的学术质疑列全，每一条给出公开可查的事实依据、该怎么回应、以及哪些地方必须坦承「数据还不成熟」。\n\n背景素材，请联网核实并补充：\n2025 ASCO 肺癌专场公开报道的研究要点：\n陆舜教授牵头的 SACHI 研究（MET 扩增人群，奥希替尼联合赛沃替尼对比含铂双药化疗，ITT 人群中位 PFS 8.2 个月对 4.5 个月，ORR 58%，三代 TKI 经治亚组中位 PFS 6.9 个月）；\n张力教授牵头的 OptiTROP-Lung03（芦康沙妥珠单抗对比多西他赛，ORR 45.1% 对 15.6%，中位 PFS 6.9 对 2.8 个月）；\n韩宝惠教授牵头的 CAMPASS（贝莫苏拜单抗联合安罗替尼对比帕博利珠单抗，中位 PFS 11 对 7.1 个月）；\n吴一龙教授公开质疑 NeoADAURA 把主要终点设为 MPR 而非 EFS；CheckMate 816 公布五年生存率 65.4%。\n\n产物给我一份 Word 文档，按质疑类型分节，每条一页，附上来源出处。", "prompt_en": "Physicians fire questions at me mid-visit and I lose my footing. Lately they quote Chinese experts' own criticisms from international congresses straight back at me — for example: \"You all say combination therapy is great, but didn't Prof. Yilong Wu criticize NeoADAURA at ASCO 2025 for setting MPR rather than EFS as the primary endpoint?\" or \"HERTHENA-Lung02 with the HER3 ADC fell flat too — is the ADC route even viable?\"\n\nBuild me a pocket handbook for handling objections: lay out every academic challenge I'm likely to face, and for each one give the publicly verifiable factual basis, how to respond, and where I have to concede that the data isn't mature yet.\n\nBackground material — verify and supplement via public sources:\nPublicly reported highlights from the 2025 ASCO lung cancer sessions:\nSACHI, led by Prof. Shun Lu (MET-amplified population; osimertinib plus savolitinib vs platinum-doublet chemotherapy; median PFS 8.2 vs 4.5 months in the ITT population, ORR 58%; median PFS 6.9 months in the third-generation TKI-pretreated subgroup);\nOptiTROP-Lung03, led by Prof. Li Zhang (sacituzumab tirumotecan vs docetaxel; ORR 45.1% vs 15.6%; median PFS 6.9 vs 2.8 months);\nCAMPASS, led by Prof. Baohui Han (benmelstobart plus anlotinib vs pembrolizumab; median PFS 11 vs 7.1 months);\nProf. Yilong Wu publicly questioned NeoADAURA for setting MPR rather than EFS as the primary endpoint; CheckMate 816 reported a 5-year survival rate of 65.4%.\n\nDeliver a Word document organized by objection type, one page per entry, with sources cited.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "医生当场抛来的学术质疑，帮我备一本能翻的应对手册" }, { "file": "output.md", "label": "医生当场抛来的学术质疑，帮我备一本能翻的应对手册（Markdown 版）" } ], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-rep-material-studio", "name": "腾讯健康NGES学术物料专家", "description": "医药学术物料生成", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_tkfW3IlyWHqB" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-recreator-flashcards", "title": "新人代表背不下来这些学术要点，帮我做成一摞能翻的闪卡", "title_en": "New Reps Can't Memorize This — Turn It Into Flip Cards", "subtitle": "可导入 Anki 的卡片包 + 网页版翻卡练习", "subtitle_en": "An Anki-importable card pack plus a web-based flip-card drill", "description": "新人代表上岗最怕被医生问住，可打印成册的学术资料没人背。这个案例演示了怎么把专家访谈和指南要点拆成三四十张问答闪卡：正面是医生真会问的问题，背面是答案和数字出处，配一个能翻着练的网页，再导出一份能直接导入 Anki 的卡包。适合培训、区域市场做新人上岗赋能。", "description_en": "New reps dread being stumped by a physician, yet nobody memorizes printed academic handbooks. This case shows how to break an expert interview and guideline highlights into 30–40 Q&A flashcards: a question a physician would actually ask on the front, the answer with its numeric source on the back, plus a web page for flipping through them and an export that imports straight into Anki. Built for training and regional marketing onboarding new reps.", "tags": [ "内容二创", "闪卡", "培训赋能", "肺癌", "Anki" ], "tags_en": [ "Content Recreation", "Flashcards", "Training Enablement", "Lung Cancer", "Anki" ], "categories": [ "business", "creation" ], "artifact_type": "html", "prompt": "新人医药代表上岗，最怕被医生问住。我们有一篇肺癌领域的专家访谈和几份指南要点，但打印成册没人背。我想做成一摞能翻的闪卡，手机上碎片时间刷一刷就能记住。\n\n素材如下：\n继表皮生长因子受体酪氨酸激酶抑制剂（EGFR-TKI）成为EGFR突变（Ex19del/L858R）晚期非小细胞肺癌（NSCLC）患者的标准治疗后，EGFR-TKI联合治疗策略取得了重要突破，有望进一步延长EGFR突变晚期NSCLC患者的生存[1]。在EGFR-TKI联合治疗策略发展的同时，针对其耐药机制及全程管理体系的探讨也在不断深入。特邀湖南省肿瘤医院邬麟教授，总结EGFR-TKI联合治疗策略的重要进展，并从EGFR-TKI联合治疗策略全程管理的角度发表独到见解。\n\n邬麟 教授｜专家简介：二级主任医师、博士生导师、博士后合作导师；湖南省肿瘤医院（中南大学湘雅医学院附属肿瘤医院）内科教研室主任，GCP肿瘤专业负责人、胸内二科主任；湖南省肺癌临床医学研究中心主任；中国抗癌协会(CACA)老年肿瘤专委会副主委；CSWOG肺癌专业委员会候任主委；CSCO理事；CACA临床化疗专业委员会常委；湖南省抗癌协会肿瘤化疗专业委员会主任委员等。\n\n肺癌的个体化精准诊疗是临床的热点和难点之一，能否请您谈谈如何为EGFR突变晚期NSCLC患者制定精准治疗策略？既往局限于医疗技术的落后，肺癌患者在临床中往往接受同质化的诊疗。EGFR突变的发现和EGFR-TKI疗法的兴起使肺癌治疗从经验性的化疗转向精准分子靶向治疗，自此晚期NSCLC迈向了精准诊疗时代。精准检测是实施精准靶向治疗的前提。邬麟教授介绍道，精准治疗，检测先行，晚期NSCLC患者在初诊时需进行基因检测，精准的基因检测对于晚期NSCLC的一线治疗选择具有指导作用，这已在全球形成共识，并成为标准的临床实践[1-2]。此外，基因检测在患者的全程管理中也十分重要。随着治疗时间的延长，EGFR-TKI耐药和疾病进展的出现不可避免，基因检测可以明确EGFR-TKI耐药机制，为后续精准治疗提供依据。针对EGFR-TKI一线治疗后出现进展的人群，国内外权威指南均建议根据进展模式和耐药机制指导后续治疗策略的制定[1-2]。总而言之，基因检测不仅在一线治疗决策中具有关键作用，更是肺癌全程管理中动态调整治疗方案的重要工具。基因检测指导下的精准治疗是确保EGFR突变晚期NSCLC患者获益的关键。需要强调的是，在肺癌的全程管理过程中，一线治疗方案的选择尤为重要。邬麟教授指出，一线治疗在EGFR突变晚期NSCLC患者的整体治疗进程中扮演着至关重要的角色，一线获益对于提升全程获益和患者治疗信心具有积极影响，有助于优化全程管理和患者的治疗依从性。继三代EGFR-TKI成为EGFR突变晚期NSCLC患者的一线标准治疗后，EGFR-TKI联合治疗策略的探索也已取得初步成功，进一步提高了患者获益。在选择一线用药方案时，亦需要把握全程管理的观念，从而优化整体治疗策略。\n\n参考文献：[1] 中国临床肿瘤学会（CSCO）非小细胞肺癌诊疗指南. 2024版. [2] NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Non-Small Cell Lung Cancer. Version 3.2025.\n\n帮我把里面的核心知识点拆成 30~40 张问答卡片：正面是问题，背面是答案和数字出处。给我两样东西——一个能在网页上直接翻着练的单文件 HTML（卡面能翻转），再加一份能直接导入 Anki 的 CSV。别出「什么是肺癌」这种没用的题，要出医生真会问的。", "prompt_en": "New sales reps dread being stumped by a physician. We have a lung cancer expert interview and several guideline highlights, but nobody memorizes printed booklets. I want a stack of flip cards they can drill through on their phone in spare moments.\n\nSource material:\nFollowing the establishment of EGFR tyrosine kinase inhibitors (EGFR-TKIs) as standard therapy for patients with advanced EGFR-mutated (Ex19del/L858R) non-small cell lung cancer (NSCLC), EGFR-TKI combination strategies have achieved major breakthroughs and are expected to further extend survival [1]. In parallel, investigation into resistance mechanisms and whole-course management continues to deepen. This issue features Professor Lin Wu of Hunan Cancer Hospital, summarizing key advances in EGFR-TKI combination strategies and offering perspectives on whole-course management.\n\nProfessor Lin Wu — Profile: Grade-II Chief Physician, Doctoral Supervisor; Director of the Internal Medicine Teaching and Research Section, Hunan Cancer Hospital (The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University); Head of the Thoracic Internal Medicine Department II; Director of the Hunan Lung Cancer Clinical Medical Research Center; Vice Chair of the Geriatric Oncology Committee of the Chinese Anti-Cancer Association (CACA); President-Elect of the CSWOG Lung Cancer Committee; Council Member of CSCO.\n\nIndividualized precision diagnosis and treatment of lung cancer remains a clinical hotspot and challenge. Historically, limited medical technology meant patients received largely homogeneous care. The discovery of EGFR mutations and the advent of EGFR-TKIs shifted lung cancer treatment from empirical chemotherapy to precision molecular targeted therapy. Precision testing is the prerequisite for precision targeted therapy. As Professor Wu notes, \"precision treatment begins with testing\" — genetic testing at initial diagnosis is now a global consensus and standard clinical practice for guiding first-line selection in advanced NSCLC [1-2]. Genetic testing is equally critical across the whole disease course: as treatment duration extends, EGFR-TKI resistance and disease progression become inevitable, and testing clarifies the resistance mechanism to inform subsequent precision therapy. For patients progressing after first-line EGFR-TKI, major Chinese and international guidelines recommend tailoring subsequent strategy to the pattern of progression and the resistance mechanism [1-2]. In short, genetic testing is both pivotal to first-line decision-making and an essential tool for dynamically adjusting therapy across the whole course. The choice of first-line regimen is especially consequential: first-line benefit positively influences overall benefit and patient confidence, optimizing whole-course management and adherence. Following third-generation EGFR-TKIs as first-line standard, exploration of EGFR-TKI combination strategies has achieved initial success, further improving patient benefit. First-line selection should likewise be guided by a whole-course management mindset.\n\nReferences: [1] Chinese Society of Clinical Oncology (CSCO) Guidelines for the Diagnosis and Treatment of Non-Small Cell Lung Cancer, 2024 Edition. [2] NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell Lung Cancer, Version 3.2025.\n\nBreak the core knowledge points into 30–40 Q&A cards: the question on the front, the answer with its numeric source on the back. Give me two things — a single self-contained HTML page where the cards flip on screen for practice, plus a CSV that imports straight into Anki. Skip useless questions like \"what is lung cancer\"; only ask what a physician would actually ask.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-content-recreator", "name": "腾讯健康NGES医学二创专家", "description": "医学内容二创", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_XL25qlKax91g" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-recreator-interactive-feature", "title": "这篇专家访谈太长了没人看，帮我改成能点着玩的互动页", "title_en": "This Expert Interview Is Too Long to Read — Make It Interactive", "subtitle": "长文拆成可点开的问答、时间轴和要点卡", "subtitle_en": "Break the long read into expandable Q&A, a timeline, and takeaway cards", "description": "公众号的长访谈打开率越来越低，不是内容不行，是形式不对——医生扫一眼发现要读三千字就划走了。这个案例演示了怎么把一篇访谈改成互动专题页：先看三句话结论，感兴趣再点开问答，专家观点和指南依据做成可展开的卡片，手机上点着就能读完。适合数字营销做学术内容的互动化改造。", "description_en": "Long-form interviews on WeChat keep losing open rates — not because the content is weak, but because the format is wrong: a physician sees three thousand characters and swipes away. This case shows how to rebuild an interview as an interactive feature page: three takeaway lines up top, expandable Q&A below, expert positions and guideline basis as cards that open on tap, all readable by tapping through on a phone. Built for digital marketing turning academic content interactive.", "tags": [ "内容二创", "互动页", "长文改造", "肺癌", "单文件HTML" ], "tags_en": [ "Content Recreation", "Interactive Page", "Long-form Rewrite", "Lung Cancer", "Single-file HTML" ], "categories": [ "business", "creation" ], "artifact_type": "html", "prompt": "我们公众号发了一篇肺癌专家访谈，内容是好的，但三千多字没人读完。我想把它改成一个能点着玩的互动专题页——医生点开可以先看结论，感兴趣再展开看细节，还能点开专家简介和参考文献。\n\n原文如下：\n继表皮生长因子受体酪氨酸激酶抑制剂（EGFR-TKI）成为EGFR突变（Ex19del/L858R）晚期非小细胞肺癌（NSCLC）患者的标准治疗后，EGFR-TKI联合治疗策略取得了重要突破，有望进一步延长EGFR突变晚期NSCLC患者的生存[1]。在EGFR-TKI联合治疗策略发展的同时，针对其耐药机制及全程管理体系的探讨也在不断深入。特邀湖南省肿瘤医院邬麟教授，总结EGFR-TKI联合治疗策略的重要进展，并从EGFR-TKI联合治疗策略全程管理的角度发表独到见解。\n\n邬麟 教授｜专家简介：二级主任医师、博士生导师、博士后合作导师；湖南省肿瘤医院（中南大学湘雅医学院附属肿瘤医院）内科教研室主任，GCP肿瘤专业负责人、胸内二科主任；湖南省肺癌临床医学研究中心主任；中国抗癌协会(CACA)老年肿瘤专委会副主委；CSWOG肺癌专业委员会候任主委；CSCO理事；CACA临床化疗专业委员会常委；湖南省抗癌协会肿瘤化疗专业委员会主任委员等。\n\n肺癌的个体化精准诊疗是临床的热点和难点之一，能否请您谈谈如何为EGFR突变晚期NSCLC患者制定精准治疗策略？既往局限于医疗技术的落后，肺癌患者在临床中往往接受同质化的诊疗。EGFR突变的发现和EGFR-TKI疗法的兴起使肺癌治疗从经验性的化疗转向精准分子靶向治疗，自此晚期NSCLC迈向了精准诊疗时代。精准检测是实施精准靶向治疗的前提。邬麟教授介绍道，精准治疗，检测先行，晚期NSCLC患者在初诊时需进行基因检测，精准的基因检测对于晚期NSCLC的一线治疗选择具有指导作用，这已在全球形成共识，并成为标准的临床实践[1-2]。此外，基因检测在患者的全程管理中也十分重要。随着治疗时间的延长，EGFR-TKI耐药和疾病进展的出现不可避免，基因检测可以明确EGFR-TKI耐药机制，为后续精准治疗提供依据。针对EGFR-TKI一线治疗后出现进展的人群，国内外权威指南均建议根据进展模式和耐药机制指导后续治疗策略的制定[1-2]。总而言之，基因检测不仅在一线治疗决策中具有关键作用，更是肺癌全程管理中动态调整治疗方案的重要工具。基因检测指导下的精准治疗是确保EGFR突变晚期NSCLC患者获益的关键。需要强调的是，在肺癌的全程管理过程中，一线治疗方案的选择尤为重要。邬麟教授指出，一线治疗在EGFR突变晚期NSCLC患者的整体治疗进程中扮演着至关重要的角色，一线获益对于提升全程获益和患者治疗信心具有积极影响，有助于优化全程管理和患者的治疗依从性。继三代EGFR-TKI成为EGFR突变晚期NSCLC患者的一线标准治疗后，EGFR-TKI联合治疗策略的探索也已取得初步成功，进一步提高了患者获益。在选择一线用药方案时，亦需要把握全程管理的观念，从而优化整体治疗策略。\n\n参考文献：[1] 中国临床肿瘤学会（CSCO）非小细胞肺癌诊疗指南. 2024版. [2] NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Non-Small Cell Lung Cancer. Version 3.2025.\n\n帮我设计一个适合手机阅读的互动页面：顶部是三句话结论，中间是能点开的问答模块，专家观点和指南依据用可展开的卡片呈现，底部放专家简介和参考文献。所有资源内联在一个 HTML 文件里，能离线打开、能直接分发。", "prompt_en": "We published a lung cancer expert interview on our WeChat account. The content is good, but nobody finishes three thousand characters. I want to rebuild it as an interactive feature page — the physician sees the conclusions first, expands into detail if interested, and can also open the expert profile and references.\n\nSource article:\nFollowing the establishment of EGFR tyrosine kinase inhibitors (EGFR-TKIs) as standard therapy for patients with advanced EGFR-mutated (Ex19del/L858R) non-small cell lung cancer (NSCLC), EGFR-TKI combination strategies have achieved major breakthroughs and are expected to further extend survival [1]. In parallel, investigation into resistance mechanisms and whole-course management continues to deepen. This issue features Professor Lin Wu of Hunan Cancer Hospital, summarizing key advances in EGFR-TKI combination strategies and offering perspectives on whole-course management.\n\nProfessor Lin Wu — Profile: Grade-II Chief Physician, Doctoral Supervisor; Director of the Internal Medicine Teaching and Research Section, Hunan Cancer Hospital (The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University); Head of the Thoracic Internal Medicine Department II; Director of the Hunan Lung Cancer Clinical Medical Research Center; Vice Chair of the Geriatric Oncology Committee of the Chinese Anti-Cancer Association (CACA); President-Elect of the CSWOG Lung Cancer Committee; Council Member of CSCO.\n\nIndividualized precision diagnosis and treatment of lung cancer remains a clinical hotspot and challenge. Historically, limited medical technology meant patients received largely homogeneous care. The discovery of EGFR mutations and the advent of EGFR-TKIs shifted lung cancer treatment from empirical chemotherapy to precision molecular targeted therapy. Precision testing is the prerequisite for precision targeted therapy. As Professor Wu notes, \"precision treatment begins with testing\" — genetic testing at initial diagnosis is now a global consensus and standard clinical practice for guiding first-line selection in advanced NSCLC [1-2]. Genetic testing is equally critical across the whole disease course: as treatment duration extends, EGFR-TKI resistance and disease progression become inevitable, and testing clarifies the resistance mechanism to inform subsequent precision therapy. For patients progressing after first-line EGFR-TKI, major Chinese and international guidelines recommend tailoring subsequent strategy to the pattern of progression and the resistance mechanism [1-2]. In short, genetic testing is both pivotal to first-line decision-making and an essential tool for dynamically adjusting therapy across the whole course. The choice of first-line regimen is especially consequential: first-line benefit positively influences overall benefit and patient confidence, optimizing whole-course management and adherence. Following third-generation EGFR-TKIs as first-line standard, exploration of EGFR-TKI combination strategies has achieved initial success, further improving patient benefit. First-line selection should likewise be guided by a whole-course management mindset.\n\nReferences: [1] Chinese Society of Clinical Oncology (CSCO) Guidelines for the Diagnosis and Treatment of Non-Small Cell Lung Cancer, 2024 Edition. [2] NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell Lung Cancer, Version 3.2025.\n\nDesign an interactive page that works on a phone: three takeaway lines at the top, an expandable Q&A module in the middle, expert positions and guideline basis as cards that open on tap, and the expert profile plus references at the bottom. Inline everything into one HTML file that opens offline and can be distributed as-is.", "preview": "output.html", "artifacts": [], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-content-recreator", "name": "腾讯健康NGES医学二创专家", "description": "医学内容二创", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_XL25qlKax91g" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "nges-recreator-podcast", "title": "把一篇学术长文做成一期 8 分钟的医生通勤播客", "title_en": "Turn an Academic Article Into an 8-Minute Commute Podcast", "subtitle": "双人对话口播稿 + 可以直接听的音频", "subtitle_en": "A two-host conversational script plus audio you can actually listen to", "description": "医生通勤路上刷不动长文，但听得进音频。这个案例演示了怎么把一篇学术访谈改写成双人对话的口播稿——一个提问一个回答，不像一个人在念稿——再直接生成可以听的音频，最后补一页文字版给没时间听的人。适合数字营销做学术内容的多形态分发。", "description_en": "Physicians won't read long-form on their commute, but they will listen. This case shows how to rewrite an academic interview into a two-host conversational script — one asks, one answers, so it never sounds like someone reading aloud — then generate listenable audio from it, and finish with a one-page text version for people who have no time to listen. Built for digital marketing distributing academic content across formats.", "tags": [ "内容二创", "播客", "音频生成", "肺癌", "口播稿" ], "tags_en": [ "Content Recreation", "Podcast", "Audio Generation", "Lung Cancer", "Audio Script" ], "categories": [ "business", "creation" ], "artifact_type": "html", "prompt": "我们有一篇肺癌专家访谈，内容扎实但打开率低。想做成一期 8 分钟左右的音频，让医生上下班路上能听。形式是两个人对话——一个提问、一个回答，别做成一个人在念稿。\n\n原文如下：\n继表皮生长因子受体酪氨酸激酶抑制剂（EGFR-TKI）成为EGFR突变（Ex19del/L858R）晚期非小细胞肺癌（NSCLC）患者的标准治疗后，EGFR-TKI联合治疗策略取得了重要突破，有望进一步延长EGFR突变晚期NSCLC患者的生存[1]。在EGFR-TKI联合治疗策略发展的同时，针对其耐药机制及全程管理体系的探讨也在不断深入。特邀湖南省肿瘤医院邬麟教授，总结EGFR-TKI联合治疗策略的重要进展，并从EGFR-TKI联合治疗策略全程管理的角度发表独到见解。\n\n邬麟 教授｜专家简介：二级主任医师、博士生导师、博士后合作导师；湖南省肿瘤医院（中南大学湘雅医学院附属肿瘤医院）内科教研室主任，GCP肿瘤专业负责人、胸内二科主任；湖南省肺癌临床医学研究中心主任；中国抗癌协会(CACA)老年肿瘤专委会副主委；CSWOG肺癌专业委员会候任主委；CSCO理事；CACA临床化疗专业委员会常委；湖南省抗癌协会肿瘤化疗专业委员会主任委员等。\n\n肺癌的个体化精准诊疗是临床的热点和难点之一，能否请您谈谈如何为EGFR突变晚期NSCLC患者制定精准治疗策略？既往局限于医疗技术的落后，肺癌患者在临床中往往接受同质化的诊疗。EGFR突变的发现和EGFR-TKI疗法的兴起使肺癌治疗从经验性的化疗转向精准分子靶向治疗，自此晚期NSCLC迈向了精准诊疗时代。精准检测是实施精准靶向治疗的前提。邬麟教授介绍道，精准治疗，检测先行，晚期NSCLC患者在初诊时需进行基因检测，精准的基因检测对于晚期NSCLC的一线治疗选择具有指导作用，这已在全球形成共识，并成为标准的临床实践[1-2]。此外，基因检测在患者的全程管理中也十分重要。随着治疗时间的延长，EGFR-TKI耐药和疾病进展的出现不可避免，基因检测可以明确EGFR-TKI耐药机制，为后续精准治疗提供依据。针对EGFR-TKI一线治疗后出现进展的人群，国内外权威指南均建议根据进展模式和耐药机制指导后续治疗策略的制定[1-2]。总而言之，基因检测不仅在一线治疗决策中具有关键作用，更是肺癌全程管理中动态调整治疗方案的重要工具。基因检测指导下的精准治疗是确保EGFR突变晚期NSCLC患者获益的关键。需要强调的是，在肺癌的全程管理过程中，一线治疗方案的选择尤为重要。邬麟教授指出，一线治疗在EGFR突变晚期NSCLC患者的整体治疗进程中扮演着至关重要的角色，一线获益对于提升全程获益和患者治疗信心具有积极影响，有助于优化全程管理和患者的治疗依从性。继三代EGFR-TKI成为EGFR突变晚期NSCLC患者的一线标准治疗后，EGFR-TKI联合治疗策略的探索也已取得初步成功，进一步提高了患者获益。在选择一线用药方案时，亦需要把握全程管理的观念，从而优化整体治疗策略。\n\n参考文献：[1] 中国临床肿瘤学会（CSCO）非小细胞肺癌诊疗指南. 2024版. [2] NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Non-Small Cell Lung Cancer. Version 3.2025.\n\n帮我写一版口语化的双人对话稿，把学术内容讲成人话但不丢严谨度，关键数字和指南出处要在对话里自然带出来。再基于这版稿子生成一份可以直接听的音频。最后把关键结论整理成一页文字版，方便没时间听的人扫一眼。", "prompt_en": "We have a lung cancer expert interview that is solid but gets few opens. I want to turn it into an 8-minute audio piece a physician can listen to on the commute. The format should be a two-person conversation — one asks, one answers — not a single voice reading a script.\n\nSource article:\nFollowing the establishment of EGFR tyrosine kinase inhibitors (EGFR-TKIs) as standard therapy for patients with advanced EGFR-mutated (Ex19del/L858R) non-small cell lung cancer (NSCLC), EGFR-TKI combination strategies have achieved major breakthroughs and are expected to further extend survival [1]. In parallel, investigation into resistance mechanisms and whole-course management continues to deepen. This issue features Professor Lin Wu of Hunan Cancer Hospital, summarizing key advances in EGFR-TKI combination strategies and offering perspectives on whole-course management.\n\nProfessor Lin Wu — Profile: Grade-II Chief Physician, Doctoral Supervisor; Director of the Internal Medicine Teaching and Research Section, Hunan Cancer Hospital (The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University); Head of the Thoracic Internal Medicine Department II; Director of the Hunan Lung Cancer Clinical Medical Research Center; Vice Chair of the Geriatric Oncology Committee of the Chinese Anti-Cancer Association (CACA); President-Elect of the CSWOG Lung Cancer Committee; Council Member of CSCO.\n\nIndividualized precision diagnosis and treatment of lung cancer remains a clinical hotspot and challenge. Historically, limited medical technology meant patients received largely homogeneous care. The discovery of EGFR mutations and the advent of EGFR-TKIs shifted lung cancer treatment from empirical chemotherapy to precision molecular targeted therapy. Precision testing is the prerequisite for precision targeted therapy. As Professor Wu notes, \"precision treatment begins with testing\" — genetic testing at initial diagnosis is now a global consensus and standard clinical practice for guiding first-line selection in advanced NSCLC [1-2]. Genetic testing is equally critical across the whole disease course: as treatment duration extends, EGFR-TKI resistance and disease progression become inevitable, and testing clarifies the resistance mechanism to inform subsequent precision therapy. For patients progressing after first-line EGFR-TKI, major Chinese and international guidelines recommend tailoring subsequent strategy to the pattern of progression and the resistance mechanism [1-2]. In short, genetic testing is both pivotal to first-line decision-making and an essential tool for dynamically adjusting therapy across the whole course. The choice of first-line regimen is especially consequential: first-line benefit positively influences overall benefit and patient confidence, optimizing whole-course management and adherence. Following third-generation EGFR-TKIs as first-line standard, exploration of EGFR-TKI combination strategies has achieved initial success, further improving patient benefit. First-line selection should likewise be guided by a whole-course management mindset.\n\nReferences: [1] Chinese Society of Clinical Oncology (CSCO) Guidelines for the Diagnosis and Treatment of Non-Small Cell Lung Cancer, 2024 Edition. [2] NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell Lung Cancer, Version 3.2025.\n\nWrite a conversational two-host script that puts the academic content into plain language without losing rigor, working the key numbers and guideline sources into the dialogue naturally. Then generate listenable audio from that script. Finally, condense the key conclusions into a one-page text version for people who don't have time to listen.", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "把一篇学术长文做成一期 8 分钟的医生通勤播客" }, { "file": "output.md", "label": "把一篇学术长文做成一期 8 分钟的医生通勤播客（Markdown 版）" } ], "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [ { "id": "med-content-recreator", "name": "腾讯健康NGES医学二创专家", "description": "医学内容二创", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_XL25qlKax91g" }, { "id": "nges-team-lead", "name": "腾讯健康NGES医药营销专家团", "description": "整合HCP客户洞察、互动病例生成、学术物料生成、医学内容二创、合规审核五大能力的医药营销专家团", "expertType": "team", "source": "expert-marketplace", "expert_id": "ex_EISUIXwarcDL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T12:56:35+08:00", "updated_at": "2026-08-31T16:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "novel-mystery-island", "title": "雾屿灯塔 · 案件卷宗", "title_en": "The Foglight Isle · Case File", "subtitle": "把一部本格推理小说，做成可翻阅的侦探卷宗", "subtitle_en": "Turn a whodunit novel into a browsable detective dossier", "description": "写好一部封闭孤岛推理，却不知怎么呈现才够味？用「开放小说创作助手」按设定→大纲→章节规格→正文→评审全流程，产出悬疑小说《雾屿灯塔》前三章真实正文（含公平线索与身份诡计反转），再排成一份牛皮纸打字机质感的侦探卷宗：嫌疑人档案卡（照片位/编号/嫌疑标签）、贴满便签与红线连结的线索板、证物清单、案发时间线，正文以「档案节选」嵌入。适合推理作者、剧本杀设计者、沉浸式叙事策划。", "description_en": "Wrote a locked-island whodunit but not sure how to present it with flavor? Use the Open Novel Writing skill to run the full worldbuild → outline → spec → prose → review pipeline, producing the first three real chapters of the mystery 'The Foglight Isle' (fair-play clues plus an identity-swap twist), then lay it out as a kraft-paper, typewriter-style detective dossier: suspect dossier cards (photo slots / case numbers / suspicion flags), a corkboard of sticky notes linked by red string, an exhibit list, and a crime timeline — with the prose embedded as 'file excerpts'. Great for mystery authors, murder-mystery-game designers, and immersive-narrative planners.", "categories": [ "creation" ], "tags": [ "小说创作", "悬疑推理", "本格", "案件卷宗", "剧本杀" ], "tags_en": [ "novel writing", "mystery", "whodunit", "case file", "murder mystery game" ], "artifact_type": "html", "cover_mode": "landscape", "skills": [ { "id": "open-novel-writing", "skill_id": "skill_2053082516611272704", "name": "开放小说创作助手", "description": "中文长篇小说创作助手，支持从想法到成书的模块化全流程" } ], "experts": [], "mcps": [], "prompt": "用开放小说创作助手帮我写一部封闭孤岛的本格推理小说《雾屿灯塔》：大雾封岛、灯塔看守人坠亡、五名被困者各有隐瞒，真相是一场跨越二十年的身份顶替诡计。按设定→大纲→章节规格→正文→评审的完整流程，产出前三章真实正文，要求公平线索、读者可与侦探同步推理，并做一次五角色评审。\n\n然后把它做成单文件 HTML 的「侦探卷宗」：牛皮纸＋打字机质感，包含嫌疑人档案卡（照片位/编号/嫌疑标签）、贴便签＋红线连结的线索板、证物清单表格、案发时间线，正文以「档案节选」嵌入。内联无外链。", "prompt_en": "Use the Open Novel Writing skill to write a locked-island fair-play mystery, 'The Foglight Isle': the island is fogbound, the lighthouse keeper falls to his death, five stranded people each hide something, and the truth is a twenty-year identity-swap. Run the full worldbuild → outline → spec → prose → review pipeline, produce the first three real chapters with fair clues the reader can solve alongside the detective, plus a five-role review.\n\nThen build it as a single-file HTML 'detective dossier': kraft paper + typewriter feel, with suspect dossier cards (photo slot / case number / suspicion flag), a corkboard of sticky notes linked by red string, an exhibit table, and a crime timeline — with the prose embedded as 'file excerpts'. Inline with no external links.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "scenario": "", "created_at": "2026-07-20T00:00:00Z", "updated_at": "2026-07-20T00:00:00Z", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "雾屿灯塔 · 案件卷宗" } ], "preview_en": "output.html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "novel-scifi-memory", "title": "潮汐记忆 · 沉浸式阅读器", "title_en": "Mnemotide · Immersive Reader", "subtitle": "深色全屏阅读体验，让长篇科幻只剩下「读」", "subtitle_en": "A dark full-screen reading experience — sci-fi with nothing but the page", "description": "想把一部原创长篇小说做成让人一口气读下去的作品，而不是花哨的卡片展示？用「开放小说创作助手」按设定→大纲→章节规格→正文→评审的完整流程，产出科幻小说《潮汐记忆》前三章真实正文，再呈现为深色沉浸式全屏阅读器：顶部细如发丝的进度条、章节间优雅过场、大段正文随滚动渐入。适合小说作者、网文创作者、想做电子书样章的独立出版人。", "description_en": "Want to turn an original novel into something readers devour, not a flashy card gallery? Use the Open Novel Writing skill to run the full worldbuild → outline → chapter-spec → prose → review pipeline, producing the first three real chapters of the sci-fi novel 'Mnemotide', then present it as a dark, immersive full-screen reader: a hairline top progress bar, elegant chapter interstitials, and long prose fading in as you scroll. Great for novelists, web-fiction writers, and indie publishers building sample chapters.", "categories": [ "creation" ], "tags": [ "小说创作", "科幻", "沉浸式阅读", "长篇正文", "电子书样章" ], "tags_en": [ "novel writing", "sci-fi", "immersive reading", "long-form prose", "ebook sample" ], "artifact_type": "html", "cover_mode": "landscape", "skills": [ { "id": "open-novel-writing", "skill_id": "skill_2053082516611272704", "name": "开放小说创作助手", "description": "中文长篇小说创作助手，支持从想法到成书的模块化全流程" } ], "experts": [], "mcps": [], "prompt": "用开放小说创作助手帮我从零创作一部科幻长篇《潮汐记忆》：设定一个海水淹没大陆、记忆可被打捞交易的世界，主角是靠捞取记忆换水的「拾忆人」。按设定→大纲→章节规格→正文→评审的完整流程，产出前三章真实正文（每章数千字中文原文），并做一次五角色评审。\n\n然后把它做成单文件 HTML 的沉浸式全屏阅读器：深色海洋背景、顶部极细进度条、封面首屏、章节之间优雅过场、正文随滚动渐入，重「阅读」而非卡片展示。内联无外链。", "prompt_en": "Use the Open Novel Writing skill to create an original sci-fi novel 'Mnemotide' from scratch: a world where the sea has drowned the continents and memory can be salvaged and traded, following a 'memory-salvager' who trades recovered memories for water. Run the full worldbuild → outline → chapter-spec → prose → review pipeline and produce the first three chapters of real prose (several thousand Chinese characters each), plus a five-role review.\n\nThen turn it into a single-file HTML immersive full-screen reader: dark ocean background, a hairline top progress bar, a cover hero screen, elegant chapter interstitials, and prose that fades in on scroll — focused on reading, not card displays. Inline with no external links.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "scenario": "", "created_at": "2026-07-20T00:00:00Z", "updated_at": "2026-07-20T00:00:00Z", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "潮汐记忆 · 沉浸式阅读器" } ], "preview_en": "output.html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "novel-urban-congee", "title": "晚粥灯火 · 深夜食帖", "title_en": "Late-Night Congee · Night Kitchen Zine", "subtitle": "把一部治愈系都市言情，做成一本温暖的美食杂志", "subtitle_en": "Turn a healing urban romance into a warm food magazine", "description": "想让一部温柔的治愈系小说被人一眼就想读下去？用「开放小说创作助手」按设定→大纲→章节规格→正文→评审全流程，产出都市治愈言情《晚粥灯火》前三章真实正文（深夜粥铺、双向治愈、不狗血），再排成一本温暖的绘本·美食杂志：暖色圆角、创刊号刊头、招牌与「今夜菜单」元素、每章配一幅内联 SVG 场景插画（雨夜巷口/八张桌/老砂锅）、故事里的粥做成食谱卡。适合言情/治愈系作者、内容策划、想做「小说＋美食」跨界样章的创作者。", "description_en": "Want a gentle healing novel that readers want to dive into at first glance? Use the Open Novel Writing skill to run the full worldbuild → outline → spec → prose → review pipeline, producing the first three real chapters of the healing urban romance 'Late-Night Congee' (a midnight congee shop, mutual healing, no melodrama), then lay it out as a warm picture-book food magazine: soft warm tones and rounded corners, a debut-issue masthead, signboard and 'tonight's menu' elements, an inline-SVG scene illustration per chapter (rainy alley / eight tables / the old clay pot), and the story's congee rendered as a recipe card. Great for romance/healing authors, content planners, and creators building 'fiction + food' crossover samples.", "categories": [ "creation" ], "tags": [ "小说创作", "治愈言情", "都市", "美食杂志", "绘本插画" ], "tags_en": [ "novel writing", "healing romance", "urban", "food magazine", "picture-book illustration" ], "artifact_type": "html", "cover_mode": "landscape", "skills": [ { "id": "open-novel-writing", "skill_id": "skill_2053082516611272704", "name": "开放小说创作助手", "description": "中文长篇小说创作助手，支持从想法到成书的模块化全流程" } ], "experts": [], "mcps": [], "prompt": "用开放小说创作助手帮我写一部治愈系都市言情《晚粥灯火》：一家只在深夜营业的粥铺，老板看客人脸色决定煮什么粥，女主是刚失去母亲、不敢回家的美食撰稿人。按设定→大纲→章节规格→正文→评审的完整流程，产出前三章真实正文，双向治愈、温暖不狗血，并做一次五角色评审。\n\n然后把它做成单文件 HTML 的「温暖绘本·美食杂志」：暖色调、柔和圆角、创刊号刊头，配招牌与今夜菜单元素，每章一幅内联 SVG 场景插画（雨夜巷口 / 八张桌 / 老砂锅），故事里的粥做成食谱卡。内联无外链。", "prompt_en": "Use the Open Novel Writing skill to write a healing urban romance, 'Late-Night Congee': a congee shop open only at night, where the owner reads a guest's face to decide what to cook, and the heroine is a food writer who just lost her mother and can't face going home. Run the full worldbuild → outline → spec → prose → review pipeline, produce the first three real chapters with mutual healing, warm and free of melodrama, plus a five-role review.\n\nThen build it as a single-file HTML 'warm picture-book food magazine': warm tones, soft rounded corners, a debut-issue masthead, signboard and 'tonight's menu' elements, one inline-SVG scene illustration per chapter (rainy alley / eight tables / the old clay pot), and the story's congee as a recipe card. Inline with no external links.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "scenario": "", "created_at": "2026-07-20T00:00:00Z", "updated_at": "2026-07-20T00:00:00Z", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "晚粥灯火 · 深夜食帖" } ], "preview_en": "output.html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "nursing-pm-workbench", "title": "医院护理信息项目管理工作台", "title_en": "Nursing Informatics PM Workbench", "subtitle": "把医院护理信息化建设装进一个可审计的项目指挥台", "subtitle_en": "A single-file command desk for hospital nursing informatics projects", "description": "护理信息化项目多、口径杂、跨年周期长，靠表格和群聊追进度很容易失焦。\n\n这个案例把一家医院护理信息专科（NIC）十余年的真实运转搬进资料库：六张在线数据表（任务台账 779 条、统计事务 95 条、年度总结 19 份、周报 18 期、到会统计与趋势）作为唯一数据源，页面直连读写、双向同步；数据治理内建——受控词表约束科室与状态口径，DAMA 六维数据质量评估直接看分；源表文件上传后离线解析、差异比对，合并结果自动写回在线表，全站指标实时重算。每一项分析都能点回原始记录。\n\n适合：医院信息科/护理部管理信息化项目群、想参考「真存表 + 数据治理内建 + 分析可审计」工作台范式的数字化负责人。", "description_en": "Nursing informatics projects are numerous, inconsistently coded, and span years — tracking them in spreadsheets and group chats loses focus fast.\n\nThis case moves a hospital nursing informatics committee's (NIC) decade of real operations into the Library: six online database tables (a 779-row task ledger, 95 statistics log entries, 19 annual summaries, 18 weekly reports, plus attendance stats and trends) serve as the single source of truth, with the page reading and writing directly and syncing both ways; governance is built in — a controlled vocabulary enforcing department/status coding plus a DAMA six-dimension quality score; source files are parsed offline in-browser, diffed, and merged results auto-written back to the online tables with site-wide metrics recomputed live. Every analysis links back to raw records.\n\nBest for: hospital IT / nursing departments running informatics project portfolios, and digital leads looking for a workbench pattern with real table storage, governance, and auditability built in.", "categories": [ "workspace" ], "tags": [ "护理信息", "项目管理", "数据治理", "DAMA", "医院工作台" ], "tags_en": [ "Nursing Informatics", "Project Management", "Data Governance", "DAMA", "Hospital Workbench" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "医院护理信息项目管理工作台（中英双语单文件 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/h7jjTmfehPVtKZZRRR1Xha，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「医院护理信息项目管理工作台」，左侧固定导航 + 顶部工具栏，约 13 个视图，数据存资料库六张在线数据表（页面通过 database SDK 直接读写、双向同步）：① 护理项目任务表（约 779 条：编号/年度/优先级/需求类别/提交部门/业务系统/问题描述/处理方案/状态/所处阶段/提交·计划·完成日期/历时天数/跨年长尾）；② 数据统计事务表（约 95 条：含提交人/安全风险/事务类型/伦理审批）；③ 年度总结归档表（约 19 条：系列/年度/标题/作者/全文）；④ 周报归档表（约 18 期：期号/周期/栏目数据/备注/计数）；⑤ NIC到会统计表（各院区到会矩阵）；⑥ NIC到会趋势表（历年各院区到会率）。视图：① 总览 KPI 卡与趋势图；② 项目台账多维筛选表格（科室/年份/状态/关键字），行点开看详情；③ 看板按状态分列；④ 统计分析多维聚合图表；⑤ 数据治理：受控词表管理 + DAMA 六维（完整性、唯一性、一致性、及时性、准确性、有效性）数据质量评分卡；⑥ 年度总结按年份归档全文查看；⑦ 到会统计矩阵与趋势曲线；⑧ 数据导入与同步：上传源表（xlsx/docx）离线解析 → 差异比对（新增/更新/移除）→ 合并后自动写回在线数据表，全站指标重算。页面加载时从六张表分页拉全量（>200 条递归 startCursor 翻页），任何设备打开均为最新数据。视觉：医疗蓝白配色、密集表格风格、PC 优先兼容移动端。\n\n交付：资料库在线工作台——在线 HTML 页面 + 14 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/h7jjTmfehPVtKZZRRR1Xha with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Nursing Informatics PM Workbench\" with a fixed left nav and top toolbar, around 13 views, backed by six online Library database tables (the page reads and writes through the database SDK with two-way sync): ① Nursing Project Tasks (~779 rows: id/year/priority/category/department/system/description/solution/status/stage/submit·plan·done dates/duration/carried-over); ② Statistics Log (~95 rows: submitter/security risk/type/ethics review); ③ Annual Summaries (~19 rows: series/year/title/author/full text); ④ Weekly Report Archive (~18 issues: issue no./period/section data/notes/counts); ⑤ NIC Attendance matrices per campus; ⑥ NIC Attendance Trend (annual rates by campus). Views: ① overview KPI cards and trend charts; ② project ledger with multi-dimensional filters (department/year/status/keyword) and expandable row details; ③ status kanban; ④ multi-dimensional analytics; ⑤ data governance: controlled-vocabulary management plus a DAMA six-dimension quality scorecard (completeness, uniqueness, consistency, timeliness, accuracy, validity); ⑥ annual summaries archived by year with full-text view; ⑦ attendance matrices and trend curves; ⑧ data import & sync: upload source files (xlsx/docx) parsed offline in-browser → diff (add/update/remove) → merge auto-written back to the online tables, with site-wide metrics recomputed. On load, page pulls all tables with cursor pagination (recursive startCursor beyond 200 rows) so every device sees the latest data. Visual: medical blue-and-white palette, dense table style, desktop-first with mobile compatibility.\n\nDeliverable: an online Library workbench — an online HTML page plus 14 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "🍄蘑菇大🍄", "name_en": "🍄蘑菇大🍄", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "community", "usage_count": 0, "created_at": "2026-09-10T12:50:00+08:00", "updated_at": "2026-09-10T12:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "office-computer-procurement-acceptance-payment-risk", "title": "办公电脑采购验收风险", "title_en": "Office Computer Procurement Risk Review", "subtitle": "围绕配置不符、设备故障、验收异议与暂缓付款的互动法律分析", "subtitle_en": "Interactive legal analysis for configuration variance, device defects, acceptance objections, and payment withholding", "description": "痛点：办公电脑到货后发现配置不符、设备故障，而供应商要求先付全款再售后，采购、IT 与法务很难在验收期限和付款节点内统一行动。演示：本案例以可点击验收流程、付款情形判断、合同条款矩阵、责任路径、函件要点和证据清单，展示如何形成可执行的初步分析。适用人群：企业采购负责人、IT 资产管理人员、法务与内控团队，以及需要处理设备采购履约争议的业务管理者。", "description_en": "Pain point: Office computers arrive with configuration discrepancies and device failures while the supplier demands full payment before after-sales support, leaving procurement, IT, and legal teams struggling to align within inspection and payment windows. Demonstration: This case uses a clickable acceptance workflow, payment scenarios, contract matrix, liability routes, notice essentials, and evidence checklist to demonstrate an actionable preliminary review. For: Corporate procurement leads, IT asset managers, legal and internal-control teams, and business managers handling equipment procurement performance disputes.", "tags": [ "办公电脑采购", "验收", "质量异议", "暂缓付款", "买卖合同", "证据留存" ], "tags_en": [ "office computer procurement", "acceptance", "quality objection", "payment withholding", "sale contract", "evidence preservation" ], "prompt": "公司向供应商采购了一批办公电脑，合同约定指定品牌、型号及关键配置。货到验收时发现部分电脑配置不符，另有数台无法开机、频繁死机或存在硬件故障；供应商要求先支付全部货款再处理售后。请基于现行有效法律和合同约定，对以下事项形成一份可供管理层决策的初步法律分析报告：1）按合同付款节点、验收约定、未通过验收、部分瑕疵与重大违约分别判断公司能否暂缓付款及合理范围；2）设计从到货签收至异议通知的验收和证据留存流程；3）提出整改、维修、换货、退货、减价、赔偿、违约责任或解除的适用条件与行动顺序；4）列明应审查的合同条款、函件要点和待核实事实。报告须明确其基于假设事实，关键结论待合同原件、检测记录和送达证据核验。", "prompt_en": "The company purchased a batch of office computers from a supplier under a contract specifying a designated brand, model, and key configurations. At delivery inspection, some computers have configurations inconsistent with the agreement, while several cannot boot, crash repeatedly, or have hardware defects; the supplier demands full payment before providing after-sales service. Based on currently effective law and the contract, prepare a preliminary legal analysis for management decision-making that: (1) separately assesses whether and to what reasonable extent payment may be withheld under the contract payment milestones, acceptance terms, failed acceptance, partial defects, and material breach; (2) designs an acceptance and evidence-preservation workflow from receipt through objection notice; (3) sets out the conditions and sequence for remediation, repair, replacement, return, price reduction, damages, breach liability, or termination; and (4) identifies contract clauses, notice essentials, and facts to verify. The report must state that it is based on assumed facts and that key conclusions await verification against the executed contract, testing records, and proof of delivery.", "artifact_type": "html", "preview": "办公电脑采购合同履约、验收与暂缓付款法律分析报告.html", "artifacts": [ { "file": "办公电脑采购合同履约、验收与暂缓付款法律分析报告.html", "label": "互动法律分析报告" } ], "categories": [ "business" ], "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "资深合同法律顾问，支持合同起草、签署、审查、对比与法规检索", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "created_at": "2026-08-13T12:10:31+08:00", "updated_at": "2026-08-13T15:40:12+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "office-lease", "title": "办公场地房屋租赁合同", "title_en": "Office Lease Agreement", "subtitle": "把租金、押金、用途与提前解约写得清楚", "subtitle_en": "Make rent, deposit, permitted use and early termination unambiguous.", "description": "租办公室不只是写租金，物业、用途、维修和提前解约往往才是后续争议焦点。\n\n本案例把 200 平方米办公场地租赁需求生成可编辑的《办公场地房屋租赁合同》，并以信息卡方式展示租期、租金、押金、用途与违约责任。\n\n适合：行政负责人、创业公司、园区招商团队、商业物业管理方。", "description_en": "Office leases involve more than rent: property fees, permitted use, repairs and early termination often become the real dispute points.\n\nThis case turns a 200-square-meter office lease brief into an editable Office Lease Agreement, with key cards for term, rent, deposit, use and breach liability.\n\nSuitable for: administrative leads, startups, park leasing teams and commercial property managers.", "categories": [ "business" ], "tags": [ "房屋租赁", "办公场地", "押金", "提前解约" ], "tags_en": [ "Lease Agreement", "Office Space", "Deposit", "Early Termination" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请起草一份办公场地房屋租赁合同。出租方为[甲方名称]，承租方为[乙方名称]；房屋位于[地址]，面积[面积]平方米，仅作[用途]使用。租期[起止日期]，月租金[金额]元，押金[金额]元，租金按[支付周期]支付，物业费由[承担方]承担。请明确房屋权属与交付、维修、转租限制、用途限制、解除条件、逾期付款、提前解约及争议解决。请输出可直接签署的中文 Word 合同。", "prompt_en": "Please draft an office lease agreement. The lessor is [甲方名称] and the lessee is [乙方名称]. The premises are at [地址], with [面积] square meters, used only for [用途]. The term is [起止日期], monthly rent is [金额] RMB, deposit is [金额] RMB, rent is paid [支付周期], and property fees are borne by [承担方]. Clearly define title and handover, repairs, sublease limits, permitted use, termination conditions, late payment, early termination and dispute resolution. Produce a signable Chinese Word contract.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-11T21:04:09+08:00", "updated_at": "2026-08-12T13:37:31+08:00", "artifacts": [ { "file": "output.html", "label": "V3.0 互动合同起草页面" }, { "file": "办公场地房屋租赁合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "oleme-job-buddy", "artifact_type": "html", "title": "O了么·求职搭子工作台", "title_en": "Oleme Job-Hunt Buddy Workbench", "subtitle": "大一到大四的求职手账：进度、情报、行动、复盘、资产五个页签，15 张数据表接住全过程", "subtitle_en": "A job-hunt journal from year 1 to year 4: progress, intel, action, review and assets across 15 database tables", "description": "大一大二不知道该准备什么、大三投递全凭感觉、大四手里几个 Offer 又算不清该选哪个——信息散在备忘录、群聊和十几个招聘网站里，没走两步就断线。这个工作台把求职拆成五本手账，用 15 张数据表接住全过程：进度页用「今天要处理」把逾期任务、快到期投递和今日三件事拢到一处，大一到大四时间轴一点就换阶段，任务打勾自动记一笔投入，打卡、月历和近 14 天投入图看得出节奏；情报页把校园招聘、事业单位和实习公告分成三个栏目，关键词命中的岗位自动置顶，一键生成投递包、随简历实时刷新润色建议；行动页是能拖拽的六列投递看板，配高频题库、STAR 故事库、人脉内推 CRM 和七项加权的 Offer 决策表；复盘页给出转化率与渠道效果，四栏便签复盘可一键转任务；我的页沉淀多版本简历、成就清单和材料包。演示数据是一份完整的大三实习求职实录：9 条任务、14 个真实在招岗位、7 张投递卡、6 道高频题、2 版简历和 1 份 Offer 对比表。适合不知道怎么规划的大一新生、正在攒实习的大二大三学生，以及手里有多个 Offer 要做选择的大四应届生。", "description_en": "Year 1 and 2 pass without knowing what to prepare, year 3 applies on gut feeling, and by year 4 there are several offers but no clear way to choose — the information is scattered across notes, group chats and a dozen job sites, and the thread breaks after two steps. This workbench splits the job hunt into five journals backed by 15 database tables: the Progress tab gathers overdue tasks, imminent deadlines and today's top three into one \"To handle today\" strip, switches stages with a tap on the Year 1 → Year 4 timeline, logs time automatically when a task is ticked, and shows your rhythm through check-ins, a month calendar and a 14-day investment chart; the Intel tab splits campus recruitment, public-institution and internship notices into three sections, pins keyword matches to the top, builds a one-click application pack and refreshes resume polish tips as you edit; the Action tab is a draggable six-column application board with a frequently-asked question bank, a STAR story bank, a referral CRM and a seven-criteria weighted offer comparison; the Review tab reports conversion rates and channel performance, and the four-column sticky-note review converts into tasks in one tap; My Offer collects multiple resume versions, achievements and an asset pack. The demo data is a complete year-3 internship job hunt: 9 tasks, 14 live postings, 7 application cards, 6 frequently asked questions, 2 resume versions and 1 offer comparison. Best for first-years who don't know where to start, second- and third-years building internships, and final-years choosing between offers.", "tags": [ "求职工作台", "投递看板", "面试题库", "求职搭子", "在线数据表" ], "tags_en": [ "job-hunt workbench", "application board", "interview prep", "job buddy", "online tables" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/fApg8dLRt3lp3TYwyLePRt，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「O了么 · 大学生的求职搭子」求职工作台，面向大一到大四在校生，把「规划 → 情报 → 行动 → 复盘 → 资产沉淀」串成一条主线。手账贴纸风：虚线卡片、和纸胶带、便签纸、手绘抖动圆角，五套治愈系配色（玫瑰粉灰 / 奶茶棕 / 鼠尾草绿 / 雾霾蓝 / 雾霾紫）可一键切换；PC 左侧导航、移动端底部标签栏，两栏自适应。\n\n数据表（共 15 张，均落资料库数据表）：\n- 求职任务：任务（文本）、年级（单选 大一/大二/大三/大四）、优先级（单选 普通/重要/紧急）、截止日期（日期）、已完成（复选框）、排序（数字）\n- 每日打卡：日期（日期）、打卡习惯（文本，多个习惯 id 逗号连接）、投递分钟 / 简历分钟 / 笔试面试分钟 / 行业学习分钟（数字）、焦虑值 / 能量值（数字）\n- 投入记录：事项（文本）、日期（日期）、分钟（数字）、分类（单选 投递/简历/笔试面试/行业学习/其他）\n- 日程事件：日期（日期）、事项（文本）\n- 岗位情报：公司、岗位、类型（单选 校园招聘/事业单位/实习）、城市、截止日期（日期）、来源、可信度、匹配度（数字）、竞争热度、所需附件、下一步、简介、更新批次（文本）、官网链接（链接）、已收藏（复选框）\n- 匹配清单：公司、岗位、类型（单选）、求职要点、已投递（复选框）、岗位标识\n- 投递看板：公司、岗位、渠道、阶段（单选 想投/已投/笔试/面试/Offer/归档）、截止日期（日期）、下一步、联系人\n- 面试备战：题目或标题、类型（单选 高频题/STAR故事/模拟面试/模板）、要点、情境S、任务T、行动A、结果R、得分（数字）、模拟日期（日期）\n- 人脉内推：姓名、身份、公司、内推码、跟进事项\n- Offer决策：公司、月薪K（数字）、城市、户口分 / 成长分 / 加班分 / 稳定分 / 违约月（数字）、已选（复选框）\n- 成就清单：名称、分类（单选 成绩记录/获奖记录/实习记录/学生工作/社团活动/社会活动）、量化成果、标签、情境S、任务T、行动A、结果R\n- 材料包：名称、类型（单选 成绩单/证书/作品集/证件照/图片）、图片（图片）\n- 简历版本：版本名、更新日期（日期）、当前使用（复选框）、姓名、学校、电话、邮箱、教育背景、实践经历、技能特长\n- 复盘周报：周起始、保持、问题、尝试、下周预案\n- 我的档案：姓名、专业、年级、城市、求职意向、目标行业、霍兰德、MBTI、职业锚、价值观、技能盘点、订阅关键词、周目标天（数字）、今日三件事（文本 JSON）、Offer权重（文本 JSON）、焦虑值 / 能量值（数字）、当前阶段\n\n五个页签：\n- offer进度：顶部「今天要处理」自动汇总逾期任务、3 天内截止的投递与今日三件事（逾期标红、未完成自动顺延）；大一→大四求职时间轴点击节点即可切换阶段；本阶段任务支持新增、勾选完成（自动记一条投入记录）、上下移、拖拽排序、删除；每日打卡含本周进度环、连续天数、习惯增删与补卡；今日三件事可勾选增删；月历按 DDL / 日程 / 打卡三种点标记，点某天弹出当天事项并可直接补记投入分钟；近 14 天按投递 / 简历 / 笔试面试 / 行业学习分目标堆叠柱状图；最近 10 条投入记录可改说明、可删；任务 DDL 可一键导出 ICS 日历。大一 / 大二阶段额外出现「搭子建议」卡，按内置建议库结合测评结果、订阅关键词与近 7 天投入自动调整，可一键加入任务。\n- offer情报：订阅关键词命中岗位自动置顶；分「校园招聘 / 事业单位公告 / 实习公告」三个栏目展示岗位卡（类型、城市、截止倒计时、来源、可信度、匹配度、竞争热度、所需附件、下一步），卡片可收藏、可跳官网、可一键加入投递看板；「我的精准匹配清单」按简历与关键词关联显示待投岗位，支持标记投递、生成一键投递包（含简历版本与所需附件清单）、删除；「岗位简历润色建议」随当前简历内容实时生成待补关键词 / 量化 / 经历 / 证书提醒，可标 ✓ 完成或 ✗ 忽略并记住状态；底部推荐攻略视频每次随机 3 个、可换一批、可弹层预览或跳原站。\n- offer行动：投递看板六列（想投 / 已投 / 笔试 / 面试 / Offer / 归档），卡片拖拽或点状态字切换，点卡片可改状态、截止、下一步、联系人并删除；面试备战含高频题库（增删、去重）、STAR 故事库、模拟面试与录音自评记录；人脉与内推 CRM 记姓名 / 身份 / 公司 / 内推码 / 跟进事项，可标记已跟进；Offer 决策表按薪资 / 城市 / 户口 / 成长 / 加班 / 稳定 / 违约成本七项加权评分，可加对比项、算最优；模板库存感谢信、跟进信、Argue 信、自我介绍等，可复制全文。\n- offer复盘：情绪能量（焦虑值 / 能量值滑块，配微行动 / 拖延阻断 / 同辈压力脱敏提示条）；数据看板给出投递→笔试、投递→面试、投递→Offer 转化率、各渠道效果表、14 天投入堆叠柱状图与目标进度环；本周四栏复盘（保持 / 问题 / 尝试 / 下周预案）以便签形式填写，预案可一键转成本阶段任务；历史周自动归档，含失败归档与成功经验标签。\n- 我的offer：个人基本信息（姓名 / 专业 / 年级 / 城市 / 求职意向 / 目标行业）；职业偏好测试（霍兰德 / MBTI / 职业锚 / 价值观 / 技能盘点）可编辑，改动后实时联动匹配分析与岗位匹配度；多版本简历可切换当前版本、在线编辑并生成完整简历文本；成就清单按 STAR 记录，可一键生成简历 bullet；材料包支持登记文件与上传图片（上传自动压缩到 800px 宽、最多 9 张，图片随数据一起云端存储）。\n\n交付：资料库在线工作台，PC 与移动端自适应。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/fApg8dLRt3lp3TYwyLePRt with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows:\n\nBuild an \"Oleme · A Job-Hunting Buddy for College Students\" job-hunt workbench for students from year 1 to year 4, threading \"plan → intel → act → review → asset building\" into one storyline. Paper-journal sticker style: dashed cards, washi tape, sticky notes, hand-drawn wobbly rounded corners, five soothing colour themes (rose grey / milk-tea brown / sage green / haze blue / haze purple) switchable in one tap; left sidebar on desktop, bottom tab bar on mobile, responsive two-column layout.\n\nData tables (15 in total, all stored in Library database tables):\n- Job tasks: task (text), year (select: Year 1/2/3/4), priority (select: Normal/Important/Urgent), due date (date), done (checkbox), order (number)\n- Daily check-in: date (date), habits (text, habit ids joined by commas), apply minutes / resume minutes / test & interview minutes / industry-learning minutes (number), anxiety / energy (number)\n- Time log: item (text), date (date), minutes (number), category (select: Apply/Resume/Tests & interviews/Industry learning/Other)\n- Calendar events: date (date), item (text)\n- Job intel: company, role, type (select: Campus recruitment/Public institution/Internship), city, deadline (date), source, credibility, match score (number), competition level, required files, next step, summary, update batch (text), official link (url), saved (checkbox)\n- Shortlist: company, role, type (select), key points, applied (checkbox), role id\n- Application board: company, role, channel, stage (select: To apply/Applied/Written test/Interview/Offer/Archived), deadline (date), next step, contact\n- Interview prep: title, type (select: Frequently asked/STAR story/Mock interview/Template), key points, situation S, task T, action A, result R, score (number), mock date (date)\n- Referral CRM: name, relation, company, referral code, follow-up\n- Offer decision: company, salary k (number), city, hukou / growth / overtime / stability / penalty (number), chosen (checkbox)\n- Achievements: name, category (select: Academic record/Award/Internship/Student leadership/Club activity/Community activity), metrics, tags, situation S, task T, action A, result R\n- Asset pack: name, type (select: Transcript/Certificate/Portfolio/ID photo/Image), image (image)\n- Resume versions: version name, updated (date), current (checkbox), name, school, phone, email, education, experience, skills\n- Weekly review: week starting, keep, problem, try, next-week plan\n- My profile: name, major, year, city, target role, target industries, Holland, MBTI, career anchor, values, skill inventory, subscribed keywords, weekly goal days (number), top three today (text JSON), offer weights (text JSON), anxiety / energy (number), current stage\n\nFive tabs:\n- Offer Progress: a \"To handle today\" strip at the top gathers overdue tasks, applications due within 3 days and today's top three (overdue in red, unfinished items roll over); the Year 1 → Year 4 timeline switches the active stage when you tap a node; stage tasks support adding, ticking off (which writes a time-log entry automatically), moving up/down, drag reordering and deleting; daily check-in shows a weekly progress ring, streak, habit add/remove and make-up check-ins; the top three can be ticked and removed; the month calendar marks deadlines, events and check-ins, and tapping a day opens that day's items with an inline shortcut to log minutes; a 14-day stacked bar chart splits minutes by apply / resume / tests & interviews / industry learning; the latest 10 time-log entries can be edited or deleted; task deadlines export to an ICS calendar. In year 1 and 2 an extra \"buddy suggestions\" card appears, tuned by a built-in suggestion library using test results, subscribed keywords and the last 7 days of invested time, with one-tap \"add as task\".\n- Offer Intel: jobs matching subscribed keywords are pinned to the top; three sections (campus recruitment / public-institution notices / internship notices) show job cards (type, city, deadline countdown, source, credibility, match score, competition level, required files, next step) that can be saved, opened on the official site or added to the application board in one tap; \"My shortlist\" is matched from the resume and keywords and supports marking as applied, generating a one-click application pack (resume version plus required-file checklist) and deleting; \"Resume polish tips for this role\" is generated live from the current resume with missing keyword / metrics / experience / certificate reminders that can be marked ✓ done or ✗ ignored with the state remembered; recommended guide videos show 3 at random each time, can be shuffled, previewed in a modal or opened on the source site.\n- Offer Action: a six-column application board (To apply / Applied / Written test / Interview / Offer / Archived) where cards can be dragged or moved by tapping a stage, and opening a card lets you edit its stage, deadline, next step and contact or delete it; interview prep holds a frequently-asked question bank (add, delete, dedupe), a STAR story bank, and mock-interview self-review records; the referral & network CRM stores name / relation / company / referral code / follow-up and can mark a contact as followed up; the offer decision table weights seven criteria (salary, city, hukou, growth, overtime, stability, penalty cost), accepts extra rows and picks the best; the template library ships thank-you notes, follow-up letters, an offer-negotiation letter and self-introductions with copy-full-text.\n- Offer Review: mood & energy (anxiety / energy sliders with micro-action, procrastination-blocker and peer-pressure tips); the dashboard shows apply→test, apply→interview and apply→offer conversion rates, a channel performance table, a 14-day stacked bar chart and a goal progress ring; the four-column weekly review (keep / problem / try / next-week plan) is written on sticky notes and the plan converts into stage tasks in one tap; past weeks are archived automatically together with a failure log and success tags.\n- My Offer: personal profile (name / major / year / city / target role / target industries); career preference tests (Holland / MBTI / career anchor / values / skill inventory) are editable and feed the match analysis and role scores live; multiple resume versions can be switched, edited online and rendered as full resume text; achievements are recorded as STAR stories and can generate resume bullets in one tap; the asset pack registers files and accepts uploaded images (auto-compressed to 800px wide, up to 9, stored in the cloud with the data).\n\nDeliverable: a Library online workbench, responsive on desktop and mobile.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "O了么·求职搭子工作台（自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd9b7744a331", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "大鹿鹿学姐", "name_en": "Dalulu", "avatar": "", "badge": "community" }, "source": "community", "skill_id": null, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 97, "created_at": "2026-09-12T21:29:33+08:00", "updated_at": "2026-09-12T21:29:33+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-drug-target-evidence", "title": "药物靶点外推证据评估", "title_en": "Drug Target Extrapolation Evidence Review", "subtitle": "FGFR2 扩增胃癌用 FGFR 抑制剂？证据链逐环审查", "subtitle_en": "FGFR inhibitor for FGFR2-amplified gastric cancer? A link-by-link evidence audit", "description": "测序发现一个可成药的扩增靶点，临床同事马上问：能不能用那个已经上市的抑制剂？「药能抑制靶点」和「药对这个病人有效」之间隔着一条外推链，链条断在哪一环往往没人说得清。这个案例审查一项真实常见的临床提议——「FGFR2 扩增（CNV=8）的胃癌患者 off-label 使用胆管癌获批的 FGFR 抑制剂 pemigatinib/futibatinib」：把提议拆成靶点检测、靶点-疾病、标志物对齐、药物-靶点、适应症外推五个环节，用 PubMed + KEGG DRUG 真实检索逐环核查（ToGA III 期作对照链），定位证据断裂点（癌种 + 变异类型双重跨线外推、无 RCT 支撑），输出分层用药路径建议与可检验假设。适合：精准医学方向研究者、肿瘤分子病理/临床基因组学人员、靶向治疗方案讨论参与者。", "description_en": "Sequencing finds a druggable amplification, and the clinician immediately asks: can we use that already-approved inhibitor? Between \"the drug inhibits the target\" and \"the drug helps this patient\" lies an extrapolation chain — and no one can usually say which link breaks. This case audits a real-world clinical proposal: \"off-label use of the cholangiocarcinoma-approved FGFR inhibitors pemigatinib/futibatinib in an FGFR2-amplified (CNV=8) gastric cancer patient.\" It decomposes the proposal into five links — target detection, target-disease, biomarker alignment, drug-target, and indication extrapolation — checks each against real PubMed and KEGG DRUG retrieval (with the ToGA phase-III trial as the comparison chain), locates the breaking point (dual cross-line extrapolation of both cancer type and variant class, with no RCT support), and outputs tiered treatment-path recommendations plus testable hypotheses. Suitable for: precision-medicine researchers, molecular pathology / clinical genomics staff, and anyone joining targeted-therapy plan discussions.", "categories": [ "data", "search" ], "tags": [ "WGS", "靶点评估", "适应症外推", "FGFR2" ], "tags_en": [ "WGS", "Target Assessment", "Indication Extrapolation", "FGFR2" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "药物靶点外推证据评估报告（自包含 HTML）" } ], "prompt": "我们的胃癌（GC）样本 WGS + 转录组联合分析发现 P13 携带 FGFR2 高水平扩增（CNV=8）且 mRNA 过表达，临床团队提议 off-label 使用 FGFR 抑制剂（pemigatinib / futibatinib，两者获批适应症均为 FGFR2 改变的胆管癌 CCA）。请帮我做靶点外推证据评估，产出单文件自包含 HTML 报告（图片 base64 内嵌、样式内联、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航，可离线打开）。\n\n输入（以下为演示用模拟数据，需在报告中明确标注；真实场景由用户提供）：\n- 13 例 GC 样本，P13 的 FGFR2 CNV=8、log TPM+1=6.8（队列均值的 4.5 倍），其余 12 例 CNV 1-4\n- 待评估外推链：FGFR2 扩增（GC）→ FGFR 抑制剂（CCA 获批）\n- 对照链：ERBB2 扩增（GC）→ 曲妥珠单抗（ToGA III 期 RCT）\n\n要求：\n1. 把临床提议拆成显式外推链：① 靶点检测 → ② 靶点-疾病（FGFR2 扩增与 GC 预后）→ ③ 标志物对齐（扩增与 FGFR2b 过表达一致性）→ ④ 药物-靶点（FGFR2 抑制剂）→ ⑤ 适应症外推（CCA 获批药用于 GC），逐环节评估证据。\n2. 用 PubMed（NCBI E-utilities）与 KEGG REST API（find/drug、get/hsa:2263 等）真实检索证据：FGFR2 扩增在 GC 的发生率与预后、FGFR2b 过表达与扩增对齐、bemarituzumab FIGHT II 期试验、futibatinib/pemigatinib 的获批适应症边界、ToGA III 期试验；检索命令与日期写入复现说明；禁止编造 PMID 或 KEGG ID——检索不到就如实写「未找到」。\n3. 输出证据表（环节 | 支持/部分支持/矛盾/未找到 | 来源 PMID/KEGG ID | 强度）+ 已知结论 + 外推断裂点分析（明确指出哪一环证据不足及原因）+ 分层用药路径建议（临床试验入组 / 同情用药 / 标准化疗的优先级排序）+ 方法学风险（CNV calling 平台差异、扩增≠驱动、标志物人群不重合、个案发表偏倚）+ 重点关注 Top-3；报告须明确标注「不构成诊疗建议」。\n4. 演示图（13 例样本 CNV 与表达联动、外推链各环节证据强度条形图、off-label 用药决策路径图）用 matplotlib 固定随机种子生成后 base64 内嵌。\n5. 报告包含分析代码段与数据合规声明（演示数据与真实检索的边界写清楚）。", "prompt_en": "Our gastric cancer (GC) WGS + transcriptome joint analysis found that P13 carries a high-level FGFR2 amplification (CNV=8) with mRNA overexpression, and the clinical team proposes off-label use of FGFR inhibitors (pemigatinib / futibatinib, both approved for FGFR2-altered cholangiocarcinoma, CCA). Please run a target-extrapolation evidence review and produce a single self-contained HTML report (figures embedded as base64, styles inlined, no external resources, content visible by default, no hash-anchor navigation, openable offline).\n\nInput (simulated demo data below — must be clearly labeled in the report; in real use these come from the user):\n- 13 GC samples; P13 has FGFR2 CNV=8 and log TPM+1=6.8 (4.5x the cohort mean); the other 12 have CNV 1-4\n- Extrapolation chain under review: FGFR2 amplification (GC) → FGFR inhibitors (CCA-approved)\n- Comparison chain: ERBB2 amplification (GC) → trastuzumab (ToGA phase-III RCT)\n\nRequirements:\n1. Decompose the clinical proposal into an explicit extrapolation chain: (1) target detection → (2) target-disease (FGFR2 amplification and GC prognosis) → (3) biomarker alignment (amplification vs FGFR2b overexpression concordance) → (4) drug-target (FGFR2 inhibitors) → (5) indication extrapolation (CCA-approved drugs for GC), and assess the evidence link by link.\n2. Retrieve real evidence via PubMed (NCBI E-utilities) and the KEGG REST API (find/drug, get/hsa:2263, etc.): FGFR2 amplification frequency and prognosis in GC, FGFR2b overexpression-amplification concordance, the bemarituzumab FIGHT phase-II trial, the approved-indication boundaries of futibatinib/pemigatinib, and the ToGA phase-III trial; record retrieval commands and date in the reproduction notes; never fabricate PMIDs or KEGG IDs — write \"not found\" when retrieval comes up empty.\n3. Output an evidence table (link | supported/partially-supported/contradicted/not-found | source PMID/KEGG ID | strength) + verified conclusions + breaking-point analysis (state explicitly which link lacks evidence and why) + tiered treatment-path recommendations (priority order of clinical-trial enrollment / compassionate use / standard chemotherapy) + methodological risks (CNV-calling platform differences, amplification ≠ driver, non-overlapping biomarker populations, case-report publication bias) + Top-3 priorities; the report must clearly state that it \"does not constitute medical advice\".\n4. Generate demo figures (CNV and expression across the 13 samples, evidence-strength bars per extrapolation link, off-label decision-path diagram) with matplotlib at a fixed random seed and embed them as base64.\n5. Include analysis code blocks and a data-compliance statement clarifying the boundary between demo data and real retrieval.", "skills": [], "experts": [ { "id": "omics-interpretation-expert", "name": "组学结果知识解读专家", "description": "在完成组学分析并产出结果后，借助公共数据库、BiOmics 知识图谱与学术文献为结论寻找支持证据，区分已知知识与潜在新发现，评估结果可靠性并标出值得重点关注的发现", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_IRWCzdVUMc2i" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-29T16:13:56+08:00", "updated_at": "2026-08-29T21:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-hpc-expert-case-1", "title": "生成 RNA-seq 分析流程代码", "title_en": "Generate RNA-seq Analysis Pipeline Code", "subtitle": "AI 写好流程脚本，先看架构再动手", "subtitle_en": "AI-generated pipeline scripts — review the architecture before running", "description": "想跑 RNA-seq 流程，但不确定脚本怎么设计、担心直接跑出问题？这个案例演示了如何让 HPC 专家自动发现集群资源，生成数据模拟、QC 统计、TPM 定量、SLURM 投递四个脚本，并输出一份含 DAG 调用图、脚本清单卡与可折叠预览的「代码设计文档」HTML，让你在提交前先把架构看清楚。\n\n适合：想先预览 RNA-seq 流程代码设计、不依赖任何外部生物信息工具（零依赖纯 Python3）、希望在 HPC 集群上自动化运维的生信用户。", "description_en": "Worried about script design before running an RNA-seq pipeline? This case shows how the HPC expert auto-discovers cluster resources and generates four scripts (data simulation, QC, TPM quantification, SLURM submission), then produces a professional 'Code Design Document' HTML with a DAG flow diagram, script inventory cards, and collapsible code previews — so you can review the architecture before submitting anything.\n\nSuitable for: bioinformatics users who want to preview RNA-seq pipeline code design, avoid external tool dependencies (zero-dependency pure Python3), and automate operations on HPC clusters.", "categories": [ "dev" ], "tags": [ "HPC集群", "RNA-seq", "SLURM", "脚本生成", "流程设计" ], "tags_en": [ "HPC Cluster", "RNA-seq", "SLURM", "Script Generation", "Pipeline Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "RNA-seq 代码设计文档" } ], "prompt": "请为我生成一套 RNA-seq 分析流程的代码（本步只写代码、不执行）。\n\n先自动发现资源：列出所有运行中集群，选第一个 SLURM 集群，找到第一个可写的 CFS 挂载点作为工作目录。然后在工作目录生成 4 个脚本 + 1 个说明：\n\n- gen_rnaseq_data.py：模拟 5 条转录本（1000–5000bp，丰度 40/25/20/10/5%）和 100 万条 SE reads（reads 名内嵌转录本 ID），并输出真值表 ground_truth.json\n- run_pipeline.py：QC 统计（reads 数/长度/GC/N 占比）+ 按 reads 名的转录本 ID 归组计数 + 计算 TPM，结果写入 results.json\n- make_report.py：读 results.json 生成自包含 HTML 报告（图表用内联 SVG，不依赖任何外部库）\n- run_rnaseq.sh：SLURM 投递脚本（#SBATCH：compute 队列/1h/4G），依次串跑上面三个\n- README.md：执行说明\n\n硬性要求：所有脚本只用 Python3 标准库，不得依赖 hisat2/salmon/samtools 等需安装的工具。\n\n最后生成一份「代码设计文档」HTML，风格专业美观、自包含内联 SVG：渐变紫色 Hero + 顶部 KPI 卡（脚本数/流程阶段数/外部依赖数=0/总代码行数）+ 脚本调用关系流程图（DAG）+ 每个脚本一张产物清单卡（文件名/用途/输入/输出/依赖/行数）+ 测试数据规格卡 + 可折叠脚本预览 + 执行准备清单。只写代码不跑分析。", "prompt_en": "Please generate a complete RNA-seq analysis pipeline code for me (this step only writes code, does not execute).\n\nFirst auto-discover resources: list all running clusters, select the first SLURM cluster, find the first writable CFS mount point as the working directory. Then generate 4 scripts + 1 README in the working directory:\n\n- gen_rnaseq_data.py: simulate 5 transcripts (1000–5000bp, abundance 40/25/20/10/5%) and 1 million SE reads (read names embed transcript IDs), output ground_truth.json\n- run_pipeline.py: QC statistics (read count/length/GC/N ratio) + group count by transcript ID from read names + calculate TPM, write results to results.json\n- make_report.py: read results.json and generate a self-contained HTML report (charts as inline SVG, no external libraries)\n- run_rnaseq.sh: SLURM submission script (#SBATCH: compute queue/1h/4G), run the above three scripts sequentially\n- README.md: execution instructions\n\nHard requirement: all scripts use only Python3 standard library, no external tools like hisat2/salmon/samtools.\n\nFinally generate a 'Code Design Document' HTML, professional and self-contained with inline SVG: gradient purple Hero + top KPI cards (script count/pipeline stages/external dependencies=0/total code lines) + script call flow diagram (DAG) + one artifact card per script (filename/purpose/input/output/dependencies/lines) + test data spec card + collapsible script previews + execution checklist. Write code only, do not run analysis.", "skills": [], "experts": [ { "id": "omics-hpc-expert", "name": "腾讯组学HPC集群运维与作业管理专家", "description": "远程组学HPC一站式运维，管控节点、队列、存储全生命周期，适配SLURM/SGE调度，自然语言交互降低门槛", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_oVOV90lFvBz5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-29T17:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-iggm-expert-case-1", "title": "从头设计（De Novo Design）", "title_en": "De Novo Antibody Design", "subtitle": "用 IgGM 生成式模型对 PD-L1 靶点进行抗体从头设计", "subtitle_en": "Antibody de novo design against PD-L1 using IgGM generative model", "description": "痛点：抗体药物研发的核心难题是从零起点快速获得高质量候选序列——传统方法依赖实验筛选，耗时耗力，成本高。本案例演示如何用腾讯 IgGM 生成式大模型（ICLR 2025），针对肿瘤免疫热门靶点 PD-L1 执行一次完整的抗体从头设计流程：提交 IgGM-WDL 任务、跟踪运行直至完成，最终以单文件 HTML 报告交付候选抗体序列表（VH/VL + 可开发性评分）与代表性候选的 3D 交互视图（链着色、自动旋转、测距截图）。适合抗体研究员、计算生物学家、想快速生成多样化先导序列的研发团队。", "description_en": "Pain point: Generating high-quality antibody candidates from scratch is a major bottleneck in drug discovery—traditional methods rely on costly experimental screening. This case demonstrates how to use Tencent's IgGM generative model (ICLR 2025) to perform a complete de novo antibody design workflow against PD-L1: submitting an IgGM-WDL job, tracking until completion, and delivering a single-file HTML report with candidate sequence table (VH/VL + developability scores) and interactive 3D structure view. Ideal for antibody researchers and computational biologists seeking diverse lead candidates.", "tags": [ "抗体设计", "De Novo Design", "IgGM", "PD-L1", "候选序列" ], "tags_en": [ "Antibody Design", "De Novo Design", "IgGM", "PD-L1", "Candidate Sequences" ], "categories": [ "bioinformatics", "drug-discovery" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "抗体从头设计报告（含 3D 结构）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "Prompt 1 ｜从头设计（De Novo Design）\n\n我是做肿瘤免疫方向的抗体研究员，靶点为 PD-L1。请使用 IgGM(WDL) 公共应用，完全按默认参数（不改动抗原、模型与资源配置），执行一次抗体从头设计（De Novo Design），并完成分析：\n\n1. 提交前给我任务确认单（设计模式｜靶点｜关键参数｜WDL 引擎｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给根因与修复建议；\n3. 完成后以单文件 HTML 报告交付（内嵌 3Dmol.js + PDB 坐标）：候选抗体序列表（排名｜VH/VL｜可开发性评分）；代表性候选的 3D 交互视图按「3D 交付标准」执行（链着色、默认自动旋转、拖拽/缩放、链显隐、表示切换、测距、截图）。", "prompt_en": "Prompt 1 ｜De Novo Design\n\nI am an antibody researcher working on tumor immunology, targeting PD-L1. Please use the IgGM(WDL) public application with fully default parameters (do not modify the antigen, model, or resource configuration), perform one antibody de novo design (De Novo Design), and complete the analysis:\n\n1. Before submission, give me a task confirmation sheet (design mode | target | key parameters | WDL engine | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and fix recommendations based on the error knowledge base;\n3. Upon completion, deliver a single-file HTML report (with embedded 3Dmol.js + PDB coordinates): candidate antibody sequence table (rank | VH/VL | developability score); representative candidate 3D interactive view per \"3D Delivery Standard\" (chain coloring, auto-rotation by default, drag/zoom, chain visibility toggle, representation switch, distance measurement, screenshot).", "experts": [ { "id": "omics-iggm-expert", "name": "IgGM抗体药物研发专家", "expertType": "plugin", "source": "expert-marketplace", "description": "展示如何用腾讯 IgGM 生成式模型对 PD-L1 靶点执行抗体从头设计（De Novo Design）——无需提供初始序列，模型直接生成完整 VH/VL 候选序列并附可开发性评分；任务完成后以单文件 HTML 交付候选序列表与 3D 交互视图，支持链着色、自动旋转、链显隐与测距截图。适合抗体研究员快速生成多样化候选序列，作为 Lead 发现的起点。", "expert_id": "ex_iNZHjvbngGc9" } ], "skills": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-26T10:09:35+08:00", "updated_at": "2026-08-27T14:15:00+08:00", "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "omics-iggm-expert-case-2", "title": "CDR 区重设计", "title_en": "CDR Region Redesign", "subtitle": "保持框架区不变，用 IgGM 对先导抗体 CDR 区重设计并输出 3D 复合物视图", "subtitle_en": "Keep the framework region intact, redesign the CDR region of a lead antibody with IgGM, and output an interactive 3D complex view", "description": "手里有一株先导抗体，框架区已人源化，但 CDR 区的结合特性还需要打磨？每次重设计都要手动整理残基对比和序列 logo，费时费力。\n\n这个案例演示了如何用 IgGM 抗体药物研发专家对抗 PD-L1 先导抗体执行 CDR Redesign：专家先确认设计区域（默认 CDR-H3）与参数摘要，确认后提交并全程跟踪，完成后输出重设计前后 CDR 对比表（IMGT 编号｜原残基→新残基）、序列 logo 与突变概率柱状图，以及与抗原 5C3T_A 复合物的 3D 交互视图。\n\n适合：持有先导抗体、希望定向优化 CDR 亲和力的研究者，抗体改造方向的生信工程师，以及需要可视化 CDR 重设计结果的药物发现团队。", "description_en": "Got a lead antibody with a humanized framework but CDR regions that still need work? Manually aligning residue comparisons and generating sequence logos after each redesign round is tedious.\n\nThis case demonstrates how to run CDR Redesign on an anti-PD-L1 lead antibody with the IgGM Antibody R&D Expert: the expert confirms the design region (default CDR-H3) and parameter summary, submits and tracks the job to completion, then outputs a before/after CDR comparison table (IMGT numbering | original → new residue), sequence logo and mutation probability histogram, and an interactive 3D view of the redesigned antibody in complex with antigen 5C3T_A.\n\nIdeal for: researchers with a lead antibody seeking targeted CDR affinity optimization, bioinformatics engineers in antibody engineering, and drug discovery teams needing visual CDR redesign results.", "categories": [ "data" ], "tags": [ "CDR重设计", "IgGM", "亲和力优化", "抗体改造", "3D结构可视化" ], "tags_en": [ "CDR Redesign", "IgGM", "Affinity Optimization", "Antibody Engineering", "3D Structure Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "CDR 区重设计分析报告（含序列 logo、突变概率图与 3D 复合物视图）" } ], "prompt": "Prompt 2 ｜CDR 区重设计\n\n我有一株抗 PD-L1 的先导抗体，框架区已人源化且稳定，需保持框架区不变，对 CDR 区重设计以改善结合特性。请使用 IgGM(WDL) 公共应用按默认参数执行 CDR Redesign，并完成分析：\n\n1. 提交前确认设计区域（默认 CDR-H3）与参数摘要，确认后提交；\n2. 跟踪任务直至完成；\n3. 结果以单文件 HTML 报告交付：重设计前后 CDR 对比表（IMGT 编号｜原残基→新残基）；运行 Merge_output 后处理生成序列 logo 与突变概率柱状图（设计组 vs 原始组）；重设计产物的 3D 交互视图按「3D 交付标准」执行（与抗原 5C3T_A 复合物视图，链着色、默认自动旋转、链显隐、抗原表面、测距、截图）。", "prompt_en": "Prompt 2 ｜ CDR Region Redesign\n\nI have an anti-PD-L1 lead antibody with a humanized and stable framework region. I need to keep the framework region unchanged and redesign the CDR region to improve binding properties. Please use the IgGM(WDL) public application with default parameters to perform CDR Redesign and complete the analysis:\n\n1. Before submission, confirm the design region (default CDR-H3) and parameter summary, then submit after confirmation;\n2. Track the task until completion;\n3. Deliver the results as a single-file HTML report: before/after CDR comparison table (IMGT numbering | original residue → new residue); run Merge_output post-processing to generate sequence logo and mutation probability histogram (design group vs. original group); interactive 3D view of the redesigned product following the \"3D delivery standard\" (complex view with antigen 5C3T_A, chain coloring, auto-rotation by default, chain show/hide, antigen surface, distance measurement, screenshot).", "skills": [], "experts": [ { "id": "omics-iggm-expert", "name": "IgGM抗体药物研发专家", "description": "演示如何在保持框架区不变的前提下，用 IgGM 对先导抗体的 CDR 区进行重设计以改善结合特性：输出重设计前后 CDR 残基对比表（IMGT 编号）、序列 logo 与突变概率柱状图，并展示与 PD-L1 抗原复合物的 3D 交互视图。适合已有先导抗体、希望定向优化 CDR 亲和力的研究者。", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_iNZHjvbngGc9" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T07:12:16Z", "updated_at": "2026-08-29T17:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-iggm-expert-case-3", "title": "CDR-H3 条件设计", "title_en": "CDR-H3 Conditional Design", "subtitle": "以 PD-L1 抗原为约束条件，用 IgGM 条件生成 CDR-H3 区并输出候选序列汇总与 3D 视图", "subtitle_en": "Use PD-L1 antigen as constraint for IgGM conditional CDR-H3 generation, and output a candidate sequence summary with interactive 3D view", "description": "想让 CDR-H3 的设计紧扣靶点结合需求，而不是漫无目的地随机生成？条件设计能约束生成方向，但结果分析和可视化还是让人头疼。\n\n这个案例演示了如何以 PD-L1 抗原为约束，用 IgGM 抗体药物研发专家执行 CDR-H3 条件设计：先确认任务单（设计模式、设计区域、靶抗原、资源与预计时长），确认后提交并全程跟踪，最终输出候选 CDR-H3 序列汇总表（序列、出现频率、位点保守度）、序列 logo 与逐位点保守度柱状图，以及 Top 候选结构的 3D 交互视图。\n\n适合：需要靶向约束 CDR-H3 设计的抗体研究者、想对比多个候选 CDR-H3 序列保守度的生信工程师、以及希望快速可视化条件生成结果的药物发现团队。", "description_en": "Want CDR-H3 design to stay tightly aligned with target binding requirements rather than generating randomly? Conditional design can steer generation, but analyzing and visualizing the results is still a chore.\n\nThis case demonstrates how to run CDR-H3 conditional design with PD-L1 antigen as the constraint using the IgGM Antibody R&D Expert: confirm a task sheet (design mode, design region, target antigen, resources, estimated duration), submit and track to completion, then output a candidate CDR-H3 sequence summary table (sequence, frequency, positional conservation), sequence logo and per-position conservation histogram, and an interactive 3D view of top candidates.\n\nIdeal for: antibody researchers who need targeted CDR-H3 design constraints, bioinformatics engineers comparing positional conservation across multiple CDR-H3 candidates, and drug discovery teams seeking rapid visualization of conditional design results.", "categories": [ "data" ], "tags": [ "CDR-H3条件设计", "IgGM", "序列保守度", "抗体设计", "3D结构可视化" ], "tags_en": [ "CDR-H3 Conditional Design", "IgGM", "Sequence Conservation", "Antibody Design", "3D Structure Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "CDR-H3 条件设计分析报告（含序列 logo、保守度图与 3D 交互视图）" } ], "prompt": "Prompt 3 ｜CDR-H3 条件设计\n\n针对默认 PD-L1 体系，请使用 IgGM(WDL) 公共应用按默认参数（不改动任何字段）执行一次 CDR-H3 条件设计：以默认抗原作为约束条件，对 CDR-H3 区做条件生成，并完成分析：\n\n1. 提交前给任务确认单（设计模式｜设计区域 CDR-H3｜靶抗原｜WDL 引擎｜资源｜预计时长），确认后提交；\n2. 跟踪任务直至完成；\n3. 结果以单文件 HTML 报告交付：候选 CDR-H3 序列汇总表（序列｜出现频率｜位点保守度）；Merge_output 后处理生成 CDR-H3 序列 logo 与逐位点保守度柱状图；Top 候选 1-2 个的 3D 交互视图按「3D 交付标准」执行（链着色、默认自动旋转、链显隐、测距、截图）。", "prompt_en": "Prompt 3 ｜ CDR-H3 Conditional Design\n\nFor the default PD-L1 system, please use the IgGM(WDL) public application with all default parameters (do not change any field) to perform one round of CDR-H3 conditional design: use the default antigen as the constraint condition for conditional generation of the CDR-H3 region, and complete the analysis:\n\n1. Before submission, provide a task confirmation sheet (design mode | design region CDR-H3 | target antigen | WDL engine | resources | estimated duration), then submit after confirmation;\n2. Track the task until completion;\n3. Deliver results as a single-file HTML report: candidate CDR-H3 sequence summary table (sequence | frequency | positional conservation); run Merge_output post-processing to generate CDR-H3 sequence logo and per-position conservation histogram; interactive 3D view of top 1-2 candidates following the \"3D delivery standard\" (chain coloring, auto-rotation by default, chain show/hide, distance measurement, screenshot).", "skills": [], "experts": [ { "id": "omics-iggm-expert", "name": "IgGM抗体药物研发专家", "description": "演示如何以 PD-L1 抗原为约束条件，用 IgGM 对 CDR-H3 区执行条件生成：提交确认单并全程跟踪，最终输出候选 CDR-H3 序列汇总表（序列、出现频率、位点保守度）、序列 logo 与保守度柱状图，以及 Top 候选结构的 3D 交互视图。适合需要靶向约束 CDR-H3 设计的抗体研究者。", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_iNZHjvbngGc9" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T07:12:16Z", "updated_at": "2026-08-29T17:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-interpretation-expert", "title": "组学分析结果交付与证据化审阅", "title_en": "Omics Analysis Delivery and Evidence Review", "subtitle": "交付级 scRNA-seq 分析报告，附 PubMed/KEGG 逐条证据核查", "subtitle_en": "Delivery-ready scRNA-seq report with item-by-item PubMed/KEGG evidence checks", "description": "单细胞分析跑完了，代码、图、结论散落各处；写进报告的结论哪些有文献支撑、哪些只是猜想，自己也说不清，组会上被追问就只能现场翻文献。这个案例把一份结直肠癌 scRNA-seq 分析（QC、聚类注释、DEG、KEGG 富集）打包成单文件自包含报告：完整 scanpy 代码、关键结果表、4 张结果图全部内嵌；并用 PubMed + KEGG 对 LGR5、FAP、SPP1、CXCL8 等关键基因结论逐条核查，输出「支持/部分支持/矛盾/未找到」证据表、可放心写入讨论的已证实结论、以及值得追查的潜在新发现假设。适合：单细胞/组学方向研究生、生信分析师、论文补充材料准备者、组会汇报人。", "description_en": "Your scRNA-seq analysis is done, but the code, figures and conclusions are scattered everywhere — and you can't say which claims are literature-backed and which are just hunches. This case packages a colorectal cancer scRNA-seq analysis (QC, clustering annotation, DEG, KEGG enrichment) into a single self-contained report: full scanpy code, key result tables, and 4 embedded figures. It then checks every key gene claim (LGR5, FAP, SPP1, CXCL8, etc.) against PubMed and KEGG, producing a supported/partially-supported/contradicted/not-found evidence table, verified conclusions safe for the discussion section, and novel hypotheses worth chasing. Suitable for: graduate students in single-cell/omics, bioinformatics analysts, supplementary-material writers, and lab-meeting presenters.", "categories": [ "data", "search" ], "tags": [ "scRNA-seq", "证据核查", "结果交付", "组学解读" ], "tags_en": [ "scRNA-seq", "Evidence Review", "Result Delivery", "Omics Interpretation" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "组学分析交付与证据化审阅报告（自包含 HTML）" } ], "prompt": "我刚刚做完一份结直肠癌（CRC）单细胞 RNA-seq 分析，结果如下。请帮我生成一份可直接交付的报告（单文件 HTML，自包含），所有图片必须以 base64 data URI 内嵌，不能是相对路径链接，所有样式内联、无任何外部资源引用，内容默认可见，不使用 hash 锚点导航。\n\n分析背景：\n- 物种/样本：人，肿瘤 vs 正常，3 v 3（模拟数据演示用）\n- 流程：scanpy QC → 降维聚类 → 细胞注释 → Wilcoxon DEG → gseapy KEGG 富集\n- 关键结论：6 大细胞群；肿瘤组 CAF 与单核/巨噬比例升高、NK 比例下降；1882 个 DEG；Wnt/ECM/黏着斑通路显著富集\n- 关键基因：LGR5、FZD5、CTNNB1、DKK1、MYC、COL1A1、FAP、SPP1、MMP9、CXCL8\n\n要求：\n1. 文档结构：标题 + 摘要（结论先行）+ 数据设计 + 完整 scanpy 代码 + DEG / 富集结果表 + 4 张主要图（UMAP 聚类、DEG 火山图、KEGG 富集条形图、细胞比例堆叠图）+ 解读与待验证点 + 复现说明。\n2. 图片使用 matplotlib 模拟生成，随机种子固定为 42，保证可复现；生成后以 base64 data URI 直接内嵌进 HTML，不引用外部 images/ 文件夹。\n3. 用 PubMed + KEGG 对 LGR5、FZD5、COL1A1/FAP、SPP1/MMP9/CXCL8 等关键基因-疾病/通路关系做一轮证据化核查，输出「主张清单 + 证据表（支持/部分/矛盾/未找到）+ 潜在新发现假设 + 方法学风险」。\n4. 删除元教程、规范说明、自查清单等对用户价值不大的章节，只保留实质交付内容。\n5. 最终产物必须是单个自包含 .html 文件，可离线打开、拷贝、汇报。", "prompt_en": "I have just finished a colorectal cancer (CRC) single-cell RNA-seq analysis with the results below. Please generate a directly deliverable report (single self-contained HTML file): all images must be embedded as base64 data URIs (no relative-path links), all styles inlined, no external resource references, content visible by default, and no hash-anchor navigation.\n\nAnalysis background:\n- Species/samples: human, tumor vs normal, 3 vs 3 (simulated data for demonstration)\n- Pipeline: scanpy QC → dimensionality reduction and clustering → cell annotation → Wilcoxon DEG → gseapy KEGG enrichment\n- Key findings: 6 major cell populations; elevated CAF and monocyte/macrophage proportions and reduced NK proportion in the tumor group; 1,882 DEGs; significant enrichment of Wnt/ECM/focal-adhesion pathways\n- Key genes: LGR5, FZD5, CTNNB1, DKK1, MYC, COL1A1, FAP, SPP1, MMP9, CXCL8\n\nRequirements:\n1. Document structure: title + summary (conclusions first) + data design + full scanpy code + DEG/enrichment result tables + 4 main figures (UMAP clustering, DEG volcano plot, KEGG enrichment bar chart, cell-proportion stacked bar) + interpretation and points to verify + reproduction notes.\n2. Generate figures with matplotlib in simulation mode, random seed fixed at 42 for reproducibility; embed them into the HTML as base64 data URIs without referencing an external images/ folder.\n3. Run an evidence-based check of the gene-disease/pathway relationships for key genes (LGR5, FZD5, COL1A1/FAP, SPP1/MMP9/CXCL8) against PubMed and KEGG, outputting a claim list + evidence table (supported/partial/contradicted/not found) + potential novel hypotheses + methodological risks.\n4. Remove meta-tutorials, specification notes, and self-check lists that add little user value; keep only substantive deliverable content.\n5. The final artifact must be a single self-contained .html file that can be opened offline, copied, and presented.", "skills": [], "experts": [ { "id": "omics-interpretation-expert", "name": "组学结果知识解读专家", "description": "在完成组学分析并产出结果后，借助公共数据库、BiOmics 知识图谱与学术文献为结论寻找支持证据，区分已知知识与潜在新发现，评估结果可靠性并标出值得重点关注的发现", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_IRWCzdVUMc2i" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-27T22:03:55+08:00", "updated_at": "2026-08-30T11:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-ori-expert-case-1", "title": "结构预测（Structure Prediction）", "title_en": "Structure Prediction", "subtitle": "USMFold 一键预测蛋白折叠结构，pLDDT 置信度分析 + 可交互 3D 报告", "subtitle_en": "One-click USMFold protein structure prediction with pLDDT analysis and interactive 3D report", "description": "蛋白质工程师手上经常有一批新序列，却苦于没有快速上手的结构预测工具。\n\n这个案例演示了通过 ORI 专家的 USMFold Predict 子应用，对示例序列一键提交结构预测任务，自动获取 pLDDT 置信度统计（整体均值 / N 端 / 核心区域），并生成内嵌可交互 3D 结构的 HTML 报告——卡通 + 半透明表面叠加、rwb 渐变着色、默认自动旋转，打开即动态呈现。\n\n适合：酶工程研究员、结构生物学实验室、生信从业者。", "description_en": "Protein engineers often have a batch of new sequences but struggle to find a quick, accessible structure prediction tool.\n\nThis case demonstrates how to use the ORI expert's USMFold Predict sub-application to submit a one-click structure prediction job, automatically obtain pLDDT confidence statistics (overall mean / N-terminal / core region), and generate an interactive 3D HTML report — cartoon + semi-transparent surface overlay, rwb gradient coloring, auto-rotation on by default, rendering dynamically when opened.\n\nIdeal for: enzyme engineering researchers, structural biology labs, and bioinformatics practitioners.", "categories": [ "data", "education" ], "tags": [ "蛋白质结构预测", "USMFold", "ORI", "pLDDT置信度", "3D可视化" ], "tags_en": [ "Protein Structure Prediction", "USMFold", "ORI", "pLDDT Confidence", "3D Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "蛋白结构预测交互 HTML 报告" } ], "prompt": "Prompt 1 ｜ 结构预测（Structure Prediction）\n\n我是做酶工程方向的研究员，拿到一批新序列想快速看折叠结构。请使用 ORI 公共应用的 usmfold-predict 子流程，完全按默认参数（不改动输入、模型与资源配置），对默认示例序列执行一次结构预测，并完成分析：\n\n1. 提交前给我任务确认单（子流程｜输入摘要｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：pLDDT 置信度统计（整体均值｜N 端｜核心区域），并生成结构预测 HTML 报告——报告内嵌可交互 3D 结构（3Dmol.js，PDB 数据直接内嵌文件中）：**卡通 + 半透明分子表面叠加表示**、按 pLDDT rwb 渐变着色（红=低置信 → 蓝=高置信）、**默认开启自动旋转**（Y 轴，0.5 速），并提供三个控制按钮（自动旋转 / 隐藏表面 / 复位视角）+ 鼠标拖动旋转 / 滚轮缩放提示，打开报告即动态呈现结构。", "prompt_en": "Prompt 1 | Structure Prediction\n\nI am a researcher in enzyme engineering and have a batch of new sequences I want to quickly visualize for folding structure. Please use the ORI public application's usmfold-predict sub-workflow, following all default parameters (no changes to input, model, or resource configuration), and run a structure prediction on the default example sequence, then complete the analysis:\n\n1. Before submission, provide me with a task confirmation sheet (sub-workflow | input summary | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and fix suggestions based on the error knowledge base;\n3. Upon completion, output: pLDDT confidence statistics (overall mean | N-terminal | core region), and generate a structure prediction HTML report — with an embedded interactive 3D structure (3Dmol.js, PDB data directly embedded in the file): **cartoon + semi-transparent molecular surface overlay**, colored by pLDDT rwb gradient (red=low confidence → blue=high confidence), **auto-rotation enabled by default** (Y-axis, speed 0.5), with three control buttons (auto-rotate / hide surface / reset view) + mouse drag rotation / scroll-to-zoom tips, rendering dynamically when the report is opened.", "skills": [], "experts": [ { "id": "omics-ori-expert", "name": "ORI蛋白设计专家", "description": "腾讯ORI蛋白设计专家，基于ORI蛋白质设计系统，提供从头序列设计、USMFold结构预测、溶解性/热稳定性/信号肽预测五大能力，并支持3D结构可视化", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_HrdscGoeTkwq" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T07:58:16Z", "updated_at": "2026-08-27T13:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-scprotein-expert-case-1", "title": "肽级不确定性定量（Stage1）", "title_en": "Peptide-Level Uncertainty Quantification (Stage1)", "subtitle": "识别低置信肽段，为后续分析打牢数据质量基础", "subtitle_en": "Identify low-confidence peptides and build a solid data quality foundation", "description": "单细胞蛋白组数据里哪些肽段定量可靠、哪些需要复核？人工逐条排查太费劲了。这个案例演示了如何用 scPROTEIN Stage1 评估每条肽段的定量可靠性，生成不确定性分布直方图（含 P90 阈值线）、Top20 高不确定性肽段复核表与高方差肽段热图，交互式 HTML 报告一键展示全部质控结果。\n\n适合：需要在单细胞蛋白组分析前做数据质控、筛查低置信肽段的蛋白质组学研究人员。", "description_en": "Which peptides in your single-cell proteomics data are reliably quantified and which need review? Manual inspection is exhausting. This case demonstrates how to use scPROTEIN Stage1 to assess quantification reliability for every peptide, producing an uncertainty distribution histogram (with P90 threshold), a Top20 high-uncertainty peptide review table, and a high-variance peptide heatmap — all in an interactive HTML report.\n\nSuitable for: proteomics researchers who need to perform data quality control and screen low-confidence peptides before single-cell proteomics analysis.", "categories": [ "data" ], "tags": [ "单细胞蛋白组", "不确定性估计", "质量控制", "scPROTEIN", "数据降噪" ], "tags_en": [ "Single-cell Proteomics", "Uncertainty Estimation", "Quality Control", "scPROTEIN", "Data Denoising" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肽级不确定性定量交互报告" } ], "prompt": "我有一份单细胞蛋白组肽段定量数据，需评估肽段定量的可靠性。请使用 scPROTEIN 应用，按默认参数执行 Stage1 肽级不确定性定量。\n\n1. 提交前确认子应用、数据规模与参数摘要，我确认后提交；\n2. 跟踪任务至完成，失败给根因诊断；\n3. 结果可视化：不确定性分布直方图（含 P90 阈值线）、Top20 高不确定性肽段图与复核表、高方差肽段热图，HTML 交互报告在内置浏览器打开。", "prompt_en": "I have a single-cell proteomics peptide quantification dataset and need to assess peptide quantification reliability. Please use the scPROTEIN application to run Stage1 peptide-level uncertainty quantification with default parameters.\n\n1. Confirm the sub-application, data scale, and parameter summary before submission; submit after my confirmation;\n2. Track the task to completion; provide root cause diagnosis on failure;\n3. Visualize results: uncertainty distribution histogram (with P90 threshold line), Top20 high-uncertainty peptide chart and review table, high-variance peptide heatmap — open the HTML interactive report in the built-in browser.", "skills": [], "experts": [ { "id": "omics-scprotein-expert", "name": "腾讯scPROTEIN单细胞蛋白组建模专家", "description": "腾讯 scPROTEIN表征模型，实现对单细胞蛋白组的多肽不确定性估计，去除批次效应噪声以及单细胞蛋白组的细胞类型注释", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_SlvtyVitPzY8" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-spatial-tme-evidence", "title": "空间转录组TME结果解读", "title_en": "Spatial Transcriptomics TME Interpretation", "subtitle": "Stereo-seq 边界生态位结论的 PubMed/KEGG 证据化核查", "subtitle_en": "Evidence check of Stereo-seq boundary-niche conclusions against PubMed/KEGG", "description": "空间转录组分析跑完了，结论写得很漂亮——「边界区形成免疫排斥生态位」——但组会被问「这个说法有人发过吗」就卡壳：哪些是文献共识、哪些是自己数据的首次观察，说不清楚。这个案例把一份结直肠癌 Stereo-seq 分析（空间域识别、边界生态位、配受体通讯）做成证据化审阅报告：先拆成 5 条原子主张，再用 PubMed + KEGG 逐条检索真实证据（Nat Commun 2022 的 SPP1+ 巨噬-CAF 互作、HCC 免疫屏障旁证等），输出「支持/部分支持/未找到」证据表、可写入讨论的已知结论、值得追查的新发现假设（SPP1-CD44 维持巨噬-CAF 配对结构）与验证实验设计。适合：空间组学方向研究生、Stereo-seq 平台用户、写论文讨论段的作者、组会汇报人。", "description_en": "Your spatial transcriptomics analysis produced a beautiful conclusion — \"an immunosuppressive niche forms at the tumor boundary\" — but when someone asks at lab meeting \"has this been published before?\" you freeze: you can't tell literature consensus from your own first observation. This case turns a colorectal cancer Stereo-seq analysis (spatial domain detection, boundary niche, ligand-receptor communication) into an evidence-based review: it first decomposes the claim into 5 atomic assertions, then retrieves real evidence from PubMed and KEGG for each (the SPP1+ macrophage-CAF interaction in Nat Commun 2022, the HCC immune-barrier analogy, etc.), producing a supported/partially-supported/not-found evidence table, verified conclusions safe for the discussion section, novel hypotheses worth chasing (SPP1-CD44 maintaining the macrophage-CAF pairing), and validation experiment designs. Suitable for: spatial-omics graduate students, Stereo-seq platform users, discussion-section writers, and lab-meeting presenters.", "categories": [ "data", "search" ], "tags": [ "Stereo-seq", "空间转录组", "TME 解读", "证据核查" ], "tags_en": [ "Stereo-seq", "Spatial Transcriptomics", "TME Interpretation", "Evidence Review" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "空间转录组 TME 证据化审阅报告（自包含 HTML）" } ], "prompt": "我做完了一张结直肠癌（CRC）组织切片的 Stereo-seq 空间转录组分析，请对结果做证据化审阅，产出单文件自包含 HTML 报告（图片 base64 内嵌、样式内联、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航，可离线打开）。\n\n分析结果（以下为演示用模拟数据，需在报告中明确标注；真实场景中由用户提供）：\n- bin50 聚合后识别 4 个空间域：肿瘤核心 D1 / 边界区 D2 / 间质边缘 D3 / 正常组织 D4\n- 边界区 SPP1+ 巨噬细胞（26%）与 POSTN+ CAF（24%）显著共定位，T 细胞在边界带（±3 bin）密度下降约 75%\n- 空间约束配受体分析：CXCL12–CXCR4（得分 3.4）与 SPP1–CD44（3.1）在边界区最高\n- 核心结论主张：「SPP1+ 巨噬与 POSTN+ CAF 在肿瘤边界形成免疫排斥生态位/免疫屏障」\n\n要求：\n1. 把复合结论拆解为可独立检索的原子主张清单（基因-细胞-空间定位、细胞-功能、空间结构、通路、潜在新发现）。\n2. 用 PubMed（NCBI E-utilities）与 KEGG REST API 真实检索每条主张的证据（如 SPP1+ 巨噬-CAF 互作、CXCL12 介导 T 细胞排斥、肿瘤边界免疫屏障），检索命令与日期写入报告复现说明；禁止编造 PMID——检索不到就如实写「未找到」。\n3. 输出证据表（主张 | 支持/部分支持/矛盾/未找到 | 来源 PMID/通路 ID | 强度）+ 已知结论（为何可信）+ 潜在新发现假设（可证伪陈述 + 支持线索 + 证据缺口 + 建议验证实验）+ 方法学风险（边界定义循环论证、bin 混合、配受体库偏差、多重检验、单切片代表性）+ 重点关注 Top-3。\n4. 演示图（空间域散点图、边界区放大与沿轴密度曲线、配受体热图、空间域组成堆叠图）用 matplotlib 固定随机种子生成后 base64 内嵌。\n5. 报告包含分析代码段（squidpy / SpaGCN / CellPhoneDB 风格，API 真实）与数据合规声明（演示数据与真实检索的边界写清楚）。", "prompt_en": "I have finished a Stereo-seq spatial transcriptomics analysis of a colorectal cancer (CRC) tissue section. Please perform an evidence-based review of the results and produce a single self-contained HTML report (figures embedded as base64, styles inlined, no external resources, content visible by default, no hash-anchor navigation, openable offline).\n\nAnalysis results (simulated demo data below — must be clearly labeled in the report; in real use these come from the user):\n- 4 spatial domains identified on bin50 aggregation: tumor core D1 / interface D2 / stromal edge D3 / normal D4\n- SPP1+ macrophages (26%) and POSTN+ CAFs (24%) co-localize in the interface; T-cell density drops ~75% within the boundary band (±3 bins)\n- Spatially constrained ligand-receptor analysis: CXCL12–CXCR4 (score 3.4) and SPP1–CD44 (3.1) highest in the interface\n- Core claim: \"SPP1+ macrophages and POSTN+ CAFs form an immunosuppressive niche / immune barrier at the tumor boundary\"\n\nRequirements:\n1. Decompose the composite conclusion into an atomic-claim checklist (gene-cell-spatial location, cell-function, spatial structure, pathway, potential novel finding).\n2. Retrieve real evidence for each claim via PubMed (NCBI E-utilities) and the KEGG REST API (e.g., SPP1+ macrophage-CAF interaction, CXCL12-mediated T-cell exclusion, tumor-boundary immune barrier); record the retrieval commands and date in the reproduction notes; never fabricate PMIDs — write \"not found\" when retrieval comes up empty.\n3. Output an evidence table (claim | supported/partially-supported/contradicted/not-found | source PMID/pathway ID | strength) + verified conclusions (why they are trustworthy) + novel-finding hypotheses (falsifiable statement + supporting clues + evidence gap + suggested validation experiments) + methodological risks (circular boundary definition, bin mixing, ligand-receptor database bias, multiple testing, single-section representativeness) + Top-3 priorities.\n4. Generate demo figures (spatial-domain scatter, boundary zoom with along-axis density curves, ligand-receptor heatmap, domain-composition stacked bar) with matplotlib at a fixed random seed and embed them as base64.\n5. Include analysis code blocks (squidpy / SpaGCN / CellPhoneDB style, real APIs) and a data-compliance statement clarifying the boundary between demo data and real retrieval.", "skills": [], "experts": [ { "id": "omics-interpretation-expert", "name": "组学结果知识解读专家", "description": "在完成组学分析并产出结果后，借助公共数据库、BiOmics 知识图谱与学术文献为结论寻找支持证据，区分已知知识与潜在新发现，评估结果可靠性并标出值得重点关注的发现", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_IRWCzdVUMc2i" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-29T16:13:56+08:00", "updated_at": "2026-08-29T21:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-tfold-expert-case-1", "title": "抗体结构预测 → CDR 质量评估", "title_en": "Antibody Structure Prediction → CDR Quality Assessment", "subtitle": "tFold-Ab 预测 7ox3 Fv 结构，CDR-H3 构象高亮可视化", "subtitle_en": "tFold-Ab predicts 7ox3 Fv structure with CDR-H3 conformation highlighted in 3D", "description": "做抗体药物研发，最头疼的就是光靠序列判断不了 CDR-H3 的构象质量，人源化改造没有结构依据根本没法决策。\n\n这个案例用 tFold-Ab 对 7ox3 体系的 VH/VL 序列做 Fv 三维结构预测，输出 lDDT-Ca/pTM/ipTM 置信度表，并在 HTML 报告里内嵌 3Dmol.js 交互可视化——重链蓝色、轻链橙色、CDR-H3 红色 stick 高亮，默认自动旋转，同时从 RCSB 加载 7ox3 实验结构做对比。\n\n适合：抗体药物研发科学家、结构生物学研究员、抗体人源化改造工程师、生物医药 AI 应用开发者。", "description_en": "When developing antibody drugs, judging CDR-H3 conformation quality from sequence alone is nearly impossible — humanization decisions have no structural basis without a 3D model.\n\nThis case uses tFold-Ab to predict the Fv structure of the 7ox3 VH/VL system, outputs an lDDT-Ca/pTM/ipTM confidence table, and embeds an interactive 3Dmol.js visualization in an HTML report — heavy chain in blue, light chain in orange, CDR-H3 highlighted in red sticks with auto-rotation enabled, alongside the RCSB 7ox3 experimental structure for comparison.\n\nSuitable for: antibody drug R&D scientists, structural biologists, antibody humanization engineers, and biomedical AI application developers.", "categories": [ "data", "search" ], "tags": [ "抗体结构预测", "tFold-Ab", "CDR-H3", "3D可视化", "结构生物学" ], "tags_en": [ "Antibody Structure Prediction", "tFold-Ab", "CDR-H3", "3D Visualization", "Structural Biology" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "抗体结构预测报告（3D可视化）" } ], "prompt": "Prompt 1｜抗体结构预测（tFold-Ab）→ CDR 质量评估\n\n我是抗体药物研发方向的分子设计科学家，需要候选抗体（7ox3 体系 VH/VL）的 Fv 三维结构，重点评估 CDR-H3 构象质量，支撑人源化改造决策。请使用 tFold 公共应用的 tfold-ab 子流程，按默认参数直接运行并完成分析：\n\n1. 提交前先给任务确认单（子流程｜设计意图｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：置信度表（lDDT-Ca｜pTM｜ipTM）的 HTML 报告页；**3D 展示**用 3Dmol.js 将预测结构**内嵌至报告的「3D 结构展示」容器内（画布限定在该容器内、不悬停覆盖报告其他内容；容器需 position:relative + overflow:hidden 约束）**——cartoon 按链着色（H 蓝 / L 橙）+ 半透明分子表面，CDR-H3 环红色 stick 高亮，**默认自动旋转**，三个按钮 ▶ 自动旋转 / ◆ 隐藏表面 / ↻ 复位视角，图例\"链 H 重链（蓝）｜链 L 轻链（橙）｜CDR-H3（红）\"，提示\"鼠标拖拽旋转·滚轮缩放\"（.pth 输出实为 PDB 文本，复制为 .pdb 后内嵌）；从 RCSB 加载 7ox3 实验结构对比。", "prompt_en": "Prompt 1｜Antibody Structure Prediction (tFold-Ab) → CDR Quality Assessment\n\nI am a molecular design scientist in antibody drug development. I need the Fv three-dimensional structure of a candidate antibody (7ox3 system VH/VL), with a focus on evaluating CDR-H3 conformation quality to support humanization decisions. Please use the tfold-ab sub-workflow of the tFold public application, run with default parameters, and complete the analysis:\n\n1. Before submission, provide a task confirmation sheet (sub-workflow | design intent | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and repair suggestions based on the error knowledge base;\n3. After completion, output: an HTML report page with a confidence table (lDDT-Ca | pTM | ipTM); **3D display** using 3Dmol.js to embed the predicted structure **inside the report's '3D Structure Display' container (canvas confined within the container, not hovering over other report content; container requires position:relative + overflow:hidden constraints)** — cartoon colored by chain (H blue / L orange) + semi-transparent molecular surface, CDR-H3 loop highlighted in red sticks, **auto-rotation enabled by default**, three buttons ▶ Auto-rotate / ◆ Hide surface / ↻ Reset view, legend 'Chain H Heavy Chain (Blue) | Chain L Light Chain (Orange) | CDR-H3 (Red)', tooltip 'Drag to rotate · Scroll to zoom' (.pth output is actually PDB text, copy as .pdb then embed); load 7ox3 experimental structure from RCSB for comparison.", "skills": [], "experts": [ { "id": "omics-tfold-expert", "name": "tFold抗体结构预测专家", "description": "以腾讯 tFold-Ab 模型预测候选抗体 Fv 三维结构，重点评估 CDR-H3 构象质量，输出 lDDT-Ca/pTM/ipTM 置信度报告并内嵌 3D 结构可视化，与 RCSB 实验结构对比，为人源化改造决策提供结构依据。", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_5VzRWJsCFEio" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "source": "official", "created_at": "2026-08-25T06:27:13Z", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-tfold-expert-case-2", "title": "抗体-抗原复合物预测 → 表位定位", "title_en": "Antibody-Antigen Complex Prediction → Epitope Mapping", "subtitle": "tFold-Ag 预测 RBD 复合物结构，双模式评分对比", "subtitle_en": "tFold-Ag predicts RBD complex structure with dual-mode scoring comparison", "description": "抗病毒抗体研发中，最难的不是跑预测，而是不知道抗体到底结合在抗原哪个位置——没有表位/对位的结构信息，突变筛选和亲和力改造都是瞎猜。\n\n这个案例用 tFold-Ag（MSA 标准模式）预测抗 SARS-CoV-2 RBD 抗体 Fab 与 RBD 抗原的复合物结构，输出带 lDDT-Ca/ipTM/pTM 置信度表的 HTML 报告，内嵌 3Dmol.js 可视化（表位紫色 stick、对位红色 stick 双重高亮），并自动级联 ICF 模式跑一遍，最终输出双模式评分对比表，让你一眼判断哪套模式更可靠。\n\n适合：结构生物学负责人、抗病毒抗体研发科学家、表位分析与亲和力改造工程师。", "description_en": "In antiviral antibody development, the hardest part isn't running predictions — it's not knowing where the antibody actually binds on the antigen. Without epitope/paratope structural information, mutation screening and affinity engineering are just guesswork.\n\nThis case uses tFold-Ag (MSA standard mode) to predict the complex structure of an anti-SARS-CoV-2 RBD antibody Fab with the RBD antigen, outputs an HTML report with lDDT-Ca/ipTM/pTM confidence tables, embeds 3Dmol.js visualization (epitope in purple sticks, paratope in red sticks), and automatically cascades an ICF mode run for comparison — producing a dual-mode scoring comparison table so you can immediately assess which mode is more reliable.\n\nSuitable for: structural biology leads, antiviral antibody R&D scientists, epitope analysis and affinity engineering researchers.", "categories": [ "data", "search" ], "tags": [ "抗体-抗原复合物", "tFold-Ag", "表位预测", "3D可视化", "亲和力改造" ], "tags_en": [ "Antibody-Antigen Complex", "tFold-Ag", "Epitope Prediction", "3D Visualization", "Affinity Engineering" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "复合物结构预测报告（双模式对比）" } ], "prompt": "Prompt 2｜抗体-抗原复合物预测（tFold-Ag）→ 表位定位与双模式对照\n\n我是抗病毒抗体项目的结构生物学负责人，需要抗 SARS-CoV-2 RBD 抗体 Fab 与 RBD 抗原的复合物结构，定位表位（epitope）与对位（paratope），指导突变与亲和力改造。请使用 tFold 公共应用的 tFold-Ag 子流程（MSA 标准模式），按默认参数直接运行并完成分析：\n\n1. 提交前先给任务确认单（子流程｜设计意图｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：置信度表（整体 lDDT-Ca｜ipTM｜pTM｜抗原链 lDDT）的 HTML 报告页；**3D 展示**用 3Dmol.js 将复合物**内嵌至报告的「3D 结构展示」容器内（画布限定在该容器内、不悬停覆盖报告其他内容；容器需 position:relative + overflow:hidden 约束）**——默认界面高亮视图（cartoon 按链着色：RBD 抗原绿 / H 蓝 / L 橙；表位紫+stick、对位红+stick）+ 半透明分子表面，**默认自动旋转**，三个按钮 ▶ 自动旋转 / ◆ 隐藏表面 / ↻ 复位视角，图例\"链 A RBD 抗原（绿）｜链 H 重链（蓝）｜链 L 轻链（橙）｜表位（紫）｜对位（红）\"，提示\"鼠标拖拽旋转·滚轮缩放\"；再自动级联 ICF 模式（默认参数）对照，输出双模式评分对比表。", "prompt_en": "Prompt 2｜Antibody-Antigen Complex Prediction (tFold-Ag) → Epitope Mapping and Dual-Mode Comparison\n\nI am the structural biology lead for an antiviral antibody project. I need the complex structure of an anti-SARS-CoV-2 RBD antibody Fab with the RBD antigen, to locate the epitope and paratope and guide mutation and affinity engineering. Please use the tFold-Ag sub-workflow (MSA standard mode) of the tFold public application, run with default parameters, and complete the analysis:\n\n1. Before submission, provide a task confirmation sheet (sub-workflow | design intent | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and repair suggestions based on the error knowledge base;\n3. After completion, output: an HTML report page with a confidence table (overall lDDT-Ca | ipTM | pTM | antigen chain lDDT); **3D display** using 3Dmol.js to embed the complex **inside the report's '3D Structure Display' container (canvas confined within the container, not hovering over other report content; container requires position:relative + overflow:hidden constraints)** — default interface highlight view (cartoon colored by chain: RBD antigen green / H blue / L orange; epitope purple+stick, paratope red+stick) + semi-transparent molecular surface, **auto-rotation enabled by default**, three buttons ▶ Auto-rotate / ◆ Hide surface / ↻ Reset view, legend 'Chain A RBD Antigen (Green) | Chain H Heavy Chain (Blue) | Chain L Light Chain (Orange) | Epitope (Purple) | Paratope (Red)', tooltip 'Drag to rotate · Scroll to zoom'; then automatically cascade ICF mode (default parameters) for comparison and output a dual-mode scoring comparison table.", "skills": [], "experts": [ { "id": "omics-tfold-expert", "name": "tFold抗体结构预测专家", "description": "以腾讯 tFold-Ag 模型预测抗 SARS-CoV-2 RBD 抗体 Fab 与 RBD 抗原的复合物结构，精确定位表位与对位，并自动级联 ICF 模式进行双模式评分对比，输出带 3D 互动可视化的结构报告，指导突变筛选与亲和力改造。", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_5VzRWJsCFEio" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "source": "official", "created_at": "2026-08-25T06:27:13Z", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-tfold-expert-case-3", "title": "TCR-pMHC 结构预测 → 识别界面验证", "title_en": "TCR-pMHC Structure Prediction → Recognition Interface Validation", "subtitle": "tFold-TCR 预测 6zkw TCR 结构，CDR3 构象与 pMHC 界面分析", "subtitle_en": "tFold-TCR predicts 6zkw TCR structure with CDR3 conformation and pMHC interface analysis", "description": "TCR 治疗研究中，验证 CDR3 环构象与 pMHC 识别界面是关键瓶颈——光凭序列根本无法判断 TCR 能否正确识别靶标多肽，需要高精度的三维结构预测。\n\n这个案例用 tFold-TCR 对结核 InhA 多肽-HLA-E 体系（6zkw）预测 TCR 三维结构，输出 lDDT-Ca/pTM/ipTM 置信度报告，并内嵌 3Dmol.js 交互可视化（α 链蓝色、β 链橙色、CDR3 区红色 stick 高亮），同时从 RCSB 加载 6zkw 实验结构做对比，为 TCR 识别机制研究提供直观的结构支撑。\n\n适合：TCR 治疗方向研究员、免疫学研究者、T 细胞工程化改造工程师、结构生物信息学分析人员。", "description_en": "In TCR therapy research, validating CDR3 loop conformation and the pMHC recognition interface is a critical bottleneck — sequence alone cannot determine whether a TCR can correctly recognize the target peptide; high-precision 3D structure prediction is essential.\n\nThis case uses tFold-TCR to predict the TCR three-dimensional structure for the Mycobacterium tuberculosis InhA peptide-HLA-E system (6zkw), outputs an lDDT-Ca/pTM/ipTM confidence report, and embeds an interactive 3Dmol.js visualization (α chain in blue, β chain in orange, CDR3 region highlighted in red sticks), while loading the 6zkw experimental structure from RCSB for comparison — providing direct structural support for TCR recognition mechanism research.\n\nSuitable for: TCR therapy researchers, immunologists, T cell engineering researchers, and structural bioinformatics analysts.", "categories": [ "data", "search" ], "tags": [ "TCR结构预测", "tFold-TCR", "CDR3构象", "pMHC界面", "3D可视化" ], "tags_en": [ "TCR Structure Prediction", "tFold-TCR", "CDR3 Conformation", "pMHC Interface", "3D Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "TCR 结构预测报告（3D可视化）" } ], "prompt": "Prompt 3｜TCR-pMHC 结构预测（tFold-TCR）→ 识别界面验证\n\n我是 TCR 治疗方向的研究员，需要结核 InhA 多肽-HLA-E 体系（6zkw）中 TCR 的预测结构，分析 CDR3 构象与 pMHC 识别界面。请使用 tFold 公共应用的 tfold-tcr 子流程，按默认参数直接运行并完成分析：\n\n1. 提交前先给任务确认单（子流程｜设计意图｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：置信度表（lDDT-Ca｜pTM｜ipTM）的 HTML 报告页；**3D 展示**用 3Dmol.js 将预测结构**内嵌至报告的「3D 结构展示」容器内（画布限定在该容器内、不悬停覆盖报告其他内容；容器需 position:relative + overflow:hidden 约束）**——cartoon 按链着色（α 蓝 / β 橙）+ 半透明分子表面，CDR3 环红色 stick 高亮（α 89–97 / β 96–104 示意区间），**默认自动旋转**，三个按钮 ▶ 自动旋转 / ◆ 隐藏表面 / ↻ 复位视角，图例\"链 α TRAV（蓝）｜链 β TRBV（橙）｜CDR3（红）\"，提示\"鼠标拖拽旋转·滚轮缩放\"；从 RCSB 加载 6zkw 实验结构对比。", "prompt_en": "Prompt 3｜TCR-pMHC Structure Prediction (tFold-TCR) → Recognition Interface Validation\n\nI am a researcher in the TCR therapy field. I need the predicted structure of the TCR in the Mycobacterium tuberculosis InhA peptide-HLA-E system (6zkw), to analyze the CDR3 conformation and pMHC recognition interface. Please use the tfold-tcr sub-workflow of the tFold public application, run with default parameters, and complete the analysis:\n\n1. Before submission, provide a task confirmation sheet (sub-workflow | design intent | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and repair suggestions based on the error knowledge base;\n3. After completion, output: an HTML report page with a confidence table (lDDT-Ca | pTM | ipTM); **3D display** using 3Dmol.js to embed the predicted structure **inside the report's '3D Structure Display' container (canvas confined within the container, not hovering over other report content; container requires position:relative + overflow:hidden constraints)** — cartoon colored by chain (α blue / β orange) + semi-transparent molecular surface, CDR3 loop highlighted in red sticks (α 89–97 / β 96–104 indicative range), **auto-rotation enabled by default**, three buttons ▶ Auto-rotate / ◆ Hide surface / ↻ Reset view, legend 'Chain α TRAV (Blue) | Chain β TRBV (Orange) | CDR3 (Red)', tooltip 'Drag to rotate · Scroll to zoom'; load 6zkw experimental structure from RCSB for comparison.", "skills": [], "experts": [ { "id": "omics-tfold-expert", "name": "tFold抗体结构预测专家", "description": "以腾讯 tFold-TCR 模型预测结核 InhA 多肽-HLA-E 体系（6zkw）中 TCR 的三维结构，重点分析 CDR3α/β 环构象与 pMHC 识别界面，输出置信度报告与 3D 结构可视化，并与 RCSB 实验结构对比，为 T 细胞受体识别机制研究提供结构支撑。", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_5VzRWJsCFEio" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "source": "official", "created_at": "2026-08-25T06:27:13Z", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-unipert-expert-case-1", "title": "基因扰动统一表征编码", "title_en": "Unified Encoding of Gene Perturbation Representations", "subtitle": "256 维统一表征，揭示基因扰动的模态邻域关系", "subtitle_en": "256-dim unified representation revealing gene perturbation modal neighborhood", "description": "想知道基因敲除数据在统一表征空间里长什么样，与蛋白质、化合物扰动的关系如何？这个案例演示了如何用 UniPert anndata 子应用将基因扰动数据编码为 256 维统一表征，分析激活维度与跨模态邻域，输出含表征指纹热图（峰值激活维度标注）、跨模态相似度对比图与稀疏激活解读的 ECharts 交互报告。\n\n适合：需要理解基因扰动表征特性、探索跨模态邻域关系的计算生物学研究人员与药物靶点发现团队。", "description_en": "Curious about what gene knockout data looks like in a unified representation space and how it relates to protein or compound perturbations? This case demonstrates how to use the UniPert anndata sub-application to encode gene perturbation data into 256-dimensional unified representations, analyze activation dimensions and cross-modal neighborhoods, and output an ECharts interactive report with representation fingerprint heatmap (peak activation dimensions annotated), cross-modal similarity comparison, and sparse activation interpretation.\n\nSuitable for: computational biology researchers and drug target discovery teams who need to understand gene perturbation representation characteristics and explore cross-modal neighborhood relationships.", "categories": [ "data" ], "tags": [ "基因扰动", "统一表征", "UniPert", "虚拟细胞", "跨模态分析" ], "tags_en": [ "Gene Perturbation", "Unified Representation", "UniPert", "Virtual Cell", "Cross-modal Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "基因扰动表征编码分析报告" } ], "prompt": "我需要将基因扰动数据编码为 UniPert 256 维统一表征并评估其模态邻域关系。请使用 UniPert-WDL 合集的 anndata 子应用，按默认输入模板（anndata_genetic.h5ad，genetic 扰动，无需我提供数据，通过 --input 显式传入该模板参数）执行扰动表征编码。\n\n1. 提交前确认子应用、同名检查结果与参数摘要，我确认后提交；\n2. 跟踪任务至完成，失败给根因诊断（debug 取证 + 错误知识库匹配 + 修复建议）；\n3. 结果固化为一份可复用的 ECharts 交互式 HTML 报告（如 unipert_anndata_report.html），在内置浏览器打开预览，须包含：概览卡片（扰动数 / 表征维度 / 激活维度数 / 稀疏度 / 最近邻扰动 / 邻域相似度）、256 维表征指纹热图（标注峰值激活维度）、跨模态相似度对比图（基因 / 蛋白 / 化合物）、稀疏激活模式解读与结论。", "prompt_en": "I need to encode gene perturbation data into UniPert 256-dimensional unified representations and evaluate cross-modal neighborhood relationships. Please use the anndata sub-application from the UniPert-WDL collection, with the default input template (anndata_genetic.h5ad, genetic perturbation, no need for me to provide data, pass the template via --input explicitly) to execute perturbation representation encoding.\n\n1. Confirm the sub-application, duplicate name check result, and parameter summary before submission; submit after my confirmation;\n2. Track the task to completion; provide root cause diagnosis on failure (debug evidence + error knowledge base matching + fix suggestions);\n3. Consolidate results into a reusable ECharts interactive HTML report (e.g., unipert_anndata_report.html), open preview in built-in browser, must include: overview cards (perturbation count / representation dimensions / active dimensions / sparsity / nearest neighbor perturbation / neighborhood similarity), 256-dim representation fingerprint heatmap (peak activation dimensions annotated), cross-modal similarity comparison chart (gene / protein / compound), sparse activation pattern interpretation and conclusions.", "skills": [], "experts": [ { "id": "omics-unipert-expert", "name": "腾讯UniPert虚拟细胞专家", "description": "基于腾讯UniPert模型，构建虚拟细胞，实现对于内外源扰动源统一表征，并且实现基因和药物扰动预测，包括细胞特异性药物响应", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_w7OMyXyVUKQh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-30T11:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "opc-studio-workbench", "title": "OPC 个人工作室", "title_en": "OPC Studio Workbench", "subtitle": "从客户到项目再到回款，一个人也能把工作室经营得井井有条", "subtitle_en": "Run clients, projects and cash flow like a studio — even when you work solo", "description": "一个人经营工作室，客户、报价、项目任务、工时和回款往往散在不同表格与聊天窗口里。最容易出问题的不是不会做，而是同一件事重复填写、老客户复购还得从头建单、项目做完却忘了追尾款。\n\n这个案例演示了一套单文件 OPC 个人工作室：8 个菜单覆盖首页、可拖动看板、客户、订单、项目、收付款、数据复盘和模板设置。客户页能查看合作与回款并发起“再次合作”；订单和项目可以复制、存为模板或从模板创建；订单确认后生成项目，任务和工时独立跟踪，分期应收自动补齐剩余报价；收入、支出、复购率、来源占比和类型时薪集中复盘。全部数据本地自动保存，支持 V1 手动迁移、JSON 备份和五工作表 Excel 导出。\n\n适合：一人公司与个人工作室经营者；同时管理多个客户和项目的自由职业者；需要把复购、模板和交付流程标准化的创意服务者；希望数据自己掌握、离线也能工作的独立顾问。", "description_en": "Running a one-person studio often means clients, quotes, project tasks, hours and payments are scattered across spreadsheets and chat threads. The biggest risk is not lacking skill — it is re-entering the same information, rebuilding repeat-client work from scratch, and finishing delivery without following up on the final payment.\n\nThis case demonstrates a self-contained OPC studio workbench with eight sections: home, a draggable board, clients, orders, projects, finance, insights, and templates/settings. Client records combine relationship and payment history with a “work together again” action. Orders and projects can be duplicated, saved as templates, or created from templates. Confirmed orders become projects; tasks and time entries are tracked independently; new installment rows automatically fill the unallocated quote balance. Income, expenses, repeat-client rate, source share and category hourly value are reviewed in one place. Everything autosaves locally, with optional V1 migration, JSON backup, and a genuine five-sheet Excel export.\n\nIdeal for: one-person companies and solo studios; freelancers managing several clients and projects; creative service providers standardizing repeat work and delivery; and independent consultants who want offline operation and full ownership of their data.", "categories": [ "workspace", "business", "office", "productivity" ], "tags": [ "一人公司", "客户管理", "项目看板", "模板复用", "经营复盘" ], "tags_en": [ "one-person company", "client management", "project board", "reusable templates", "business review" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/I4YOLlBwyjpd3qvGrneFah），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n把一个简单的接单工具升级成适用于 OPC（一人公司）和个人工作室的多菜单工作台，仍然交付为一个可离线打开的。不要加入线索、合同、内容运营、自动化、客户门户或报税。\n\n导航包括：首页、看板、客户、订单、项目、收付款、数据复盘、模板与设置。首页突出今日任务、逾期应收、待跟进客户和经营摘要；看板按洽谈、待开始、进行中、待验收、已完成展示，并支持拖动与状态选择；客户页显示联系资料、合作历史、累计实收、待收和“再次合作”；订单支持服务项、分期应收、复制、存为模板和转项目；项目支持任务、工时、交付清单、修改轮次、复制和模板；财务与复盘展示收付款、支出、月收入、来源占比、复购率和类型时薪。\n\n在订单和项目页直接提供“从模板新建”。新增分期时，默认金额自动等于报价减去已填分期合计。数据自动保存到浏览器本地，支持 JSON 备份、真实五工作表 Excel 导出和手动 V1 数据迁移。视觉采用清新可爱但可信赖的奶油白、薄荷绿和浅蓝风格，手机和电脑都能完整使用，无网络依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/I4YOLlBwyjpd3qvGrneFah) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nUpgrade a simple order tracker into a multi-section workbench for an OPC (one-person company) or solo studio, while still delivering it as one self-contained HTML file that works offline. Do not add lead management, contracts, content operations, automation, client portals, or tax filing. The navigation should include Home, Board, Clients, Orders, Projects, Finance, Insights, and Templates & Settings. Home highlights today’s tasks, overdue receivables, clients to follow up with, and operating metrics. The board shows Negotiating, To Start, In Progress, Awaiting Acceptance, and Completed, with both drag-and-drop and a status selector. Client records include contact details, work history, total cash received, outstanding balances, and a “work together again” action. Orders support service items, installment receivables, duplication, saving as templates, and conversion into projects. Projects support tasks, time entries, deliverables, revision counts, duplication, and templates. Finance and insights cover payments, expenses, monthly income, source share, repeat-client rate, and category hourly value. Put “Create from template” directly on the Orders and Projects pages. When adding an installment, default its amount to the quote minus the total of installments already entered. Autosave all data in browser storage, with JSON backup, a genuine five-sheet Excel export, and manual V1 data migration. Use a fresh, friendly but trustworthy cream, mint and pale-blue visual style. Desktop and mobile must both retain the full workflow, and no network dependency.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 0, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-01T17:15:46+08:00", "updated_at": "2026-08-01T17:15:46+08:00", "artifacts": [ { "file": "output.html", "label": "OPC 个人工作室（单文件 HTML）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "opportunity-radar", "title": "本周该跟谁？商机雷达帮你排出高意向客户", "title_en": "Opportunity Radar", "subtitle": "线索行为自动扫描，高意向客户排好序", "subtitle_en": "Lead behavior scanned automatically, high-intent customers ranked", "description": "自动扫描客户线索行为，按意向分排出本周高意向商机清单，标注信号来源和建议跟进动作。", "description_en": "It automatically scans customer-lead behavior and ranks this week's high-intent opportunities, annotating the signal sources and suggested follow-up actions.", "categories": [ "data" ], "tags": [ "商机雷达", "线索筛选", "销售跟进", "B2B贸易" ], "tags_en": [ "Opportunity Radar", "Lead Qualification", "Sales Follow-up", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "商机雷达清单（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易销售专家。请帮我扫描本周的客户线索，排出高意向商机清单。\n\n【数据来源（按我的实际情况）】\n- 我已绑定业务系统：用【找钢龙虾MCP】取商机/客户行为数据；\n- 我上传了客户/询价记录表：请读取并映射到客户、需求、询价频次、互动等口径；\n- 我暂时没有数据：说明需要哪些字段并给示例清单。\n\n【要求】\n1. 查询【行业信息资产库】里这个行业的商机意向评分口径，综合询价频次、是否索要材质单/比价、复购周期、互动信号等给每条线索打意向分；\n2. 按意向分从高到低排出清单：意向分、客户、需求品类、预估量、信号来源、建议动作（今日必跟/3天内/本周培育）；\n3. 底部给出本周重点客户与跟进建议。\n\n若数据源未就绪或返回不足，请以通用销售线索管理方法论为主产出并说明。请生成完整商机雷达清单。", "skills": [ { "id": "opportunity-radar", "name": "商机雷达", "description": "TradeBuddy把“哪些询价还没接住、该先跟谁、识别准不准”的诉求，做成按状态清晰分层、可直接照着干的商机清单。", "skill_id": "skill_2091155187457040384" } ], "experts": [ { "id": "xiaoshou-expert", "name": "B2B贸易销售专家", "description": "懂B2B贸易的销售专家：发货盘、写文案、出报价、盯商机、经营客户、看业绩、谈价成单，助你多成单。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_42GwnCqPUv7a" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "optics-module-workbench", "title": "光模块产业分析工作台", "title_en": "Optical Module Industry Analysis Workbench", "subtitle": "A股16家光模块公司 · 个股分析 / 行业解析 · 易中天三巨头专题", "subtitle_en": "16 A-share optical-module names · Stock analysis / Industry insights · EPT trio focus", "description": "AI算力爆发带火光模块，板块估值普遍百倍以上，该不该追、追哪家、风险在哪，普通投资者很难一次看清。本案例组合「NeoData 金融搜索服务」与「WeStock Data」：用 NeoData 查产业链公司与财务、宏观数据，用 WeStock Data 拉行情、财报、资金流和技术指标，把 A 股光模块产业链拆成两个页面。页面一是「个股分析」：顶部搜索栏可按公司名/代码/关键词检索，16 家公司以卡片列表呈现（光模块整机 / 上游光器件光芯片 / 配套材料封装 / 易中天三巨头可筛选、可排序），点击卡片进入详情页，看公司概况与估值、近 5 个单季营收/净利/毛利率趋势图、ROE 等关键财务指标、行情与主力资金（52 周区间、主力净流入、北向持仓、融资余额、MACD/KDJ 信号）、同业对比、投资亮点与风险，最后给一段综合评价（非投资建议）。页面二是「行业分析」：板块一是 16 家公司最新业绩（含业绩预告）核心指标对比表，未披露的标注；板块二是需求/供给/业绩/估值/资金/风险六维行业解析，重点呈现中际旭创、新易盛、天孚通信「易中天」三巨头对比专题。页面托管在云端、长期有效、手机电脑双端适配，内置使用手册。适合跟踪 AI 算力产业链的 A 股投资者、想快速横向对比光模块公司的研究员，以及需要一份可分享的板块观察面板的人。", "description_en": "AI-compute demand has set optical-module names on fire, with sector PEs mostly above 100x — hard for ordinary investors to see at a glance whether to chase, which name to pick, and where the risks sit. This case combines NeoData Financial Search with WeStock Data: NeoData supplies chain companies, financials and macro; WeStock Data supplies quotes, reports, capital flows and technical indicators. The A-share optical-module chain is split into two pages. Page 1 'Stock Analysis': a top search bar matches by name/code/keyword; 16 companies render as cards (filterable by module makers / upstream components & chips / materials & packaging / EPT trio, sortable), and clicking a card opens a detail view — profile & valuation, last-5-quarter revenue/net-profit/gross-margin trend chart, ROE and key metrics, quotes & main capital (52W range, main inflows, northbound holding, margin balance, MACD/KDJ signals), peer comparison, highlights & risks, and an overall assessment (not investment advice). Page 2 'Industry Analysis': Part 1 is a results-comparison table of all 16 names (incl. earnings forecasts, undisclosed cells marked); Part 2 is a six-dimension analysis (demand/supply/earnings/valuation/capital/risk) with a focus section on the 'EPT trio' — Innolight, Eoptolink and TFC Communication. Cloud-hosted, long-term valid, responsive on mobile and desktop, with a built-in user guide. Best for: A-share investors tracking the AI-compute chain, analysts who want fast horizontal comparison of optical-module names, and anyone needing a shareable sector watchboard.", "categories": [ "finance" ], "tags": [ "光模块", "AI算力", "个股分析", "行业研究", "投资工具" ], "tags_en": [ "Optical Modules", "AI Compute", "Stock Analysis", "Industry Research", "Investment Tool" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "光模块产业分析工作台（单体 HTML）" } ], "prompt": "帮我搭建一个「光模块产业分析工作台」，覆盖 A 股上市光模块公司，可参照此版本（https://workbuddy.link/p/85VTCQZrWfis9cMXpoF0Ms?source=2）搭建。项目使用资料库能力发布到我的资料库，并生成发布态网页链接返回。操作适配手机和电脑，并内置使用手册。用 NeoData 金融搜索服务查财务与宏观，用 WeStock Data 查行情、财报、资金流和技术指标。\n\n先筛选出 A 股光模块公司，汇总最新业绩情况，做两个页面：\n\n第一个页面「个股分析」：把筛选出的所有公司按卡片样式列表呈现，点击卡片进入详情页，显示公司基本情况与财务业绩——公司概况与估值、近四季营收/净利润/毛利率趋势、ROE 等关键财务指标、行情与主力资金（价格区间/主力净流入/北向持仓/技术信号）、同业公司对比、投资亮点与风险，最后给一段综合评价（非投资建议）。首页顶部设置搜索栏，输入公司名称、代码、关键词后点击搜索，结果按卡片列表排列。\n\n第二个页面「行业分析」：板块一，把所有光模块企业的最新业绩（含业绩预告）核心指标做列表对比，未披露业绩的显示「暂未披露」；板块二，结合最新业绩与行业市场情况，对行业做多维度全面解析，重点突出易中天（中际旭创/新易盛/天孚通信）三家公司的数据情况。\n\n另提供「CSV 工具」页：可导出公司核心指标与最新业绩对比 CSV，并支持用模板按股票代码批量更新线上数据库中的行情与业绩字段。\n\n过程中可以完善更多功能细节。视觉走金融工作台风格（深蓝渐变 header，A 股涨红跌绿配色），中文界面，所有 CSS/JS 全内联、无外部 CDN 依赖。", "prompt_en": "Build an 'Optical Module Industry Analysis Workbench' covering A-share listed optical-module companies, referencing this version (https://workbuddy.link/p/85VTCQZrWfis9cMXpoF0Ms?source=2). Use the library (WorkBuddy 资料库) publishing capability to publish the project to my library and return the published page link. It must adapt to both mobile and desktop, with a built-in user guide. Use NeoData Financial Search for financials and macro, and WeStock Data for quotes, earnings, capital flows and technical indicators.\n\nFirst screen the A-share optical-module names and aggregate the latest results, then build two pages:\n\nPage 1 'Stock Analysis': render all screened companies as a card list; clicking a card opens a detail page with company profile & valuation, recent-quarter revenue/net-profit/gross-margin trends, ROE and other key metrics, quotes & main capital (price range / main net inflow / northbound holding / technical signals), peer comparison, investment highlights and risks, ending with an overall assessment (not investment advice). Add a search bar at the top of the home page: enter a company name, code or keyword and hit search; results render as a card list.\n\nPage 2 'Industry Analysis': Part 1 compares the latest results (incl. earnings forecasts) of all optical-module names on core metrics, marking undisclosed companies as 'not disclosed yet'; Part 2 gives a multi-dimensional industry analysis combining the latest results and market conditions, with a focus on the 'EPT trio' (Innolight / Eoptolink / TFC Communication).\n\nAlso add a 'CSV Tools' page: export company core metrics and latest-results comparison as CSV, and support batch-updating quotes/earnings fields in the online database by stock code using a template.\n\nFeel free to refine more functional details along the way. Visual: finance-workbench style (deep-blue gradient header, A-share red-up/green-down colors). Chinese-only UI. All CSS/JS inline, no external CDN dependencies.", "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" }, { "id": "westockdata", "name": "WeStock Data", "description": "A股/港股/美股行情数据查询（K线、财报、资金流、技术指标、板块排行）", "description_en": "A-share/HK/US stock data: K-line, financials, capital flow, technicals & more", "skill_id": "skill_2053083170922696704" } ], "experts": [], "mcps": [], "creator": { "name": "星月聊财经", "name_en": "星月聊财经", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 97, "source": "community", "created_at": "2026-08-26T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "oracle-bone-character-derivation-wei", "title": "「为」字为什么这么写", "title_en": "Why is 「为」 written that way?", "subtitle": "看甲骨字形推本义——字库的释义为依据，必要时用联网检索补一段学界看法", "subtitle_en": "Read the glyph, infer the origin — source from the library, supplement with scholarly views when needed", "description": "讲古文字最迷人的环节就是「为什么这么写」，但也最容易翻车：很多自媒体言之凿凿地讲「『为』就是大象的形状，因为古文字里『为』通『象』」，学界其实没这么简单。有些字库能直接给一句「本义为……」的字源依据，有些则要靠学者论文才能讲清来龙去脉。\n\n这个案例演示的是「腾讯探元甲骨字查询助手」回答字形取象类问题的能力：用户问「『为』字为什么这么写」，专家先按字头查字库（query=「为的甲骨文」），命中字库释义就直接挂图说明；字库释义未涉及字形取象则如实告知「字库对此没有进一步说明」，并改走联网检索补一段学界看法、标注来源。整个过程不臆造字形、不强行把「为」解释为「大象」，区分字库依据与公开资料。\n\n适合：古文字课讲「会意字 vs 形声字」、博物馆展板讲文字学小知识、文博自媒体做甲骨文字源栏目、字源类公众号做图文、对外汉语教师备字源课。", "description_en": "The most captivating beat in any paleography lesson is «why it was written that way» — and the easiest one to get wrong. Many casual writers confidently explain 「为」 as «the shape of an elephant, because 『为』 originally meant 『象』», but academia is more cautious. Sometimes the library gives a one-line dictionary entry for character origin; sometimes only scholarly papers can fill in the full backstory.\n\nThis case shows how the TanYuan Oracle Bone Script Lookup assistant answers glyph-origin questions: the user asks «why is 「为」 written that way?» The expert first queries the library (query: 「为的甲骨文」). If the library's interpretation covers the origin, the answer hangs on the glyph directly. If the library's interpretation does not address origin, the assistant honestly notes that the library has no further comment and switches to web search to add a calibrated scholarly view, properly credited. The answer never fabricates a glyph, never forces 「为」 to mean 「大象」, and keeps library sources separate from public sources.\n\nBest for: paleography courses on compound-ideograph vs phono-semantic compounds, museum exhibit panels offering a paleography sidebar, cultural-heritage media running a character-origin column, character-origin WeChat accounts and overseas-Chinese teachers preparing character-origin lessons.", "categories": [ "education", "search" ], "tags": [ "甲骨文", "字形取象", "字源", "通识问答", "学界并存说法" ], "tags_en": [ "Oracle bone script", "Glyph origin", "Character etymology", "General knowledge Q&A", "Multiple scholarly views" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "「为」字源解读" } ], "prompt": "你是「腾讯探元甲骨字查询助手」。\n\n# 任务\n回答用户：**「『为』字为什么这么写」**，按混搭档位（D + 字卡）回答：先出字卡（图 → 表格 → 释义），再补一段字形取象的解读（字库够用就用字库，库不足就联网检索补一段）。\n\n# 走法（严格按助手内置规则）\n1. 字卡部分：\n - 调一次 `search_oracle_bone_character`，query 严格按 \"为的甲骨文\"。\n - 过滤：仅保留**字头等于「为」**的记录，其余丢弃（不论释义里出现「为」还是其它看似相关）。\n - 命中且释义涉及「字形取象」/「象形」「本义为……」「像……之形」等表述 → 直接挂在卡片下做一段解读，依据来自字库，**不**外加未在释义中出现的解释。\n2. 字形取象部分：\n - 库内释义已说明 → 一段引用，标「> 来源：腾讯探元甲骨文字库」。\n - 库内释义未说明 → **如实告知**用户「字库对此没有进一步说明」，改走联网检索找学界主流看法（如「『为』本义大象 vs 本义母猴 vs 为+象字形混合」等说法均有学界讨论），分点简述、保留「待考」「或说」「(某某说)」等审慎表述，不选边站、不把一家之言说成定论，标「> 来源：联网检索」。\n3. 字形图：仅在有编码时出图，URL 严格走 `https://static.jgw.qq.com/jgwpng/png/<编码>.png`。\n4. 字库来源与联网来源在各段末尾**分别标注**，不合并成一句笼统的来源。\n\n# 输出要求\n- 一次回复给齐：字卡 + 字形取象解读（库依据或联网依据）。\n- 不臆造字形、不强行解释为「大象」或其它单一说法；学界分歧如实并存。\n- 全文不出现工具名、置信度、SQL、原始 JSON、检索过程叙述、甲骨字编码。\n\n用户输入：「为」字为什么这么写？", "prompt_en": "You are the 'TanYuan Oracle Bone Script Lookup Assistant'.\n\n# Task\nAnswer the user: **Why is 「为」 written that way?** Use the mixed mode (card + origin discussion): first emit the standard card (image → table → interpretation), then add a paragraph of glyph-origin discussion (use the library if it suffices, otherwise fall back to web search for a calibrated scholarly view).\n\n# Workflow (strictly follow the assistant's built-in rules)\n1. Card part:\n - Call `search_oracle_bone_character` exactly once, with canonical query \"为的甲骨文\".\n - Filter: keep ONLY records whose character head equals 「为」; everything else is discarded (no matter if the interpretation happens to mention 「为」).\n - If a retained interpretation mentions glyph origin (e.g. «象形», «本义为…», «像…之形»), use it directly as the discussion anchored to the card; do NOT add anything the interpretation does not contain.\n2. Origin-discussion part:\n - Library interpretation sufficient → quote it directly, then close that block with `> 来源：腾讯探元甲骨文字库`.\n - Library interpretation does NOT cover origin → honestly tell the user «the library has no further comment on this», and switch to web search for the scholarly mainstream (e.g. competing views that 「为」 originally meant elephant, or a mother-monkey, or a hybrid elephant-representation, all of which have academic discussion). Present the views in short bullets, preserve cautious markers like «待考», «或说», «(某某说)», never pick a side, never assert any single view as established, and close with `> 来源：联网检索`.\n3. Glyph image: render only when an internal glyph code is available. URL MUST be `https://static.jgw.qq.com/jgwpng/png/.png`.\n4. Library source and web source are credited separately at the end of their respective paragraphs — do not merge them into one generic line.\n\n# Output requirements\n- One single reply contains: the card + the origin discussion (either library-backed or web-backed).\n- Do NOT fabricate glyph shapes, do NOT force a «great elephant» reading, do NOT pick a single academic side; present conflicting scholarly views honestly when they exist.\n- NEVER expose the tool name, confidence scores, raw SQL/JSON, retrieval process narration, or the internal glyph code in the answer.\n\nUser input: Why is the character 「为」 written that way?", "skills": [], "experts": [ { "id": "tanyuan-oracle-bone-expert", "name": "甲骨文找字专家", "description": "腾讯 SSV 数字文化实验室出品，依托腾讯探元甲骨文字库，提供现代汉字的甲骨字形图、读音、释义与通识问答", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b514405bc2ef", "expert_id": "ex_dMGPC1t0ekJ5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:36:22+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "oracle-bone-multiple-interpretations-huang", "title": "「黄」字为什么有三种解释", "title_en": "Why the character huang has three competing explanations", "subtitle": "构形三说并存 + 卜辞用法 + 六书归类——一个字看懂古文字研究里的分歧", "subtitle_en": "Competing etymologies, oracle-bone usage and the six-categories class, all from one verified character", "description": "查甲骨字最容易踩的坑，是把某一家之言当成定论。同一个字形，学界往往并存几种互相说服不了的解释，教材却常常只印其中一种。\n\n这个案例以「黄」字为例，演示「腾讯探元甲骨字查询助手」如何忠实呈现分歧：字库同时收录三种构形解释——佩玉说（像人佩玉之形，为珩或璜之初文）、由「矢」分化说（「黄」与「寅」皆由「矢」分化，借作珩或璜，与构形无关）、「尪」之初文说（从大，像人仰面向天、腹部膨大，假借为颜色之「黄」）——并用「或说」「也有人认为」明确标注，不替用户拍板。除此之外，还给出卜辞中的实际用法（多用作人名、地名或颜色，还可作神名）与六书归类（初文属合体象形）。\n\n适合：博物馆社教做字源讲解、文博自媒体做科普内容、高校古文字入门课、书法篆刻教学、编辑校对古文字类稿件。", "description_en": "The most common trap in looking up oracle-bone forms is mistaking one scholar's reading for the settled answer. A single glyph often carries several competing explanations that none of its proponents can settle, yet textbooks usually print just one.\n\nThis case uses the character 「黄」 to show how the TanYuan Oracle Bone Script Lookup assistant represents that disagreement faithfully: the library records three coexisting etymologies — the pendant-jade reading, the derivation-from-arrow reading, and the wang-prototype reading — each explicitly hedged with formulaic wording rather than resolved into a single verdict. It also supplies attested oracle-bone usage and the six-categories classification.\n\nBest for: museum educators, cultural-heritage media editors, introductory paleography courses, calligraphy and seal-engraving tutors, and editors fact-checking classical-script copy.", "categories": [ "education", "creation" ], "tags": [ "甲骨文", "古文字查询", "构形分析", "卜辞解读", "字源科普" ], "tags_en": [ "Oracle bone script", "Ancient script lookup", "Etymology", "Oracle-bone text interpretation", "Paleography primer" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "「黄」字多维解读" } ], "prompt": "你是「腾讯探元甲骨字查询助手」。\n\n# 任务\n帮我查「黄」字的甲骨字，重点要讲清楚它的构形分歧、卜辞用法和六书归类。\n\n# 走法（按助手内置规则）\n1. 调用 `search_oracle_bone_character`，query 严格按 \"黄的甲骨文\" 形态。\n2. 检索结果里**字头必须等于**「黄」，否则一律丢弃。\n3. 按档位 A 出卡：字头标题 → 字形图 → 读音/繁体/简体表格 → 文字释义。\n4. 在释义之后，把字库给出的内容按三个维度拆开讲：**构形解释**（有几种说法就列几种，保留「或说」「也有人认为」的表述，不替学界下结论）、**卜辞用法**、**六书归类**。\n5. 字形图走 `https://static.jgw.qq.com/jgwpng/png/<编码>.png`，编码仅内部使用、不写进回答。\n6. 末尾标注「> 来源：腾讯探元甲骨文字库」。\n\n# 输出要求\n- 字库没写的内容一律不补，不得引入字库之外的学者归属、年代或通假推断。\n- 全文不允许出现工具名、置信度、SQL、原始 JSON、检索过程叙述、甲骨字编码。\n\n用户输入：黄的甲骨文", "prompt_en": "You are the 'TanYuan Oracle Bone Script Lookup Assistant'.\n\n# Task\nLook up the oracle-bone form of 「黄」, covering its competing etymologies, attested oracle-bone usage, and six-categories classification.\n\n# Workflow\n1. Call `search_oracle_bone_character` with the query in the canonical form \"黄的甲骨文\".\n2. Keep ONLY records whose character head exactly equals 「黄」.\n3. Emit one standard card: character-head title → glyph image → reading / traditional / simplified table → interpretation paragraph.\n4. After the interpretation, split the library content into three dimensions: **etymology** (list every competing reading, preserving hedged wording, never adjudicating), **oracle-bone usage**, and **six-categories classification**.\n5. Glyph images use `https://static.jgw.qq.com/jgwpng/png/.png`; the internal code must never leak into the answer.\n6. Close with .\n\n# Output requirements\n- Never supplement beyond what the library records: no scholar attributions, dating, or loan-graph inferences from outside the library.\n- NEVER expose the tool name, confidence scores, raw SQL/JSON, retrieval narration, or the internal glyph code.\n\nUser input: 黄的甲骨文", "skills": [], "experts": [ { "id": "tanyuan-oracle-bone-expert", "name": "甲骨文找字专家", "description": "腾讯 SSV 数字文化实验室出品，依托腾讯探元甲骨文字库，提供现代汉字的甲骨字形图、读音、释义与通识问答", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b514405bc2ef", "expert_id": "ex_dMGPC1t0ekJ5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:36:22+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "oracle-bone-single-lookup-xiang", "title": "「象」字甲骨文图鉴", "title_en": "Single-glyph card for 「象」", "subtitle": "字形图 + 读音 + 释义 + 异体说明——一张图讲清楚一个字的来龙去脉", "subtitle_en": "Glyph + reading + interpretation + variant notes — full provenance in a single card", "description": "讲课时最常卡壳的不是冷僻字，反而是最熟悉的那个——「象」字甲骨文什么样、跟今天的写法怎么对应、为什么这样写，老师张口就能答上两句，但要给学生一张能贴课堂笔记的卡，常常就不知道去哪儿翻资料。网上图片要么是设计演绎、要么把字头画错，硬贴进去还要担心误人子弟。\n\n这个案例演示的是「腾讯探元甲骨字查询助手」单字深挖能力：给出「象」一个，专家调用字库、抽取本义与字形取象、必要时合并异体、按档位 A 出卡（图 → 表格 → 释义），结尾标注字库来源；用户拿来即可贴教案、做展板、配推文，不用二次校对。\n\n适合：博物馆社教讲「象」字引申用法、小学语文讲会意字取象、中学历史讲商代象形文字、出版社古文字专栏撰稿、篆刻家入印前的字形参考。", "description_en": "The awkward moment in class is rarely over an obscure glyph — it is the familiar one. Teachers can usually say a sentence about how 「象」 looks in oracle bone script, how it maps onto today's form, and why it was written that way, but producing a print-ready card students can paste into their notebooks is a different matter: every free image online is either a designer's re-drawing or carries an outright wrong glyph head, and reposting means risking misinformation.\n\nThis case demonstrates the TanYuan Oracle Bone Script Lookup assistant's single-glyph depth: feed it 「象」 and the expert queries the library, extracts the original meaning and visual etymology, merges variants where appropriate, and emits one tidy card (image → table → interpretation) with the library credited at the end. The output drops directly into lesson plans, exhibit panels and articles with no second-round proofreading.\n\nBest for: museum docents covering the figurative extensions of 「象」, elementary-school Chinese teachers introducing compound-ideograph character origin, middle-school history teachers teaching Shang-period pictographs, publishers working on ancient-script columns, and seal-engraving artists cross-checking a glyph before carving.", "categories": [ "education", "creation" ], "tags": [ "甲骨文", "单字详解", "字形图鉴", "会意字", "字源教学" ], "tags_en": [ "Oracle bone script", "Single-glyph deep dive", "Glyph card", "Compound-ideograph", "Character etymology teaching" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "「象」甲骨字卡（含异体）" } ], "prompt": "你是「腾讯探元甲骨字查询助手」。\n\n# 任务\n查「象」字的甲骨字并出一张完整字卡。\n\n# 走法（按助手内置规则）\n1. 仅一次调用 `search_oracle_bone_character`，query 严格按 \"象的甲骨文\"。\n2. 检索结果逐条筛：**字头（繁体或简体）必须等于「象」**，其余一律丢弃。\n3. 同一字头命中多条记录：\n - 同一个甲骨字编码 → 视为同一异体，合并为一段。\n - 不同甲骨字编码 → 视为不同异体，在同一字卡下分「异体 1 / 异体 2」小节，各出各自的字形图与释义。\n4. 字段渲染顺序：字头标题 → 字形图 → 读音 / 繁体 / 简体 表格 → 文字释义。\n5. 字形图全部走 `https://static.jgw.qq.com/jgwpng/png/<编码>.png`。\n6. 释义忠于字库原文，可改标点、改编号，但**不得增补**库外解释；库中未涉及的字段直接不出，不要写「暂无」「未知」「—」。\n7. 末尾标注「> 来源：腾讯探元甲骨文字库」。\n\n# 输出要求\n- 单次回复完成整张卡，含字形图、表格、释义与异体（如有）。\n- 全文不出现工具名、置信度、SQL、原始 JSON、检索过程叙述、甲骨字编码。\n- 不展开金文/小篆/隶书演变（字库不提供这些字段），若想讲演变改走「联网检索 + 标源」。\n\n用户输入：「象」的甲骨字是什么样子？", "prompt_en": "You are the 'TanYuan Oracle Bone Script Lookup Assistant'.\n\n# Task\nLook up 「象」 and produce one complete oracle-bone-script card.\n\n# Workflow (follow the assistant's built-in rules)\n1. Call `search_oracle_bone_character` exactly once, using the canonical query \"象的甲骨文\".\n2. Filter each result record strictly: the character head (simplified or traditional) MUST equal 「象」; everything else is discarded.\n3. If multiple records share the same oracle-bone code → treat as one variant, merge into a single section. If they have different codes → render them as distinct variants under the same character-head card, with separate subsections 「异体 1 / 异体 2」 each carrying its own glyph image and interpretation.\n4. Field rendering order: head title → glyph image → reading / traditional / simplified table → interpretation paragraph.\n5. All glyph images use `https://static.jgw.qq.com/jgwpng/png/.png`.\n6. The interpretation must track the library's wording. Punctuation and enumeration may be tidied, but do NOT add any explanation beyond the library's own. Fields absent from the library stay absent — never write \"暂无\", \"未知\" or \"—\" as placeholders.\n7. Close with `> 来源：腾讯探元甲骨文字库`.\n\n# Output requirements\n- Deliver the entire card in a single reply, including glyph image, table, interpretation and any variants.\n- NEVER expose the tool name, confidence scores, raw SQL/JSON, retrieval process narration, or the internal glyph code in the answer.\n- Do NOT extend into bronze / seal / clerical-script evolution (those fields are not in the library). If the user wants evolution, switch to web search and credit the source.\n\nUser input: What does the oracle-bone-script form of 「象」 look like?", "skills": [], "experts": [ { "id": "tanyuan-oracle-bone-expert", "name": "甲骨文找字专家", "description": "腾讯 SSV 数字文化实验室出品，依托腾讯探元甲骨文字库，提供现代汉字的甲骨字形图、读音、释义与通识问答", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b514405bc2ef", "expert_id": "ex_dMGPC1t0ekJ5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:36:22+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "orange-launch-sales-poster", "title": "橙心果园开售战报海报", "title_en": "Orange Launch Sales Poster", "subtitle": "一整条橙子皮游过画面，托起开售当天的数字", "subtitle_en": "A single unbroken peel winds across the frame and holds up launch-day numbers", "description": "开售战报最容易变成“热烈祝贺、销量破万”加一堆红色爆炸贴，数字被淹没，看着也不像真的。\n\n这个案例给虚构的脐橙果园「橙心果园」做了一张横版战报：一条完整的橙子皮从右上角蜿蜒游过画面，尽头是一只剥好的橙子；黑色大字“3,860 箱”靠左最先被看见，下面是“开售 3 小时售出，还剩 640 箱”和“明早 9:00 第二批开售”，右下角留着二维码的位置。\n\n适合水果、生鲜等限量开售的商家，也适合想参考“用商品本身的形状带出视线”的海报设计者。", "description_en": "Launch reports tend to become congratulations, record sales and a pile of red burst stickers, which swallows the figures and does not look believable.\n\nThis case makes a landscape sales poster for a fictional navel orange orchard. One unbroken orange peel winds across the frame from the top right and ends at a peeled orange. “3,860 boxes” in big black type on the left is seen first, followed by “sold in 3 hours, 640 boxes left” and “second batch tomorrow at 9:00”, with a spot at the lower right reserved for a QR code.\n\nUseful for fruit and fresh-produce merchants with limited-quantity launches, and for poster designers exploring how the product's own shape can lead the eye.", "categories": [ "creation", "business" ], "tags": [ "战报海报", "水果", "开售", "微信小店" ], "tags_en": [ "Sales Poster", "Fruit", "Launch", "WeChat Shop" ], "artifact_type": "image", "preview": "output.png", "artifacts": [ { "file": "output.png", "label": "橙心果园开售战报海报（PNG）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "为脐橙果园「橙心果园」做一张微信小店开售战报海报，横版 16:9，用于店铺动态、群聊和大屏展示。数字如下：开售 3 小时售出 3,860 箱，还剩 640 箱，明早 9:00 第二批开售；统计至 10 月 14 日 21:00，数据来源微信小店后台。\n\n创意不要宫格、卡片、居中堆叠这类规整排版，用商品自己的形状带出视线：一条完整的橙子皮从右上角蜿蜒游过画面，尽头放一只剥好的橙子，背景是干净的奶白色，画面用写实棚拍质感。最大的黑色粗体数字“3,860 箱”靠左，下面两行“开售 3 小时售出、还剩 640 箱”，其下一行绿色小字“明早 9:00 第二批开售”；右下角留一块白色方形空白放二维码，底部一行小字写数据来源和统计时间。\n\n战报数字先请商家确认；海报只呈现商家后台的真实数字，不编造销量或评价；二维码由程序贴在预留的空白处。\n\n交付：一张 PNG 横版海报。", "prompt_en": "Create a WeChat Shop launch sales poster for a navel orange orchard, Orange Heart Orchard, in landscape 16:9, for shop updates, group chats and large screens. Figures: 3,860 boxes sold in the first 3 hours, 640 left, second batch tomorrow at 9:00; as of 14 October 21:00, source: WeChat Shop back office.\n\nAvoid tidy grids, cards or centred stacks; let the product's own shape lead the eye. One unbroken orange peel winds across the frame from the top right and ends at a peeled orange, on a clean milk-white background in a realistic studio-photo look. The biggest bold black figure, “3,860 boxes”, sits on the left, with two lines below, “sold in 3 hours, 640 left”, and a small green line beneath, “second batch tomorrow at 9:00”. Leave a white square at the lower right for a QR code and one small line at the bottom for the source and time.\n\nConfirm the merchant's figures first; the poster only shows real figures from the merchant's back office and never invents sales or reviews; the QR code is placed programmatically on the blank area reserved for it.\n\nDeliverable: one landscape PNG poster.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 97, "created_at": "2026-10-15T15:30:00+08:00", "updated_at": "2026-10-15T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "orange-orchard-promo-page", "title": "张记果园宣传页", "title_en": "Orange Orchard Promo Page", "subtitle": "一个巨大的「甜」字开场，用糖度计讲清为什么等到 11 月", "subtitle_en": "A giant “Sweet” opens the page; a Brix gauge explains why they wait until November", "description": "水果页很容易写成“新鲜、香甜、产地直发”，每家看起来都一样，买家分不出谁家的橙子真的更甜。\n\n这个案例给虚构的江西寻乌脐橙果园「张记果园」做了一张手机优先的宣传页：首屏是满屏橙色的大字“甜”和一枚糖度牌；往下，一支糖度计随着滚动从 5.2 升到 13.8，对应 9、10、11 月三个时点的糖度；再往下是 11.11 选一箱，点 5 斤、9 斤、15 斤，木箱和价格一起变。页面同时有中文和英文两个版本。\n\n适合想用一个具体数字讲清产品好在哪的农产品商家，也适合想参考“用一个大字撑起首屏”的设计者。", "description_en": "Fruit pages tend to say fresh, sweet and shipped from the orchard, so every shop looks the same and buyers cannot tell whose oranges are really sweeter.\n\nThis case builds a mobile-first promo page for a fictional navel orange orchard in Xunwu, Jiangxi. The first screen is one huge orange character for “sweet” and a Brix tag. Scrolling on, a gauge rises from 5.2 to 13.8 for September, October and November, then the 11.11 box picker lets visitors choose 2.5, 4.5 or 7.5 kg while the crate and price change together. The page comes in Chinese and English.\n\nUseful for farm merchants who want one concrete number to show what makes the product good, and for designers exploring a first screen carried by a single giant word.", "categories": [ "creation", "business" ], "tags": [ "宣传页", "水果", "农产品", "滚动动效" ], "tags_en": [ "Promo Page", "Fruit", "Farm Produce", "Scroll Motion" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "单页 HTML 宣传页" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "为江西寻乌脐橙果园「张记果园」（自家 40 亩）做一张手机优先的商品宣传页，用于 11.11 前把人引到店铺。卖点只有一个：不打蜡、不催熟，挂在树上等到 11 月，今年糖度 13.8。\n\n首屏用一个占满画面的橙色大字“甜”，右上角贴一枚“今年糖度 13.8”的标牌，橙子图在右下，滚动时橙子转起来；下面一句“不打蜡，不催熟，挂在树上等到 11 月”。接一条写着“树上熟透才摘、坏果包赔、当天摘当天发”的跑马灯带。\n\n中段是一支贴着屏幕一侧的糖度计，随滚动从 5.2 升到 13.8，旁边三个月份依次出现：9 月还酸着（8.6）、10 月开始转甜（11.2，白天 24°C 夜里 11°C）、11 月可以摘了（13.8，11 月 3 日第一批）。\n\n再往下是 11.11 选一箱：5 斤 ¥49、9 斤 ¥79、15 斤 ¥119，点选后木箱示意图变大变小、价格数字滚动到新数字；写明果径 80–85 mm 一级果、顺丰包邮（新疆西藏除外）、坏一个赔一个、11 月 1 日起每天摘 600 箱；最后一个去店铺的按钮，页尾写商家信息。\n\n风格：奶皮米色、橙色和深海军蓝，大字体量。商品图为示意，正式发布前用商家实拍替换；页面只做展示和引流，下单在店铺完成。商家、数字均为演示，不编造真实销量、评价或功效。\n\n交付：一个开箱即用的单页 HTML 宣传页，手机和电脑都能打开，页面里没有需要保存的数据。", "prompt_en": "Create a mobile-first product promo page for Zhang's Orchard, a navel orange orchard in Xunwu, Jiangxi (40 mu of its own), to bring people to the shop before 11.11. There is one selling point: no wax, no forced ripening, fruit left on the tree until November, this year's Brix 13.8.\n\nThe first screen is one orange character for “sweet” filling the frame, a tag in the top right reading “Brix this year 13.8”, and an orange photo at the bottom right that spins as the visitor scrolls. Beneath it: “No wax, no forced ripening, left on the tree until November.” A moving ticker band follows: picked only when fully ripe on the tree, bad fruit replaced, picked and shipped the same day.\n\nThe middle section has a gauge fixed to one side of the screen that rises from 5.2 to 13.8 as the visitor scrolls, with three months appearing beside it: September still sour (8.6), October turning sweet (11.2, 24°C by day and 11°C at night), November ready to pick (13.8, first batch on 3 November).\n\nNext is 11.11 Pick a box: 2.5 kg ¥49, 4.5 kg ¥79, 7.5 kg ¥119. Choosing one resizes the crate illustration and rolls the price to the new number. State grade-A fruit of 80–85 mm, free SF Express shipping (except Xinjiang and Tibet), one bad fruit replaced with one good one, and 600 boxes picked a day from 1 November. End with one button to the shop and shop details in the footer.\n\nStyle: cream peel, orange and deep navy, oversized type. Product photos are illustrative and should be replaced with the merchant's own before publishing. The page only showcases and drives traffic; orders are placed in the shop. Shop and figures are demos; never invent real sales, reviews or efficacy claims.\n\nDeliverable: a ready-to-use single-page HTML promo page that opens on phones and desktops; the page keeps no data of its own.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 92, "created_at": "2026-10-15T15:30:00+08:00", "updated_at": "2026-10-15T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "order-fulfillment-tracking", "title": "订单发到哪了？一张看板盯住全部履约", "title_en": "Order Fulfillment Tracking", "subtitle": "发货进度、交期状态一板盯全，超期预警", "subtitle_en": "Shipping progress and delivery status on one board, overdue alerts included", "description": "接入或上传订单，自动生成履约跟踪看板，实时显示发货进度、交期状态，超期订单一眼预警。", "description_en": "Connect or upload orders and it generates a fulfillment-tracking dashboard showing shipping progress and delivery status in real time, with overdue orders flagged at a glance.", "categories": [ "productivity" ], "tags": [ "订单履约", "交期跟踪", "内勤", "B2B贸易" ], "tags_en": [ "Order Fulfillment", "Delivery Tracking", "Back Office", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "订单履约跟踪看板（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易内勤专家。请帮我把在跟订单整理成一张履约跟踪看板。\n\n【数据来源】\n- 我已绑定业务系统/ERP：读取在跟订单数据；\n- 我上传了订单表 / 用【OCR识别】识别发货单磅单签收单：请归集进度；\n- 我暂时没有数据：说明需要哪些字段并给示例看板。\n\n【要求】\n1. 顶部KPI：在跟订单数、待发货、运输中、超期预警数；\n2. 订单明细表：订单号、客户、货物、数量、约定交期、履约进度、状态（待发货/运输中/已签收/已超期）；\n3. 履约提示：重点标注超期或临近交期的订单，给出协调/催发/通知客户的建议；把台账记进【腾讯文档】。\n\n若数据源未就绪或返回不足，请以通用订单履约管理方法论为主产出并说明。请生成完整履约跟踪看板。", "skills": [ { "id": "order-fulfillment-tracking", "name": "订单履约跟踪", "description": "TradeBuddy把“手上一堆单不知道哪个该催”的内勤诉求，做成按履约节点清晰、异常置顶、能直接照着盯的订单看板。", "skill_id": "skill_2091156322115547136" } ], "experts": [ { "id": "neiqin-expert", "name": "B2B贸易内勤专家", "description": "懂B2B贸易的内勤专家＝AI版录单员+档案室：识单、发票管理、智能对账、出表出单、盯订单进度。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Hu9b9KuHQwa" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "order-time-heatmap-insights", "title": "订单成交时段热力洞察", "title_en": "Order Conversion Time Heatmap Insights", "subtitle": "近30天支付订单成交时段与经营节奏分析", "subtitle_en": "A practical Weimob playbook for order conversion time heatmap insights", "description": "做订单成交时段热力洞察时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“订单成交时段热力洞察”展开：分析近30天支付订单的成交时段、GMV和订单量分布，识别高峰与低谷时段，并给出活动排期、客服排班和推送时机建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、经营数据概览运营人员、经营分析团队。", "description_en": "The hardest part of order conversion time heatmap insights is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates order conversion time heatmap insights through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Business Overview specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "成交时段", "订单热力", "GMV", "活动排期", "客服排班" ], "tags_en": [ "Order Conversion Time Heatmap Insights", "Business Diagnosis", "Business Overview", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "分析近30天支付订单的成交时段、GMV和订单量分布，识别高峰与低谷时段，并给出活动排期、客服排班和推送时机建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a order conversion time heatmap insights covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "org-biweekly-work-summary", "title": "机构双周工作小结", "title_en": "Biweekly Organization Summary", "subtitle": "近 14 天跨周期对比 + 月捐留存矩阵，双周会复盘一步到位", "subtitle_en": "14-day cross-period comparison plus monthly-donor retention matrix — biweekly review in one file", "description": "某基金会每两周向项目组同步数据，7 天窗口太短看不出趋势、30 天又太迟。\n\n本案由「机构工作小结」专家以近 14 天数据为底，对 15,208 名捐款用户做跨周期对比：本周期 vs 上一周期逐日趋势叠加（虚线基线），量化「总额环比 +16.3%、峰值 26.80 万」的节奏变化，并通过月捐留存矩阵（近 6 个月 91.2%→95.1%）判断用户基本盘是否在变好。交付含上期对比基线的单文件 HTML 看板。\n\n适合：双周会、活动中期复盘。", "description_en": "One foundation syncs data with project teams every two weeks — 7 days is too short to see a trend, 30 days is too late.\n\nThe Biweekly Summary Expert reads 14 days of data covering 15,208 donors and runs a cross-period comparison: this period vs. the previous one, with an overlaid daily-trend chart (dashed baseline) that quantifies the rhythm shift (\"WoW +16.3%, peak day ¥268K\"). A monthly-donor retention matrix (6-month view: 91.2% → 95.1%) tells you whether the user base is genuinely improving. Delivered as a single-file HTML dashboard with the previous period baked in as baseline.\n\nFor: biweekly reviews, mid-campaign retrospectives.", "categories": [ "data", "office" ], "tags": [ "公益", "工作小结", "双周报", "数据看板" ], "tags_en": [ "Public Welfare", "Biweekly Summary", "Ops Report", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我生成本机构最近 14 天的工作小结报告。", "prompt_en": "Please generate a work summary for our organization covering the last 14 days.", "skills": [], "experts": [ { "id": "ex_II2sGyfokFOC", "name": "腾讯公益机构工作小结专家", "description": "基于机构后台真实数据生成周报 / 双周报 / 月报，含资金、渠道、用户结构的多维复盘", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cb492d3cc351", "expert_id": "ex_II2sGyfokFOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "机构双周小结（脱敏样例）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "org-monthly-work-summary", "title": "机构上月工作小结", "title_en": "Monthly Organization Summary", "subtitle": "完整自然月 + 深色 SaaS 看板，理事会与月度会直接用", "subtitle_en": "Full calendar month in a dark-mode SaaS dashboard — board-ready, month-close-ready", "description": "每月 1 号，机构负责人需要一份覆盖完整自然月的复盘报告，用于月度总结与向上汇报。\n\n本案由「机构工作小结」专家以 2026 年 7 月全月 31 天数据为底，对 31,240 名捐款用户做完整盘点：628.63 万总额的资金产品切分、渠道 TOP8 集中度（小程序+H5 占 69.8%）、用户结构 4 卡、月捐留存矩阵与省份分布，并附口径说明（窗口/来源/算法/盲区）。交付深色 SaaS 风格的单文件 HTML，可印刷、可投屏。\n\n适合：直接用于理事会或月度总结会。", "description_en": "On the 1st of every month, org leaders need a full-month review for internal close-out and upward reporting.\n\nThe Monthly Summary Expert reads a full 31-day window (July 2026) covering 31,240 donors and produces a complete audit: ¥6.2863M total broken down by fund product, top-8 channel concentration (mini-program + H5 combined 69.8%), 4-card user structure, monthly-donor retention matrix, and province distribution — plus a methodology note (window / source / algorithm / blind spots). Delivered as a single-file HTML in a dark-mode SaaS style, printable and projectable.\n\nFor: board meetings and monthly close-out reviews.", "categories": [ "data", "office" ], "tags": [ "公益", "工作小结", "月报", "数据看板" ], "tags_en": [ "Public Welfare", "Monthly Summary", "Ops Report", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我生成本机构上个月的工作小结报告。", "prompt_en": "Please generate a work summary for our organization covering last month.", "skills": [], "experts": [ { "id": "ex_II2sGyfokFOC", "name": "腾讯公益机构工作小结专家", "description": "基于机构后台真实数据生成周报 / 双周报 / 月报，含资金、渠道、用户结构的多维复盘", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cb492d3cc351", "expert_id": "ex_II2sGyfokFOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "机构月度小结（2026 年 7 月脱敏样例）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "org-weekly-work-summary", "title": "机构本周工作小结", "title_en": "Weekly Organization Summary", "subtitle": "近 7 天 4 大 KPI 巡检 + 单文件 HTML 看板，周一晨会直接投屏", "subtitle_en": "7-day KPI check on 4 metrics, delivered as a single-file HTML dashboard ready for Monday's stand-up", "description": "公益机构运营每周一要快速回答「上周善款涨没涨、人稳不稳」，但手工拉表常要半天。\n\n本案由「机构工作小结」专家以平台近 7 天数据为底，对 8,942 名捐款用户做 4 个核心 KPI 巡检：捐赠总额环比、新老用户占比、月捐次月留存与净增，并用「资金/渠道/用户」三个结构模块定位波动来源，如「总额环比 +12.4% 主要来自 8 月 15 日活动单次捐放量」。交付一份带 AI 摘要的单文件 HTML 看板，零外链、可离线、周一晨会直接投屏。\n\n适合：周会汇报、日常巡检、小额专项基金。", "description_en": "Every Monday, nonprofit ops leads need a quick answer to \"did donations grow last week, is the user base stable?\" — but pulling reports by hand eats half a day.\n\nThe Weekly Summary Expert reads 7 days of platform data covering 8,942 donors and runs a check on 4 core KPIs: WoW total donations, new vs. returning donor mix, monthly-donor next-month retention, and net additions. It uses a three-module structure — Funds / Channels / Users — to pinpoint where the swing came from (e.g. \"WoW +12.4%, driven by the Aug 15 campaign one-off donation surge\"). Delivered as a single-file HTML dashboard with an AI-written summary — zero external links, fully offline, ready to project on Monday's stand-up screen.\n\nFor: weekly reviews, routine checks, small dedicated funds.", "categories": [ "data", "office" ], "tags": [ "公益", "工作小结", "周报", "数据看板" ], "tags_en": [ "Public Welfare", "Weekly Summary", "Ops Report", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我生成本机构最近 7 天的工作小结报告。", "prompt_en": "Please generate a work summary for our organization covering the last 7 days.", "skills": [], "experts": [ { "id": "ex_II2sGyfokFOC", "name": "腾讯公益机构工作小结专家", "description": "基于机构后台真实数据生成周报 / 双周报 / 月报，含资金、渠道、用户结构的多维复盘", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cb492d3cc351", "expert_id": "ex_II2sGyfokFOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "机构本周小结（脱敏样例）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "out-of-stock-alternative-selling-guide", "title": "缺货时还能卖什么", "title_en": "Out-of-Stock Alternative Selling Guide", "subtitle": "缺货商品替代推荐与导购防流失话术", "subtitle_en": "A practical Weimob playbook for out-of-stock alternative selling guide", "description": "做缺货时还能卖什么时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“缺货时还能卖什么”展开：当热销商品缺货或库存不足时，为门店准备可替代的商品推荐思路和导购说法，尽量避免客户流失。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、门店库存协同运营人员、经营分析团队。", "description_en": "The hardest part of out-of-stock alternative selling guide is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates out-of-stock alternative selling guide through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Store Inventory Collaboration specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "缺货商品", "替代推荐", "库存协同", "导购话术", "客户防流失" ], "tags_en": [ "Out-of-Stock Alternative Selling Guide", "Store and Sales Associate Operations", "Store Inventory Collaboration", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "当热销商品缺货或库存不足时，帮我为门店准备可替代的商品推荐思路和导购说法，尽量避免客户流失。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a out-of-stock alternative selling guide covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "palace-museum-digital-preservation", "title": "故宫博物院数字化保护技术的最新进展", "title_en": "Latest Progress of Palace Museum Digital Preservation Technology", "subtitle": "「数字故宫」二十年：从三维扫描到 AI 修复的科技赋能路径", "subtitle_en": "Two decades of Digital Palace Museum: from 3D scanning to AI-assisted restoration", "description": "想弄清故宫博物院数字化保护到了哪一步?单看新闻常被「数字故宫」「智慧博物馆」「云游故宫」这些词刷屏,但到底落到哪些技术、哪些项目、哪些可考量的进展?\n\n这个案例演示了如何借助腾讯探元文物数据库与知识库的并行检索,主代理亲自做 3-4 路网络调研(最新进展/典型案例/影响成效/发展脉络),覆盖三维激光扫描、高精度纹理采集、AI 辅助文物修复、数字孪生与云展览、可移动文物数字化保护标准、活化利用与文创开发等维度,最终交付包含 3-6 张配图与 1-2 张趋势图表的 Markdown / HTML / DOCX / PDF 四格式趋势综述报告。\n\n适合:博物馆数字化保护从业者、考古与文物保护工程技术人员、文博策展人与社教人员、文化遗产政策研究者、博物馆科技企业与文创开发从业者、文物与博物馆学专业研究生。", "description_en": "Want to know how far the Palace Museum's digital preservation has advanced? News headlines throw around buzzwords like \"Digital Palace Museum,\" \"smart museum,\" and \"Cloud Tour of the Palace,\" but it's hard to tell which technologies, projects, and measurable advances are real. This case shows how to leverage Tencent Tanyuan relic and knowledge bases for parallel retrieval, plus 3-4 concurrent main-agent WebSearch passes (latest progress / benchmark cases / impact / historical lines), covering 3D laser scanning, high-resolution texture acquisition, AI-assisted restoration, digital twins and cloud exhibitions, movable-relic digitization standards, plus cultural-product development, delivering a 3-6 image + 1-2 trend-chart, Markdown/HTML/DOCX/PDF four-format trend review report.\n\nBest for: museum digitization practitioners, archaeology and conservation engineers, curators and docents, cultural-heritage policy researchers, museum-tech and cultural-product professionals, museum and heritage graduate students.", "categories": [ "business" ], "tags": [ "故宫博物院", "数字故宫", "三维扫描", "AI文物修复", "数字孪生" ], "tags_en": [ "Palace Museum", "Digital Palace Museum", "3D Scanning", "AI Restoration", "Digital Twin" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含趋势图表)" } ], "prompt": "请按「腾讯探元学术研究-趋势综述」专家的工作流,综述「故宫博物院数字化保护技术的最新进展」。\n\n研究范围(覆盖五维):\n- 三维激光扫描与高精度纹理采集:古建筑本体(太和殿、养心殿、灵沼轩砖雕等)、可移动文物(家具、织绣、书画)的微米级三维数据获取\n- AI 辅助文物修复:残损虚拟复原、书画接笔、瓷器补配、彩绘增强,以及「文物指纹」标准建立\n- 数字孪生与云展览:「数字故宫」「全景故宫」「云游故宫」平台、海内外巡展数字化呈现\n- 可移动文物数字化保护标准:行业标准编制、数据规范、保护装备研发\n- 活化利用与文创开发:数字 IP、文创产品、研学课程与公众教育\n\n交付要求(强制):\n1. 探元两库 MCP(search_movable_relics + search_relics)并行检索 + 主代理 3-4 路 WebSearch 多源核验,优先近 1-2 年(2025/2026)信源\n2. 配图 ≥ 3 张(Wikimedia 公共领域 / 故宫博物院官网 / 公开新闻图源),必须实图、与主题相关\n3. 趋势图表 ≥ 1 张(建议 2-3 张:数字化进展时间线/技术体系桑基图/海外影响力雷达图等),matplotlib 渲染为 SVG,严格避免标签重叠\n4. 四格式报告:Markdown 母版 → HTML(嵌入 base64 图片) → DOCX → PDF,综述结构(领域背景→发展现状→典型案例→影响成效→趋势展望→结论→证据与局限→参考来源)\n5. 报告日期用真实当前日期;参考文献按时间倒序;只写结论不含建议;主要趋势判断同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Trend Review' expert workflow to survey the 'Latest Progress of Palace Museum Digital Preservation Technology'.\n\nResearch scope (five dimensions):\n- 3D laser scanning and high-resolution texture acquisition: ancient building bodies (Hall of Supreme Harmony, Hall of Mental Cultivation, Lingzhao Xuan brick carvings), movable relics (furniture, textiles, calligraphy), micrometer-grade 3D data acquisition\n- AI-assisted restoration: virtual reconstruction of damaged works, calligraphy-painting stroke continuation, porcelain patching, polychrome enhancement, plus 'cultural-relic fingerprint' standards\n- Digital twins and cloud exhibitions: 'Digital Palace Museum,' 'Panoramic Palace Museum,' 'Cloud Tour of the Palace' platforms, digitised touring exhibitions\n- Movable-relic digitization standards: industry standard drafting, data specifications, conservation equipment R&D\n- Activation and cultural-product development: digital IP, cultural products, research courses, public education\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 3-4 main-agent WebSearch passes for cross-source verification, prioritizing sources from the last 1-2 years (2025/2026)\n2. ≥ 3 images (Wikimedia public domain / Palace Museum official site / public news sources), each verified real and topic-relevant\n3. ≥ 1 trend chart (2-3 recommended: digitization timeline / technology sankyaku / global impact radar), rendered to SVG via matplotlib with strict label overlap avoidance\n4. Four formats: Markdown master → HTML (base64-inlined images) → DOCX → PDF, trend-review structure (background → status → cases → impact → outlook → conclusion → limits → references)\n5. Report date is the true current date; references sorted reverse-chronologically; conclusions only, no recommendations; key trend judgments sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-trend", "name": "文博趋势综述专家", "description": "宽泛主题广谱调研,3-4 路网络调研强调时效性,案例卡片化呈现,交付四格式趋势综述报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "2e5f2aab7036", "expert_id": "ex_M24IASAeVHLO" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:20:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pan-finance-workbench", "title": "PAN 的个人理财工作台", "title_en": "PAN's Personal Finance Workbench", "subtitle": "持仓、预算、账户、目标加半年报，一个工作台看管理财全景", "subtitle_en": "Holdings, budgets, accounts, goals and interim reports in one finance workbench", "description": "钱分散在券商、银行、零钱和基金里，每次想看看整体情况都要开五六个 App 对着算。\n\n这个案例是 PAN 的个人理财工作台：首页一眼看到总资产、本月收支和储蓄率，配资产配置环形图和近 6 个月收支趋势；点开收支管理能看预算执行进度和分类支出，随手「记一笔」就写进资料库数据表；账户与现金流自动汇总各账户余额；理财目标按购车基金、应急备用金这样的家庭大事跟踪进度；再往深了走还有自选股行情、组合持仓与再平衡建议、提醒中心、投资笔记、2026 半年报专题（行业对标、ROE 散点、财务三表）、智能筛选和 AI 研报问答。所有记录落在资料库的结构化表里，手机电脑同步。\n\n适合：想把散落各处的钱收进一个界面统一看的人、认真做预算和储蓄率管理的人、喜欢自己跟财报季和持仓做决策的个人投资者。", "description_en": "Money scattered across brokers, banks, wallets and funds means checking five apps just to answer \"how am I doing?\".\n\nThis case is PAN's personal finance workbench: the home page shows total assets, monthly income and spending, and the savings rate at a glance, with an asset-allocation ring and a 6-month income-vs-spending trend. The Income & Spending module tracks budget progress and spending by category, and one tap on \"Add Entry\" writes straight to a library data table. Accounts & Cash Flow aggregates balances across accounts; Financial Goals follows family milestones like the car fund and the emergency reserve. Go deeper and you get a watchlist with live quotes, portfolio positions with rebalancing advice, an alert center, investment notes, a 2026 interim-reports hub (industry benchmarking, ROE scatter, three statements), a smart screener and an AI research chat. All records live in structured library tables that sync across devices.\n\nBest for: anyone who wants every scattered account in one view, people serious about budgeting and savings discipline, and hands-on investors who track earnings season and their own positions.", "categories": [ "workspace" ], "tags": [ "个人理财", "投资组合", "预算管理", "半年报分析", "数据同步" ], "tags_en": [ "Personal Finance", "Portfolio", "Budgeting", "Interim Reports", "Data Sync" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "PAN 个人理财工作台（单体 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/u9ZfIAPjZRLnvjx7FtIhLl?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n帮我做一个「个人理财工作台」，基于资料库能力搭建在线版工作台，所有记录存储于资料库数据表并支持双向同步。\n\n包含：① 理财全景首页：总资产/本月收支/储蓄率 KPI，资产配置环形图（股票/基金/理财/现金/年金/数字资产），近 6 月收支趋势折线；② 收支管理：预算执行进度、分类支出，「记一笔」按钮把记录写入资料库数据表；③ 账户与现金流：各账户余额、月度流水、近期账单；④ 理财目标：购车基金、应急备用金等目标进度与缺口分析；⑤ 自选股与个股分析：行情卡片、个股详情抽屉（财务指标、K 线、资金流向）；⑥ 组合管理：持仓明细、目标仓位与偏离、再平衡建议；⑦ 提醒中心与投资笔记；⑧ 半年报专题：行业对标、ROE 散点（营收增速 × 净利率）、重点公司业绩排行；⑨ 智能筛选（自然语言选股）与 AI 研报问答。\n\n数据结构（资料库数据表）：名称 / 类型（select：持仓·预算·账户·目标·提醒·笔记）/ 状态（select：进行中·逾期·关注·已完成）/ 金额 / 日期 / 分类 / 备注；预置 12 条演示记录（持仓、月度预算、账户、目标、提醒、复盘笔记各若干）。\n\n视觉：浅色金融工作台风，主色 #165DFF 深蓝、涨红跌绿，左侧导航 + 顶部搜索，ECharts 图表（K 线、散点、环形、折线）。无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 9 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/u9ZfIAPjZRLnvjx7FtIhLl?source=2, with online data storage and two-way sync. Prefer the one-click clone path, if unavailable build as follows: Build a \"Personal Finance Workbench\" as an online workbench powered by the library capability, with all records stored in library data tables and two-way sync. Include: ① a finance overview home with total assets / monthly income & spending / savings-rate KPIs, an asset-allocation ring (stocks / funds / wealth products / cash / annuity / digital assets) and a 6-month income-vs-spending trend; ② income & spending with budget progress, spending by category, and an \"Add Entry\" button that writes records to a library data table; ③ accounts & cash flow with balances, monthly flows and recent bills; ④ financial goals with progress and gap analysis for milestones like the car fund and emergency reserve; ⑤ watchlist and stock analysis with quote cards and a stock drawer (financial metrics, K-line, capital flow); ⑥ portfolio management with positions, target weights, drift and rebalancing advice; ⑦ an alert center and investment notes; ⑧ an interim-reports hub with industry benchmarking, an ROE scatter (revenue growth × net margin) and top-company rankings; ⑨ a natural-language smart screener and an AI research chat. Data structure (library data table): 名称 (name) / 类型 (type, select: position · budget · account · goal · alert · note) / 状态 (status, select: active · overdue · watch · done) / 金额 (amount) / 日期 (date) / 分类 (category) / 备注 (note); pre-seed 12 demo records (positions, monthly budgets, accounts, goals, alerts and review notes). Visual: light finance-workbench look, primary #165DFF deep blue, red-up/green-down quotes, left sidebar navigation + top search, ECharts charts (K-line, scatter, ring, line). No external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus 9 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "PAN", "name_en": "PAN", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "community", "usage_count": 0, "created_at": "2026-08-20T14:30:00+08:00", "updated_at": "2026-08-20T14:46:00+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "paper-review-mechanism-methylation", "title": "中心法则甲基化机制3D可视化", "title_en": "Central Dogma Methylation 3D Visualization", "subtitle": "一站式交互探索DNA/RNA/蛋白甲基化的书写器-擦除器-阅读器系统", "subtitle_en": "Interactive exploration of the writer-eraser-reader system across DNA, RNA, and protein methylation", "description": "阅读综述论文时，最头疼的是机制描述分散在几十页的文字中，难以建立整体认知。\n\n这个案例演示了如何将一篇46页的Nature子刊甲基化综述转化成可交互的3D可视化网页。场景中包含细胞核/线粒体/内质网等3D细胞器、DNA/RNA/蛋白三类甲基化的粒子动画、12项实时HUD指标条，以及6个疾病进展阶段（从正常稳态到癌症/神经系统疾病/衰老，再到治疗干预）。点击任意分子标签立即弹出功能详情面板。\n\n适合：生物医学研究人员、学术论文作者制作补充材料、医学生理解表观遗传机制、科普内容创作者。", "description_en": "The most tedious part of reading review papers is piecing together mechanisms scattered across dozens of pages.\n\nThis case demonstrates transforming a 46-page Nature journal methylation review into an interactive 3D visualization. The scene features 3D organelles (nucleus/mitochondria/ER), particle animations for DNA/RNA/protein methylation, 12 real-time HUD metrics, and 6 disease progression stages—from normal homeostasis through cancer/neurological disorders/aging to therapeutic intervention. Click any molecular label for instant detail panels.\n\nFor: biomedical researchers, paper authors creating supplementary materials, medical students learning epigenetic mechanisms, science communicators.", "categories": [ "education", "search", "data" ], "tags": [ "表观遗传学", "DNA甲基化", "RNA甲基化", "3D可视化", "综述论文" ], "tags_en": [ "Epigenetics", "DNA Methylation", "RNA Methylation", "3D Visualization", "Review Paper" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "甲基化机制3D交互可视化网页" } ], "prompt": "请从以下论文PDF构建一个交互式甲基化机制可视化网页：\n\n论文：Liu et al., 2023, Signal Transduction and Targeted Therapy 8:310, doi:10.1038/s41392-023-01528-y\n论文标题：Methylation across the central dogma in health and diseases: new therapeutic strategies\n\n要求：\n1. 构建基于Three.js的3D交互场景，包含细胞膜（半透明圆角长方体）、细胞核（含染色质/组蛋白/核仁）、线粒体、内质网等细胞器。\n2. 使用粒子系统分别展示DNA甲基化（蓝色，核内）、RNA甲基化（紫色，胞质）、蛋白甲基化（青色，胞质）和SAM甲基供体（金色）。\n3. 右上角HUD面板显示12项实时指标：DNA甲基化、RNA甲基化、蛋白甲基化、SAM水平、CpG岛高甲基化、全基因组低甲基化、H3K4me3/H3K27me3/H3K9me3/H3K36me3、疾病进展、治疗响应。\n4. 底部6个阶段按钮：正常甲基化稳态 → 早期表观遗传扰动 → 癌症甲基化重塑 → 神经系统疾病 → 衰老相关甲基化改变 → 治疗干预策略。阶段切换时参数平滑插值过渡。\n5. 3D标签标注关键分子（DNMT1/3A/3B、TET1/2、METTL3/14、FTO、ALKBH5、EZH2、SUV39H1、SETD2、LSD1、MeCP2、p53、FMR1等），点击弹出详情面板。\n6. 粒子行为驱动动画、键盘快捷键（Space/←→/R）。\n7. 最终输出为单文件HTML，所有JS/CSS内联，支持file://离线打开。", "prompt_en": "Build an interactive methylation mechanism visualization webpage from the following paper PDF:\n\nPaper: Liu et al., 2023, Signal Transduction and Targeted Therapy 8:310, doi:10.1038/s41392-023-01528-y\nTitle: Methylation across the central dogma in health and diseases: new therapeutic strategies\n\nRequirements:\n1. Build a Three.js-based 3D interactive scene with cell membrane (translucent rounded box), nucleus (with chromatin/histones/nucleolus), mitochondria, and endoplasmic reticulum.\n2. Use particle systems to separately display DNA methylation (blue, nuclear), RNA methylation (purple, cytoplasmic), protein methylation (cyan, cytoplasmic), and SAM methyl donor (gold).\n3. Top-right HUD panel showing 12 real-time metrics: DNA methylation, RNA methylation, protein methylation, SAM level, CpG island hypermethylation, global hypomethylation, H3K4me3/H3K27me3/H3K9me3/H3K36me3, disease progression, therapy response.\n4. Bottom stage buttons for 6 stages: Normal methylation homeostasis → Early epigenetic perturbation → Cancer methylation remodeling → Neurological disorders → Aging-related methylation changes → Therapeutic intervention. Smooth parameter interpolation between stages.\n5. 3D labels for key molecules (DNMT1/3A/3B, TET1/2, METTL3/14, FTO, ALKBH5, EZH2, SUV39H1, SETD2, LSD1, MeCP2, p53, FMR1, etc.), click to open detail panels.\n6. Behavior-driven particle animations, keyboard shortcuts (Space/←→/R).\n7. Final output as single-file HTML, all JS/CSS inlined, supports offline file:// opening.", "skills": [ { "id": "paper-review-mechanism-viewer", "name": "论文机制交互可视化", "description": "将综述/研究论文PDF自动构建为交互式3D科研机制可视化网页", "skill_id": "skill_2085355597942149120" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-04T03:22:56Z", "updated_at": "2026-08-04T03:22:56Z", "is_recommended": false, "recommended_order": 0 }, { "id": "partnership-agreement", "title": "创业项目合伙协议", "title_en": "Startup Partnership Agreement", "subtitle": "把出资比例、决策机制与退出规则前置", "subtitle_en": "Set investment shares, decision rules and exit terms upfront.", "description": "合伙创业最容易在出资、决策权限和退出结算上留下口头约定，后续成本很高。\n\n本案例演示腾讯电子签合同法务专家如何将三方数字化咨询项目的出资、治理、利润分配、知识产权、竞业与退伙规则生成可编辑的《创业项目合伙协议》。\n\n适合：联合创始人、创业团队、项目合伙人、早期投资协作方。", "description_en": "Startup partnerships often leave capital contributions, decision authority and exit settlement as verbal understandings, creating high downstream costs.\n\nThis case shows how the Tencent eSign Contract Legal Expert turns a three-party digital consulting venture into an editable Startup Partnership Agreement covering capital, governance, profit sharing, IP, non-compete and exit terms.\n\nSuitable for: co-founders, startup teams, project partners and early-stage investment collaborators.", "categories": [ "business" ], "tags": [ "合伙协议", "创业项目", "出资比例", "退出机制" ], "tags_en": [ "Partnership Agreement", "Startup", "Capital Contributions", "Exit Mechanism" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请起草一份创业项目合伙协议。合伙人包括[合伙人名称]，共同开展[项目名称]；首期出资总额[金额]元，各方出资额和比例为[出资比例]。请明确执行事务合伙人、重大事项表决门槛、利润与亏损分配、知识产权归属、竞业限制、份额转让、退伙与清算、违约责任和争议解决。请输出可直接签署的中文 Word 协议，重点避免出资与退出争议。", "prompt_en": "Please draft a startup partnership agreement. The partners are [合伙人名称] and jointly operate [项目名称]. Initial total capital is [金额] RMB, with contributions and percentages of [出资比例]. Clearly define the managing partner, voting threshold for material matters, profit and loss allocation, IP ownership, non-compete, equity transfers, withdrawal and liquidation, breach liability and dispute resolution. Produce a signable Chinese Word agreement that minimizes capital and exit disputes.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-11T21:04:09+08:00", "updated_at": "2026-08-12T13:37:31+08:00", "artifacts": [ { "file": "output.html", "label": "V3.0 互动合同起草页面" }, { "file": "创业项目合伙协议.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "patent-claim-authoring-review-1", "title": "批任务调度架构 — 权利要求书起草", "title_en": "Batch Job Scheduler — Patent Claim Drafting", "subtitle": "1 条独立方法权利要求 + 10 条从属 + 覆盖对照表 + 待确认清单", "subtitle_en": "1 independent method claim + 10 dependent claims + cross-reference matrix + open-question checklist", "description": "最头疼的就是手写软件方法类专利的权利要求——既要覆盖核心创新点，又要每一条都能在交底书里找到依据，遇到文档没写清楚的地方要么硬补要么留漏洞。\n\n这个案例把一份公开的「批任务调度架构」技术说明（GitHub raw, MIT 许可, 2.1KB 偏架构设计而非标准交底书）作为唯一事实来源，按 patent-claim-authoring-review skill 的「起草分支」流程跑通一遍：自动判断走起草分支、严格贴着交底书文字撰写、文档未提及的具体算法/接口名保持上位写法并显式列入待确认事项；最终产出 1 条独立方法权利要求（覆盖资源画像聚合+匹配打分+限频重排队+派发四步全链路） + 10 条从属权利要求（含滑动窗口平滑、间隔与分数变化双阈值、心跳契约假死检测、亲和性等可追溯细节），并配套「权利要求↔交底书依据」覆盖对照表与「待确认事项」清单（分 P0/P1/P2 三级优先级）。\n\n适合：专利代理人（处理偏架构/非标准格式的交底书时）；技术发明人（需要快速从零散的设计说明中提炼权利要求骨架时）；法务/合规岗（需要审查权利要求是否过度推测、是否有交底书支撑时）。", "description_en": "Drafting software-method patent claims is painful: you need to cover the core invention while ensuring every single feature traces back to the disclosure — and when the disclosure is silent, you either over-imagine or leave loopholes.\n\nThis case takes a public \"batch job scheduler architecture\" note (GitHub raw, MIT-licensed, ~2.1KB, architecture-flavored rather than a standard disclosure) as the sole source of truth, and runs the full \"drafting branch\" workflow of the patent-claim-authoring-review skill: auto-determine the branch, write strictly against the document text, keep specific algorithm/interface names as upper-level wording or explicitly list them as open questions when the document is silent. Final outputs: 1 independent method claim (covering the full four-step pipeline: resource profile aggregation → matching scoring → rate-limited requeue → dispatch) + 10 dependent claims (including sliding-window smoothing, dual-threshold rate-limited requeue, heartbeat-contract death-detection, affinity, and other traceable details) + a \"claim ↔ disclosure\" cross-reference matrix + a P0/P1/P2 prioritized open-question checklist.\n\nBest for: patent attorneys handling architecture-flavored / non-standard disclosures; technical inventors who need to quickly extract a claim skeleton from scattered design notes; legal/compliance reviewers who need to audit whether claims are over-speculative or properly grounded.", "categories": [ "productivity" ], "tags": [ "专利撰写", "权利要求", "软件方法", "批任务调度", "交底书" ], "tags_en": [ "patent drafting", "claims", "software method", "batch job scheduler", "disclosure" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "单文件 HTML 展示页（含核心流程 SVG + 权利要求全文 + 覆盖对照表 + 待确认清单 + 复现指引）" } ], "prompt": "请按 patent-claim-authoring-review skill 的「起草分支」流程，依据以下公开技术说明文档起草一份软件方法类发明专利的权利要求书：\n\n- 文档 URL：https://raw.githubusercontent.com/handsomestWei/patent-disclosure-skill/main/examples/example_batch_job_scheduler/knowledge/docs/architecture.md\n- 文档许可：MIT（GitHub raw，可直接下载）\n- 作者声明：本文档仅供本仓库专利技能示例使用，不对应任何真实产品（虚构示例）\n\n工作流要求：\n1. 先判定走「起草分支」，并先下载/阅读 architecture.md 全文与同目录口径一致的 sample_architecture_review.docx、sample_scheduler_deck.pptx（docx/pptx 内嵌的抽象模块/队列示意 PNG 不补充新特征，仅作交底书附图参考）。\n2. 写 1 条独立方法权利要求，覆盖「资源画像聚合 + 匹配打分 + 限频重排队 + 派发」核心流程；并把 §2 痛点（异构集群大任务占满高性能节点/小任务长尾等待、静态优先级无法反映实时画像）与 §3 目标（可接受分钟级调度延迟、避免秒级频繁重排导致震荡）作为限定条件融入「限频重排队」步骤的「间隔阈值 + 分数变化阈值」两级限定。\n3. 依据文档可追溯细节逐条续写从属权利要求，至少覆盖：多维资源指标（CPU/内存/IO）、滑动窗口平滑（具体算法作并列上位写法）、任务需求向量×节点画像匹配打分（具体运算式作并列上位写法）、限频重排队两级阈值（分钟级时间间隔+相对/绝对分数变化量，AND 同时满足）、仅对队列窗口子集重排（非全量）、静态优先级入队+基于匹配分数二次排序、失败重试（交底书写「略」，具体策略作推测并列入待确认）、心跳契约假死检测与任务迁移触发（对应交底书 §4 表格明文标注的「从属特征」）、任务迁移的实现方式（重新入队/重新选择节点并列）、亲和性标签与异构节点属性维度（算力/IO/存储/地域，后者为推测）、饥饿缓解+分钟级延迟。\n4. 对文档未提及的任何具体算法/接口名/数值，保持上位写法或列入待确认事项；凡是基于通用技术原理的合理推测必须显式标注；不得替发明人脑补技术细节。\n5. 一并给出：(a) 权利要求书工作稿 Markdown（含每条权利要求的起草思路）；(b) 「权利要求 ↔ 交底书依据」覆盖对照表（每项特征标注 ✅ 直接 / 🟡 可追溯 / ⚠️ 推测 + 原文引用位置）；(c) 待确认事项清单（按 P0 撰写前必确认 / P1 影响从属 / P2 暂不影响 分级）。\n6. 全套终稿（3 份 Markdown）一次性写到 ./output/ 目录。\n\n真实数据/来源声明：交底书 architecture.md 为公开 MIT 许可开源示例材料，URL 已在上面给出；草拟过程中所有技术特征都需回到该文档找原文依据；不得使用任何「示意/模拟/合成的虚假技术数据」代替交底书内容。", "prompt_en": "Following the \"drafting branch\" workflow of the patent-claim-authoring-review skill, draft a software-method invention patent claim set based on the following public technical note:\n\n- Document URL: https://raw.githubusercontent.com/handsomestWei/patent-disclosure-skill/main/examples/example_batch_job_scheduler/knowledge/docs/architecture.md\n- License: MIT (GitHub raw, directly downloadable)\n- Author's notice: This document is a fictional example for this repository's patent-skill demo and does NOT correspond to any real product.\n\nWorkflow requirements:\n1. First determine to take the \"drafting branch\"; download and read architecture.md in full, plus the same-directory sample_architecture_review.docx and sample_scheduler_deck.pptx (their embedded module/queue diagram PNGs are abstract and add NO new features — use them only as disclosure-figure references).\n2. Write 1 independent method claim covering the core flow: \"resource profile aggregation + matching scoring + rate-limited requeue + dispatch\". Fold §2 pain points (heterogeneous cluster: large tasks monopolize high-perf nodes, small tasks long-tail waiting; static priority cannot reflect real-time profile) and §3 goals (accept minute-level dispatch latency, avoid second-level requeue thrashing) into the \"rate-limited requeue\" step as a dual-threshold (time interval + score change) limitation.\n3. Sequentially draft dependent claims based on traceable details from the document. Cover at least: multi-dimensional resource metrics (CPU/memory/IO); sliding-window smoothing (specific algorithm as parallel upper-level wording); task-demand vector × node-profile matching (specific formula as parallel upper-level wording); rate-limited requeue dual-threshold (minute-level time interval + relative/absolute score change, AND-gated); queue-window subset requeue (not full requeue); static-priority enqueue + match-score based secondary sorting; failure retry (the document writes \"(略)\", treat specific strategy as speculation and list as open question); heartbeat-contract death-detection and task migration trigger (corresponding to the \"从属特征\" row in §4 table); migration implementation (re-enqueue / re-select-and-dispatch as parallel); affinity tags and heterogeneous-node attribute dimensions (compute/IO/storage/region — region is speculation); starvation mitigation + minute-level latency.\n4. For any specific algorithm/interface/numeric value not mentioned in the document, keep upper-level wording OR list as an open question; any reasonable speculation based on common technical principles MUST be explicitly marked; never fabricate technical details for the inventor.\n5. Also produce: (a) claim-set working draft in Markdown (with drafting rationale per claim); (b) \"claim ↔ disclosure\" cross-reference matrix (mark each feature as ✅ direct / 🟡 traceable / ⚠️ speculation + source location); (c) open-question checklist (prioritize as P0 must-confirm before drafting / P1 affects dependents / P2 does not affect claims yet).\n6. Write all final deliverables (3 Markdown files) into ./output/ in one batch.\n\nData-source declaration: the disclosure document architecture.md is public, MIT-licensed, open-source example material (URL above); every technical feature in the draft must trace back to the document; do NOT use any \"示意/模拟/synthesized fake technical data\" in place of disclosure content.", "skills": [ { "id": "patent-claim-authoring-review", "name": "专利权利要求起草与审查", "description": "依据技术交底书与附图起草、续写并审查专利申请的权利要求", "skill_id": "skill_2090718524817666048" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-21T14:34:28+08:00", "updated_at": "2026-08-21T14:34:28+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "patent-claim-authoring-review-2", "title": "注塑机锁模力自适应调节专利权利要求书", "title_en": "Patent Claims for Adaptive Injection Molding Clamping Force", "subtitle": "从技术交底书起草 1 条独权 + 6 条从权，逐条标注出处与起草思路", "subtitle_en": "One independent plus six dependent claims drafted from a technical disclosure, each with source and rationale", "description": "写专利权利要求最怕两件事：一是漏掉交底书里的关键细节，二是替发明人脑补交底书没写的内容。这个案例演示了如何把一份《注塑机锁模力自适应调节_技术交底书》交给 AI，起草出 1 条独立方法权利要求加 6 条从属权利要求的完整权利要求书，每条都标注交底书出处与起草思路，并配套「权利要求 ↔ 交底书依据」覆盖对照表和 8 项待确认事项清单，温度补偿公式、采样周期、阈值等细节全部溯源到原文。\n\n适合：专利代理人、企业 IPR、发明人，以及一切需要从技术交底书准备专利申请文件的场景。", "description_en": "Drafting patent claims is scary in two ways: missing key details in the technical disclosure, or inventing details the disclosure never mentions. This case shows how to hand a technical disclosure on adaptive clamping force control to AI and get a complete claim set — one independent method claim plus six dependent claims — with each claim sourced back to the disclosure, together with a claim-to-disclosure coverage table and a list of 8 items to confirm with the inventor. Every detail, from the temperature-compensation formula to the sampling period and thresholds, is traceable to the source text.\n\nBest for: patent attorneys, corporate IPR staff, inventors, and anyone preparing patent application documents from a technical disclosure.", "categories": [ "productivity" ], "tags": [ "专利", "权利要求", "技术交底书", "专利撰写", "知识产权" ], "tags_en": [ "patent", "claims", "technical disclosure", "patent drafting", "intellectual property" ], "artifact_type": "html", "preview": "output.html", "prompt": "请根据附件《注塑机锁模力自适应调节_技术交底书.md》，为我起草一份完整的专利申请权利要求书，把全套终稿一次性写入指定输出目录（默认 ./output/）：\n\n1. 先写 1 条独立方法权利要求，再依据交底书「进一步地」部分逐条续写从属权利要求，编号与引用关系清晰；\n2. 给出每条从属权利要求对应的交底书出处与起草思路；\n3. 输出三件套：权利要求书工作稿、权利要求↔交底书依据覆盖对照表、待确认事项清单；\n4. 严格以交底书为唯一事实来源：交底书未提及的技术细节保持上位写法或列入待确认事项，不要替发明人脑补；交底书中的采样周期、阈值等示意值按原样写入并在待确认清单中标注；\n5. 交底书中含糊处（如偏差值计算口径）如实标注为待确认，不臆断。\n\n附件：[粘贴/上传技术交底书 Markdown 全文]", "prompt_en": "Based on the attached technical disclosure document 《注塑机锁模力自适应调节_技术交底书.md》, draft a complete set of patent claims and write all final deliverables to the designated output directory (default ./output/):\n\n1. Write 1 independent method claim first, then continue with dependent claims one by one following the \"further\" (进一步地) sections of the disclosure, with clear numbering and reference relationships;\n2. Provide the disclosure source and drafting rationale for each dependent claim;\n3. Deliver three documents: the claims working draft, a claim-to-disclosure coverage mapping table, and a list of items to confirm with the inventor;\n4. Treat the technical disclosure as the sole source of truth: keep unmentioned technical details at a higher level or list them as to-confirm items; do not invent details on behalf of the inventor; keep sample values such as sampling period and thresholds as written and flag them in the to-confirm list;\n5. Where the disclosure is ambiguous (e.g. how the deviation value is computed), mark it honestly as to-confirm rather than guessing.\n\nAttachment: [paste/upload the full technical disclosure in Markdown]", "skills": [ { "id": "patent-claim-authoring-review", "name": "专利申请权利要求起草与审查", "description": "依据技术交底书与附图起草、续写并审查专利申请的权利要求", "skill_id": "skill_2090718524817666048" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 96, "source": "official", "artifacts": [ { "file": "output.html", "label": "权利要求书 HTML 版（案例预览）" } ], "created_at": "2026-08-21T06:41:18Z", "updated_at": "2026-08-21T06:50:00Z", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "patent-claim-authoring-review-3", "title": "权利要求技术细节核实", "title_en": "Patent Claim Detail Verification", "subtitle": "核实从属项「步进电机驱动丝杆」限定是否有交底书依据，区分原文与推测并给出修订建议", "subtitle_en": "Verify whether the \"stepper motor & lead screw\" limitation is supported by the disclosure, separate facts from speculation, and advise on claim revision", "description": "写专利权利要求最怕的就是把交底书里没有的细节写进去——从属项限定得越具体，越容易因缺乏依据被质疑，保护范围也越收越窄。这个案例演示了如何把一条「步进电机驱动丝杆微调夹持开度」的从属项限定，对照交底书逐层核实：先检索正文原文依据，再从 docx 二进制检查嵌入附图（本例确无图片，如实标注「未提取到附图」），最后把「原文结论」与「基于通用技术原理的合理推测」明确分开，给出「保持上位写法」还是「列入待确认事项」的处理建议，全程不替发明人脑补技术细节。适合：专利代理人、企业 IPR、发明人，以及一切需要核对权利要求与交底书一致性的场景。", "description_en": "The worst fear when drafting patent claims is writing in details that the invention disclosure never mentions — the more specific a dependent claim gets, the more it can be challenged for lack of support, and the narrower the protection becomes. This case shows how to take a dependent-claim limitation like \"stepper motor driving a lead screw to fine-tune the clamping opening\" and verify it layer by layer against the disclosure: first search the original text for supporting evidence, then inspect the docx binary for embedded figures (none in this example, honestly marked as \"no embedded image extracted\"), and finally separate \"original-text findings\" from \"reasonable speculation based on general technical principles\", with a clear recommendation to keep the broader claim language or list it as a to-be-confirmed item — never inventing technical details on behalf of the inventor. Suitable for: patent attorneys, corporate IPR staff, inventors, and anyone who needs to check consistency between claims and the disclosure.", "categories": [ "productivity" ], "tags": [ "专利", "权利要求", "交底书核实", "技术细节", "知识产权" ], "tags_en": [ "Patent", "Claims", "Disclosure Verification", "Technical Detail", "IP" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请对照《静脉输液滴速控制_交底书.docx》核实一个技术细节：权利要求的从属项想限定「步骤S3 中夹持机构的调节方式是『步进电机驱动丝杆微调夹持开度』」，请帮我确认这一限定在交底书文字里是否有依据；如果文字不足以说明，请尝试从 docx 二进制中提取嵌入的附图（图1）进一步核实（若 docx 中确无嵌入图片，请如实说明「未提取到附图」）。请明确答复：交底书是否有该依据；若没有，请把「原文结论」与「基于通用技术原理的合理推测」分开写清，并给出这条从属权利要求该怎么处理的建议（保持上位写法 or 列入待确认事项）。请把核实结论与修订建议输出为一份 Markdown 报告 + 一份单文件 HTML 报告（样式简洁专业、资源全内联），HTML 为最终展示主产物。", "prompt_en": "Verify a technical detail against the disclosure document 静脉输液滴速控制_交底书.docx: a dependent claim wants to limit that \"in step S3, the clamping mechanism is adjusted by a stepper motor driving a lead screw to fine-tune the clamping opening\". Please confirm whether this limitation has any basis in the disclosure text; if the text is insufficient, try to extract the embedded figure (Fig. 1) from the docx binary for further verification (if the docx truly contains no embedded image, honestly state \"no embedded image extracted\"). Give a clear answer on whether the disclosure supports this limitation; if not, clearly separate the \"original-text findings\" from \"reasonable speculation based on general technical principles\", and advise how to handle this dependent claim (keep the broader wording or list it as a to-be-confirmed item). Output the verification conclusion and revision advice as a Markdown report plus a single-file HTML report (clean professional style, all resources inlined), with the HTML as the final showcase artifact.", "artifacts": [ { "file": "output.html", "label": "权利要求技术细节核实报告（HTML 展示版）" } ], "skills": [ { "id": "patent-claim-authoring-review", "name": "专利申请权利要求起草与审查", "description": "依据技术交底书与附图起草、续写并审查专利申请的权利要求", "skill_id": "skill_2090718524817666048" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-08-21T14:35:00+08:00", "updated_at": "2026-08-21T14:42:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "patient-edu-six-dimension-review", "title": "六维度患教内容审核", "title_en": "Six-Dimension Patient Education Content Review", "subtitle": "粘贴患教正文，一键输出分级审核意见", "subtitle_en": "Paste patient education content, get graded review comments in one pass", "description": "医学科普内容发布前最怕的就是漏审——数据和文献原文差一个百分点、把复合终点写成单一终点、推介药物却没写安全性提示，这些问题人工逐条核对一篇要花半天，还容易看漏。\n\n这个案例把一篇克罗恩病患教文章交给 AI 做六维度全量审核，审核意见直接标回原文。产物顶部是审核结论（8 项需修改、7 处标注建议优化、六个维度各自结论），下方左栏是文章原文、问题句带底色高亮和序号，右栏是逐条审核意见，点高亮能定位到对应意见、点意见能跳回原文，文末再按需修改与建议优化分组展开全部 18 条明细。每条意见都指到具体段落和原文——比如查出文中「47%-55% 维持临床缓解」实为文献 Table 3 的复合终点（第12周应答且第48周缓解），「34%-37% 内镜缓解」的下限在 Figure 4 中实为 33%，临床缓解标准把原文「每日稀便≤3次」误写为「每周较基线增加≤3次」。\n\n适合：医学内容编辑、药企市场与医学事务人员、患教内容审核员、健康科普自媒体运营、医学编辑团队负责人。", "description_en": "The biggest fear before publishing medical content is a missed check — a percentage point that diverges from the cited paper, a composite endpoint written up as a single endpoint, a drug promoted without any safety notice. Verifying one article by hand takes half a day and things still slip through.\n\nThis case hands a Crohn's disease patient education article to AI for a full six-dimension review, writing comments back onto the original text. The deliverable opens with the verdict (8 must-fix items, 7 highlighted suggestions, and a conclusion for each of the six dimensions); below it, the article sits on the left with problem sentences highlighted and numbered, and the review comments on the right — click a highlight to jump to its comment, click a comment to jump back to the text — followed by all 18 findings expanded at the end, grouped into must-fix and suggested improvements. Every comment points to a specific paragraph and the original wording — for example, it found that \"47%-55% maintained clinical remission\" actually corresponds to the composite endpoint in Table 3 (week-12 response AND week-48 remission), that the lower bound of \"34%-37% endoscopic remission\" is in fact 33% in Figure 4, and that the clinical remission criterion misstated the source's \"daily stool frequency ≤3\" as \"weekly increase of ≤3 over baseline\".\n\nSuitable for: medical content editors, pharma marketing and medical affairs staff, patient education reviewers, health science communication operators, and medical editorial team leads.", "categories": [ "education", "business" ], "tags": [ "患教审核", "医学科普", "内容合规", "文献核查", "医学编辑" ], "tags_en": [ "Patient Education Review", "Medical Communication", "Content Compliance", "Reference Verification", "Medical Editing" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "审核报告网页版（头部结论 + 原文标注 + 批注联动）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是医学科普内容审核助手，请对我提供的患教文章做一次完整的六维度审核，并把审核意见标注回原文，输出为一份可直接打开的审核报告网页（不要分多轮追问，一次输出到底）。\n\n待审核内容：{粘贴患教文章正文，含参考文献列表}\n\n请按以下六个维度依次审核：\n1. 严重问题：政治法律红线、低俗内容、医疗高风险、商业推广倾向、版权风险\n2. 科学性：医学概念与数据准确性，逐项核对文中数据与所引参考文献原文是否一致（含终点定义、分析集、单位、数值区间），检查绝对化与夸大表述\n3. 出版规范：标题层级与编号连续性、段落样式一致性、标点符号、图表质量与引用标注位置、参考文献著录格式\n4. 语言风格：面向患者的可读性，专业术语与英文缩写首次出现是否给出全称和通俗解释\n5. 医学逻辑：因果关系是否成立、数据能否支持结论、推荐方案与指南是否一致、关键安全信息（不良反应/禁忌/用药前筛查）是否缺失\n6. 参考文献：来源级别、时效性、著录格式、与正文标注体系是否对应、具体数据与观点是否标注引用\n\n网页需包含以下四部分：\n- 头部审核结论：需修改与建议优化的总数，以及六个维度各自的结论概述\n- 文章原文：完整保留原文，有问题的句子用底色高亮并编号（红色=需修改，黄色=建议优化）\n- 审核意见栏：与原文并排，逐条给出严重程度、所属维度、问题描述、修改建议\n- 文末明细：按「需修改」「建议优化」分组，逐条展开完整意见\n\n交互与呈现要求：\n- 点击正文高亮可定位到对应审核意见，点击意见可跳回原文位置\n- 文中的评分量表等表格请还原为网页表格，不要用截图，保证文字清晰可选中\n- 单文件 HTML，CSS/JS/图片等资源全部内联，可离线打开\n\n审核意见撰写要求：\n- 涉及数据与文献不一致时，必须写明文献中的原始位置（如 Table 3 / Figure 4）与原始数值，并给出按原文改写后的建议表述\n- 每条修改建议须具体可执行，不要只说「建议核实」\n- 保留原文全部内容，只做标注，不要改写正文\n- 只发现问题、给修改建议，不做「可以发布」的发布决策", "prompt_en": "You are a medical science communication content reviewer. Please run a complete six-dimension review on the patient education article I provide, write the review comments back onto the original text, and output a ready-to-open review report web page (do not split this across multiple rounds of questions — produce it in one pass).\n\nContent to review: {paste the patient education article, including the reference list}\n\nReview across the following six dimensions in order:\n1. Serious issues: political/legal red lines, vulgar content, high medical risk, commercial promotion bias, copyright risk\n2. Scientific accuracy: accuracy of medical concepts and data — verify item by item whether figures in the text match the cited source (including endpoint definitions, analysis sets, units, and value ranges); check for absolute or exaggerated claims\n3. Publishing standards: heading hierarchy and numbering continuity, paragraph style consistency, punctuation, table/figure quality and citation placement, reference formatting\n4. Language style: readability for patients; whether technical terms and English abbreviations are given full names and plain-language explanations on first use\n5. Medical logic: whether causal claims hold, whether data supports the conclusions, whether recommendations align with guidelines, and whether critical safety information (adverse reactions / contraindications / pre-treatment screening) is missing\n6. References: source level, timeliness, citation format, correspondence with in-text citation markers, and whether specific data and claims carry citations\n\nThe page must contain these four parts:\n- Header verdict: total counts of must-fix and suggested items, plus a summary conclusion for each of the six dimensions\n- Article text: the full original text preserved, with problematic sentences highlighted and numbered (red = must fix, amber = suggested improvement)\n- Review comment pane: side by side with the text, listing severity, dimension, problem description, and recommendation one by one\n- Detailed findings at the end: grouped into \"Must fix\" and \"Suggested improvements\", each expanded in full\n\nInteraction and presentation requirements:\n- Clicking a highlight in the text jumps to its corresponding comment; clicking a comment jumps back to its position in the text\n- Render scoring scales and other tables as real HTML tables rather than screenshots, so the text stays sharp and selectable\n- A single-file HTML with all CSS/JS/image assets inlined, openable offline\n\nRequirements for the comments themselves:\n- When data diverges from the cited source, state the exact location in the source (e.g. Table 3 / Figure 4) and the original value, and provide suggested wording rewritten to match the source\n- Every recommendation must be specific and actionable — do not merely say \"please verify\"\n- Preserve all original content; only annotate — do not rewrite the body text\n- Only identify problems and give revision advice; do not make any \"ready to publish\" decision", "skills": [], "experts": [ { "id": "patient-education-content-review-word-assistant", "name": "腾讯健康药箱患教审核助手", "description": "六维度审核患教内容并核对数据文献一致性，审核意见直接标注在原文上输出。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-19T02:35:46Z", "updated_at": "2026-08-20T20:22:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "patient-registration", "title": "导诊挂号", "title_en": "Triage & Registration", "subtitle": "看病第一步不卡壳：帮你选对科室 + 附近医院推荐 + 就诊全流程须知 + 权威专家，一页理清", "subtitle_en": "Never stuck at the first step of seeing a doctor: the right department, nearby hospitals, full visit guidance, and top specialists in one page", "description": "人生病时最常见的卡壳，往往不是病本身，而是“第一步不知道往哪迈”：该挂什么科、去哪家医院、找哪位医生——信息不对称让不少人挂错号、跑错科室，耽误了时间也消磨了耐心。\n\n以「打喷嚏、流鼻涕、夜间鼻塞」为例：把症状说给微小医听，它会一步步帮你理清该挂什么科，推荐杭州 5 家附近的医院（等级、地址、特色都写明），讲清就诊前中后要留意的事，还附全国耳鼻喉权威专家的推荐——一页理清，存图带着去看病，少走冤枉路。", "description_en": "When people get sick, the most common stumbling block isn't the illness itself — it's not knowing where to take the first step: which department, which hospital, which doctor. That information gap leads many to the wrong registration desk, wasting both time and patience.\n\nTake \"sneezing, runny nose, nighttime congestion\" as an example: tell Weixiaoyi your symptoms, and it gently walks you to the right department, recommends 5 nearby hospitals in Hangzhou with their grades, addresses and strengths, explains what to keep in mind before, during and after the visit, and adds top national ENT specialists — all on one page you can save and bring along, so you don't waste a single trip.", "categories": [ "business" ], "tags": [ "导诊挂号", "耳鼻喉科", "过敏性鼻炎", "微医", "HTML生成" ], "tags_en": [ "Triage & Registration", "ENT", "Allergic Rhinitis", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "patient-registration.html", "artifacts": [ { "file": "patient-registration.html", "label": "鼻部不适导诊挂号指南 HTML 页面" } ], "prompt": "我最近一直打喷嚏流鼻涕，晚上睡觉鼻子都是堵的。我该看什么科室？离我比较近的医院有哪些？就诊有什么注意事项？全国范围比较权威的医生有哪些，我想了解医生尽量详细的信息。将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203），logo裁成圆形。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果；5. 重点内容用流程图、指标卡片等特殊样式进行说明。", "prompt_en": "I've been sneezing with a runny nose lately, and my nose is completely blocked when I sleep at night. Which department should I see? Which hospitals are nearby? What should I pay attention to during the visit? Who are the most authoritative doctors nationwide — I'd like detailed information about them. Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for the sharing scenario; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203), with the logo cropped into a circle. HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation; 5. Key content should be illustrated with special styles such as flowcharts and metric cards.", "skills": [ { "id": "we-health-skill", "name": "微医健康管家", "description": "微医AI健康管家（官方Skill），覆盖报告解读、就医推荐、用药咨询、健康咨询等全部医疗健康场景", "slug": "we-health-skill" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-20T18:10:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "ped-chronic-drug-edu-review", "title": "慢病用药患教长文六维审核", "title_en": "Six-Dimension Review of Chronic-Disease Medication Patient-Education Articles", "subtitle": "把一篇糖尿病患教长文交给专家，按六维度逐段审核输出结构化报告", "subtitle_en": "Hand a long diabetes patient-education article to the assistant; get a structured six-dimension review report with issue-by-issue findings and revision suggestions", "description": "私域里发的糖尿病、高血压患教长文，最怕科学性出错——用药剂量、禁忌、停药说法一旦写错，患者照做就有风险，运营也担责任。\n\n本案例把一篇 2800 字的糖尿病患教长文交给「腾讯健康药箱-私域患教内容审核助手」，它按「科学性准确性 / 引用规范性 / 医学逻辑性 / 数据来源可靠性 / 特殊人群覆盖 / 合规风险」六维度逐段审核，输出结构化 Markdown 审核报告：每处问题标注位置、类型、风险等级（Critical/High/Medium/Low）、参考依据（NMPA说明书/指南/法规）和修改建议，最终给出 Approved/Rejected 判定。\n\n适合: 药企/互联网医院私域运营、药箱社群内容编辑、慢病管理产品经理、医学编辑。", "description_en": "Patient-education long-reads on chronic diseases (diabetes, hypertension) published in private domains carry the highest risk of scientific errors — wrong dosage, missing contraindications, or incorrect stop-medication advice can directly harm patients and expose operators to liability.\n\nThis case hands a ~2,800-word diabetes patient-education article to the Tencent Health Pill-Box Private-Domain Patient-Education Content Review Assistant. It reviews paragraph by paragraph across six dimensions (Scientific Accuracy / Citation Norms / Medical Logic / Data Source Reliability / Special-Population Coverage / Compliance Risk), outputting a structured Markdown review report: each issue is tagged with location, type, risk level (Critical/High/Medium/Low), reference basis (NMPA package insert / guideline / regulation), and revision suggestion, ending with an Approved/Rejected verdict.\n\nSuitable for: private-domain operations at pharma companies / internet hospitals, content editors for pill-box communities, chronic-disease product managers, medical editors.", "categories": [ "business", "productivity" ], "tags": [ "患教审核", "慢病用药", "六维度审核", "合规审查" ], "tags_en": [ "Patient Education Review", "Chronic Medication", "Six-Dimension Review", "Compliance Review" ], "artifact_type": "html", "preview": "output.html", "prompt": "我们运营糖尿病患者社群，市场部写了篇患教长文准备推送。我不是医学背景，看不出用药说法有没有问题，但发出去几千人会照着做，出事我担不起。帮我逐条查一遍，指出哪里错了、多严重、怎么改。\n\n以下是文中几个我拿不准的段落：\n\n1. 「格泰平® 二甲双胍缓释片采用先进缓释技术，临床数据显示 85% 的患者三个月内血糖达标（MULTI-CARE 研究，2023）。」\n\n2. 「服用方法：每日一次随晚餐服用。如果当天忘记吃，第二天可以一次吃两片补上，保证疗程完整。」\n\n3. 「连续 3 个月空腹血糖低于 6.0、糖化血红蛋白低于 6.5%，说明胰岛功能已恢复，可以自行逐步减量直至停药。我们社群的张阿姨停药 8 个月血糖依然稳定，说明糖尿病是可以通过规范用药实现逆转的。」\n\n4. 「特殊人群：孕妇孕期禁用所有口服降糖药；老年患者代谢慢，剂量可自行酌情减半；轻中度肝损伤一般可继续正常服用。」\n\n5. 「饮食建议：主食换成粗粮可以放心多吃，粗粮不升血糖；标注『无糖』就代表不含糖，可以随意吃；水果尽量不吃。」\n\n6. 「运动建议：餐后立即运动效果最好，可以直接消耗刚吃进去的糖分。」\n\n麻烦按严重程度排序，明确区分必须改和建议改，最后给我一份能直接交回市场部的整改清单。", "prompt_en": "We run a diabetes patient community, and our marketing team drafted a patient-education long-read for distribution. I don't have a medical background — I can't tell whether the medication claims hold up, but thousands of people will follow them once published, and I can't carry that risk. Please review each item, tell me what's wrong, how serious it is, and how to fix it.\n\nHere are the passages I'm unsure about:\n\n1. \"Getaiping® Metformin ER Tablets use advanced extended-release technology; clinical data show 85% of patients reach glycemic targets within three months (MULTI-CARE study, 2023).\"\n\n2. \"Dosing: once daily with dinner. If you miss a dose, take two tablets the next day to make it up and keep the course complete.\"\n\n3. \"If fasting glucose stays below 6.0 and HbA1c below 6.5% for three consecutive months, islet function has recovered and you may taper down on your own until you stop. Aunt Zhang in our community has been off medication for eight months with stable glucose — proof that diabetes can be reversed through proper medication.\"\n\n4. \"Special populations: all oral hypoglycemics are contraindicated during pregnancy; elderly patients metabolize slowly so the dose may be halved at your own discretion; mild-to-moderate liver injury generally allows continued normal dosing.\"\n\n5. \"Diet: switch staples to whole grains and eat freely — whole grains don't raise blood glucose; anything labeled 'sugar-free' contains no sugar and can be eaten at will; avoid fruit entirely.\"\n\n6. \"Exercise: working out immediately after meals is most effective — it burns off the sugar you just ate.\"\n\nPlease rank issues by severity, clearly separate must-fix from nice-to-fix, and finish with a remediation checklist I can hand straight back to marketing.", "skills": [], "experts": [ { "id": "patient-education-content-review-word-assistant", "name": "腾讯健康药箱-私域患教内容审核助手", "description": "基于国家药监局权威数据源，按六维度（科学性/引用规范/医学逻辑/数据来源/特殊人群/合规风险）审核患教内容，输出结构化审核报告。能力来源于官方产品截图（六维度审核/药品说明书查询/参考文献P0-P5分级/科学性三问审核）。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "", "promptSourceNote": "官方 prompt 文件 HTTP 404 缺失；本次基于 9 张官方产品截图提取能力基线执行", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "source": "official", "usage_count": 0, "artifacts": [ { "file": "output.html", "label": "糖尿病患教长文六维审核报告（20项问题，Rejected）" }, { "file": "output.md", "label": "糖尿病患教长文六维审核报告（20项问题，Rejected）（Markdown 版）" } ], "cover_image": "cover.jpg", "cover_mode": "landscape", "cover_source": "官方参考图片 image0（六维度患教内容审核封面）", "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-31T12:04:40+08:00", "theme_label": "健康绿 · 慢病用药六维审核", "theme_color": "#0a9b6e", "is_recommended": false, "recommended_order": 0 }, { "id": "personal-ip-growth-hub", "title": "个人IP成长台", "title_en": "Personal IP Growth Hub", "subtitle": "把定位、作品、涨粉、变现和成长路线拧成一个手绘风工作台", "subtitle_en": "A hand-drawn workspace that ties your positioning, works, growth and monetization together", "description": "做自媒体最怕的就是「一腔热血开号，三个月后不知道自己在往哪走」——定位飘忽、作品散落各处、涨粉停滞也说不清卡在哪、变现更是一笔糊涂账。\n\n这个案例做出来的是一个手绘可爱风格的单文件网页工作台：把「我是谁（定位）」「我做了什么（作品库）」「涨了多少粉、互动怎么样（数据看板 + 手绘成长曲线）」「靠什么赚钱（变现记录）」「下一步往哪走（成长路线图）」全塞进一个页面。数据全存在浏览器本地，关掉再打开还在；PC 端是侧边栏、手机端自动变底部 Tab，可以直接「添加到主屏幕」当 App 用；还内置了中英双语，链接后面加 ?lang=en 就整页切英文。\n\n适合：刚起号想理清定位的自媒体新人、需要复盘涨粉数据的内容创作者、有多平台账号要统一管理的博主、想把副业IP做成长期资产的斜杠青年、以及带学员做个人品牌的运营教练。", "description_en": "The scariest part of building a personal brand online is starting with passion and, three months later, having no idea where you're heading — a fuzzy positioning, works scattered everywhere, stalled follower growth you can't explain, and monetization that's a total mess.\n\nThis case produces a single-file web workspace in a cute hand-drawn style: it packs everything into one page — who you are (positioning), what you've made (works library), how much you've grown and how engagement looks (a dashboard plus a hand-drawn growth curve), how you make money (monetization log), and where to go next (a growth roadmap). All data lives in your browser's local storage, so it's still there after you close and reopen; on desktop it's a sidebar, on mobile it turns into a bottom tab bar you can \"add to home screen\" and use like an app. It also ships with built-in bilingual support — just append ?lang=en to the URL to flip the whole page to English.\n\nGreat for: creators who just launched an account and want to nail their positioning, content makers who need to review follower data, multi-platform bloggers who want one place to manage it all, slashies turning a side-hustle IP into a long-term asset, and coaches guiding students through personal branding.", "categories": [ "workspace", "business", "creation" ], "tags": [ "个人IP", "自媒体成长", "涨粉复盘", "内容变现", "工作台" ], "tags_en": [ "Personal IP", "Creator growth", "Follower review", "Content monetization", "Workspace" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "个人IP成长台" } ], "preview_en": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/GfhrA29pYRrt1I6k9Qx17b），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个「个人IP成长台」——给做自媒体/个人品牌的人用的单文件网页工作台，直接用浏览器打开就能用，数据存本地浏览器（localStorage），关掉再打开不丢。\n\n要包含这几个模块：\n1. 首页概览：今日激励语 + 核心数据卡片（总粉丝、总作品、本月涨粉、变现总额），一眼看清现状；\n2. 我的定位：记录账号定位、目标人群、内容方向、一句话人设；\n3. 作品库：登记发过的内容（标题/平台/发布日期/数据表现/链接），能增删改；\n4. 涨粉复盘：多平台粉丝数据看板 + 用内联 SVG 手绘一条粉丝成长曲线，能编辑每月数据点；\n5. 变现记录：记录每一笔收入（来源/金额/日期/备注），自动汇总；\n6. 成长路线图：分阶段的成长目标与打卡，标记完成进度。\n\n设计要求：整体走「手绘可爱风」——不规则圆角、粗墨描边、贴纸式硬阴影、卡片轻微倾斜、有弹性/摇摆的小动画，配色明亮活泼但不刺眼；PC 端用侧边栏导航，手机端自动切换成底部 Tab，触控友好，可添加到手机主屏当 App。\n\n技术约束：单个 HTML 文件、CSS/JS 全部内联、不依赖任何外部 CDN 或联网资源、可完全离线打开；提供浅色/深色主题切换。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/GfhrA29pYRrt1I6k9Qx17b) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nHelp me build a \"Personal IP Growth Hub\" — a single-file web workspace for people building a personal brand or doing content creation. It should run by simply opening it in a browser, store data locally in the browser (localStorage), and keep the data after closing and reopening. Include these modules:\n1. Home overview: a daily motivational line plus core metric cards (total followers, total works, followers gained this month, total earnings), so the current state is clear at a glance;\n2. My positioning: record the account positioning, target audience, content direction, and a one-line persona;\n3. Works library: log published content (title / platform / publish date / performance / link), with add/edit/delete;\n4. Growth review: a multi-platform follower dashboard plus a hand-drawn follower growth curve using inline SVG, with editable monthly data points;\n5. Monetization log: record every income entry (source / amount / date / note) with automatic totals;\n6. Growth roadmap: staged growth goals with check-ins and completion progress. Design requirements: an overall \"cute hand-drawn\" style — irregular rounded corners, thick ink strokes, sticker-like hard shadows, slightly tilted cards, springy/wobbly micro-animations, and a bright, lively but not harsh color scheme; use a sidebar for navigation on desktop, and automatically switch to a bottom tab bar on mobile, touch-friendly and addable to the home screen as an app. Technical constraints: a single HTML file no dependency on any external CDN or online resources, fully openable offline; provide light/dark theme switching.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-08T11:20:00+08:00", "updated_at": "2026-07-08T11:20:00+08:00", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "personal-office-workbench", "title": "个人办公工作台", "title_en": "Personal Office Workbench", "subtitle": "待办、灵感、项目跟进、周复盘四合一，数据存本地，断网可用", "subtitle_en": "Tasks, notes, project tracking and weekly review in one offline-first single-file app", "description": "待办清单换了七八个，还是没坚持下来。要么昨天没做完的第二天就消失了，要么项目和任务是两张互不相干的表——项目卡在哪儿只能靠脑子记，到了周末想复盘，连这周做完几件事都数不清。想找个简单点的工具，结果都要注册登录、数据存在别人服务器上。\n\n本案例做出的是一个单文件 HTML 工作台：左侧导航分「今日焦点 / 今天待办 / 灵感速记 / 项目跟进 / 每周复盘 / 数据备份」六个区，顶部常驻「今天要处理的」把逾期任务和卡住的项目一并标红。待办按四象限分优先级，昨天没做完的会自动带到今天并累计「顺延 N 次」（顺延有成本，自己才会不好意思拖）；未来任务单独归到「即将到来」按日期分组。项目与待办通过关联打通——任务可挂到项目下，项目卡上直接显示进度条和未完成清单、能就地勾选完成；删项目时会先问你保留还是一并删除待办，不静默丢数据。周复盘统计完成数、优先级分布、精力花在哪些项目上，以及卡点汇总。全部数据存在浏览器本地并自动保存，攒满 20 条会提醒导出备份，配「导出 / 导入」实现换设备搬家，手机端按钮加大到 44px 以上。\n\n适合：想要一个不用注册、数据自己掌握的轻量工作台的人；同时要盯多个项目又容易漏掉卡点的产品/运营；习惯用待办清单但总是坚持不下来、需要「顺延成本」提醒的自我管理者；需要每周复盘产出的独立工作者。", "description_en": "You've tried half a dozen to-do apps and none of them stuck. Either yesterday's unfinished tasks silently disappear, or projects and tasks live in two unrelated lists — you keep blockers in your head, and by the weekend you can't even count how many things you actually finished. You just want something simple, but everything requires signup and stores your data on someone else's server.\n\nThis case produces a single-file HTML workbench: a left sidebar with six sections (Today's Focus / Tasks / Quick Notes / Project Tracking / Weekly Review / Backup), plus a persistent \"needs attention today\" banner that flags overdue tasks and blocked projects in red. Tasks use four-quadrant priorities; anything unfinished carries over to today automatically with a running \"postponed N times\" counter (postponing has a visible cost, so you stop doing it). Future tasks live in a separate \"Upcoming\" view grouped by date. Projects and tasks are properly linked — tasks attach to a project, project cards show a progress bar plus open items you can check off in place, and deleting a project asks whether to keep or remove its tasks instead of silently dropping data. The weekly review reports completion counts, priority distribution, which projects consumed your effort, and a blocker summary. Everything persists in browser local storage with autosave, prompts a backup export every 20 new entries, and ships export/import for moving between devices; mobile buttons are sized 44px+.\n\nIdeal for: anyone wanting a lightweight workbench with no signup and full data ownership; product and ops folks juggling several projects who keep missing blockers; self-managers who need a visible \"cost of postponing\" to actually stick with a task list; independent workers who owe themselves a weekly review.", "categories": [ "workspace", "office", "productivity" ], "tags": [ "任务管理", "项目跟进", "每周复盘", "本地存储", "离线可用" ], "tags_en": [ "task management", "project tracking", "weekly review", "local storage", "offline ready" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/EJmKTgqQ0IiFe4NHkjCCCl），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个个人办公工作台，左边是导航菜单，右边是内容。\n\n要这几块：\n· 今天待办：填任务、选轻重缓急、填截止日期，做完打勾；过期没做的标红，一键顺延到明天；未来的任务也要能看到，按日期分组\n· 灵感速记：随手记想法，能打标签、按标签找\n· 项目跟进：记项目名、现在到哪一步、下一步做什么、卡在哪儿；要能和待办关联——任务可以挂到项目下，项目上显示关联待办的进度\n· 每周复盘：这周做完几件事、按优先级分布、精力花在哪些项目上、卡点汇总\n\n最上面放「今天要处理的」：逾期没做的任务、卡住的项目列出来标红。\n昨天没做完的自动挪到今天，别消失，并且记录顺延了几次。\n白底简洁，蓝绿色，商务精致一点，数据要有可视化（图表用内联 SVG/CSS 实现）。可以放一张克制的免版权配图当页头，不要喧宾夺主。\n\n另外这几点帮我做好：\n1. 我填的东西自动保存，关掉页面再打开还在\n2. 页面上放「导出备份」和「导入」两个按钮，换手机的时候能搬过去\n3. 攒到 20 条的时候提醒我导出一次，免得哪天清了浏览器就没了\n4. 手机和电脑都能用，手机上按钮大一点方便点\n5. 做成一个 HTML 文件，所有不要用网上的现成组件，断网也能打开\n6. 先放几条示例内容让我看效果，再给个「清空示例」按钮\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/EJmKTgqQ0IiFe4NHkjCCCl) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nHelp me build a personal office workbench, with a navigation menu on the left and content on the right. I need these sections:\n· Today's tasks: enter a task, pick its priority, set a due date, check it off when done; flag overdue items in red with one-click postpone to tomorrow; future tasks should also be visible, grouped by date\n· Quick notes: jot down ideas, tag them, find them by tag\n· Project tracking: record the project name, where it stands now, what's next, and what's blocking it; it should link with tasks — tasks can attach to a project, and the project shows the progress of its linked tasks\n· Weekly review: how many things I finished this week, the priority distribution, which projects consumed my effort, and a blocker summary At the very top put a \"needs attention today\" area: list overdue tasks and blocked projects, flagged in red.\nAnything unfinished from yesterday moves to today automatically instead of disappearing, and record how many times it has been postponed. Clean white background, teal accent, refined and business-like, with data visualized (implement charts with inline SVG/CSS). A restrained royalty-free image as a page header is fine, but it must not steal focus. Also please take care of these:\n1. Whatever I enter saves automatically and is still there after I close and reopen the page\n2. Put \"Export backup\" and \"Import\" buttons on the page so I can move everything when I switch phones\n3. Remind me to export once I've accumulated 20 entries, so I don't lose everything if the browser gets cleared\n4. Works on both phone and desktop, with larger buttons on mobile for easier tapping\n5. Make it a single HTML file — no components pulled from the internet, and it must open offline\n6. Put in a few sample entries first so I can see how it looks, then give me a \"Clear samples\" button\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 0, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-01T09:08:34Z", "updated_at": "2026-08-01T10:33:14Z", "artifacts": [ { "file": "output.html", "label": "个人办公工作台（单文件 HTML）" } ], "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "personal-portfolio-rebalance", "title": "个人投资组合再平衡仪表盘", "title_en": "Personal Portfolio Rebalancing Dashboard", "subtitle": "一眼看清资产偏离，给出可执行的再平衡操作", "subtitle_en": "See allocation drift at a glance with an actionable rebalancing plan", "description": "个人投资者最头疼的是：账户涨涨跌跌后配置早已偏离目标，却不知道该买卖什么、买卖多少。本案例由「理财财」财富管理专家生成一份个人投资组合再平衡仪表盘——直观对比当前与目标配置、标出每类资产的偏离度，并给出精确到金额的买入/卖出建议。适合有一定资产、希望定期做组合体检的个人投资者。", "description_en": "The biggest headache for individual investors: after markets swing, the portfolio has drifted from its targets — but they don't know what to buy or sell, or how much. Generated by the Wealth Management Expert, this personal portfolio rebalancing dashboard visually compares current vs target allocation, flags each asset's drift, and gives buy/sell recommendations down to the exact amount. Ideal for individual investors who want to run a regular portfolio checkup.", "categories": [ "finance" ], "tags": [ "投资组合", "再平衡", "资产配置", "个人理财" ], "tags_en": [ "Portfolio", "Rebalancing", "Asset Allocation", "Personal Finance" ], "artifact_type": "html", "prompt": "我是稳健型个人投资者王先生，35 岁，投资目标是退休储备。帮我生成一个个人投资组合再平衡仪表盘。\n\n当前组合（总资产 ¥1,280,000）：A股股票基金 40%、港美股 20%、债券基金 15%、黄金/商品 10%、货币基金/现金 15%。要看到总资产与收益 KPI、当前 vs 目标配置对比、再平衡操作建议表、再平衡后的预期配置，以及再平衡节奏与费税提示。", "prompt_en": "I'm Mr. Wang, a 35-year-old conservative individual investor saving for retirement. Help me generate a personal portfolio rebalancing dashboard.\n\nCurrent portfolio (total assets ¥1,280,000): A-share equity funds 40%, HK/US equities 20%, bond funds 15%, gold/commodities 10%, money-market/cash 15%. Show total-assets and return KPIs, current vs target allocation comparison, a rebalancing action table, the expected post-rebalance allocation, and notes on rebalancing cadence and fees/taxes.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "再平衡仪表盘" } ], "skills": [], "experts": [ { "id": "WealthManagementExpert", "name": "理财财", "description": "财富管理工具集：客户回顾、财务规划、投资组合再平衡、税损收割、投资提案和客户报告生成", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "06802d3ccfd4", "name_en": "Harlan · Wealth Management Expert", "description_en": "Wealth management toolkit: client review, financial planning, portfolio rebalance, tax-loss harvesting, investment proposal, and client reporting", "expert_id": "ex_wMZ6zxKnk1uI" } ], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-16T19:30:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pet-care-workbench", "title": "小林coding的家有宠物工作台", "title_en": "Pet Family Workbench", "subtitle": "喂食喂药医疗消费，毛孩子的事一个工作台全管住", "subtitle_en": "Feeding, medication, vet visits and expenses — one workbench for all your pets", "description": "多宠家庭的日常记录总是散落在聊天记录、备忘录和脑子里：今天喂了几顿、驱虫什么时候到期、上次看病花了多少钱，越想越乱。\n\n这个案例展示了一个「家有宠物」管理工作台：首页宠物卡片墙 + 今日待办（喂食/喂药打卡/医疗到期自动汇总）+ 最近动态流；宠物档案带头像和自动年龄；喂食有近 14 天进食量柱状图和食物类型占比；喂药按疗程管理，进度条 + 按频次打卡 + 漏喂标红；医疗记录带下次到期提醒，填了费用会自动同步一笔「医疗」消费；体重有趋势折线；消费是按月账本带分类占比。中英双语一键切换，数据落在资料库结构化表里，多端同步。\n\n适合：多宠家庭、经常忘记驱虫疫苗时间的铲屎官、想给毛孩子建一份完整健康账本的人。", "description_en": "In multi-pet homes, daily care records end up scattered across chats, memos and memory: how many meals today, when is deworming due, how much was the last vet visit — it gets messy fast.\n\nThis case shows a Pet Family workbench: a home page with pet cards, today's to-dos (feeding / medication check-ins / upcoming vet reminders auto-aggregated) and a recent-activity feed; pet profiles with photo upload and auto-calculated age; feeding logs with a 14-day intake bar chart and food-type breakdown; medication managed by course with progress bar, per-dose check-ins and missed-dose alerts; medical records with next-due reminders that auto-sync any cost into a \"Medical\" expense entry; weight trends on a line chart; and a monthly expense ledger with category shares. Switch between Chinese and English with one tap, and all data lives in structured Library tables that sync across devices.\n\nBest for: multi-pet families, owners who keep forgetting vaccine and deworming dates, and anyone who wants a complete health and expense ledger for their pets.", "categories": [ "workspace" ], "tags": [ "宠物养护", "健康管理", "生活工作台", "数据同步", "家庭记账" ], "tags_en": [ "Pet Care", "Health Management", "Life Workbench", "Data Sync", "Family Ledger" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "家有宠物工作台（单体 HTML，中英双语）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/wFhUfzUQhc7Sp744yZ6uK3，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「家有宠物」管理工作台。家里有三只宠物：奶糖（银渐层猫，4 岁 10 个月，母，已绝育）、小弟（美短猫，6 岁 1 个月，公，已绝育）、小七（比熊犬，1 岁 6 个月）。做成单个完整 HTML 文件。\n\n包含：① 首页宠物卡片墙 + 今日待办（喂食次数、喂药打卡、医疗到期提醒自动汇总）+ 最近动态流；② 宠物档案（头像本地上传压缩、品种/生日/性别/绝育状态/毛色，年龄自动计算）；③ 喂食记录（近 14 天进食量柱状图 + 近 30 天食物类型占比）；④ 喂药疗程管理（疗程进度条、每日按频次打卡、漏喂标红、疗程编辑联动同步剂量记录）；⑤ 医疗记录（疫苗/驱虫/就诊/体检/手术，下次到期日提醒，填了费用自动同步一条「医疗」分类的消费记录）；⑥ 体重记录（趋势折线图 + 较上次变化）；⑦ 消费记录（月度账本、分类占比、按宠物筛选）。\n\n视觉走移动端优先的微信风格：浅灰底 + 白色圆角卡片 + 微信绿主色，底部 Tab 导航（首页/档案/＋快捷记录/健康/我的）。制作完成后：① 将 HTML 存入我的资料库，并创建 6 张资料库子数据表（宠物档案/喂食/喂药/医疗/体重/消费）绑定数据，实现医疗-消费联动；② 发布为公开链接并返回给我CSS/JS 无外部 CDN 依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 9 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the Library's one-click clone capability to make me an exact copy of https://www.workbuddy.cn/space/d/wFhUfzUQhc7Sp744yZ6uK3 — data must be stored online with two-way sync. Prefer the one-click clone path; if that capability is unavailable, build it as follows: Build a \"Pet Family\" management workbench. I have three pets: Candy (Silver Shaded cat, 4y 10m, female, neutered), Didi (American Shorthair cat, 6y 1m, male, neutered) and Seven (Bichon Frise dog, 1y 6m). Deliver a single complete HTML file. Include: ① a home page with a pet card wall + today's to-dos (feeding counts, medication check-ins and upcoming medical reminders auto-aggregated) + a recent-activity feed; ② pet profiles (local photo upload with compression, breed / birthday / gender / neuter status / coat color, auto-calculated age); ③ feeding records (14-day intake bar chart + 30-day food-type breakdown); ④ medication course management (course progress bar, per-dose daily check-ins, missed doses flagged red, course edits cascading to dose records); ⑤ medical records (vaccine / deworming / clinic visit / checkup / surgery, next-due reminders, and any cost entered auto-syncs a \"Medical\" expense entry); ⑥ weight records (trend line chart + change vs last); ⑦ expense records (monthly ledger, category shares, filter by pet). Visual style: mobile-first WeChat-style look — light gray background, white rounded cards, WeChat-green primary color, bottom tab bar (Home / Profiles / ＋ quick log / Health / Me). When done: ① save the HTML to my WorkBuddy Library and create 6 bound Library child tables (pets / feeding / medication / medical / weight / expenses) with the medical-to-expense automation; ② publish it as a public link and return the link to me. Fall back to localStorage when the SDK is unavailable. Inline all CSS/JS with no external CDN dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus 9 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "小林coding", "name_en": "小林coding", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-17T10:55:00+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "is_recommended": false, "recommended_order": 0 }, { "id": "pia-cross-border-ecommerce", "title": "跨境电商个人信息保护影响评估", "title_en": "Cross-Border E-Commerce PIA Report", "subtitle": "一条 Prompt 自动排查 5 类触发场景，产出带 8 张专业表格、可直接申报的五模块 PIA 报告", "subtitle_en": "One prompt auto-scans all five trigger scenarios and produces a filing-ready five-module PIA report with eight structured tables", "description": "企业处理海量用户个人信息，最怕的就是被监管认定违规——人脸支付、AI 推荐、数据出海，每一样都可能踩中个保法的红线，一罚就是上千万甚至按年营收的 5%。\n\n这个案例把一家跨境电商平台（500 万用户、AI 精准营销、人脸支付、向新加坡母公司传输数据）交给 AI，一键产出五模块 PIA 报告：自动排查 PIPL 第 55 条全部 5 类触发场景，还附带 8 张可直接抄作业的专业表格——字段清单、合法性基础识别、风险评估分级、缓解措施对照（含优先级和整改期限）、合规要求对照，以及把散落各处的「待补充」信息缺口汇总成一张表。跨境场景还会自动叠加 GDPR DPIA 交叉判断 + 新加坡 PDPA 合规义务对照，可直接作为标准合同备案或数据出境安全评估申报的前置材料。\n\n适合：数据合规律师、企业法务、隐私合规负责人、出海电商运营者。", "description_en": "Handling millions of users' personal data is a compliance minefield — face payment, AI recommendations, and cross-border transfers each risk breaching China's PIPL, with fines up to RMB 50 million or 5% of annual revenue.\n\nThis case hands a cross-border e-commerce platform (5M users, AI-driven marketing, face payment, data transfer to a Singapore parent) to AI, which produces a five-module PIA report: it auto-scans all five PIPL Article 55 trigger scenarios and ships eight ready-to-reuse tables — a data-field inventory, a lawfulness-basis check, a risk-grading matrix, a mitigation roadmap (with priorities and deadlines), a compliance checklist, and a consolidated table of every 'to-be-supplied' information gap. The cross-border scenario also auto-appends a GDPR DPIA cross-check and a Singapore PDPA obligations mapping — ready to serve as the prerequisite for a standard contractual clause (SCC) filing or data-export security assessment.\n\nFor: data-compliance lawyers, in-house counsel, privacy officers, and cross-border e-commerce operators.", "categories": [ "business", "productivity" ], "tags": [ "个人信息保护影响评估", "PIA报告", "PIPL合规", "数据出境", "跨境电商" ], "tags_en": [ "PIA Report", "PIPL Compliance", "Data Export", "Cross-Border E-Commerce", "Privacy Impact Assessment" ], "artifact_type": "doc", "preview": "pia-report.docx", "artifacts": [ { "file": "pia-report.docx", "label": "PIA 评估报告（DOCX）" }, { "file": "pia-report.md", "label": "PIA 评估报告（Markdown）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "preview_en": "", "cover_image_en": "", "prompt": "请调用「个人信息保护影响评估」技能，对跨境电商平台「星云跨境（深圳）科技有限公司」的个人信息处理活动开展个人信息保护影响评估（PIA），按《个人信息保护法》第 55/56 条和网信办指引格式出具五模块 PIA 报告。\n\n【处理活动信息】\n• 处理主体：星云跨境（深圳）科技有限公司，跨境电商社交平台，覆盖约 500 万注册用户\n• 处理活动：AI 个性化推荐与精准营销、人脸支付、订单履约、风控与反欺诈\n• 涉及字段：手机号、真实姓名、收货地址、浏览与购买行为、设备信息、设备位置信息（敏感·行踪轨迹）、人脸特征（敏感·生物识别）、支付账号/银行卡号（敏感·金融账户）\n• 处理目的：①订单履约（必要）②AI 个性化推荐与精准营销（辅助）③人脸支付（辅助）④风控与反欺诈（辅助）\n• 共享/委托：向境外新加坡母公司传输用于统一风控与推荐模型训练；委托云服务商、第三方物流、支付机构处理\n• 安全措施：传输/存储加密（AES）、RBAC 访问控制、脱敏、日志审计、人脸特征单独加密存储、应急预案（未演练）\n\n【评估要求】\n1. 输出五模块报告（M1 处理目的方式 → M2 合法性正当性必要性 → M3 权益影响风险 → M4 保护措施 → M5 结论），严格按网信办指引格式\n2. 逐项排查 PIPL 第 55 条 5 类触发场景，命中场景在对应模块嵌入专项分析\n3. 命中跨境提供场景时，增加跨境多法域叠加分析（GDPR DPIA 交叉适用判断 + 新加坡 PDPA 叠加分析）\n4. 输出格式：DOCX（正式申报用）+ Markdown（内部参考）\n5. 所有法条引用标注「以申报时生效版本为准」，信息缺口标注「待补充」", "prompt_en": "Invoke the Personal Information Protection Impact Assessment (PIA) skill to run a PIA on the personal-information processing activities of the cross-border e-commerce platform Xingyun Cross-Border (Shenzhen) Technology Co., Ltd., and produce a five-module PIA report following PIPL Articles 55/56 and the CAC guideline format.\n\n[Processing activity details]\n• Controller: Xingyun Cross-Border (Shenzhen) Technology Co., Ltd., a cross-border social e-commerce platform with ~5 million registered users\n• Activities: AI-based personalized recommendation and targeted marketing, face payment, order fulfillment, risk control and anti-fraud\n• Data fields: mobile number, real name, delivery address, browsing and purchase behavior, device information, device location (sensitive · location tracking), facial features (sensitive · biometric), payment account / bank card number (sensitive · financial account)\n• Purposes: (1) order fulfillment (necessary); (2) AI recommendation and targeted marketing (auxiliary); (3) face payment (auxiliary); (4) risk control and anti-fraud (auxiliary)\n• Sharing / entrustment: transfer to an overseas Singapore parent company for unified risk control and recommendation model training; entrust cloud service providers, third-party logistics, and payment institutions\n• Security measures: AES encryption in transit and at rest, RBAC access control, de-identification, log auditing, separate encrypted storage of facial features, emergency response plan (not yet drilled)\n\n[Assessment requirements]\n1. Output a five-module report (M1 purpose and method → M2 lawfulness/legitimacy/necessity → M3 rights impact and risk → M4 protection measures → M5 conclusion), strictly following the CAC guideline format\n2. Review all five PIPL Article 55 trigger scenarios one by one, embedding a dedicated analysis of each triggered scenario in the corresponding module\n3. Since the cross-border scenario is triggered, add a cross-jurisdiction overlay analysis (GDPR DPIA cross-applicability check + Singapore PDPA overlay analysis)\n4. Output format: DOCX (for official filing) + Markdown (internal reference)\n5. Tag every legal citation with \"subject to the version in effect at the time of filing\" and every information gap with \"to be supplemented\"", "skills": [], "experts": [ { "id": "data-privacy-ai-compliance-expert", "name": "数据隐私与AI合规专家", "description": "面向数据合规律师与DPO，覆盖个人信息合规、PIA评估、数据出境与AI治理，内嵌执业安全红线。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "882dfd2c5f4a", "expert_id": "ex_ww996Cmrx1T3" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-09-07T09:58:00+08:00", "updated_at": "2026-09-07T09:58:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "picky-cat-food-promo-page", "title": "挑嘴研究所宣传页", "title_en": "Picky Cat Food Promo Page", "subtitle": "让一只嫌弃脸橘猫当主角，把试吃报告做成引流页", "subtitle_en": "A grumpy tabby stars in a promo page built on a tasting report", "description": "宠物食品页最容易做成“鸡胸肉含量、配料表、下单按钮”的说明书，主人看完没有任何记忆点。\n\n这个案例给虚构的猫零食店「挑嘴研究所」做了一张手机优先的宣传页：首屏是一只一脸嫌弃的橘猫，两条红黑条带斜着拉过画面；往下是 50 只挑嘴猫的两周试吃报告，横条逐项拉满；还有一个拖动猫体重就能算出每天吃几克、一罐吃几天的小工具；最后是 11.11 囤两罐的说明。页面同时有中文和英文两个版本。\n\n适合想让宣传页有性格的小微商家，也适合想参考“从商品独有的一件事出发”如何做首屏的设计者。", "description_en": "Pet food pages tend to read like a spec sheet of meat content, ingredients and an order button, and nobody remembers them.\n\nThis case builds a mobile-first promo page for a fictional cat-treat shop, The Picky Lab. The first screen is a grumpy orange tabby with two red and black ribbons slanting across it. Below come a two-week tasting report on 50 picky cats with bars that fill one by one, a small tool that turns your cat's weight into grams a day and days per jar, and the 11.11 two-jar offer. The page comes in Chinese and English.\n\nUseful for small merchants who want a promo page with character, and for designers looking at how to start a first screen from the one thing only this product has.", "categories": [ "creation", "business" ], "tags": [ "宣传页", "宠物食品", "商品推广", "创意设计" ], "tags_en": [ "Promo Page", "Pet Food", "Product Promotion", "Creative Design" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "单页 HTML 宣传页" } ], "cover_image": "cover-portrait.png", "cover_image_en": "cover-portrait-en.png", "cover_mode": "portrait", "prompt": "为猫零食店「挑嘴研究所」（杭州）做一张手机优先的商品宣传页，用于 11.11 前把人引到店铺。商品是 120 克一罐的冻干鸡胸，鸡胸肉含量 97%，不加诱食剂；11.11 两罐 ¥89（原价 ¥118，第二罐半价，可混搭鸡胸 / 鸭胸 / 三文鱼口味），不爱吃拆封 7 天内整罐退。\n\n创意从这家店独有的一件事出发：让一只一脸嫌弃的橘猫当页面角色，口吻是“挑嘴的猫，吃了都说还行”。首屏用整幅粉色底的猫图，两条红黑条带斜着横穿画面并随滚动被拉动，猫随滚动放大，不要罐头和价格铺满首屏。\n\n往下依次是：50 只挑嘴猫的两周试吃报告（闻了就吃 44 只、当顿吃完 38 只、主动讨要 31 只、闻完走开 4 只，横条进入视野后逐项拉满）；喂食量计算器（拖动猫的体重 2 到 9 公斤，实时显示每天克数、碗里的量和一罐能吃几天）；11.11 囤两罐的优惠说明和一个下单按钮（按下有猫爪印飘起，只做演示反馈）；页尾写商家信息。\n\n风格：奶油粉底、番茄红和墨黑，字体活泼，每一屏只用一个动效。商品图为示意，正式发布前用商家实拍替换；页面只做展示和引流，真正下单在店铺完成。商家、数字均为演示，不编造真实销量、评价或功效。\n\n交付：一个开箱即用的单页 HTML 宣传页，手机和电脑都能打开，页面里没有需要保存的数据。", "prompt_en": "Create a mobile-first product promo page for The Picky Lab, a cat-treat shop in Hangzhou, to bring people to the shop before 11.11. The product is a 120 g jar of freeze-dried chicken breast, 97% chicken, no attractants. For 11.11, two jars cost ¥89 (normally ¥118, second jar half price, flavours can be mixed between chicken, duck and salmon), and an uneaten jar can be returned within 7 days of opening.\n\nStart from the one thing only this shop has: a grumpy orange tabby as the page's character, with the line “Even picky cats said not bad.” The first screen is the full pink cat photo with two red and black ribbons slanting across it and pulled along as the visitor scrolls, the cat growing as they scroll. Do not fill the first screen with the jar and price.\n\nThen, in order: a two-week tasting report on 50 picky cats (44 ate on first sniff, 38 finished the meal, 31 begged for more, 4 sniffed and left, with bars filling one by one as they come into view); a feeding calculator (drag the cat's weight from 2 to 9 kg and see grams per day, the amount in the bowl and how many days a jar lasts); the 11.11 two-jar offer with an order button (pressing it floats up paw prints, as a demo response only); and shop details in the footer.\n\nStyle: cream pink, tomato red and ink black, playful lettering, one motion effect per screen. Product photos are illustrative and should be replaced with the merchant's own before publishing. The page only showcases and drives traffic; real orders are placed in the shop. Shop and figures are demos; never invent real sales, reviews or efficacy claims.\n\nDeliverable: a ready-to-use single-page HTML promo page that opens on phones and desktops; the page keeps no data of its own.", "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "ex_Weo3x2EZuSDm", "name": "商家宣传设计师", "description": "给商家做宣传页和战报海报：宣传页用创意突出品牌或商品特点，引人进店；海报按店铺数据构思，不套模板。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "ff113452d619", "expert_id": "ex_Weo3x2EZuSDm" } ], "skills": [], "mcps": [], "creator": { "name": "小超", "name_en": "小超", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 94, "created_at": "2026-10-15T15:30:00+08:00", "updated_at": "2026-10-15T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "plant-growth-changes", "title": "植物的生长变化（种子发芽）", "title_en": "植物的生长变化（种子发芽）", "subtitle": "小学三年级科学 · 植物的生长变化（种子发芽）", "subtitle_en": "小学三年级科学 · 植物的生长变化（种子发芽）", "description": "三年级科学《植物的生长变化》互动课件，用延时发芽动画和观察日记，讲清种子发芽的条件与过程。", "description_en": "三年级科学《植物的生长变化》互动课件，用延时发芽动画和观察日记，讲清种子发芽的条件与过程。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "plant-growth-changes.html", "artifacts": [ { "file": "plant-growth-changes.html", "label": "植物的生长变化（种子发芽）｜互动教学网页" } ], "prompt": "帮我做一个小学三年级科学《植物的生长变化》的互动教学网页，要有种子发芽的延时动画和观察日记。", "prompt_en": "帮我做一个小学三年级科学《植物的生长变化》的互动教学网页，要有种子发芽的延时动画和观察日记。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "planting-problem-trees-and-intervals", "title": "植树问题（棵数与间隔数关系）", "title_en": "植树问题（棵数与间隔数关系）", "subtitle": "小学五年级数学广角 · 植树问题（棵数与间隔数关系）", "subtitle_en": "小学五年级数学广角 · 植树问题（棵数与间隔数关系）", "description": "五年级数学《植树问题》互动课件，用磁贴和动态种树动画，对比三种端点情形下的棵数与间隔数关系。", "description_en": "五年级数学《植树问题》互动课件，用磁贴和动态种树动画，对比三种端点情形下的棵数与间隔数关系。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "planting-problem-trees-and-intervals.html", "artifacts": [ { "file": "planting-problem-trees-and-intervals.html", "label": "植树问题（棵数与间隔数关系）｜互动教学网页" } ], "prompt": "帮我做一个小学五年级数学《植树问题》的互动教学网页，要能动态种树、对比两端都栽和两端不栽的区别。", "prompt_en": "帮我做一个小学五年级数学《植树问题》的互动教学网页，要能动态种树、对比两端都栽和两端不栽的区别。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "playwright-catalog-inspector", "title": "商品目录情报面板", "title_en": "Catalog Intelligence Dashboard", "subtitle": "在真实浏览器里抓取公开商品目录，并把结果整理成轻量数据看板", "subtitle_en": "Capture a public product directory in a real browser and turn the result into a lightweight data dashboard", "description": "目录页采集最怕两类问题：一类是采回来一堆原始文本，后面还得手工清洗；另一类是只要表格不要页面证据，别人根本无法判断采集当时页面长什么样。这个案例把采集结果和页面现场一起交付。\n\n本案例通过 Playwright Browser Automation 的官方 Playwright API 真实访问 https://books.toscrape.com/，提取前 10 个商品的标题、价格、库存状态、评分和商品链接，保存目录页截图并输出结构化 extracted.json。最终交付单文件内联 output.html，采用明亮的数据展板风格，并支持右上角中英切换与 ?lang=zh / ?lang=en。\n\n适合：电商运营、竞品采集团队、测试工程师、增长分析同学，以及任何需要把“打开目录页 → 提取商品字段 → 产出结构化看板”沉淀成可复用 Playbook 的团队。", "description_en": "Directory extraction usually fails in two ways: you end up with raw text that still needs manual cleanup, or you get a table without page evidence so nobody can tell what the page looked like at extraction time. This case delivers both the structured result and the visible page state together.\n\nUsing the official Playwright API behind Playwright Browser Automation, this case truly visits https://books.toscrape.com/, extracts the first 10 products with title, price, availability, rating, and product links, stores a catalog-page screenshot, and outputs a structured extracted.json file. The final deliverable is a fully inlined single-file output.html in a bright data-dashboard style, with top-right language switching and support for ?lang=zh / ?lang=en.\n\nBest for ecommerce operators, competitive-intelligence collectors, test engineers, growth analysts, and any team that wants to turn “open directory page → extract product fields → deliver a structured board” into a reusable Playbook.", "categories": [ "dev", "data", "agent" ], "tags": [ "Playwright", "目录采集", "数据提取", "商品情报", "浏览器自动化" ], "tags_en": [ "Playwright", "catalog extraction", "data extraction", "product intelligence", "browser automation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "双语数据面板" }, { "file": "catalog.png", "label": "目录页截图" }, { "file": "extracted.json", "label": "结构化商品数据" }, { "file": "execution-notes.txt", "label": "执行备注" } ], "external_url": "", "prompt": "请用 Playwright Browser Automation 的官方 Playwright API 真实访问公开商品目录页，优先使用 https://books.toscrape.com/。提取至少 10 个商品的标题、价格、库存状态、评分和商品链接，保存目录页截图 catalog.png，并输出 extracted.json。然后基于真实采集结果生成一个单文件内联的 output.html：明亮的数据展板风格，并清楚展示 KPI、评分分布、商品表格、页面截图和可复现说明。", "prompt_en": "Use the official Playwright API behind Playwright Browser Automation to truly visit a public product directory page, preferably https://books.toscrape.com/. Extract at least 10 products with title, price, availability, rating, and product links, save a catalog screenshot as catalog.png, and output extracted.json. Then generate a fully inlined single-file output.html from the real extraction result: a bright data-dashboard style, and clear sections for KPIs, rating distribution, product rows, page evidence, and reproducibility notes.", "skills": [ { "id": "playwright-browser-automation", "name": "Playwright Browser Automation", "description": "直接调用 Playwright API 实现浏览器自动化（无需 MCP）", "skill_id": "skill_2053082580726448128" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-07-28T09:17:04.226Z", "updated_at": "2026-07-28T09:31:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "playwright-docs-archive", "title": "Playwright 文档归档巡检", "title_en": "Playwright Docs Archive Inspection", "subtitle": "真实浏览器访问、截图、PDF 导出与导航提取的档案馆式案例包", "subtitle_en": "An archive-style case pack with real browser visit, screenshot, PDF export, and navigation extraction", "description": "文档站巡检最怕两件事：一是只留下口头结论，没有页面证据；二是为了省事改成静态抓取，结果和真实浏览器看到的页面不一致。这个案例直接面向“需要保留现场”的自动化场景，把一次真实访问落成可复查的本地案例包。\n\n本案例用 Playwright Browser Automation 对应的官方 Playwright API 真实访问 Playwright 官方文档页 https://playwright.dev/docs/intro，完成页面加载、顶部导航提取、至少 6 个页面要点摘录、全页截图、PDF 导出与结构化 JSON 提取。最终交付单文件内联 output.html，采用明亮色系的杂志 / 档案馆叙事风，并支持右上角中英切换与 ?lang=zh / ?lang=en 语言参数。\n\n适合：浏览器自动化工程师、QA、开发者关系、技术文档运营，以及任何需要把“打开页面 → 提取关键信息 → 生成截图 / PDF 归档”做成可复用 Playbook 的团队。", "description_en": "Documentation inspections usually fail in two ways: they leave only verbal conclusions with no page evidence, or they fall back to static scraping and no longer match what a real browser saw. This case is built for automation scenarios that need preserved evidence, turning one real visit into a locally reviewable case pack.\n\nUsing the official Playwright API behind Playwright Browser Automation, this case really visits the Playwright documentation page at https://playwright.dev/docs/intro and completes page loading, top-navigation extraction, at least 6 key-point captures, a full-page screenshot, PDF export, and structured JSON extraction. The final deliverable is a fully inlined single-file output.html in a bright magazine / archive narrative style, with a top-right Chinese / English switch and support for ?lang=zh / ?lang=en.\n\nBest for browser automation engineers, QA teams, developer-relations roles, technical documentation operators, and any team that wants to turn “open page → extract key information → archive screenshot / PDF” into a reusable Playbook.", "categories": [ "dev", "productivity", "agent" ], "tags": [ "Playwright", "浏览器自动化", "文档归档", "截图导出", "PDF 导出" ], "tags_en": [ "Playwright", "browser automation", "documentation archive", "screenshot export", "PDF export" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "双语归档页面" }, { "file": "page.png", "label": "全页截图" }, { "file": "page.pdf", "label": "页面 PDF" }, { "file": "extracted.json", "label": "结构化提取结果" }, { "file": "execution-notes.txt", "label": "执行备注" } ], "external_url": "", "prompt": "请用 Playwright Browser Automation 的官方 Playwright API 真实访问 Playwright 官方文档页 https://playwright.dev/docs/intro；若该页暂时不可达，可切换到同站稳定文档页，但必须记录实际访问 URL。完成页面访问、等待主内容加载、提取顶部导航、提取至少 6 个页面要点，保存全页截图 page.png 与页面 PDF page.pdf，并输出 extracted.json。再基于真实抓取结果生成一个单文件内联的 output.html：明亮色系、杂志 / 档案馆叙事风。最后补齐 execution-notes.txt、cover-portrait.png、cover-portrait-en.png，并保持整个案例包可本地直接预览。", "prompt_en": "Use the official Playwright API behind Playwright Browser Automation to truly visit the Playwright documentation page at https://playwright.dev/docs/intro. If that page is temporarily unreachable, switch to another stable page on the same site, but you must record the actual visited URL. Complete page access, wait for the main content to load, extract the top navigation, capture at least 6 key page points, save a full-page screenshot as page.png and a page PDF as page.pdf, and output extracted.json. Then generate a fully inlined single-file output.html from the real extraction result: bright visual tone and a magazine / archive narrative style. Finally add execution-notes.txt, cover-portrait.png, and cover-portrait-en.png so the whole case pack can be previewed locally.", "skills": [ { "id": "playwright-browser-automation", "name": "Playwright Browser Automation", "description": "直接调用 Playwright API 实现浏览器自动化（无需 MCP）", "skill_id": "skill_2053082580726448128" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-07-28T09:17:04.226Z", "updated_at": "2026-07-28T09:31:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "playwright-login-qa", "title": "公开站点登录验证台", "title_en": "Public Site Login Validation Board", "subtitle": "用真实浏览器执行公开演示登录流程，并把断言结果整理成 QA 控制台式案例页", "subtitle_en": "Execute a public demo login in a real browser and package the assertions into a QA control-board case page", "description": "登录自动化最容易失真：脚本写完了，却没有留下页面证据，也说不清到底是凭据错误、页面变更还是断言没等到。这个案例直接把“跑一次登录”变成可回看的 QA 页面。\n\n本案例通过 Playwright Browser Automation 的官方 Playwright API 真实打开公开演示登录站点，填写账号密码、提交表单、等待成功状态出现，并保存登录成功截图与结构化断言结果。最终交付单文件内联 output.html，采用明亮的 QA 控制台风格，右上角支持中英切换与 ?lang=zh / ?lang=en。\n\n适合：QA 工程师、自动化测试同学、前端 / 全栈开发、测试平台建设者，以及需要把“登录链路是否正常”固化成可复用 Playbook 的团队。", "description_en": "Login automation often becomes fuzzy: the script exists, but there is no visual evidence and no clear story about whether the failure came from credentials, page changes, or timing assertions. This case turns a single login run into a QA page that can be reviewed afterwards.\n\nUsing the official Playwright API behind Playwright Browser Automation, this case truly opens a public demo login site, fills credentials, submits the form, waits for a success state, and stores both the success screenshot and structured assertion output. The final deliverable is a fully inlined single-file output.html in a bright QA control-board style, with top-right language switching and support for ?lang=zh / ?lang=en.\n\nBest for QA engineers, automation testers, frontend / full-stack developers, test-platform builders, and any team that wants to convert “is the login path healthy?” into a reusable Playbook.", "categories": [ "dev", "agent" ], "tags": [ "Playwright", "登录测试", "断言验证", "浏览器自动化", "QA" ], "tags_en": [ "Playwright", "login testing", "assertion validation", "browser automation", "QA" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "双语 QA 报告页" }, { "file": "login-result.png", "label": "登录成功截图" }, { "file": "extracted.json", "label": "结构化断言结果" }, { "file": "execution-notes.txt", "label": "执行备注" } ], "external_url": "", "prompt": "请用 Playwright Browser Automation 的官方 Playwright API 真实执行一个公开演示站点的登录流程，优先使用 https://www.saucedemo.com/（账号 standard_user，密码 secret_sauce）；若该站点暂时不可用，可切换到另一公开练习登录页，但必须记录真实站点、凭据和实际访问 URL。完成打开登录页、填写账号密码、提交表单、等待成功状态出现，并输出 login-result.png 与 extracted.json。再基于真实执行结果生成一个单文件内联的 output.html：明亮的 QA 控制台风格，并清楚展示执行步骤、断言结果、截图证据和可复现说明。", "prompt_en": "Use the official Playwright API behind Playwright Browser Automation to truly execute a login flow on a public demo site, preferably https://www.saucedemo.com/ (username standard_user, password secret_sauce). If that site is temporarily unavailable, switch to another public practice login page, but you must record the actual site, credentials, and visited URL. Open the login page, fill the credentials, submit the form, wait for the visible success state, and output login-result.png plus extracted.json. Then generate a fully inlined single-file output.html from the real execution result: a bright QA control-board style, and clear sections for execution steps, assertion results, screenshot evidence, and reproducibility notes.", "skills": [ { "id": "playwright-browser-automation", "name": "Playwright Browser Automation", "description": "直接调用 Playwright API 实现浏览器自动化（无需 MCP）", "skill_id": "skill_2053082580726448128" } ], "experts": [], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 87, "usage_count": 0, "created_at": "2026-07-28T09:17:04.226Z", "updated_at": "2026-07-28T09:31:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-ab-test-report", "title": "结算页改版 A/B 实验结果分析报告", "title_en": "Checkout Redesign A/B Test Result Analysis Report", "subtitle": "对照组 vs 实验组：转化率提升 +21.9% 的显著性检验与上线决策", "subtitle_en": "Control vs Variant: significance testing and launch decision for a +21.9% conversion lift", "description": "做完 A/B 实验，最难的不是跑实验，而是判断「这个提升到底是真的还是随机波动」并给出可落地的上线建议。本案由「产品通」专家以实验设计与指标体系方法论，针对虚构电商 App 结算页改版实验，完成了一份完整的结果分析报告：从假设设定、样本量功效核算、主指标双比例 Z 检验，到转化漏斗双组对比、置信区间可视化、分群洞察，再到护栏指标与上线决策、实验局限。交付物是一份单文件 HTML 分析报告，数据彼此勾稽、统计自洽。适合产品经理、增长与数据分析师，作为撰写实验复盘的结构与表达范本。", "description_en": "The hard part of an A/B test isn't running it — it's deciding whether a lift is real or noise, then turning that into an actionable launch call. Here the Product Management Expert applies experiment-design and metrics methodology to a fictional e-commerce checkout-redesign test, producing a full result report: hypothesis, sample-size power check, a two-proportion Z-test on the primary metric, dual-group funnel comparison, confidence-interval visualization, segment insights, guardrail metrics, the launch decision, and study limitations. The deliverable is a single-file HTML report with internally consistent numbers — a solid template for product, growth and data analysts writing experiment readouts.", "categories": [ "business" ], "tags": [ "A/B实验", "转化率分析", "显著性检验", "数据报告" ], "tags_en": [ "A/B Testing", "Conversion Analysis", "Significance Testing", "Data Report" ], "artifact_type": "html", "prompt": "我们电商 App 刚跑完一个结算页改版的 A/B 实验（对照组是原三步式结算，实验组是单页聚合+优惠券自动勾选），帮我做一份精美的单文件 HTML 实验结果分析报告。要有：实验设计（假设、分流、样本量功效核算、周期）、核心指标对照表（转化率 A 3.2% vs B 3.9%、提升、p值、置信区间）、转化漏斗双组对比、置信区间可视化、次级指标（客单价、加购率）、分群洞察、结论与上线建议、实验局限。数据要彼此勾稽、统计上自洽，数据实验报告风，浅色简净+数据高亮，涨红跌绿，要响应式。", "prompt_en": "We just finished an A/B test on our e-commerce app's checkout redesign (control = original 3-step checkout, variant = single-page checkout with auto-selected coupon). Build a polished single-file HTML result-analysis report. Include: experiment design (hypothesis, traffic split, sample-size power calculation, duration), a core-metrics comparison table (conversion A 3.2% vs B 3.9%, lift, p-value, confidence interval), dual-group funnel comparison, a confidence-interval visualization, secondary metrics (AOV, add-to-cart rate), segment insights, conclusions with launch recommendations, and study limitations. Numbers must be internally consistent and statistically sound; use a clean light data-report style with highlighted figures, red-up/green-down, and responsive layout.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "结算页改版 A/B 实验结果分析报告" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-competitor-radar", "title": "竞品功能雷达 + 机会矩阵", "title_en": "Competitor Feature Radar & Opportunity Matrix", "subtitle": "消费级 App 横评：雷达图 + 机会矩阵一页看清", "subtitle_en": "Consumer-app benchmark: radar chart + opportunity matrix at a glance", "description": "产品经理做竞品分析常常只产出一份枯燥的对比文档，看不出格局也看不出机会。本案例由「产品通」产品管理专家以「价值判断官」视角，对 5 款消费级 AI 笔记 App 做可视化横评——SVG 六维雷达图、2×2 机会矩阵（含机会空位标注）、功能对比表与差异化结论一页呈现。适合做竞品研究、想直观看清市场格局与切入点的产品经理。", "description_en": "PMs doing competitive analysis often produce a dull comparison doc that reveals neither the landscape nor the opportunity. Acting as the Product Management Expert with a 'value-judge' lens, this case visually benchmarks 5 consumer AI note-taking apps — an SVG 6-axis radar, a 2×2 opportunity matrix (with the open-gap marked), a feature table, and differentiation takeaways on one page. Ideal for PMs who want to see the market landscape and entry points clearly.", "categories": [ "office" ], "tags": [ "竞品分析", "产品管理", "数据可视化", "雷达图" ], "tags_en": [ "Competitive Analysis", "Product Management", "Data Viz", "Radar Chart" ], "artifact_type": "html", "prompt": "帮我做一个消费级 AI 笔记 App 的竞品分析可视化页面（单文件 HTML）。对比 Notion、Obsidian、Mem、Craft、Apple Notes 五款产品，维度含 AI 能力、知识关联、协作、跨端同步、易用性、价格友好度。要直观看到：六维雷达图对比、2×2 机会矩阵（标出机会空位）、功能对比表、各产品定位与差异化机会结论。靛蓝网格科技感，界面中文。", "prompt_en": "Help me build a competitive-analysis visualization page for consumer AI note-taking apps (single-file HTML). Compare Notion, Obsidian, Mem, Craft, and Apple Notes across dimensions like AI capability, knowledge linking, collaboration, cross-device sync, ease of use, and price friendliness. Show a six-axis radar comparison, a 2×2 opportunity matrix (mark the open gap), a feature comparison table, and per-product positioning with differentiation takeaways. Indigo grid tech aesthetic; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "竞品功能雷达 + 机会矩阵" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-19T15:00:00Z", "updated_at": "2026-06-19T15:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-feature-prioritization-board", "title": "产品需求池与功能优先级看板", "title_en": "Product Backlog & Feature Prioritization Board", "subtitle": "价值×工作量四象限 + 四列看板，让迭代排期不再拍脑袋", "subtitle_en": "Value × Effort quadrant plus a four-column board for evidence-based sprint planning", "description": "迭代规划时需求一堆、谁先做谁后做总靠拍脑袋，争议难以收敛。本案由「产品战略团队」专家按数据驱动 SOP，把 13 条候选需求统一评分，落到「价值×工作量」四象限并铺成「待评估/已排期/进行中/已上线」四列看板：每张卡片标注价值分、工作量与 P0–P3 / RICE 优先级，再给出本迭代承诺 vs 储备的容量分配和四条决策原则。交付物是一份单文件 HTML 看板，含散点四象限图、看板列与脚注。适合产品经理、项目负责人与需要对齐排期口径的研发团队。", "description_en": "Sprint planning often turns into a who-goes-first argument with no clear rule. Here the Product Strategy Team applies a data-driven SOP: it scores 13 candidate features, plots them on a Value × Effort quadrant, and lays them across a four-column board (To Evaluate / Scheduled / In Progress / Shipped). Each card carries a value score, effort, and a P0–P3 / RICE priority, alongside a committed-vs-backlog capacity split and four decision principles. The deliverable is a single-file HTML board with a scatter quadrant chart, kanban columns, and a footnote. Built for product managers, project leads, and engineering teams aligning on sprint scope.", "categories": [ "business" ], "tags": [ "需求优先级", "迭代规划", "看板", "RICE评分" ], "tags_en": [ "Prioritization", "Sprint Planning", "Kanban", "RICE Scoring" ], "artifact_type": "html", "prompt": "帮我做一个单文件 HTML 的产品需求池与功能优先级看板（虚构 SaaS 产品的双周迭代规划）。要有：顶部 hero 大图压标题；需求池概览指标（总数、各状态计数）；一个「价值×工作量」四象限散点图（SVG，坐标轴容纳所有点、标注速赢/战略投入/填充/谨慎四个象限，放 10 条以上带名称的需求点）；待评估/已排期/进行中/已上线四列看板，每张卡片带价值分、工作量、P0–P3 或 RICE 优先级标签；本迭代承诺 vs 储备的容量分配；最后是决策原则脚注。浅灰底+彩色标签列+卡片的高效团队看板风，响应式，界面中文，无 emoji 无外链。", "prompt_en": "Build a single-file HTML product backlog and feature prioritization board for a fictional SaaS product's bi-weekly sprint. Include: a hero image with overlaid title; backlog overview metrics (total and per-status counts); a Value × Effort quadrant scatter chart (inline SVG, axes covering all points, four labeled quadrants — Quick Wins / Big Bets / Fill-ins / Money Pit — with 10+ named feature dots); a four-column kanban (To Evaluate / Scheduled / In Progress / Shipped) where each card shows value score, effort, and a P0–P3 or RICE priority tag; a committed-vs-backlog capacity split; and a footer of decision principles. Use a clean light-gray, color-tagged team-board aesthetic, responsive, Chinese UI, no emoji, no external resources.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "功能优先级看板" } ], "skills": [], "experts": [ { "id": "ProductStrategyTeam", "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师、用户研究员、竞品分析师、数据分析师和路线图规划师，覆盖从构思到上线的完整产品生命周期", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "6f3c88c704c9", "name_en": "Product Strategy Team", "description_en": "A 5-role product team led by a product director: requirement analyst, user researcher, competitive analyst, data analyst, roadmap planner", "expert_id": "ex_LqCvsVsLOp3J" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-gtm-launch-plan", "title": "新产品上市 GTM 发布计划一页纸", "title_en": "Product Go-To-Market Launch Plan (One-Pager)", "subtitle": "协作白板 Pro 上市：定位 / 定价 / 渠道 / 时间线 / 指标一图说清", "subtitle_en": "Whiteboard Pro launch: positioning, pricing, channels, timeline & metrics on one page", "description": "新模块要上市，团队却把定位、定价、渠道、时间线、成功指标散落在十几份文档里，高管、销售、工程各看各的，发布前对不齐。本案由「产品通」专家把这些要素合成为一张内部自洽的 GTM 一页纸：先讲清目标市场与定位语句、三类客户画像，再到三档定价、四条渠道的预算矩阵、预热→公测→GA 三阶段泳道时间线，最后落到北极星（每周活跃协作团队）+ 驱动/健康指标与可勾稽的增长漏斗，并附风险应对。交付物是一份可直接对齐评审的精美单文件 HTML，适合产品经理、增长与市场负责人在发布启动会上使用。", "description_en": "When a new module ships, positioning, pricing, channels, timeline and success metrics often scatter across a dozen docs—execs, sales and engineering never align before launch. Here the Product Management Expert synthesizes them into one self-consistent GTM one-pager: target market and positioning statement, three buyer personas, three pricing tiers, a four-channel budget matrix, a three-phase (pre-launch → beta → GA) swimlane timeline, and a North Star (weekly active collaborating teams) plus input/health metrics with a reconciled growth funnel and risk mitigations. The deliverable is a polished single-file HTML ready for a launch kickoff review—built for PMs, growth and marketing leads.", "categories": [ "business" ], "tags": [ "GTM上市", "发布计划", "渠道矩阵", "指标体系" ], "tags_en": [ "Go-To-Market", "Launch Plan", "Channel Matrix", "Metrics Framework" ], "artifact_type": "html", "prompt": "我们一个 B2B SaaS 新模块「协作白板 Pro」要上市，帮我做一份 Go-To-Market 发布计划的一页纸，做成精美的单文件 HTML（长滚动网页，不要 PPT）。要包含：目标市场与定位语句、目标客户画像、核心价值主张、三档定价与套餐、内容/社群/销售/合作四条渠道的预算矩阵、预热→公测→GA 三阶段泳道时间线、北极星与成功指标（激活率/转化/MRR，数字要能勾稽）、风险与应对。现代产品团队风、浅色配靛蓝+活力橙强调、卡片化排版，配一张产品流程便签墙的 hero 大图压标题，要做好手机端适配。", "prompt_en": "We're launching a new B2B SaaS module, 'Whiteboard Pro'. Build me a Go-To-Market launch plan one-pager as a polished single-file HTML (long-scroll web page, not slides). Include: target market & positioning statement, buyer personas, core value props, three pricing tiers, a four-channel (content/community/sales/partnership) budget matrix, a three-phase pre-launch→beta→GA swimlane timeline, North Star & success metrics (activation/conversion/MRR with reconciled numbers), and risks with mitigations. Modern product-team style, light theme with indigo + vibrant orange accents, card-based layout, a hero image of a product-flow sticky-note wall with the title overlaid, and solid mobile responsiveness.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "scenario_cover_en": "scenario-cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "协作白板 Pro · GTM 上市发布计划一页纸" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-north-star-metrics", "title": "北极星指标体系拆解看板", "title_en": "North Star Metric Tree Dashboard", "subtitle": "北极星 → 驱动 → 健康指标，一张图讲清指标体系", "subtitle_en": "North Star → driver → health metrics, the whole tree on one board", "description": "很多团队定了北极星指标却讲不清它如何被拆解、由哪些驱动指标支撑、哪里在掉链子。本案例由「产品通」产品管理专家搭建一套可视化指标体系看板——SVG 指标树把北极星拆到驱动与健康指标、12 周趋势折线、达成率进度条，并附指标治理说明（含虚荣指标反例）。适合需要建立或对齐指标体系的产品与运营团队。", "description_en": "Many teams set a North Star metric but can't explain how it decomposes, which driver metrics support it, or where it's slipping. Built by the Product Management Expert, this case lays out a visual metric-system dashboard — an SVG metric tree from North Star down to driver and health metrics, a 12-week trend line, attainment progress bars, plus governance notes (with vanity-metric counter-examples). Ideal for product and ops teams establishing or aligning a metric system.", "categories": [ "office" ], "tags": [ "指标体系", "北极星指标", "数据看板", "产品管理" ], "tags_en": [ "Metric System", "North Star", "Dashboard", "Product Management" ], "artifact_type": "html", "prompt": "帮我做一个 B2B SaaS 产品的北极星指标体系可视化看板（单文件 HTML）。北极星为周活跃创作用户数 WAU。要直观看到：指标拆解树（北极星 → 获取/激活/留存/收入/推荐 5 类驱动指标 → 各自的健康指标）、顶部 KPI 卡、12 周北极星趋势折线、每个指标的定义/当前/目标/健康状态，并给一段指标治理说明（为什么选这个北极星、虚荣指标反例）。深色数据仪表盘风，界面中文。", "prompt_en": "Help me build a North Star metric-system dashboard for a B2B SaaS product (single-file HTML). The North Star is Weekly Active Creators (WAU). Show a metric decomposition tree (North Star → five driver groups: acquisition/activation/retention/revenue/referral → their health metrics), top KPI cards, a 12-week North Star trend line, and each metric's definition/current/target/health status, plus a governance note (why this North Star, vanity-metric counter-examples). Dark data-dashboard style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "北极星指标体系拆解看板" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "created_at": "2026-06-19T15:00:00Z", "updated_at": "2026-06-19T15:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-product-roadmap-quarters", "title": "年度产品路线图（主题轨 × 四季度）", "title_en": "Annual Product Roadmap (Themes × Quarters)", "subtitle": "增长 / 体验 / 平台 / AI 四泳道横跨 Q1–Q4，对齐目标与里程碑", "subtitle_en": "Growth / Experience / Platform / AI swimlanes across Q1–Q4, aligned to goals and milestones", "description": "产品团队常苦于「需求堆叠、目标失焦」，路线图沦为承诺清单。本案由「产品通」专家用路线图管理方法论，把虚构 SaaS「Lumina 协作云」的全年规划梳理成「主题轨 × 时间」双视图：四条主题轨各锚定一个北极星指标，16 个季度 Epic 标注交付状态，配 7 个关键里程碑时间线、北极星指标进展表、依赖与风险清单，并用 Now-Next-Later 管理远期不确定性。交付物是一份可滚动浏览的单文件 HTML 路线图，适合产品经理、研发负责人与高管对齐全年节奏。", "description_en": "Product teams often struggle with feature pile-ups and blurred goals, turning roadmaps into mere commitment lists. Here the Product Management Expert applies roadmap methodology to a fictional SaaS, Lumina, presenting the year as a Theme × Time view: four swimlanes each anchored to a north-star metric, 16 quarterly epics with delivery status, a 7-milestone timeline, a metrics progress table, a dependency-and-risk list, and a Now-Next-Later backlog. The deliverable is a single-file scrollable HTML roadmap for PMs, eng leads, and execs to align on the year.", "categories": [ "business" ], "tags": [ "产品路线图", "Now-Next-Later", "北极星指标", "里程碑规划" ], "tags_en": [ "Product Roadmap", "Now-Next-Later", "North-Star Metric", "Milestone Planning" ], "artifact_type": "html", "prompt": "帮我做一份虚构 SaaS 产品的 2026 年度产品路线图，单文件 HTML。要「主题轨 × 时间」双视图：增长/体验/平台/AI 四条泳道横跨 Q1–Q4，每条泳道用色条画出各季度的 Epic 并标注状态（已交付/进行中/规划）。还要包含路线图说明与目标对齐、关键里程碑时间线、每条主题轨对应的北极星指标进展表、依赖与风险清单、以及 Now-Next-Later 储备。整体浅色专业风，路线图甘特/泳道样式，配一张真实主题图做 hero 大图叠渐变蒙版压标题，要响应式、界面中文、无 emoji。", "prompt_en": "Create an annual 2026 product roadmap for a fictional SaaS as a single-file HTML page. Use a Theme × Time view: four swimlanes (Growth/Experience/Platform/AI) spanning Q1–Q4, with colored bars for each quarter's epics labeled by status (delivered/in-progress/planned). Include a roadmap overview with goal alignment, a key-milestone timeline, a north-star metrics progress table per theme, a dependency-and-risk list, and a Now-Next-Later backlog. Use a clean professional light style with a gantt/swimlane look, a real hero photo with a gradient overlay behind the title, responsive layout, Chinese UI, no emoji.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "scenario_cover_en": "scenario-cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "Lumina 协作云 · 2026 年度产品路线图" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-quarterly-strategy-onepager", "title": "季度产品战略一页纸", "title_en": "Quarterly Product Strategy One-Pager", "subtitle": "一页纸讲清季度战略：支柱、指标、里程碑、Non-goals", "subtitle_en": "One page for the quarter: pillars, metrics, milestones, non-goals", "description": "季度战略常散落在十几页 PRD 和 OKR 文档里，高管和团队很难一眼对齐方向。本案例由「产品战略团队」编排，把一个季度的产品战略浓缩成一张可视化一页纸——一句话战略主张、TL;DR、核心结论卡、三大战略支柱、KPI 目标进度、里程碑时间线、行动清单与 Non-goals。适合需要对内对外快速对齐季度方向的产品负责人。", "description_en": "Quarterly strategy is often scattered across a dozen PRD and OKR docs, making it hard for execs and teams to align at a glance. Orchestrated by the Product Strategy Team, this case condenses a quarter's product strategy into a visual one-pager — a one-line strategic thesis, TL;DR, conclusion cards, three strategic pillars, KPI target progress, a milestone timeline, an action list, and non-goals. Ideal for product leads who need to align the quarter quickly, internally and externally.", "categories": [ "office" ], "tags": [ "产品战略", "季度规划", "一页纸", "OKR" ], "tags_en": [ "Product Strategy", "Quarterly Plan", "One-Pager", "OKR" ], "artifact_type": "html", "prompt": "帮我做一张季度产品战略一页纸（单文件 HTML，汇报级）。某协作 SaaS 产品 2026 Q3。要直观看到：产品名+季度+一句话战略主张、TL;DR 三行、核心结论卡（主题/优先级/影响/资源/风险）、三大战略支柱（目标+关键举措+衡量指标）、4 个 KPI 的当前→目标进度、M1-M3 里程碑时间线、行动清单表、本季 Non-goals 与风险缓解。高端深色信息图风格，界面中文。", "prompt_en": "Help me build a quarterly product strategy one-pager (single-file HTML, board-ready) for a collaboration SaaS product, 2026 Q3. Show the product name + quarter + a one-line strategic thesis, a three-line TL;DR, conclusion cards (theme/priority/impact/resources/risk), three strategic pillars (goal + key initiatives + success metric), current→target progress for 4 KPIs, an M1-M3 milestone timeline, an action-list table, and this quarter's non-goals plus risk mitigations. High-end dark infographic style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "季度产品战略一页纸" } ], "skills": [], "experts": [ { "id": "ProductStrategyTeam", "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师、用户研究员、竞品分析师、数据分析师和路线图规划师，覆盖从构思到上线的完整产品生命周期", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "6f3c88c704c9", "name_en": "Product Strategy Team", "description_en": "A 5-role product team led by a product director: requirement analyst, user researcher, competitive analyst, data analyst, roadmap planner", "expert_id": "ex_LqCvsVsLOp3J" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "usage_count": 0, "created_at": "2026-06-19T15:00:00Z", "updated_at": "2026-06-19T15:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-rice-decision-matrix", "title": "RICE 需求优先级决策矩阵", "title_en": "RICE Prioritization Decision Matrix", "subtitle": "用 RICE 打分 + 散点矩阵，把优先级讲明白", "subtitle_en": "RICE scoring + scatter matrix to make priorities defensible", "description": "需求排期最容易吵架，因为优先级常凭感觉、说不清依据。本案例由「产品通」产品管理专家用 RICE 框架做一个可视化决策工具——10 个候选需求按 R×I×C÷E 精确打分，SVG 散点矩阵按价值/成本分四象限（气泡大小映射触达），排序表标出 P0/P1/P2，并给出为什么做、为什么砍的决策说明。适合需要透明、可复盘地排优先级的产品团队。", "description_en": "Roadmap prioritization sparks the most arguments because priority is often a gut call with no clear rationale. Using the RICE framework, the Product Management Expert builds a visual decision tool here — 10 candidate features scored precisely by R×I×C÷E, an SVG scatter matrix splitting value vs cost into four quadrants (bubble size maps Reach), a ranked table marking P0/P1/P2, and rationale for what to build and what to cut. Ideal for product teams who want transparent, reviewable prioritization.", "categories": [ "office" ], "tags": [ "RICE", "优先级", "决策矩阵", "产品管理" ], "tags_en": [ "RICE", "Prioritization", "Decision Matrix", "Product Management" ], "artifact_type": "html", "prompt": "帮我做一个 RICE 需求优先级决策矩阵可视化页（单文件 HTML）。给 10 个候选需求，每个含 Reach、Impact、Confidence、Effort，按 RICE=R×I×C÷E 打分。要直观看到：SVG 散点矩阵（X=成本 Y=价值，气泡大小=触达，分四象限并标注优先做/规划/填空/砍掉）、按总分降序的排序表（含分项与 P0/P1/P2 建议，P0 高亮）、顶部 KPI，以及为什么做/为什么砍的决策说明。极简留白风格，界面中文。", "prompt_en": "Help me build a RICE prioritization decision-matrix visualization page (single-file HTML). Provide 10 candidate features, each with Reach, Impact, Confidence, Effort, scored by RICE = R×I×C÷E. Show an SVG scatter matrix (X=effort, Y=value, bubble size=reach, four quadrants labeled do-now/plan/fill-in/cut), a table ranked by score (with sub-scores and P0/P1/P2 recommendations, P0 highlighted), top KPIs, and rationale for what to build vs cut. Minimalist whitespace style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "RICE 需求优先级决策矩阵" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "created_at": "2026-06-19T15:00:00Z", "updated_at": "2026-06-19T15:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-user-journey-emotion", "title": "用户旅程情绪曲线地图", "title_en": "User Journey Emotion Curve Map", "subtitle": "8 阶段旅程 + 情绪曲线，看清痛点与机会", "subtitle_en": "8-stage journey + emotion curve to surface pain points and opportunities", "description": "产品体验问题往往藏在用户旅程的情绪低谷里，但团队常常只有零散吐槽、没有全局视图。本案例由「体验达」用户体验架构师绘制一张电商用户旅程情绪曲线地图——SVG 平滑情绪曲线贯穿 8 个阶段，对齐每阶段的触点/行为/痛点/机会，并给出痛点到机会的改进建议。适合做体验优化、想系统定位痛点的产品与设计团队。", "description_en": "Experience problems often hide in the emotional troughs of the user journey, yet teams usually have only scattered complaints and no holistic view. Drawn by the UX Architect, this case maps an e-commerce user journey emotion curve — a smooth SVG emotion line across 8 stages aligned with each stage's touchpoints/behaviors/pain points/opportunities, plus pain-to-opportunity improvement ideas. Ideal for product and design teams doing experience optimization.", "categories": [ "office" ], "tags": [ "用户旅程", "体验设计", "情绪曲线", "可视化" ], "tags_en": [ "User Journey", "UX Design", "Emotion Curve", "Visualization" ], "artifact_type": "html", "prompt": "帮我做一张电商场景的用户旅程情绪曲线地图（单文件 HTML）。覆盖发现→搜索筛选→商品详情→加购→结算→支付→物流等待→收货评价 8 个阶段。要直观看到：一条贯穿全程的情绪曲线（低谷与高峰用色区分），对齐每阶段的用户目标/触点/行为/痛点/机会卡片，以及 3-4 条痛点到机会的改进建议。暖色叙事风格，界面中文。", "prompt_en": "Help me create an e-commerce user journey emotion curve map (single-file HTML). Cover 8 stages: discovery → search & filter → product detail → add to cart → checkout → payment → shipping wait → delivery & review. Show one emotion curve across the whole journey (troughs and peaks color-coded), aligned per-stage cards for goals/touchpoints/behaviors/pain points/opportunities, and 3-4 pain-to-opportunity improvement ideas. Warm narrative style; UI in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "用户旅程情绪曲线地图" } ], "skills": [], "experts": [ { "id": "UserExperienceArchitect", "name": "体验达", "description": "为开发者提供坚实的技术基础和CSS系统，是设计与开发之间的桥梁", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "012ca07220e3", "name_en": "ArchitectUX", "description_en": "Technical architecture & UX specialist bridging design and development", "expert_id": "ex_bzSq5Zh81H2y" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-19T15:00:00Z", "updated_at": "2026-06-19T15:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pm-user-persona-cards", "title": "用户画像 Persona 卡片集", "title_en": "User Persona Card Set", "subtitle": "在线教育产品「知阶」的 4 类典型用户画像", "subtitle_en": "Four key personas for the online-learning product \"ZhiJie\"", "description": "团队手里有几十份访谈和上千份问卷，却始终拼不出「我们到底在为谁做产品」的清晰共识。本案由「产品通」专家用其用户研究综合能力，把零散的定性洞察与定量数据聚类提炼成 4 张结构统一的用户画像卡片——备考白领、在校大学生、职场宝妈、自由职业者，每张覆盖人口属性、目标动机、痛点障碍、使用场景、关键诉求与产品机会点，并附一张跨人群的需求优先级对比图。交付一份可直接贴进路线图评审、对齐团队认知的精美单文件 HTML。适合产品经理、用户研究员与增长团队。", "description_en": "Teams often hold dozens of interviews and thousands of survey responses yet still lack a shared answer to \"who are we building for?\" Here the Product Management Expert applies user-research synthesis to cluster scattered qualitative and quantitative signals into four uniformly structured persona cards — exam-prep professional, college student, working mom, freelancer — each covering demographics, goals, pain points, scenarios, key needs and product opportunities, plus a cross-persona needs-priority chart. Delivered as a polished single-file HTML ready for roadmap reviews. Ideal for PMs, user researchers and growth teams.", "categories": [ "business" ], "tags": [ "用户画像", "用户研究", "在线教育", "需求优先级" ], "tags_en": [ "User Persona", "User Research", "Online Education", "Needs Priority" ], "artifact_type": "html", "prompt": "我们是一款在线教育 App，刚做完一轮用户访谈和问卷。帮我用「产品通」专家的用户研究综合能力，输出一份用户画像 Persona 卡片集：包含 3-4 个典型用户（比如备考白领、在校大学生、职场宝妈、自由职业者），每个 persona 做成一张精致卡片，含姓名/标签/一句话画像、人口属性、目标与动机、痛点与障碍、使用场景、关键诉求、产品机会点；再配一张跨 persona 的需求优先级对比图。做成温暖人本的研究报告风、单文件 HTML，头像用 CSS 几何圆形不要外链，要响应式、简体中文。", "prompt_en": "We're an online-education app that just finished a round of user interviews and surveys. Using the Product Management Expert's user-research synthesis, produce a user-persona card set: 3-4 typical users (e.g. exam-prep professional, college student, working mom, freelancer), each as a refined card with name/labels/one-line portrait, demographics, goals & motivations, pain points & barriers, usage scenarios, key needs, and product opportunities; plus a cross-persona needs-priority comparison chart. Make it a warm, human-centered research-report style single-file HTML, with CSS geometric circular avatars (no external images), responsive, in Simplified Chinese.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "用户画像 Persona 卡片集" } ], "skills": [], "experts": [ { "id": "ProductManagementExpert", "name": "产品通", "description": "产品管理工具集：功能规格编写、路线图规划、利益相关者沟通、用户研究综合、竞品分析和指标追踪", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "8f2316c9b230", "name_en": "Product Management Expert", "description_en": "Product management toolkit: feature specs (PRD), roadmap, stakeholder comms, user research synthesis, competitive analysis, metrics tracking", "expert_id": "ex_4sjpP0QoU5h1" } ], "mcps": [], "scenario": "产品管理", "scenario_id": 106, "scenario_en": "Product", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T16:00:00Z", "updated_at": "2026-06-22T16:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-cap-diagnosis", "title": "胸部 CT 肺炎 AI 分析：一例重症 CAP 怎么量化", "title_en": "Chest CT Pneumonia AI: Quantifying a Severe CAP Case", "subtitle": "阳性样本跑一遍，看 AI 如何量化 27 处病灶与双肺弥漫浸润", "subtitle_en": "Run a positive scan and see how AI quantifies 27 lesions across both lungs", "description": "社区获得性肺炎（CAP）在 CT 上往往是多灶、多叶、双肺弥漫的，靠肉眼一处处数病灶、估体积既慢又难量化。这个案例拿一例公开数据集里的 CAP 阳性胸部 CT（168 层）真跑了一遍：AI 给出肺炎表征概率 99.8%、全肺感染占比 33.0%（左肺 41.0%、右肺 23.3%），并逐一列出 27 处病灶的肺叶位置、层号范围、体积、占比和平均 CT 值——最大一处独占左肺 41%，最小的只有 104mm³，体积跨 5 个数量级。案例把这份重症阳性结果整理成带环形图、体积对数轴条形图和病灶分布气泡图的 HTML 报告，并配 PDF 报告，全流程命令可复现。和「正常样本」案例正好一阴一阳，适合做医疗影像 AI 的产品经理、科研人员，以及想建立阳性/重症判读参照的人。", "description_en": "Community-acquired pneumonia (CAP) on CT is often multifocal, multilobar and bilaterally diffuse — counting lesions and estimating volumes by eye is slow and hard to quantify. This case runs a real CAP-positive chest CT (168 slices) from a public dataset: the AI returns a 99.8% pneumonia-sign probability, 33.0% whole-lung infection (left 41.0%, right 23.3%), and lists all 27 lesions with their lobe, slice range, volume, ratio and mean CT value — the largest alone takes 41% of the left lung, the smallest is just 104mm3, spanning five orders of magnitude. The result is turned into an HTML report with a donut chart, a log-scale volume bar chart and a lesion-distribution bubble chart, plus a PDF report, with reproducible commands. It pairs with the normal-case example as a positive counterpart — built for medical-imaging AI PMs, researchers, and anyone establishing a positive/severe reading baseline.", "artifact_type": "html", "prompt": "这是一例社区获得性肺炎（CAP）阳性受试者的胸部 CT 检查，帮我基于腾讯觅影肺炎AI评估专家跑一遍肺炎 AI 分析，重点把多灶、多叶的病灶量化清楚。\n\n影像目录：{胸部CT的DICOM目录路径，如 桌面/cap001}\nstudyId：CASE_CAP001\n需要 PDF 报告（needReport=1）\n\n请按以下步骤完成，最终输出一份 HTML 分析报告（用于灵感案例展示，排版清晰美观、含基础样式）：\n\n1. **打包并校验**：把目录内 DICOM 打包为 ZIP，只收 .dcm 文件（排除 .DS_Store、__MACOSX、隐藏文件），打包后校验包内文件数与纯净度。\n\n2. **提交分析并轮询**：提交时开启 PDF 报告并携带患者信息。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态：待处理/处理中/处理完成），不要静默等待。\n\n3. **输出报告**，包含章节：案例概览（影像来源、序列层数、studyId、taskId、分析耗时）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、并用表格完整列出每一处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值，一处都不能漏）／结果解读（病灶数量与分布模式、体积悬殊如何理解、感染负荷左右肺对比、CT 值密度特征）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）。\n\n4. **生成可视化图表**：用 matplotlib 生成两张图并嵌入报告——① 肺炎表征概率环形图 + 全肺/左肺/右肺感染占比条形图；② 病灶体积 Top10 条形图（体积跨多个数量级，须用对数轴）+ 全部病灶在整段序列中的分布气泡图（横轴层号、纵轴左右肺、气泡大小表示体积、颜色映射平均 CT 值）。设置中文字体避免乱码，图片与报告同目录、用相对路径 标签引用。肺炎/异常用红色系（符合国内医疗惯例）。\n\n要求：\n- 所有数值必须来自 AI 接口返回的真实结果，不得推测、遗漏或填充缺失值；27 处病灶要逐一列全\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗，结果不构成诊断意见", "prompt_en": "This is a chest CT of a community-acquired pneumonia (CAP) positive subject. Please run pneumonia AI analysis using the Tencent Miying Pneumonia AI Assessment Expert, focusing on cleanly quantifying the multifocal, multilobar lesions.\n\nImage folder: {path to the chest CT DICOM folder, e.g. Desktop/cap001}\nstudyId: CASE_CAP001\nNeed PDF report (needReport=1)\n\nComplete the following steps and deliver an HTML report (clean, styled, suitable for a showcase):\n\n1. **Package and validate**: zip the DICOM files, including only .dcm files (exclude .DS_Store, __MACOSX, hidden files), then verify the archive file count and purity.\n\n2. **Submit and poll**: enable the PDF report and pass patient info on submission. Poll in the background and keep reporting progress (elapsed time, current status: pending/processing/done) — do not wait silently.\n\n3. **Produce the report** with sections: Case overview (image source, slice count, studyId, taskId, elapsed time) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and a table listing every lesion with its lobe, slice range, volume, volume ratio and mean CT value — omit none) / Interpretation (lesion count and distribution pattern, how to read the wide volume spread, left-vs-right infection burden, CT density characteristics) / Health notes (general guidance only, no specific clinical advice; defer diagnosis, medication and follow-up to a physician) / Reproducible workflow.\n\n4. **Generate charts**: use matplotlib for two figures embedded in the report — (1) a donut chart of pneumonia probability plus a bar chart of whole-lung/left/right infection ratios; (2) a Top-10 lesion volume bar chart (volumes span orders of magnitude, use a log scale) plus a bubble chart of all lesions across the series (x=slice number, y=left/right lung, bubble size=volume, color=mean CT value). Configure a CJK font to avoid tofu glyphs, reference images by relative path, and use a red palette for pneumonia/abnormal cues (per China medical convention).\n\nRequirements:\n- Every number must come from the real API response; never infer, omit or fill in missing values — list all 27 lesions one by one\n- Do not write real absolute paths; render the image path as \"Desktop/{folder name}\"\n- Open with the notice: this model is licensed for research use in mainland China only, must not be used for clinical diagnosis or treatment, and the result is not a diagnostic opinion", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "社区获得性肺炎", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "Community-Acquired Pneumonia", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 分析报告 · 社区获得性肺炎（HTML）" }, { "file": "infection-ratio.png", "label": "肺炎表征概率与感染占比图" }, { "file": "lesion-distribution.png", "label": "27 处病灶体积与分布图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "usage_count": 0, "created_at": "2026-09-02T16:00:00+08:00", "updated_at": "2026-09-02T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-covid19-diagnosis", "title": "胸部 CT 肺炎 AI 分析：新冠肺炎长什么样", "title_en": "Chest CT Pneumonia AI: What COVID-19 Looks Like", "subtitle": "阳性样本跑一遍，看新冠肺炎为何概率 100% 但占比只有 2.6%", "subtitle_en": "Run a COVID-positive scan: why 100% probability yet only 2.6% infection", "description": "新冠肺炎在 CT 上和细菌性肺炎不一样——它常是双肺多灶、散在的小磨玻璃影，靠肉眼很难量化到底累及了多少肺。这个案例拿一例公开数据集里 RT-PCR 确诊的 COVID-19 胸部 CT（148 层）真跑了一遍：AI 给出肺炎表征概率 100%，全肺感染占比却只有 2.6%（左肺 1.4%、右肺 3.8%），逐一列出 17 处病灶——数量不少但单个体积普遍偏小、集中在双肺下叶。有意思的是左肺病灶更多（11 处）占比反而更低，右肺病灶少却占比高，说明数个数不等于量范围。案例把结果整理成带环形图、体积对数轴图、病灶分布气泡图和三类样本对照表的 HTML 报告，并配 PDF 报告，命令可复现。和「正常样本」「社区获得性肺炎」两例凑成一套完整对照，适合做医疗影像 AI 的产品经理、科研人员，以及想看清不同病因肺炎影像差异的人。", "description_en": "COVID-19 on CT differs from bacterial pneumonia — it is often bilateral, multifocal, scattered small ground-glass opacities, hard to quantify by eye. This case runs a real RT-PCR-confirmed COVID-19 chest CT (148 slices) from a public dataset: the AI returns a 100% pneumonia-sign probability, yet only 2.6% whole-lung infection (left 1.4%, right 3.8%), listing all 17 lesions — many in number but mostly small, clustered in both lower lobes. Notably the left lung has more lesions (11) but a lower ratio, while the right has fewer yet a higher ratio, showing lesion count is not lesion burden. The result is turned into an HTML report with a donut chart, a log-scale volume chart, a lesion-distribution bubble chart and a three-sample comparison table, plus a PDF report, with reproducible commands. Together with the normal and CAP cases it forms a full comparison set — built for medical-imaging AI PMs, researchers, and anyone wanting to see how pneumonia of different causes differs on imaging.", "artifact_type": "html", "prompt": "这是一例新冠病毒性肺炎（COVID-19）阳性受试者的胸部 CT 检查，帮我基于腾讯觅影肺炎AI评估专家跑一遍肺炎 AI 分析，重点把多灶、散在的病灶量化清楚，并讲清它和细菌性肺炎在量化上的差异。\n\n影像目录：{胸部CT的DICOM目录路径，如 桌面/P001}\nstudyId：CASE_COVID19_P001\nstudyName：胸部CT\n需要 PDF 报告（needReport=1）\n\n请按以下步骤完成，最终输出一份 HTML 分析报告（用于灵感案例展示，排版清晰美观、含基础样式）：\n\n1. 打包并校验：把目录内 DICOM 打包为 ZIP，只收 .dcm 文件（排除 .DS_Store、__MACOSX、隐藏文件），打包后校验包内文件数与纯净度。\n\n2. 提交分析并轮询：提交时开启 PDF 报告并携带患者信息。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态：待处理/处理中/处理完成），不要静默等待。\n\n3. 输出报告，包含章节：案例概览（影像来源、序列层数、studyId、taskId、分析耗时）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、并用表格完整列出每一处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值，一处都不能漏）／结果解读（病灶数量与体积特征、病灶数与感染量为何不成正比、分布规律、CT 值密度特征）／三类样本对照（如手头有正常样本与社区获得性肺炎案例，用一张表把肺炎表征概率、全肺感染占比、病灶数、分布特征放在一起对照，点明概率与占比是两个维度）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）。\n\n4. 生成可视化图表：用 matplotlib 生成两张图并嵌入报告——① 肺炎表征概率环形图 + 全肺/左肺/右肺感染占比条形图；② 病灶体积 Top10 条形图（体积跨多个数量级，须用对数轴）+ 全部病灶在整段序列中的分布气泡图（横轴层号、纵轴左右肺、气泡大小表示体积、颜色映射平均 CT 值）。设置中文字体避免乱码，图片与报告同目录、用相对路径 标签引用。肺炎/异常用红色系（符合国内医疗惯例）。\n\n要求：\n- 所有数值必须来自 AI 接口返回的真实结果，不得推测、遗漏或填充缺失值；17 处病灶要逐一列全\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗，结果不构成诊断意见", "prompt_en": "This is a chest CT of a COVID-19 positive subject. Please run pneumonia AI analysis using the Tencent Miying Pneumonia AI Assessment Expert, focusing on quantifying the multifocal, scattered lesions and explaining how they differ from bacterial pneumonia quantitatively.\n\nImage folder: {path to the chest CT DICOM folder, e.g. Desktop/P001}\nstudyId: CASE_COVID19_P001\nstudyName: Chest CT\nNeed PDF report (needReport=1)\n\nComplete the following steps and deliver an HTML report (clean, styled, suitable for a showcase):\n\n1. Package and validate: zip the DICOM files, including only .dcm files (exclude .DS_Store, __MACOSX, hidden files), then verify the archive file count and purity.\n\n2. Submit and poll: enable the PDF report and pass patient info on submission. Poll in the background and keep reporting progress (elapsed time, current status: pending/processing/done) — do not wait silently.\n\n3. Produce the report with sections: Case overview (image source, slice count, studyId, taskId, elapsed time) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and a table listing every lesion with its lobe, slice range, volume, volume ratio and mean CT value — omit none) / Interpretation (lesion count and volume characteristics, why lesion count is not proportional to infection burden, distribution pattern, CT density) / Three-sample comparison (if normal and CAP cases are available, put pneumonia probability, whole-lung infection ratio, lesion count and distribution in one table, noting probability and ratio are two dimensions) / Health notes (general guidance only, no specific clinical advice; defer diagnosis, medication and follow-up to a physician) / Reproducible workflow.\n\n4. Generate charts: use matplotlib for two figures embedded in the report — (1) a donut chart of pneumonia probability plus a bar chart of whole-lung/left/right infection ratios; (2) a Top-10 lesion volume bar chart (log scale) plus a bubble chart of all lesions across the series (x=slice number, y=left/right lung, bubble size=volume, color=mean CT value). Configure a CJK font, reference images by relative path, use a red palette for pneumonia/abnormal cues (per China medical convention).\n\nRequirements:\n- Every number must come from the real API response; never infer, omit or fill in missing values — list all 17 lesions one by one\n- Do not write real absolute paths; render the image path as \"Desktop/{folder name}\"\n- Open with the notice: this model is licensed for research use in mainland China only, must not be used for clinical diagnosis or treatment, and the result is not a diagnostic opinion", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "新冠肺炎", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "COVID-19", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 分析报告 · 新冠病毒性肺炎（HTML）" }, { "file": "infection-ratio.png", "label": "肺炎表征概率与感染占比图" }, { "file": "lesion-distribution.png", "label": "17 处病灶体积与分布图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "usage_count": 0, "created_at": "2026-09-02T16:42:00+08:00", "updated_at": "2026-09-02T16:42:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-diffuse-multifocal", "title": "胸部 CT 肺炎 AI 分析：651 层大序列 · 39 处弥漫多发", "title_en": "Chest CT Pneumonia AI: 651-Slice Series, 39 Diffuse Lesions", "subtitle": "上传大序列 DICOM，一次跑通并输出全部 39 处病灶的量化明细", "subtitle_en": "Upload a large DICOM series, run it end-to-end and quantify all 39 lesions", "description": "做胸部 CT 肺炎量化时，遇到 600 多层的大序列常让人心里没底：传得动吗？要等多久？病灶多了还数得清吗？这个案例把一份 651 层、168MB 的真实胸部 CT 完整跑了一遍——一次提交成功，AI 返回肺炎表征概率、双肺感染占比，以及全部 39 处病灶各自的肺叶位置、层号范围、体积、体积占比和平均 CT 值，整理成一份结构化报告。它同时说清了大序列的一个真实体感：排队 + 分析要等约半小时，轮询超时不等于失败，凭 taskId 继续查就行。适合需要量化多发病灶、评估双肺分布、并把结果归档的科研人员和医疗 AI 产品经理。", "description_en": "Quantifying pneumonia on a 600+ slice chest CT can feel daunting: will it upload, how long is the wait, and can you still track lesions once there are dozens? This case runs a real 651-slice, 168MB chest CT end to end — submitted successfully on the first try, with the AI returning the pneumonia probability, per-lung infection ratios, and for all 39 lesions their lobe, slice range, volume, volume ratio and mean CT value, compiled into a structured report. It also captures a real lesson about large series: queuing plus analysis takes about half an hour, a poll timeout does not mean failure, and you simply re-query with the taskId. Built for researchers and medical-AI product managers who need to quantify multifocal disease, assess bilateral distribution and archive the results.", "artifact_type": "html", "prompt": "这是一次胸部 CT 检查，帮我基于腾讯觅影肺炎 AI 评估专家分析这例检查的肺炎情况，最终输出一份 Markdown 结构化报告。\n\n影像目录：{胸部CT的DICOM目录路径，如 桌面/A619214}\n\n请按以下步骤完成：\n\n1. **读取影像元数据**：提取患者 ID、性别、年龄、模态、层数、层厚、像素间距、设备厂商与型号、卷积核、传输语法、SeriesNumber，汇总成表格。\n\n2. **打包并校验**：将目录内 DICOM 打包为 ZIP，只收 .dcm 文件（排除 .DS_Store、__MACOSX 等无关文件），打包后校验包内文件数与纯净度；确认 ZIP 不超过 300MB、序列层数 > 10。\n\n3. **提交分析并轮询**：提交时开启 PDF 报告并携带患者信息。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态），不要静默等待。若遇到大序列（数百层）导致「待处理/处理中」持续较久甚至一次轮询超时，说明这是排队而非失败——不要重复提交，直接凭 taskId 继续查询。\n\n4. **输出报告**，包含章节：案例概览（患者与检查信息、taskId、分析耗时、剩余配额）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、按体积降序完整列出每处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值，并汇总肺叶累及分布）／结果解读（弥漫 vs 集中、双肺对称性、CT 值密度特征、主灶占比）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）／要点小结。\n\n要求：\n- 所有数值必须来自肺炎 AI 接口返回的真实结果，不得推测或填充缺失值\n- 病灶较多时须完整列出全部病灶，不得省略或截断\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗", "prompt_en": "This is a chest CT scan. Please analyze this study for pneumonia using the Tencent Miying Pneumonia AI Assessment Expert and deliver a structured Markdown report.\n\nImage folder: {path to the chest CT DICOM folder, e.g. Desktop/A619214}\n\nComplete the following steps:\n\n1. **Read image metadata**: extract patient ID, sex, age, modality, slice count, slice thickness, pixel spacing, manufacturer and model, convolution kernel, transfer syntax and SeriesNumber; summarize in a table.\n\n2. **Package and validate**: zip the DICOM files in the folder, including only .dcm files (exclude .DS_Store, __MACOSX and other irrelevant files), then verify the file count and purity of the archive; confirm the ZIP is under 300MB and the series has more than 10 slices.\n\n3. **Submit and poll**: enable the PDF report and pass patient info on submission. Poll in the background and keep reporting progress (elapsed time, current status) — do not wait silently. If a large series (hundreds of slices) stays in 'pending/processing' for a long time or a poll round times out, this is queuing rather than failure — do not resubmit; simply keep querying with the taskId.\n\n4. **Produce the report** with these sections: Case overview (patient and study info, taskId, elapsed time, remaining quota) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and every lesion listed by descending volume with its lobe, slice range, volume, volume ratio and mean CT value, plus a per-lobe involvement summary) / Interpretation (diffuse vs concentrated, bilateral symmetry, CT density characteristics, dominant-lesion share) / Health notes (general guidance only — no specific clinical advice; always defer diagnosis, medication and follow-up to a physician) / Reproducible workflow / Key takeaways.\n\nRequirements:\n- All values must come from the real API response; never guess or fill in missing values\n- When there are many lesions, list all of them without omission or truncation\n- Do not write real absolute paths; always write the image path as 'Desktop/{folder name}'\n- State up front: this model is for research use in mainland China only and must not be used for clinical diagnosis or treatment", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "CT量化分析", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "CT Quantification", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 量化分析报告（651 层 · 39 处病灶）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-08-27T11:05:00+08:00", "updated_at": "2026-08-28T16:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-empty-uid-repair", "title": "胸部 CT 肺炎 AI 分析：清洗 __MACOSX + 补齐空 UID", "title_en": "Chest CT Pneumonia AI: Clean __MACOSX and Backfill Empty UIDs", "subtitle": "两步数据修复让报错的 DICOM 跑通，量化 7 处微小病灶", "subtitle_en": "Two data fixes turn a failing DICOM into a run, quantifying 7 tiny lesions", "description": "从医院导出的 DICOM 直接提交，回来一句「AI引擎上传失败」是最劝退的场景——它不告诉你到底哪里错了。这个案例把一份真实报错的胸部 CT 从头救活：先发现 ZIP 里混进了 427 个 macOS 的 __MACOSX 干扰文件，清洗后仍失败，再逐字段扫描才定位到 StudyInstanceUID/SeriesInstanceUID 整段为空，补齐 UID 后一次跑通。最终 AI 返回肺炎表征 98.6%、但全肺感染仅 0.1%、7 处微小病灶——顺带演示了「高表征概率 ≠ 高感染负荷」这个容易误读的点。适合经常要清洗、修复医院导出 DICOM 的科研人员和医疗 AI 产品经理。", "description_en": "Submitting a DICOM straight from a hospital export and getting back a bare \"AI engine upload failed\" is the most discouraging scenario — it never tells you what actually went wrong. This case rescues a real failing chest CT end to end: first it finds 427 macOS __MACOSX junk entries hidden in the ZIP; cleaning them still fails, so a field-by-field scan pins down that StudyInstanceUID/SeriesInstanceUID are entirely empty, and backfilling the UIDs makes it run on the first try. The AI returns a 98.6% pneumonia probability but only 0.1% whole-lung infection across 7 tiny lesions — incidentally illustrating that a high probability does not equal a high infection burden. Built for researchers and medical-AI product managers who routinely clean and repair hospital-exported DICOM.", "artifact_type": "html", "prompt": "这是一次胸部 CT 检查，帮我基于腾讯觅影肺炎 AI 评估专家分析这例检查的肺炎情况，最终输出一份 Markdown 结构化报告。\n\n影像目录/压缩包：{胸部CT的DICOM目录或ZIP路径，如 桌面/100577}\n\n请按以下步骤完成：\n\n1. **读取影像元数据**：提取患者 ID、性别、年龄、模态、层数、层厚、像素间距、设备厂商与型号、传输语法，以及 StudyInstanceUID、SeriesInstanceUID、SeriesNumber、SOPClassUID 是否完整，汇总成表格。\n\n2. **清洗并打包**：若来源是从 macOS 导出/中转的 ZIP，先排除 __MACOSX 目录与 ._ 开头的资源分叉文件，只收真正的 .dcm 重新打包；校验包内 __MACOSX/._ 计数为 0、.dcm 数量正常、包体不超过 300MB。\n\n3. **提交分析并轮询**：提交时开启 PDF 报告并携带患者信息（患者 ID 为空时补填）。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态），不要静默等待。\n\n4. **失败自动诊断修复**：若返回 90002「AI引擎上传失败」，按响应里的 UID 是否为空判断根因——① 响应 UID 为空 → 优先查原始 DICOM 的 StudyInstanceUID/SeriesInstanceUID/SOPClassUID 是否缺失，为全序列补齐一组新 UID（每张单独生成 SOPInstanceUID，并同步写入 file_meta），必要时补 CT Image Storage 的 SOPClassUID；② 响应 UID 有值仅报错 → 优先查 SeriesNumber 等分组字段。确认后直接自动补齐并重新提交，只改元数据、绝不改动像素数据，不要把失败抛回给我等指示。\n\n5. **输出报告**，包含章节：案例概览（患者与检查信息、taskId、分析耗时、剩余配额）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、每处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值，并汇总肺叶累及分布）／结果解读（重点说明「高表征概率」与「感染负荷」是两个维度、病灶密度特征、分布模式）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）／关键排障记录（现象、排查路径、根因、经验总结）。\n\n要求：\n- 所有数值必须来自肺炎 AI 接口返回的真实结果，不得推测或填充缺失值\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗", "prompt_en": "This is a chest CT scan. Please analyze this study for pneumonia using the Tencent Miying Pneumonia AI Assessment Expert and deliver a structured Markdown report.\n\nImage folder / archive: {path to the chest CT DICOM folder or ZIP, e.g. Desktop/100577}\n\nComplete the following steps:\n\n1. **Read image metadata**: extract patient ID, sex, age, modality, slice count, slice thickness, pixel spacing, manufacturer and model, transfer syntax, and whether StudyInstanceUID, SeriesInstanceUID, SeriesNumber and SOPClassUID are complete; summarize in a table.\n\n2. **Clean and package**: if the source is a ZIP exported/relayed via macOS, first exclude the __MACOSX directory and ._-prefixed resource-fork files, keeping only real .dcm files, and repackage; verify the archive has zero __MACOSX/._ entries, a normal .dcm count, and is under 300MB.\n\n3. **Submit and poll**: enable the PDF report and pass patient info on submission (backfill patient ID if empty). Poll in the background and keep reporting progress (elapsed time, current status) — do not wait silently.\n\n4. **Auto-diagnose and fix on failure**: if the API returns 90002 \"AI engine upload failed\", use whether the response UIDs are empty to find the root cause — (1) empty response UIDs -> first check whether the source DICOM's StudyInstanceUID/SeriesInstanceUID/SOPClassUID are missing; backfill one new UID set for the whole series (generate a separate SOPInstanceUID per slice and sync file_meta), adding the CT Image Storage SOPClassUID if needed; (2) UIDs present but still failing -> check grouping fields like SeriesNumber first. Once confirmed, backfill and resubmit automatically, touching only metadata and never modifying pixel data, instead of handing the failure back to me.\n\n5. **Produce the report** with these sections: Case overview (patient and study info, taskId, elapsed time, remaining quota) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and for each lesion its lobe, slice range, volume, volume ratio and mean CT value, plus a per-lobe involvement summary) / Interpretation (emphasize that a high probability and the infection burden are two different dimensions, plus density characteristics and distribution) / Health notes (general guidance only — no specific clinical advice; always defer diagnosis, medication and follow-up to a physician) / Reproducible workflow / Troubleshooting log (symptom, investigation path, root cause, takeaways).\n\nRequirements:\n- All values must come from the real API response; never guess or fill in missing values\n- Do not write real absolute paths; always write the image path as 'Desktop/{folder name}'\n- State up front: this model is for research use in mainland China only and must not be used for clinical diagnosis or treatment", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "数据修复", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "Data Repair", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 量化分析报告（__MACOSX + 空 UID 双重修复）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-08-27T14:00:00+08:00", "updated_at": "2026-08-28T16:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-normal-screening", "title": "胸部 CT 肺炎 AI 筛查：正常受试者怎么读", "title_en": "Chest CT Pneumonia AI Screening: Reading a Normal Case", "subtitle": "跑一例正常样本，看懂概率打分与病灶量化为何要一起看", "subtitle_en": "Run a normal scan and learn why probability and lesion metrics must be read together", "description": "很多人用肺炎 AI 只盯着那个概率数字，一看到 75% 就慌——但概率高不等于有病。这个案例拿一例公开数据集里的正常受试者胸部 CT（154 层）真跑了一遍：AI 给出的肺炎表征概率是 75.1%，量化结论却是「无肺炎感染」、全肺感染占比 0.0%，只检出一处 722mm³、占比 0.0% 的微小影。案例把这份张力讲透了——概率是整图纹理的打分，容易受噪声体位影响，真正定性要落到病灶体积、占比、CT 值和分布上。全流程真实提交、后台轮询、结果整理成带图表的 HTML 报告，命令可复现。适合做医疗影像 AI 的产品经理、科研人员，以及需要建立「正常/阴性」判读基线的人。", "description_en": "Many people fixate on the probability number when using pneumonia AI — see 75% and panic. But a high probability does not mean disease. This case runs a real chest CT of a normal subject (154 slices) from a public dataset: the AI returns a 75.1% pneumonia-sign probability, yet the quantitative verdict is \"no pneumonia infection\" with 0.0% whole-lung infection, detecting only one tiny 722mm3 focus at 0.0% ratio. The case unpacks this tension — probability scores the whole-image texture and is sensitive to noise and positioning, while the real call rests on lesion volume, ratio, CT value and distribution. The whole pipeline is a real submission, background polling, and a charted HTML report with reproducible commands. Built for medical-imaging AI PMs, researchers, and anyone establishing a normal/negative reading baseline.", "artifact_type": "html", "prompt": "这是一例正常受试者的胸部 CT 检查，帮我基于腾讯觅影肺炎AI评估专家跑一遍肺炎 AI 分析，并教我怎么正确解读结果（尤其当概率打分和量化结论看起来不一致时）。\n\n影像目录：{胸部CT的DICOM目录路径，如 桌面/normal001}\nstudyId：CASE_NORMAL001\n\n请按以下步骤完成，最终输出一份 HTML 分析报告（用于灵感案例展示，排版清晰美观、含基础样式）：\n\n1. **打包并校验**：把目录内 DICOM 打包为 ZIP，只收 .dcm 文件（排除 .DS_Store、__MACOSX、隐藏文件），打包后校验包内文件数与纯净度。\n\n2. **提交分析并轮询**：提交时开启 PDF 报告并携带患者信息。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态：待处理/处理中/处理完成），不要静默等待。\n\n3. **输出报告**，包含章节：案例概览（影像来源、序列层数、studyId、taskId、分析耗时）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、每处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值）／结果解读（重点讲清概率打分 vs 病灶量化：为什么概率高但结论可以是无感染，如何结合病灶体积/占比/CT值/分布综合判断）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）。\n\n4. **生成可视化图表**：用 matplotlib 生成两张图并嵌入报告——① 肺炎表征概率环形图 + 全肺/左肺/右肺感染占比条形图；② 病灶体积（用对数轴，并标出显著病灶常见量级作参照，直观显示本例病灶多小）+ 病灶层号在全序列中的定位。设置中文字体避免乱码，图片与报告同目录、用相对路径 标签引用。风格干净专业，正常/健康用绿色、异常提示用红色（符合国内医疗惯例）。\n\n要求：\n- 所有数值必须来自 AI 接口返回的真实结果，不得推测或填充缺失值；若结论与概率不一致，如实呈现并解释，不得为了「好看」而修改数字\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗，结果不构成诊断意见", "prompt_en": "This is a chest CT of a normal subject. Please run pneumonia AI analysis using the Tencent Miying Pneumonia AI Assessment Expert, and teach me how to read the result correctly — especially when the probability score and the quantitative verdict seem inconsistent.\n\nImage folder: {path to the chest CT DICOM folder, e.g. Desktop/normal001}\nstudyId: CASE_NORMAL001\n\nComplete the following steps and deliver an HTML report (clean, styled, suitable for a showcase):\n\n1. **Package and validate**: zip the DICOM files, including only .dcm files (exclude .DS_Store, __MACOSX, hidden files), then verify the archive file count and purity.\n\n2. **Submit and poll**: enable the PDF report and pass patient info on submission. Poll in the background and keep reporting progress (elapsed time, current status: pending/processing/done) — do not wait silently.\n\n3. **Produce the report** with sections: Case overview (image source, slice count, studyId, taskId, elapsed time) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and for each lesion its lobe, slice range, volume, volume ratio and mean CT value) / Interpretation (focus on probability score vs lesion quantification: why the probability can be high while the verdict is no infection, and how to judge by combining lesion volume/ratio/CT value/distribution) / Health notes (general guidance only, no specific clinical advice; defer diagnosis, medication and follow-up to a physician) / Reproducible workflow.\n\n4. **Generate charts**: use matplotlib for two figures embedded in the report — (1) a donut chart of pneumonia probability plus a bar chart of whole-lung/left/right infection ratios; (2) lesion volume (log scale, with the common magnitude of a significant lesion marked as reference to show how small this focus is) plus lesion slice-number positioning within the whole series. Configure a CJK font to avoid tofu glyphs, reference images by relative path, and keep the style clean and professional — green for normal/healthy, red for abnormal cues (per China medical convention).\n\nRequirements:\n- Every number must come from the real API response; never infer or fill in missing values. If the verdict conflicts with the probability, present it faithfully and explain it — do not alter numbers to look nice\n- Do not write real absolute paths; render the image path as \"Desktop/{folder name}\"\n- Open with the notice: this model is licensed for research use in mainland China only, must not be used for clinical diagnosis or treatment, and the result is not a diagnostic opinion", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "正常样本筛查", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "Normal Case Screening", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 筛查报告 · 正常受试者（HTML）" }, { "file": "infection-ratio.png", "label": "肺炎表征概率与感染占比图" }, { "file": "lesion-distribution.png", "label": "病灶体积与层号定位图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "usage_count": 0, "created_at": "2026-09-02T11:40:00+08:00", "updated_at": "2026-09-02T11:40:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "pneumonia-ct-ai-quantitative-report", "title": "胸部 CT 肺炎 AI 量化分析报告", "title_en": "Chest CT Pneumonia AI Quantitative Report", "subtitle": "上传 DICOM 影像，自动输出肺炎表征概率、病灶体积与可视化图表", "subtitle_en": "Upload DICOM images to auto-generate pneumonia probability, lesion volumes and charts", "description": "科研场景下做胸部 CT 影像量化，最耗时的往往不是分析本身，而是数据准备：DICOM 目录里混着隐藏文件、元数据字段缺失，提交上去只回一句「AI引擎上传失败」，却不告诉你到底哪里不对。这个案例把整条链路跑通了——读元数据、打包校验、提交轮询、结果整理成结构化报告，并生成两张可视化图表。更关键的是，遇到 SeriesNumber 为空导致的提交失败时，专家会自己诊断出根因、补齐字段后重新提交，不需要你介入。适合需要批量量化影像并归档结果的科研人员、医疗 AI 产品经理。", "description_en": "In research workflows, the real time sink in chest CT quantification is rarely the analysis itself — it is data preparation. DICOM folders carry hidden files, metadata fields go missing, and the API replies with a bare \"AI engine upload failed\" without telling you what is actually wrong. This case runs the whole chain: read metadata, package and validate, submit and poll, then turn the response into a structured report with two charts. More importantly, when submission fails because SeriesNumber is empty, the expert diagnoses the root cause, backfills the field and resubmits on its own. Built for researchers and medical-AI product managers who need to quantify imaging in batches and archive the results.", "artifact_type": "html", "prompt": "这是一次肺部的肺炎 CT 检查，帮我基于腾讯觅影肺炎AI评估专家分析一下这例检查的肺炎情况。\n\n影像目录：{胸部CT的DICOM目录路径，如 桌面/P002}\n\n请按以下步骤完成，最终输出一份 HTML 分析报告（用于灵感案例展示，排版需清晰美观、含基础样式）：\n\n1. **读取影像元数据**：提取患者 ID、性别、年龄、模态、层数、层厚、像素间距、设备厂商、传输语法，汇总成表格。\n\n2. **打包并校验**：将目录内 DICOM 打包为 ZIP，只收 .dcm 文件（排除 .DS_Store、__MACOSX、.WeDrive 等无关文件），打包后校验包内文件数与纯净度。\n\n3. **提交分析并轮询**：提交时开启 PDF 报告并携带患者信息。提交后后台轮询等待，轮询期间持续反馈进度（已等待时长、当前状态），不要静默等待。\n\n4. **失败自动修复**：若返回 90002「AI引擎上传失败」，先诊断 DICOM 关键元数据是否缺失——最常见原因是 SeriesNumber 为空。确认后直接自动补齐并重新提交，不要把失败抛回给我等指示。只补元数据，不得修改影像像素数据。\n\n5. **输出报告**，包含章节：案例概览（患者与检查信息、taskId、分析耗时）／AI 分析结果（肺炎表征概率、全肺与左右肺感染占比、每处病灶的肺叶位置·层号范围·体积·体积占比·平均 CT 值）／结果解读（病灶集中度、CT 值密度特征、分布模式）／健康提示（通用科普口径，不给具体临床建议，凡涉及诊断用药随访一律引导咨询医师）／完整调用流程（可复现命令）／排障记录（问题、排查路径、根因）。\n\n6. **生成可视化图表**：用 matplotlib 生成两张图并嵌入报告——① 肺炎表征概率环形图 + 全肺/左肺/右肺感染占比条形图；② 各病灶体积对比（体积差异悬殊，须用对数轴）+ 病灶 CT 层号纵向分布。设置中文字体避免乱码，图片与报告同目录、用相对路径 `` 标签引用。\n\n要求：\n- 所有数值必须来自 AI 接口返回的真实结果，不得推测或填充缺失值\n- 报告中不写真实绝对路径，影像路径统一写为「桌面/{目录名}」\n- 开篇标注：本模型仅限中国大陆地区科研用途，不得用于临床检测与诊疗", "prompt_en": "This is a chest CT scan for pneumonia. Please analyze this study using the Tencent Miying Pneumonia AI Assessment Expert.\n\nImage folder: {path to the chest CT DICOM folder, e.g. Desktop/P002}\n\nComplete the following steps and deliver an HTML report (clean, styled, suitable for a showcase):\n\n1. **Read image metadata**: extract patient ID, sex, age, modality, slice count, slice thickness, pixel spacing, manufacturer and transfer syntax; summarize in a table.\n\n2. **Package and validate**: zip the DICOM files in the folder, including only .dcm files (exclude .DS_Store, __MACOSX, .WeDrive and other irrelevant files), then verify the file count and purity of the archive.\n\n3. **Submit and poll**: enable the PDF report and pass patient info on submission. Poll in the background and keep reporting progress (elapsed time, current status) — do not wait silently.\n\n4. **Auto-fix on failure**: if the API returns 90002 \"AI engine upload failed\", first check whether key DICOM metadata is missing — the most common cause is an empty SeriesNumber. Once confirmed, backfill the field and resubmit automatically instead of handing the failure back to me. Only touch metadata; never modify pixel data.\n\n5. **Produce the report** with these sections: Case overview (patient and study info, taskId, elapsed time) / AI results (pneumonia probability, whole-lung and per-side infection ratios, and for each lesion its lobe, slice range, volume, volume ratio and mean CT value) / Interpretation (lesion concentration, CT density characteristics, distribution pattern) / Health notes (general guidance only — no specific clinical advice; always defer diagnosis, medication and follow-up to a physician) / Reproducible workflow / Troubleshooting log (symptom, investigation path, root cause).\n\n6. **Generate charts**: use matplotlib to produce two figures and embed them — (1) a donut chart of pneumonia probability plus a bar chart of whole-lung/left/right infection ratios; (2) lesion volume comparison (volumes differ by orders of magnitude, so use a log scale) plus lesion distribution across CT slice numbers. Configure a CJK font to avoid tofu glyphs, and reference the images by relative path next to the report.\n\nRequirements:\n- Every number must come from the real API response; never infer or fill in missing values\n- Do not write real absolute paths in the report; render the image path as \"Desktop/{folder name}\"\n- Open with the notice: this model is licensed for research use in mainland China only and must not be used for clinical diagnosis or treatment", "skills": [], "experts": [ { "id": "pneumonia-ai-analyst", "name": "觅影肺炎AI评估专家", "promptFileSnapshot": "e29272908e24", "expert_id": "ex_5TB0wBnvRZHH" } ], "mcps": [], "categories": [ "data" ], "tags": [ "医疗影像", "肺炎AI", "DICOM", "CT量化分析", "腾讯觅影" ], "tags_en": [ "Medical Imaging", "Pneumonia AI", "DICOM", "CT Quantification", "Tencent Miying" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "肺炎 AI 量化分析报告（HTML）" }, { "file": "infection-ratio.png", "label": "肺炎表征概率与感染占比图" }, { "file": "lesion-distribution.png", "label": "病灶体积与层号分布图" } ], "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-26T17:45:00+08:00", "updated_at": "2026-08-28T16:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "points-redemption-coupon-plan", "title": "积分兑换与发券方案", "title_en": "Points Redemption and Coupon Plan", "subtitle": "会员积分兑换、商品池与批量发券方案", "subtitle_en": "A practical Weimob playbook for points redemption and coupon plan", "description": "做积分兑换与发券方案时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“积分兑换与发券方案”展开：分析会员积分分布，设计积分兑换规则和兑换商品池，并为指定等级的会员批量发放优惠券和权益。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、积分与会员权益运营人员、经营分析团队。", "description_en": "The hardest part of points redemption and coupon plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates points redemption and coupon plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Points and Membership Benefits specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "会员积分", "积分兑换", "兑换商品池", "批量发券", "会员权益" ], "tags_en": [ "Points Redemption and Coupon Plan", "Customer Operations", "Points and Membership Benefits", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我分析会员积分分布，设计积分兑换规则和兑换商品池，并为指定等级的会员批量发放优惠券和权益。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a points redemption and coupon plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "member-ops", "name": "会员权益运营", "description": "查询积分与储值方案、流水和客户权益变动，支持按客户信息、业务编号及明确时间范围筛选。", "skill_id": "skill_2087080649681588224" }, { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "policy-chain-reaction-digest", "title": "政策与资产联动专题页", "title_en": "Policy and Asset Chain-Reaction Digest", "subtitle": "把政策、能源、商品、汇率与市场影响串成可复盘专题", "subtitle_en": "Connect policy, energy, commodities, FX and market impact in one replayable briefing", "description": "遇到政策、能源和跨资产新闻同一天爆发时，最容易犯的错就是只看 headline，不看它们之间的传导关系。结果是新闻看了不少，真正能指导判断的框架却没有搭起来。\n\n这个案例演示如何把 2026-07-27 00:00–20:00 HKT 的政策、国际宏观、能源、商品与资本市场新闻统一抓回，做时间过滤和去重后，整理成一份双语 HTML 专题页。页面会给出主线判断、联动链路图、章节化深度摘要，以及不同情景下的观察重点。\n\n适合：投研周会、投委会 briefing、市场策略复盘、行业研究内容共创、需要把“消息—逻辑—资产”一条线讲明白的业务团队。", "description_en": "When policy, energy and cross-asset headlines all break on the same day, the easiest mistake is to read the headlines one by one without mapping the transmission path between them. You end up seeing plenty of news, but without a decision-ready framework.\n\nThis case shows how to pull policy, global macro, energy, commodities and capital-market coverage for 2026-07-27 between 00:00 and 20:00 HKT, apply time filtering and deduplication, and turn the result into a bilingual HTML briefing page. The page provides a core thesis, a chain-reaction map, chapter-based deep summaries and the key checkpoints under different scenarios.\n\nBest for: research weekly meetings, investment-committee briefings, strategy reviews, collaborative industry-research content and any business team that needs to explain one full line from headline to logic to assets.", "categories": [ "finance", "search", "data" ], "tags": [ "政策链路", "能源转型", "商品", "外汇", "双语HTML" ], "tags_en": [ "Policy Chain", "Energy Transition", "Commodities", "FX", "Bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请抓取 2026-07-27 00:00–20:00 HKT 的政策、国际宏观、能源产业、商品/外汇和资本市场新闻，做时间过滤和去重后，输出一份单文件 HTML 专题页。报告需要包含：核心判断、抓取与去重统计、政策到资产的联动链路、按章节整理的重点新闻、每条新闻的时间/简述/市场含义、原文链接，以及几个可复盘的观察情景。内容必须基于真实查询结果并注明时间与来源；HTML 资源全内联。", "prompt_en": "Please pull policy, global macro, energy-industry, commodities/FX and capital-market news for 2026-07-27 between 00:00 and 20:00 HKT. After time filtering and deduplication, generate a single-file HTML briefing page. The report should include a core thesis, collection and deduplication stats, a policy-to-assets chain-reaction map, chapter-based highlights, the time/summary/market implication of each item, original article links, and several reusable observation scenarios. The content must be based on real query results with timestamps and sources noted; all HTML assets must be inlined.", "skills": [ { "id": "daily-financial-news", "name": "每日财经新闻", "description": "财经新闻多源接入与深度解读，经去重分类后逐条深度概括，附市场影响解读。", "skill_id": "skill_2076628269121765376" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "政策与资产联动双语专题页" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T19:43:22+08:00", "updated_at": "2026-07-27T20:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "postclose-sentiment-battlecard", "title": "收盘后情绪作战卡", "title_en": "Post-Close Sentiment Battlecard", "subtitle": "把指数、涨跌分布、连板梯队与次日观察压缩成一页作战面板", "subtitle_en": "Compress indices, breadth, the limit-up ladder, and next-session watchpoints into a one-page battle board", "description": "盘后最难的不是把数据抄一遍，而是判断今天这根阴线到底是情绪彻底崩掉，还是高弹性方向在杀估值、防守方向在接力。这个案例用大盘走势研判 skill 的日频情绪流路径，把三大指数、涨跌分布、板块强弱、连板梯队、机构席位和热股反馈放进一张可直接复用的收盘面板里。\n\n演示内容是一份单文件双语 HTML，不只给“今天强/弱”的口头判断，而是先给一句话结论，再拆出指数窗口、宽度结构、板块承接、连板梯队和次日观察清单，并尽量在每一段旁边标注来源与日期，方便继续做同类复盘。\n\n适合：每天收盘后要快速定调的短线交易者、负责给团队同步市场情绪的复盘同学、需要把盘面讲成结构化故事的投资内容作者、想把“局部火种还在不在”讲清楚的 A 股活跃参与者。", "description_en": "The hardest part after the close is not copying the numbers. It is deciding whether today’s selloff means sentiment fully broke down, or whether high-beta names are still being derated while defensive areas are taking over. This case uses the daily sentiment-flow path of the Market Trend Assessment skill to put the three major indices, breadth, sector strength, the limit-up ladder, institutional seats, and hot-stock feedback into one reusable post-close board.\n\nThe demo is a single-file bilingual HTML report. It does not stop at saying the market was strong or weak. It starts with a one-line call, then breaks down the index window, breadth structure, sector absorption, the ladder, and next-session watchpoints, while adding sources and dates next to each section whenever possible so the same layout can be reused for future recaps.\n\nSuitable for: short-term traders who need a fast post-close read, teammates responsible for summarizing market sentiment, investment content creators who need to turn the tape into a structured story, and active A-share participants who want to explain whether local speculative fire is still alive.", "categories": [ "finance", "data", "search" ], "tags": [ "收盘复盘", "情绪流", "连板梯队", "次日观察", "A股短线" ], "tags_en": [ "Post-Close Review", "Sentiment Flow", "Limit-Up Ladder", "Next-Session Watch", "A-Share Short-Term" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "收盘后情绪作战卡" } ], "prompt": "请用大盘走势研判 skill 做一份“收盘后情绪作战卡”A 股报告，重点是用日频情绪 + 涨停梯队 + 次日观察的路径，快速判断当日市场是修复、震荡还是继续走弱。先遵循 wb-finance-skill 的红线、时间口径与 HTML 规范，再通过 westock-data / westock-tool 的官方能力查询真实数据，至少覆盖市场涨跌分布、三大指数收盘与日内区间、板块排行、连板梯队、主力 5 日净流入、机构龙虎榜、热搜股票，以及核心观察股当日反馈。输出单文件 HTML，资源全内联、可离线打开。页面做成明亮的信息终端与作战面板风格，包含一句话结论、指数窗口、宽度结构、板块承接、连板梯队、次日观察和数据来源清单，文末加入“不构成投资建议”的合规免责声明。", "prompt_en": "Use the Market Trend Assessment skill to create an A-share report titled \"Post-Close Sentiment Battlecard\". The focus is to use the daily sentiment, limit-up ladder, and next-session-watch path to judge whether the market is repairing, churning, or weakening further. First follow the red lines, time conventions, and HTML rules of wb-finance-skill, then query real data through the official capabilities of westock-data and westock-tool. At minimum, cover market breadth, the close and intraday range of the three major indices, sector ranking, the limit-up ladder, five-day main-force inflows, institutional leaderboard data, hot stocks, and same-day feedback from the key watch names. Output a single-file HTML report with all resources inlined so it can work offline. Use a bright information-terminal and battle-board direction, and include a one-line conclusion, an index window, breadth structure, sector absorption, the limit-up ladder, next-session watchpoints, and a source manifest. End the page with a compliance disclaimer stating that the report is not investment advice.", "skills": [ { "id": "market-trend-assessment", "name": "大盘走势研判", "description": "A股大盘走势研判，三套框架：短线情绪流、周线阶段论、行为金融。", "skill_id": "skill_2076627841600368640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-07-28T17:34:30+08:00", "updated_at": "2026-07-28T17:50:00+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "ppt-ai-history-timeline", "title": "人工智能简史：从图灵到大模型", "title_en": "A Brief History of AI: From Turing to Large Models", "subtitle": "科技编年史风格的 AI 发展历程科普幻灯片", "subtitle_en": "A tech-chronicle style slide deck on the evolution of AI", "description": "面向科普分享、课堂导入与团队内训的场景：需要把人工智能七十余年的发展历程讲得清晰、准确又有感染力。本案由「布道道」开发者布道师专家操刀，以「先抛出问题、再沿时间轴讲演进」的叙事方法论，将图灵测试、达特茅斯会议、深蓝、AlexNet、AlphaGo、Transformer、GPT 到 ChatGPT 等真实里程碑串成一条发光的时间线。交付物是一份可翻页的单文件 HTML 幻灯片（深色未来感时间线视觉、自实现导航），共 14 页，含封面与算力页大图。适合教师、技术布道者、产品与运营人员快速做一场有质感的 AI 科普演示。", "description_en": "For science-popularization talks, classroom intros, and internal training where you need to explain seventy years of AI history clearly, accurately, and compellingly. Crafted with the methodology of the 'Kip' Developer Evangelist expert—pose a question first, then walk a glowing timeline—this deck threads real milestones from the Turing Test, Dartmouth Conference, Deep Blue, AlexNet, AlphaGo, Transformer, and GPT to ChatGPT. The deliverable is a single-file, page-flippable HTML slide deck (dark futuristic timeline visuals, self-built navigation) of 14 slides, with full-bleed images on the cover and compute pages. Ideal for teachers, evangelists, product and marketing people who need a polished AI overview fast.", "categories": [ "office" ], "tags": [ "幻灯片", "人工智能科普", "发展历程", "时间线" ], "tags_en": [ "Slide Deck", "AI Explainer", "History Timeline", "Tech Chronicle" ], "artifact_type": "html", "prompt": "帮我做一个关于「人工智能发展历程」的 HTML 幻灯片演示，可以左右翻页，大概 13 页左右，一屏一页 16:9。内容要覆盖从 1950 图灵测试、1956 达特茅斯会议、感知机、专家系统与 AI 寒冬、1997 深蓝、2012 AlexNet、2016 AlphaGo、2017 Transformer、GPT 预训练与 GPT-3、2022 ChatGPT 到当下的 Agent、多模态、端侧 AI。年份和事件务必真实准确。风格要科技编年史、深色未来感时间线，有发光节点和时间轴，封面和结尾要有设计感，不要用 reveal.js 或任何外链。", "prompt_en": "Make me an HTML slide deck about the history of artificial intelligence, page-flippable left/right, around 13 slides, one 16:9 slide per screen. Cover the milestones from the 1950 Turing Test, the 1956 Dartmouth Conference, the Perceptron, expert systems and the AI winters, 1997 Deep Blue, 2012 AlexNet, 2016 AlphaGo, 2017 Transformer, GPT pretraining and GPT-3, 2022 ChatGPT, through to today's Agents, multimodality, and on-device AI. Years and events must be historically accurate. Use a tech-chronicle, dark futuristic timeline style with glowing nodes and a timeline axis; the cover and closing must look designed. No reveal.js or any external resources.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "人工智能简史演示幻灯片" } ], "skills": [], "experts": [ { "id": "DeveloperEvangelist", "name": "布道道", "description": "构建发展开发者社区，推动产品在开发者群体中的采用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c93ef6310507", "name_en": "Kip", "description_en": "Builds and grows developer communities, driving product adoption among developers", "expert_id": "ex_iiIjnlq0IsVF" } ], "mcps": [], "scenario": "幻灯片", "scenario_id": 105, "scenario_en": "Slides Creation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T12:00:00Z", "updated_at": "2026-06-22T12:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ppt-journey-west-intro", "title": "《西游记》名著介绍 PPT", "title_en": "Journey to the West: A Classic Introduction Deck", "subtitle": "国风水墨卷轴风格，一部降妖伏魔、修心问道的奇幻史诗", "subtitle_en": "Chinese ink-and-scroll style intro to the fantasy epic of demon-quelling and self-cultivation", "description": "想给孩子、读书会或文化课做一份既好看又靠谱的名著介绍？市面模板要么千篇一律，要么文学常识含糊出错。本案例由「幻灯灯」(PptCreationExpert) 以专业 PPT 方法论，围绕中国古典四大名著之一《西游记》制作了一套 12 页可翻页 HTML 幻灯片：从封面题解、明代吴承恩的成书背景，到长安西天的取经主线、师徒五人（孙悟空、唐僧、猪八戒、沙僧、白龙马）的性格与法宝，再到大闹天宫等经典篇章、艺术成就与主题思想，文学事实准确、逻辑清晰。视觉采用中国古典水墨／国风卷轴：宣纸米色底、水墨远山祥云、朱砂红印章与楷书感字体，古雅有韵。交付物为单文件 HTML（自实现键盘／点击／圆点翻页，无任何外链），开箱即用。适合语文教师、读书会主持人、文化科普创作者与家长。", "description_en": "Need a beautiful yet accurate intro to a literary classic for students, book clubs, or culture classes? Most templates are generic or get the facts wrong. In this case, the expert 'Slide' (PptCreationExpert) applies professional PPT methodology to build a 12-slide, page-flippable HTML deck on Journey to the West, one of China's Four Great Classical Novels: from cover and premise, the Ming-dynasty author Wu Cheng'en and the book's origin, to the Chang'an-to-the-West pilgrimage, the five companions (Sun Wukong, Tang Sanzang, Zhu Bajie, Sha Wujing, the White Dragon Horse) with their traits and treasures, plus iconic chapters like Havoc in Heaven, artistic achievements and themes. The visual language is classical Chinese ink and scroll: rice-paper beige, ink mountains and auspicious clouds, cinnabar seals and serif/Kai-style type. Delivered as a single self-contained HTML file with self-built keyboard/click/dot navigation and no external resources. Ideal for language teachers, book-club hosts, culture creators, and parents.", "categories": [ "office" ], "tags": [ "幻灯片", "西游记", "名著科普", "国风水墨" ], "tags_en": [ "slide deck", "Journey to the West", "literary classic", "Chinese ink style" ], "artifact_type": "html", "prompt": "帮我做一个介绍《西游记》的 HTML 幻灯片演示，可以翻页，大概 12 页一屏一页。内容要文学常识准确：封面、一句话题解、作者吴承恩和成书背景、长安到西天的取经主线、孙悟空/唐僧/猪八戒/沙僧+白龙马五个人物（性格和法宝）、大闹天宫等经典篇章、艺术成就与影响、主题思想、结尾名句。风格要中国古典水墨、国风卷轴感：宣纸米色底、水墨远山祥云、朱砂红印章、楷书感字体，古雅有文化味，不要用 emoji，单文件能直接打开翻页。", "prompt_en": "Make me a page-flippable HTML slide deck introducing Journey to the West, about 12 slides, one screen per slide. The literary facts must be accurate: cover, one-line premise, author Wu Cheng'en and the book's origin, the Chang'an-to-the-West pilgrimage, the five companions (Sun Wukong / Tang Sanzang / Zhu Bajie / Sha Wujing + White Dragon Horse) with personalities and treasures, iconic chapters like Havoc in Heaven, artistic achievements and influence, themes, and a closing quote. Use a classical Chinese ink / scroll aesthetic: rice-paper beige background, ink mountains and clouds, cinnabar seals, serif/Kai-style fonts, elegant and cultured, no emoji, single self-contained file that flips pages.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "《西游记》名著介绍幻灯片" } ], "skills": [], "experts": [ { "id": "PptCreationExpert", "name": "幻灯灯", "description": "智能 PPT 生成助手，一键将想法转化为精美演示文稿", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b59304464bc5", "name_en": "Slide", "description_en": "Intelligent PPT generation assistant that turns ideas into polished presentations in one click", "expert_id": "ex_xagtUG00hwPD" } ], "mcps": [], "scenario": "幻灯片", "scenario_id": 105, "scenario_en": "Slides Creation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T12:00:00Z", "updated_at": "2026-06-22T12:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ppt-popmart-brand-intro", "title": "泡泡玛特品牌与潮玩商业模式介绍", "title_en": "POP MART: Brand & Designer-Toy Business Model", "subtitle": "盲盒、IP 飞轮与情绪消费的潮玩帝国", "subtitle_en": "Blind boxes, the IP flywheel and the rise of emotion-driven toy retail", "description": "想给同事或客户快速讲清楚泡泡玛特为什么火？本案由「幻灯灯」PPT 制作专家操刀，把「盲盒模式、头部 IP 矩阵、IP 运营飞轮、渠道与全球化、增长逻辑与风险」梳理成一份逻辑清晰、视觉统一的潮玩波普风演示文稿。交付物是一份可直接翻页浏览的单文件 HTML 幻灯片（约 11 页，16:9，支持方向键/空格/点击/圆点导航），用 CSS 绘制盲盒、概率条与 IP 卡片，无需任何外部资源。适合做品牌科普、商业模式分享、行业案例教学，也方便二次编辑。其中数字均以量级/示意呈现，避免编造精确财报。", "description_en": "Want to quickly explain why POP MART took off? Built with the methodology of the 'Slide' PPT-creation expert, this deck lays out the blind-box model, flagship IP matrix, IP operating flywheel, channels and globalization, plus growth logic and risks in a clean, visually consistent pop-art style. The deliverable is a single-file, page-flippable HTML slide deck (about 11 slides, 16:9, with arrow/space/click/dot navigation). Blind boxes, probability bars and IP cards are all drawn in CSS with zero external resources. Great for brand explainers, business-model talks and case-study teaching, and easy to edit. Figures are shown as orders of magnitude / illustrative only.", "categories": [ "office" ], "tags": [ "幻灯片", "泡泡玛特", "商业模式", "潮玩盲盒" ], "tags_en": [ "Slides", "POP MART", "Business Model", "Designer Toys" ], "artifact_type": "html", "prompt": "帮我做一个介绍泡泡玛特（POP MART）品牌和潮玩商业模式的 HTML 幻灯片演示，可以翻页，大概 11 页：封面、一句话定位、公司速览、盲盒模式解析（带抽盒概率条）、头部 IP 矩阵（Molly/Labubu/Dimoo/SkullPanda）、IP 运营飞轮、渠道矩阵、全球化扩张、增长驱动逻辑、风险与展望、结尾页。风格要潮玩波普风、明亮撞色、圆角卡片、俏皮现代，用 CSS 画盲盒和图标不要用图片，财务数字请联网搜索泡泡玛特真实财报/公告数据（注明年份与来源），不要编造。", "prompt_en": "Make me a page-flippable HTML slide deck introducing POP MART's brand and its designer-toy business model, about 11 slides: cover, one-line positioning, company snapshot, blind-box model (with draw-probability bars), flagship IP matrix (Molly/Labubu/Dimoo/SkullPanda), the IP operating flywheel, channel matrix, global expansion, growth drivers, risks & outlook, and a closing slide. Use a trendy pop-art toy style with bright clashing colors, rounded cards and a playful modern feel; draw the blind boxes and icons in CSS instead of images, and keep any financial figures as orders of magnitude or real (web-searched).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "泡泡玛特品牌介绍幻灯片" } ], "skills": [], "experts": [ { "id": "PptCreationExpert", "name": "幻灯灯", "description": "智能 PPT 生成助手，一键将想法转化为精美演示文稿", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b59304464bc5", "name_en": "Slide", "description_en": "Intelligent PPT generation assistant that turns ideas into polished presentations in one click", "expert_id": "ex_xagtUG00hwPD" } ], "mcps": [], "scenario": "幻灯片", "scenario_id": 105, "scenario_en": "Slides Creation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T12:00:00Z", "updated_at": "2026-06-22T12:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ppt-product-launch-keynote", "title": "星澜 Stellar 春季新品发布会 Keynote", "title_en": "Stellar Spring Product Launch Keynote", "subtitle": "暗色影院级 · AI 智能穿戴新品发布演示", "subtitle_en": "Cinematic dark keynote for AI wearable launch", "description": "新品发布会需要仪式感与清晰节奏：从愿景到产品、从亮点到价格，一气呵成地讲好故事。本案由「幻灯灯」专家以 Apple 风暗色影院级 keynote 方法论，为虚构品牌「星澜 Stellar」打造旗舰 AI 智能手表 Watch X 与降噪耳机 Buds Pro 的发布演示：纯黑底、超大轻量标题、产品大图聚光叠渐变蒙版压字，配端侧 AI、健康监测、续航等核心卖点与醒目价格页。交付物是一份单文件、可翻页（方向键/空格/点击/圆点）的 HTML 幻灯片，共 11 页 16:9。适合做产品发布、品牌路演与对外演讲的演示骨架。", "description_en": "A product launch needs ceremony and rhythm—from vision to product, highlights to pricing, told as one story. Using a cinematic Apple-style dark keynote methodology, the Slide expert built a launch deck for the fictional brand Stellar: pitch-black canvas, oversized light titles, spotlit product hero shots with gradient veils for legible text, plus on-device AI, health monitoring, battery life and a bold pricing page. The deliverable is a single-file, navigable (arrows/space/click/dots) HTML slide deck of 11 16:9 pages—ideal as a skeleton for product launches, brand roadshows and keynote talks.", "categories": [ "office" ], "tags": [ "幻灯片", "产品发布会", "Keynote", "AI智能穿戴" ], "tags_en": [ "Slides", "Product Launch", "Keynote", "AI Wearables" ], "artifact_type": "html", "prompt": "帮我做一份产品发布会的 HTML 幻灯片演示，可翻页（方向键/空格/点击/底部圆点），大约 12 页，一屏一页 16:9。主题是虚构品牌「星澜 Stellar」春季发布旗舰 AI 智能手表和 AI 降噪耳机。要有发布会节奏：封面、愿景开场、手表主角登场大图、三大亮点（14天续航/ECG/端侧大模型）、性能规格、第二产品耳机、耳机亮点（降噪dB/空间音频/实时翻译）、生态联动、醒目价格与发售页、One more thing 惊喜功能、品牌结尾。风格要 Apple 风暗色影院级 keynote：纯黑底、超大轻量标题、大量负空间、电光蓝高光、产品大图叠暗色渐变蒙版压字，极简高级、无 emoji。", "prompt_en": "Make me a product-launch HTML slide deck that's navigable (arrows/space/click/bottom dots), about 12 pages, one 16:9 slide per screen. Theme: the fictional brand Stellar launching a flagship AI smartwatch and AI noise-cancelling earbuds this spring. Keep a launch rhythm: cover, vision opener, watch hero shot, three highlights (14-day battery / ECG / on-device LLM), specs, the second product earbuds, earbud highlights (ANC dB / spatial audio / live translation), ecosystem, a bold pricing & release page, a One-more-thing surprise feature, and a brand closing. Style it as an Apple-like cinematic dark keynote: pitch-black canvas, oversized light titles, lots of negative space, electric-blue accents, hero product photos with dark gradient veils for legible text—minimal, premium, no emoji.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "星澜 Stellar 发布会 Keynote 幻灯片" } ], "skills": [], "experts": [ { "id": "PptCreationExpert", "name": "幻灯灯", "description": "智能 PPT 生成助手，一键将想法转化为精美演示文稿", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b59304464bc5", "name_en": "Slide", "description_en": "Intelligent PPT generation assistant that turns ideas into polished presentations in one click", "expert_id": "ex_xagtUG00hwPD" } ], "mcps": [], "scenario": "幻灯片", "scenario_id": 105, "scenario_en": "Slides Creation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T12:00:00Z", "updated_at": "2026-06-22T12:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ppt-year-end-business-review", "title": "企业年度总结与来年规划汇报 PPT", "title_en": "Year-End Business Review & Next-Year Strategy Deck", "subtitle": "云策科技 2025 年度经营回顾 + 2026 战略，可翻页 HTML 幻灯片", "subtitle_en": "YunCe Tech 2025 review + 2026 strategy, a navigable HTML slide deck", "description": "年度董事会与全员大会上，一份讲得清、看得信的经营汇报往往决定了团队的下一步共识。本案以「策必中」(Tatum) 销售提案与方案设计专家方法论，把一年业绩讲成有说服力的商业叙事：用统一的赢单主题串起核心成绩、营收趋势、业务结构与客户口碑，再以三幕式结构从复盘走向 2026 战略、量化目标与季度路线图。交付物是一份单文件、可翻页的 HTML 幻灯片（约 13 页，16:9，自带方向键/空格/点击/圆点导航），深蓝配品牌金的董事会汇报质感，KPI 卡与商务图表数字内部自洽。适合创始人、CEO 办公室、战略与运营负责人做年度复盘与战略沟通。", "description_en": "At a board meeting or all-hands, a clear and credible business review shapes the team's next consensus. Using the proposal-strategist methodology of expert 'Tatum', this case turns a year of results into a persuasive business narrative: a unified win theme threads through annual highlights, revenue trends, business mix and customer NPS, then a three-act arc moves from honest reflection into the 2026 strategy, quantified goals and a quarterly roadmap. The deliverable is a single-file, navigable HTML slide deck (~13 slides, 16:9, with arrow/space/click/dot navigation) in a boardroom-grade navy-and-gold style, with self-consistent KPI cards and business charts. Ideal for founders, the CEO office, and strategy or operations leads.", "categories": [ "office" ], "tags": [ "幻灯片", "年度总结", "经营汇报", "战略规划" ], "tags_en": [ "Slide Deck", "Year-End Review", "Business Report", "Strategy Planning" ], "artifact_type": "html", "prompt": "帮我做一份企业年度总结与来年规划的 HTML 幻灯片演示，可翻页（方向键/空格/点击/底部圆点），大约 13 页一屏一页 16:9。内容是虚构 SaaS 公司「云策科技」2025 年度经营回顾加 2026 战略：封面、议程、4 个核心 KPI 大数字、四季度营收趋势图、各业务线收入占比、关键里程碑时间线、客户与 NPS、团队组织、挑战反思、2026 三大战略方向、量化目标、季度路线图、结尾愿景。风格要沉稳专业的企业商务董事会汇报质感，深蓝配品牌金，干净可信，图表坐标轴要容纳数据、占比合计 100%，涨红跌绿，不要 emoji。", "prompt_en": "Create a single-file, navigable HTML slide deck for a company year-end review and next-year plan (arrow keys / space / click / bottom dots), about 13 slides, one screen each at 16:9. Content is a fictional SaaS company 'YunCe Tech': 2025 business review plus 2026 strategy — cover, agenda, four headline KPIs, quarterly revenue trend chart, revenue mix by business line, milestone timeline, customers & NPS, team & org, challenges & reflection, three 2026 strategic directions, quantified goals, quarterly roadmap, and a closing vision. Style it as a calm, professional boardroom business deck in navy with brand gold — clean and credible, with chart axes that fit the data, shares summing to 100%, China's convention of red-up/green-down, and no emoji.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "云策科技 2025 年度总结暨 2026 战略汇报" } ], "skills": [], "experts": [ { "id": "ProposalStrategist", "name": "策必中", "description": "精通销售提案和方案设计，将价值转化为无法拒绝的商业论证", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "36bbc2f13354", "name_en": "Tatum", "description_en": "Masters sales proposals and solution design, turning value into business cases that are hard to refuse", "expert_id": "ex_PZw8Gu81HfN4" } ], "mcps": [], "scenario": "幻灯片", "scenario_id": 105, "scenario_en": "Slides Creation", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T12:00:00Z", "updated_at": "2026-06-22T12:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "private-community-ops-hub", "title": "私域社群运营台", "title_en": "Private Community Ops Hub", "subtitle": "社群资产、成员分层、运营日历、转化漏斗和话术库一站搞定", "subtitle_en": "Community assets, member tiers, ops calendar, conversion funnel and scripts in one place", "description": "做私域运营最头疼的，是手里几个群、几百上千号人，全靠脑子记和翻聊天记录——哪个群快凉了、谁是铁粉谁快流失、这周该发什么活动、从进群到付费到底漏在哪一环，永远算不清。\n\n这个案例做出来的是一个单文件网页运营台：社群资产管理（每个群的成员数、活跃度、运营阶段一目了然）、成员分层运营（KOC/付费/活跃粉/流失预警分层打标签）、社群运营日历（可视化排期每日话题和活动）、私域转化追踪（进群→活跃→咨询→付费的漏斗，自动算每一环转化率）、月度运营复盘（含内联 SVG 成长曲线）、群话术库（欢迎语/活动通知/转化促单一键复制直接发群）。数据全存浏览器本地，支持导出备份；PC 端侧边栏、手机端底部 Tab，可加到主屏当 App；内置中英双语，链接加 ?lang=en 整页切英文。\n\n适合：管着多个微信群/知识星球的私域操盘手、做付费社群陪跑的知识博主、需要盯转化漏斗的增长运营、给客户做私域代运营的服务商、以及想把粉丝沉淀成长期资产的个人创业者。", "description_en": "The biggest headache in private-traffic ops is having a few groups and hundreds or thousands of people, all tracked by memory and by scrolling chat logs — you can never figure out which group is going cold, who's a loyal fan versus who's about to churn, what event to post this week, or exactly where people drop off between joining and paying.\n\nThis case produces a single-file web ops console: community asset management (each group's member count, activity and lifecycle stage at a glance), tiered member ops (tag KOC / paid / active fans / churn-risk), an ops calendar (visually schedule daily topics and events), private-traffic conversion tracking (a join → active → inquiry → paid funnel that auto-computes each step's rate), monthly ops reviews (with an inline SVG growth curve), and a message script library (welcome / event notice / upsell scripts you can copy and send with one click). All data lives in the browser locally with export backup; a sidebar on desktop, a bottom tab bar on mobile, addable to the home screen as an app; built-in bilingual support flips the whole page to English by appending ?lang=en.\n\nGreat for: private-traffic operators managing multiple WeChat/Zsxq groups, knowledge creators running paid coaching communities, growth ops who need to watch the conversion funnel, agencies running private-traffic ops for clients, and solo founders turning fans into a long-term asset.", "categories": [ "workspace", "business" ], "tags": [ "私域运营", "社群管理", "成员分层", "转化漏斗", "运营复盘" ], "tags_en": [ "Private traffic", "Community mgmt", "Member tiers", "Conversion funnel", "Ops review" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "私域社群运营台" } ], "preview_en": "output.html", "prompt": "请使用资料库的能力帮我一键做同款（https://workbuddy.link/p/VXLAdcLg6c4MkuE4lugnAC），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n帮我做一个「私域社群运营台」——给做私域/社群运营的人用的单文件网页工作台，浏览器直接打开就能用，数据存本地浏览器（localStorage），关掉再打开不丢，支持导出/导入备份。\n\n要包含这几个模块：\n1. 私域概览：今日焦点 + 核心数据卡片（私域总成员、活跃成员、付费转化率、流失预警人数），近期排期和各阵地健康度一览；\n2. 社群资产管理：登记每个社群阵地（平台/名称/成员数/活跃人数/运营阶段/备注），能按平台筛选、增删改；\n3. 成员分层运营：把重点成员按 KOC核心/付费用户/活跃粉/流失预警等层级打标签管理，只记值得单独运营的人；\n4. 社群运营日历：月历视图排期每日话题、每周活动、直播、答疑等，点日期即可添加，带本月排期清单；\n5. 私域转化追踪：进群→活跃→咨询→付费的可视化漏斗，自动算每一环转化率，并对比各社群转化潜力；\n6. 运营复盘：月度数据记录（新增/流失/活跃/新付费）+ 用内联 SVG 画的总成员成长曲线；\n7. 群话术库：按欢迎语/活动通知/转化促单等分类存常用话术，一键复制直接发群。\n\n设计要求：清爽的卡片式运营后台风格，绿色为主色调，信息层级清晰、数据可视化优先（漏斗、成长曲线、健康度进度条都用内联 SVG/CSS 实现）；PC 端侧边栏导航，手机端自动切换成底部 Tab，触控友好，可添加到手机主屏当 App。\n\n技术约束：单个 HTML 文件、CSS/JS 全部内联、不依赖任何外部 CDN 或联网资源、可完全离线打开；提供浅色/深色主题切换。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library skill to replicate this workspace (https://workbuddy.link/p/VXLAdcLg6c4MkuE4lugnAC) with online data storage and bidirectional sync. Prioritize the one-click replication flow; if unavailable, build from scratch:\nHelp me build a \"Private Community Ops Hub\" — a single-file web workspace for people doing private-traffic / community operations. It should run by simply opening it in a browser, store data locally (localStorage), keep the data after closing and reopening, and support export/import backup. Include these modules:\n1. Private-traffic overview: today's focus plus core metric cards (total members, active members, paid conversion rate, churn-risk count), with upcoming schedule and each community's health at a glance;\n2. Community asset management: register each community (platform / name / members / active count / lifecycle stage / note), with filtering by platform and add/edit/delete;\n3. Tiered member ops: tag and manage key members by tier such as Core KOC / Paid / Active fan / Churn risk, logging only those worth dedicated care;\n4. Ops calendar: a month view to schedule daily topics, weekly events, livestreams, Q&A, etc., add by clicking a date, with a monthly schedule list;\n5. Conversion tracking: a visual join → active → inquiry → paid funnel that auto-computes each step's rate and compares each community's conversion potential;\n6. Ops review: monthly data logs (new / lost / active / new paid) plus a total-members growth curve drawn with inline SVG;\n7. Message script library: store common scripts by category such as welcome / event notice / upsell, copy with one click and send to the group. Design requirements: a clean card-based ops-console style with green as the primary color, clear information hierarchy and data-visualization first (funnel, growth curve, health progress bars all implemented with inline SVG/CSS); sidebar navigation on desktop, automatically switching to a bottom tab bar on mobile, touch-friendly and addable to the home screen as an app. Technical constraints: a single HTML file no dependency on any external CDN or online resources, fully openable offline; provide light/dark theme switching.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "name_en": "DeskCraft", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-08T11:20:00+08:00", "updated_at": "2026-07-08T11:20:00+08:00", "scenario": "个人工作台", "scenario_id": 122, "scenario_use": "", "task_mode": "日常办公", "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "procurement-review", "title": "办公设备采购合同风险审查", "title_en": "Office Equipment Procurement Risk Review", "subtitle": "把控大额设备采购中的权属瑕疵、付款节奏与质保陷阱", "subtitle_en": "Control title defects, payment timing, and warranty pitfalls in major equipment procurement", "description": "百万元级办公设备采购涉及多台设备的质量一致性、交货时效性、付款安全性和售后保障。本案例演示专家如何从100台移动工作站采购合同中，揪出签约主体资格证明缺失、授权委托书有效性未验证、标的物权属瑕疵未排除、预付款未绑定发货节点、验收期未设定等高频风险，为采购方提供全链路风控参考。\n\n适合：采购总监、IT资产管理者、财务审批人、供应商管理人员。", "description_en": "Million-RMB office-equipment procurement involves quality consistency across units, delivery timeliness, payment security and after-sales assurance. This case demonstrates how the expert surfaces high-frequency risks in a 100-unit workstation purchase: missing party-qualification proofs, unverified authorization letters, undisclosed title defects, prepayments not tied to delivery milestones, undefined acceptance periods.\n\nSuitable for: procurement directors, IT-asset managers, finance approvers, supplier managers.", "categories": [ "business" ], "tags": [ "签约授权", "设备权属", "验收付款" ], "tags_en": [ "Signing Authority", "Asset Title", "Acceptance & Payment" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请对我提供的《办公设备采购合同》进行全面风险审查。重点审核买卖双方的签约主体资格文件（营业执照、授权书）是否齐全、标的物是否存在抵押查封等权利瑕疵、预付款30%是否应绑定到货发货节点而非仅签约日、验收合格后60日内付尾款是否过长、质保期3年上门服务的响应时间SLA是否明确、退换货条件和不可抗力范围是否合理。输出分级风险报告。", "prompt_en": "Please conduct a comprehensive risk review on my Office Equipment Procurement Agreement. Verify both parties' qualification documents are complete, check whether goods have encumbrances, assess if 30% prepayment should tie to delivery-not-signing, evaluate if 60-day post-acceptance final payment is excessive, confirm 3-year warranty SLA clarity, and validate return/exchange terms. Output a tiered report.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T16:56:11+08:00", "updated_at": "2026-08-13T18:08:23+08:00", "artifacts": [ { "file": "output.html", "label": "基于 V1.3 真实审查任务的互动审查页面" }, { "file": "办公设备采购合同_完整审查报告_V1.6.html", "label": "可视化商务版完整审查报告（HTML）" }, { "file": "办公设备采购合同_完整审查报告_V1.3.md", "label": "完整审查报告数据版（Markdown）" }, { "file": "办公设备采购合同_审查批注_V1.3.docx", "label": "腾讯电子签官方导出的 V1.3 审查批注合同" }, { "file": "办公设备采购合同_审查摘要_V1.3.xlsx", "label": "腾讯电子签官方导出的 V1.3 审查摘要 Excel" }, { "file": "真实审查结果_V1.3.json", "label": "腾讯电子签 V1.3 审查任务原始返回数据" }, { "file": "办公设备采购合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "product-launch-email-series", "title": "产品发布邮件序列策划", "title_en": "Product Launch Email Series", "subtitle": "智能客服SaaS产品的5封完整发布邮件模板", "subtitle_en": "5-Email Launch Sequence for AI Customer Service SaaS", "description": "中小企业在产品发布时常面临邮件营销缺乏系统规划、内容千篇一律、转化率低等痛点。本案例通过email-template-generator技能，为一款智能客服SaaS工具策划了完整的5封产品发布邮件序列，涵盖预热造势、正式发布、功能深度解析、限时优惠促进转化及发布后持续跟进。每封邮件均包含专业的主题行、预览文本、结构化正文和明确的行动号召，帮助企业实现高效的产品上市推广。", "description_en": "SMBs launching new products often struggle with unstructured email campaigns, generic content, and low conversion rates. This case uses the email-template-generator skill to create a complete 5-email product launch sequence for an AI-powered customer service SaaS tool, covering pre-launch teaser, launch announcement, feature deep-dive, limited-time offer reminder, and post-launch follow-up. Each email includes professional subject lines, preview text, structured body content, and clear CTAs to drive effective product launch communication.", "categories": [ "business", "productivity" ], "tags": [ "邮件营销", "产品发布", "SaaS", "客户沟通", "营销文案" ], "tags_en": [ "email marketing", "product launch", "SaaS", "customer communication", "copywriting" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "产品发布邮件序列" } ], "prompt": "帮我策划一套完整的产品发布邮件序列（Product Launch Email Series），包括预热邮件、发布公告、功能亮点介绍、限时优惠提醒和发布后跟进邮件，共5封。产品是一款面向中小企业的智能客服SaaS工具。", "prompt_en": "Help me plan a complete Product Launch Email Series, including a teaser email, launch announcement, feature highlight introduction, limited-time offer reminder, and a post-launch follow-up email, for a total of 5 emails. The product is an intelligent customer service SaaS tool aimed at small and medium-sized businesses.", "skills": [ { "id": "outbound-engine", "name": "outbound-engine", "description": "自动化外拓邮件引擎，从 ICP 定义到邮件入站全流程", "description_en": "Automated cold outbound engine from ICP definition to email delivery", "skill_id": "skill_2053082541772902400" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "product-price-band-sales-mix", "title": "商品价格带成交结构", "title_en": "Product Price-Band Sales Mix", "subtitle": "近30天商品价格带销售与退款结构分析", "subtitle_en": "A practical Weimob playbook for product price-band sales mix", "description": "做商品价格带成交结构时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“商品价格带成交结构”展开：按商品成交价划分价格带，分析近30天各价格带的销量、GMV、客单贡献和退款表现，并提出商品组合与定价优化建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、商品运营运营人员、经营分析团队。", "description_en": "The hardest part of product price-band sales mix is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates product price-band sales mix through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Product Operations specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "商品价格带", "销量", "GMV", "客单贡献", "定价优化" ], "tags_en": [ "Product Price-Band Sales Mix", "Mall Operations", "Product Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "按商品成交价将近30天支付商品划分价格带，分析各价格带的销量、GMV、客单贡献和退款表现，并提出商品组合与定价优化建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a product price-band sales mix covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "product-sales-after-sales-quadrant", "title": "商品销售售后四象限", "title_en": "Product Sales and After-Sales Quadrant", "subtitle": "销量、售后与退款金额四象限经营诊断", "subtitle_en": "A practical Weimob playbook for product sales and after-sales quadrant", "description": "做商品销售售后四象限时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“商品销售售后四象限”展开：结合近 30 天商品销量、支付订单数、售后单量和退款金额，将有成交商品划分为高销量低售后、高销量高售后、低销量低售后和低销量高售后四类。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、商品销售诊断运营人员、经营分析团队。", "description_en": "The hardest part of product sales and after-sales quadrant is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates product sales and after-sales quadrant through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Product Sales Diagnosis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "商品分析", "销量", "支付订单", "售后率", "四象限" ], "tags_en": [ "Product Sales and After-Sales Quadrant", "Business Diagnosis", "Product Sales Diagnosis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "结合近30天商品销量、支付订单数、售后单量和退款金额，把有成交商品划分为四类经营象限，找出高销量高售后和低销量高售后商品，并给出分象限处理建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a product sales and after-sales quadrant covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "product-sales-anomaly-monitoring", "title": "商品异常波动监测", "title_en": "Product Sales Anomaly Monitoring", "subtitle": "近30天商品销量、转化率与退款率异常识别", "subtitle_en": "A practical Weimob playbook for product sales anomaly monitoring", "description": "做商品异常波动监测时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“商品异常波动监测”展开：识别最近 30 天销量、转化率或退款率异常波动的商品，并按照经营影响程度安排处理优先级。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、商品销售诊断运营人员、经营分析团队。", "description_en": "The hardest part of product sales anomaly monitoring is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates product sales anomaly monitoring through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Product Sales Diagnosis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "商品监测", "销量异常", "转化异常", "退款异常", "风险排序" ], "tags_en": [ "Product Sales Anomaly Monitoring", "Business Diagnosis", "Product Sales Diagnosis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我识别最近30天销量、转化率或退款率异常波动的商品，按影响程度排序，分析原因并给出处理建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a product sales anomaly monitoring covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "professional-resume-generator", "title": "简历优化与生成", "title_en": "Professional Resume Generator", "subtitle": "输入个人经历，一键生成设计感在线简历页面", "subtitle_en": "Input your experience, instantly generate a beautifully designed online resume page", "description": "写简历是求职的第一步，但多数人卡在排版和设计上——Word 简历千篇一律，在线简历工具要么收费要么模板老旧。想自己写一个 HTML 简历页面？配色、布局、响应式适配，每一步都是坑。\n\n这个案例演示了如何用「作品集网站」Skill 为一位 3 年经验的产品经理生成专业级在线简历：深蓝+白配色的现代风格，包含个人简介、工作经历时间线（带核心成果高亮）、技能雷达图与进度条、项目亮点数据卡片、教育背景——全部打包在一个响应式 HTML 文件中，支持滚动动画和导航吸顶。\n\n适合：求职者、产品经理、设计师、开发者——任何想要快速拥有一份高颜值在线简历的人。", "description_en": "Writing a resume is the first step in any job search, but most people get stuck on formatting and design — Word resumes all look the same, and online resume builders are either paid or use outdated templates. Building your own HTML resume page? Color schemes, layouts, responsive design — every step is a pitfall.\n\nThis case demonstrates using the Portfolio Website Skill to generate a professional online resume for a product manager with 3 years of experience: modern dark-blue-and-white color scheme, featuring personal intro, work experience timeline with highlighted achievements, skills radar chart and progress bars, project highlight data cards, and education background — all packed in a single responsive HTML file with scroll animations and sticky navigation.\n\nPerfect for: job seekers, product managers, designers, developers — anyone who wants a polished online resume fast.", "categories": [ "office", "education" ], "tags": [ "简历", "求职", "产品经理", "在线简历", "职场" ], "tags_en": [ "Resume", "Job Hunting", "Product Manager", "Online CV", "Career" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "产品经理在线简历" } ], "prompt": "请为一位 3 年经验的产品经理生成一份在线简历页面（单文件 HTML，内联无外部依赖），现代风格、专业配色（深蓝+白）、响应式。\n\n个人信息：张明，27 岁，浙江大学计算机本科，现任某中型互联网公司产品经理，求职意向高级产品经理（B 端 SaaS）。\n工作经历：① 2023.07-至今 某科技公司 产品经理 — 负责企业协作工具，主导实时协作编辑上线（DAU +35%），推动续费率 72%→88%，管理 5 人小组；② 2022.03-2023.06 某创业公司 初级产品经理 — 从 0 到 1 搭建电商 SaaS 后台，MVP 上线获首批 50 家付费客户。\n技能：用户研究、数据分析（SQL/Python）、Axure/Figma 原型、敏捷开发。\n包含个人简介、工作经历时间线、技能雷达图、项目亮点数据卡片、教育背景。", "prompt_en": "Please generate an online resume page for a product manager with 3 years of experience (single-file HTML, inlined with no external dependencies), modern style, professional color scheme (navy + white), responsive.\n\nPersonal info: Zhang Ming, 27, bachelor's in Computer Science from Zhejiang University, currently a product manager at a mid-sized internet company, targeting a Senior Product Manager role (B2B SaaS).\nWork experience: ① 2023.07-present, a tech company, Product Manager — responsible for enterprise collaboration tools, led the launch of real-time collaborative editing (DAU +35%), drove renewal rate from 72% to 88%, managed a team of 5; ② 2022.03-2023.06, a startup, Junior Product Manager — built an e-commerce SaaS admin platform from 0 to 1, with the MVP launch winning the first 50 paying customers.\nSkills: user research, data analysis (SQL/Python), Axure/Figma prototyping, agile development.\nInclude a personal bio, a work-experience timeline, a skills radar chart, project-highlight data cards, and education background.", "skills": [ { "id": "minimax-pdf", "name": "PDF 文档生成", "description": "高质量 PDF 文档生成", "description_en": "Generate high-quality designed PDF documents", "skill_id": "skill_2053082401866706944" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-21T10:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "project-proposal-drafting", "title": "项目立项报告一键起草", "title_en": "Project Proposal Report Drafting", "subtitle": "按模板从零起草立项报告，机构与项目数据自动喂入", "subtitle_en": "Draft a proposal from scratch against your template, with org and project data auto-injected", "description": "公益机构在立项阶段撰写立项报告时，常常被从零起草整份文档拖住——项目背景要重新组织、需求分析要重新调研、目标计划要重新设计、预算清单要按模板逐项梳理，手动起草一份符合规范的立项报告往往要花好几天。\n\n让「项目立项报告撰写专家」接住这个活：它读取用户上传的立项报告模板，结合本机构在平台上的相似项目经验、执行数据与预算记录，自动生成一份含项目背景、需求分析、项目目标、实施计划、评估指标与预算明细的完整立项报告，章节结构严格贴合模板要求，关键数据来自后台可溯源。\n\n适合：公益机构项目立项人员、社会组织负责人，以及需要按内部或资助方模板产出规范立项报告的项目场景。", "description_en": "At the proposal stage, drafting a full report from scratch — reorganizing context, redoing needs analysis, resetting goals, itemizing the budget line-by-line to match the template — routinely takes days.\n\nLet the Project Proposal Expert take it on: reads your uploaded template, combines it with similar-project experience, execution data, and budget records from your organization's backend, and auto-generates a complete proposal covering background, needs analysis, goals, implementation plan, evaluation metrics, and budget breakdown. Section structure strictly follows the template; key figures remain traceable to backend records.\n\nFor: proposal writers at nonprofits, social organization leads, and any project needing a compliant proposal against an internal or funder template.", "categories": [ "productivity", "business" ], "tags": [ "公益", "立项报告", "项目文档", "文档生成" ], "tags_en": [ "Public Welfare", "Project Proposal", "Project Docs", "Document Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我需要为{项目名称}生成一份立项报告，请你参考本机构在平台的{项目名称}，按照我上传的立项报告模板，撰写立项报告。", "prompt_en": "Please draft a project proposal report for {project name}. Reference our organization's project {project name} on the platform, and follow the proposal template I've uploaded.", "skills": [ { "id": "grant-proposal-writing", "name": "公益申报写作助手", "description": "公益项目申报书、立项报告等文档的专业写作助手", "skill_id": "skill_2088454736381997056" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "立项报告：山乡童心乡村留守儿童成长陪伴计划（脱敏样例）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "project-proposal-polishing", "title": "立项报告语言与格式润色", "title_en": "Project Proposal Language & Format Polishing", "subtitle": "语言更专业，表格更清爽——内容原封不动、只做质感升级", "subtitle_en": "Sharper language, cleaner tables — content untouched, presentation upgraded", "description": "公益机构在完成立项报告初稿后，常常被语言和格式问题卡住——摘要开头多了一个引号、编号项缩进不齐、术语前后不统一、表头白底黑字缺乏层次、活动条目标签和正文混在一起分不清主次、列宽挤成一团，手动逐处调整既费时又容易改了语言漏了格式、改了格式丢了内容。\n\n让「项目立项报告撰写专家」接住这个活：它通读全文逐章润色，修正标点与术语规范、统一列表缩进与符号、优化句式断句；同时做格式层面的系统美化——表头改深蓝底白字加粗、活动条目标签加粗对齐、表格列宽按内容分配、标签列加底色增强可读性。润色后全部内容与预算结构不变，仅做语言和格式层面的优化，并回读校验确保文件可正常打开。\n\n适合：已完成立项报告初稿、需要提升语言专业度与排版美观度的公益机构项目人员，以及提交前需要统一格式规范的资助申请场景。", "description_en": "With the first draft of a project proposal done, the polishing pass is where nonprofits get stuck: a stray quote in the abstract, uneven numbered-list indents, inconsistent terminology, plain-white table headers, activity labels that blur into body text, columns squeezed together. Fixing them by hand is slow, and it's easy to change the language but miss the format — or vice versa.\n\nLet the Project Proposal Expert take it on: reads the full document chapter by chapter to polish punctuation and terminology, unify list indents and bullet marks, and improve sentence rhythm. In parallel, it does a systematic format upgrade — dark-blue bold table headers, bold-aligned activity labels, content-driven column widths, tinted label columns for readability. Content and budget structure remain untouched; only language and format are optimized, with a final read-through to confirm the file opens cleanly.\n\nFor: nonprofit project staff with a first-draft proposal in hand who need a professionalism and presentation boost, and any team standardizing format before grant submission.", "categories": [ "productivity", "business" ], "tags": [ "公益", "立项报告", "文档润色", "格式美化" ], "tags_en": [ "Public Welfare", "Project Proposal", "Document Polishing", "Formatting" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "这是我已经写好的立项报告，请你按照{润色要求……}帮我进行润色。", "prompt_en": "Here's my drafted project proposal. Please polish it following these requirements: {polishing requirements ...}.", "skills": [ { "id": "grant-proposal-writing", "name": "公益申报写作助手", "description": "公益项目申报书、立项报告等文档的专业写作助手", "skill_id": "skill_2088454736381997056" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T10:59:00+08:00", "updated_at": "2026-08-25T10:59:00+08:00", "artifacts": [ { "file": "output.html", "label": "立项报告润色版：山乡童心乡村留守儿童成长陪伴计划" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "projectile-motion", "title": "平抛运动的分解与轨迹方程", "title_en": "平抛运动的分解与轨迹方程", "subtitle": "高中物理 · 平抛运动的分解与轨迹方程", "subtitle_en": "高中物理 · 平抛运动的分解与轨迹方程", "description": "高中物理《平抛运动》互动课件，用可调节 v₀、h、g 的发射模拟器，演示运动分解与抛物线轨迹。", "description_en": "高中物理《平抛运动》互动课件，用可调节 v₀、h、g 的发射模拟器，演示运动分解与抛物线轨迹。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "projectile-motion.html", "artifacts": [ { "file": "projectile-motion.html", "label": "平抛运动的分解与轨迹方程｜互动教学网页" } ], "prompt": "帮我做一个高中物理《平抛运动》的互动教学网页，要能实时调初速度和高度看轨迹变化。", "prompt_en": "帮我做一个高中物理《平抛运动》的互动教学网页，要能实时调初速度和高度看轨迹变化。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "punkzhou-selfmedia-workspace", "title": "朋克周自媒体工作台", "title_en": "Punk Zhou Self-Media Workspace", "subtitle": "内容、商单、日程、健身四张表，收进一个自媒体工作台", "subtitle_en": "Content, deals, schedule, and fitness — four tables in one creator workspace", "description": "一个自媒体博主的内容、商单、日程、健身数据散在好几张表里，每天要开一堆文件才知道今天该干什么，最烦的就是来回切换、记不住进度。\n\n这个案例演示了一个「朋克周自媒体工作台」：把内容项目（40 条）、商业交付（20 条）、出差日程（13 条）、健身记录（13 条）四张表整合进一个页面——今日待办、内容流水线（文案→剪辑→审核→发布）、商单扇形看板、四周日程拖拽排期、播放数据洞察，打开即见今天该做什么。数据用资料库数据表在线存储、改动双向同步，手机电脑都能用。\n\n适合：正在做自媒体想统一管理内容与商单的创作者、需要把多张 Excel 收进一个看板的运营、想让团队协作看进度的新媒体小组、想给自己搭一个「个人 OS」工作台的效率党。", "description_en": "A self-media creator's content, deals, schedule, and fitness data are scattered across several spreadsheets — every day you open a pile of files just to figure out what to do, and switching between them while tracking progress is the most frustrating part.\n\nThis case shows a Punk Zhou self-media workspace that unifies four tables — content (40 items), commercial deals (20), travel schedule (13), and fitness records (13) — into one page. You get today's to-do, a content pipeline (copywriting → editing → review → publish), a fan-style deals board, a four-week drag-and-drop schedule, and playback insights, all visible the moment you open it. Data lives in structured Library tables with two-way sync, and it works on both desktop and mobile.\n\nBest for: creators running self-media who want content and deals in one place, operators who want to fold multiple spreadsheets into a single dashboard, new-media teams that need a shared progress view, and efficiency-minded people who want a personal-OS workbench.", "categories": [ "workspace" ], "tags": [ "自媒体工作台", "内容管理", "商单管理", "数据同步", "单文件网页" ], "tags_en": [ "Self-Media Workspace", "Content Management", "Deal Management", "Data Sync", "Single-file Web App" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "自媒体工作台单文件（内置中英切换）" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://www.workbuddy.cn/space/d/u24zeHpjinaXthG8ftedxx），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n帮我搭一个「朋克周自媒体工作台」。\n\n五个工作区：\n1. 今日（Today）：今日待办、出发提醒、内容流和商单预览\n2. 内容（Content）：内容生产流水线（文案→剪辑→审核→发布），支持新建、筛选、推进状态\n3. 商单（Deals）：商业交付项目管理，六卡扇形轮播 + 项目焦点卡 + 打开项目资料\n4. 日程（Schedule）：四周日程窗 + 按天查看 + 拖拽排期\n5. 洞察（Insights）：播放量/互动 KPI + 内容状态分布、渠道播放、反馈分布、7 日趋势图表\n\n数据存储：用资料库能力自建 4 张数据表（内容项目、商业交付、出差日程、健身记录）存数据，页面实时读资料库数据表（增删改直接写回表）；再把工作台 HTML 部署到资料库，交付一个在线链接。\n\n界面：暖白纸面 + 深色侧边栏 + 克制工业风，信息密度高，PC/手机双适配；零外链，图表用内联 SVG，图标用内联 SVG 不用 emoji。\n\n交付：资料库在线工作台——在线 HTML 页面 + 4 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/u24zeHpjinaXthG8ftedxx, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build me a Punk Zhou self-media workspace. Five workspaces:\n1. Today: daily to-dos, departure reminder, content-flow and deals previews\n2. Content: the content production pipeline (copywriting → editing → review → publish), with create, filter, and advance-status\n3. Deals: commercial delivery project management with a six-card fan carousel + project focus card + open project files\n4. Schedule: a four-week schedule window + day view + drag-and-drop scheduling\n5. Insights: plays/interaction KPIs + content-status distribution, channel plays, feedback distribution, and a 7-day trend chart Data storage: use the Library to self-create 4 tables (content projects, commercial delivery, travel schedule, fitness records) to hold the data; the page reads the Library tables in real time (writes go straight back to the tables); then deploy the workspace HTML to the Library and deliver an online link. Interface: warm paper surface + dark sidebar + restrained industrial style, high information density, responsive on desktop and mobile; fully inline with zero external links, charts as inline SVG, icons as inline SVG (no emoji).\n\nDeliverable: an online Library workbench — an online HTML page plus 4 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929" } ], "mcps": [], "creator": { "name": "朋克周", "name_en": "Punk Zhou", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-14T11:53:53+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "python-learning-path", "title": "Python 零基础学习路径", "title_en": "Python Learning Path from Zero", "subtitle": "90天系统化学习规划，从入门到能做项目", "subtitle_en": "90-day structured learning plan — from beginner to building projects", "description": "想学 Python 但资源太多不知从何下手？看了几天视频就放弃了？缺少系统规划让自学变成三天打鱼两天晒网？这个案例展示了如何用教育 Skill，一键生成 90 天 Python 学习路线图。\n\n路线图按周拆解为 5 大阶段：基础语法 → 进阶特性 → 实用库 → 实战项目 → 综合项目。每周都有明确的学习目标、核心知识点、免费推荐资源、动手练习和检查点。两周一个里程碑，让你随时知道自己在哪、下一步该做什么。还附带常见坑点提醒和坚持学习的实用技巧。\n\n适合：编程零基础学习者、转行程序员、数据分析入门者、大学生自学编程。", "description_en": "Want to learn Python but overwhelmed by too many resources? Watched tutorials for a few days then gave up? This case shows how to use the Education Skill to generate a 90-day Python roadmap.\n\nThe roadmap breaks down into 5 phases across 13 weeks: fundamentals → intermediate features → practical libraries → mini projects → capstone project. Each week includes clear objectives, key concepts, free resources, hands-on exercises, and checkpoints.\n\nBest for: complete beginners, career changers, data analysis newcomers, and college students.", "categories": [ "education" ], "tags": [ "Python", "学习路径", "编程入门", "自学规划", "90天计划" ], "tags_en": [ "Python", "Learning Path", "Programming Basics", "Self-Study Plan", "90-Day Plan" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "Python 90天学习路径" } ], "prompt": "请帮我规划一个 Python 零基础学习路径。\n\n背景：我是编程零基础，想系统学习 Python，目标是 90 天后能独立做小项目。\n\n要求：\n1. 每天投入约 2 小时\n2. 分阶段规划，从基础语法到实战项目\n3. 每周有明确学习目标和练习\n4. 推荐免费学习资源\n5. 包含实战项目练手\n6. 给出常见问题和避坑建议", "prompt_en": "Please help me map out a Python learning path for an absolute beginner.\n\nBackground: I have zero programming experience and want to learn Python systematically, with the goal of being able to independently build small projects after 90 days.\n\nRequirements:\n1. About 2 hours of study per day\n2. A phased plan, going from basic syntax to hands-on projects\n3. Clear learning objectives and exercises each week\n4. Recommended free learning resources\n5. Include hands-on projects to practice with\n6. Provide common pitfalls and tips for avoiding them", "skills": [ { "id": "education", "name": "Education", "description": "学习助手：生成学习计划、测验、抽认卡、复习材料并跟踪进度", "description_en": "Study assistant: generate plans, quizzes, flashcards, review materials and track progress", "skill_id": "skill_2053082003609702400" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "qiuzhao-dual-track-workbench", "title": "秋招投递记录工作台", "subtitle": "各大企业/央国企双赛道 ｜ 筛子筛企→投递→面试→offer 全流程一站管理", "description": "秋招投了十几家就乱成一团：哪家约了一面、哪家挂了、哪家还在等结果，全靠翻聊天记录？这个工作台把秋招全程拆成四件套：秋招筛子（批量把已开放秋招的企业扔进篓子，投过的自动打绿标）、关注清单（56 家硬件/智能硬件公司按赛道整理好招聘官网）、投递记录（状态从已投简历到挂掉共 10 档，彩色标签一眼看清，支持编辑/删除/搜索/四种排序/紧凑切换）、日程规划（笔试面试日程时间线 + 月历视图 + 月度分布 + 近 7 日投递节奏）。左右侧边栏一键切「各大企业 / 央国企」双赛道，数据互不干扰；页面数据在线存储、多设备自动同步，手机电脑都能记。演示数据内置 10 条投递记录和 56 家关注企业（公司、岗位、状态、备注齐全），打开就能上手试。适合正在打秋招/暑期实习的同学、帮学弟学妹管进度的班委、任何想把自己投递漏斗管起来的人。", "tags": [ "秋招", "求职管理", "投递追踪", "面试日程", "工作台" ], "title_en": "Autumn Application Tracker Workbench", "subtitle_en": "Dual-track (Tech & SOE) | Sieve → Apply → Interview → Offer, all in one place", "description_en": "Applied to a dozen companies and already lost track — who scheduled you for a first interview, who rejected you, who's still reviewing? This workbench breaks autumn recruiting into four tools: the Sieve (batch-drop companies with open autumn hiring, applied ones get a green tag automatically), the Watchlist (56 hardware / smart-device companies with careers-site links, organized by sector), the Application Log (10 status stages from Applied to Rejected with colored chips, plus edit / delete / search / four sort orders / compact view), and the Schedule (event timeline + month calendar + monthly breakdown + 7-day application rhythm). The sidebar toggles between the Tech Companies and SOEs tracks with fully separate data; everything is stored online and syncs across devices. Demo data ships with 10 application records and 56 watched companies (companies, roles, statuses and notes all filled in) so you can try it right away. Best for students in autumn/summer recruiting, class reps tracking peers' progress, and anyone who wants their application funnel under control.", "tags_en": [ "Autumn Hiring", "Job Hunt Management", "Application Tracking", "Interview Schedule", "Workbench" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/2Ws96lp6qNy4DJVjGdItuR，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「秋招投递记录工作台」 工作台，用资料库数据表存数据，页面运行时读写、多设备同步。四大模块：① 投递记录——卡片列表（企业/Base/岗位/状态/投递日期/备注），状态 10 档（已投简历/已过初筛/hr筛选/业务筛选/待AI面试/待线上测评/面试/评估中/等offer/挂）用彩色标签区分，支持新增（快速新增条 + 完整弹窗）、编辑、删除、按企业/岗位/base 搜索、按状态筛选、四种排序、紧凑/详细密度切换、状态图例与统计条；② 关注清单——待开放秋招的企业卡片（企业/岗位/预计开放/备注），可新增编辑删除，已投递的企业自动打「✓ 已投」绿标并可一键转投递；③ 秋招筛子——批量粘贴企业名单（支持带官网链接）入篓，逐个核实岗位后转为投递或删除，显示待筛选/已投计数；④ 日程规划——笔试/面试日程时间线（今天分隔标记、过去记录折叠）、月份日历点选筛选、月度分布统计、近 7 日投递节奏柱图。左侧边栏切换「各大企业 / 央国企」双赛道（数据按赛道隔离，用资料库表里的赛道字段区分），并排日程规划视图；顶部状态徽标显示在线同步状态。视觉走清爽求职工具风（亮色、蓝紫主色、圆角卡片、彩色状态标签），所有 CSS/JS 无外部依赖，PC/移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 4 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to recreate this workbench from https://www.workbuddy.cn/space/d/2Ws96lp6qNy4DJVjGdItuR with online data storage and two-way sync. Prefer the one-click clone flow; if unavailable, build it as follows: Build an 'Autumn Application Tracker Workbench' as a workbench that stores data in Library databases, reads/writes at runtime and syncs across devices. Four modules: 1) Application Log — card list (company / base / role / status / applied date / notes) with 10 status stages (Applied / Screened / HR screening / Dept screening / AI interview pending / Assessment pending / Interview / Under review / Awaiting offer / Rejected) shown as colored chips; supports quick-add bar + full modal, edit, delete, search by company/role/base, status filter, four sort orders, compact/detail density toggle, status legend and stats bar; 2) Watchlist — cards for companies whose autumn hiring is not yet open (company / role / expected opening / notes) with add/edit/delete; companies already applied to get a green '✓ Applied' tag and a one-click convert-to-application; 3) Sieve — batch-paste company names (with optional careers links), verify roles one by one, then convert to applications or remove; shows to-screen / applied counts; 4) Schedule — written-test/interview event timeline (a 'today' divider, past events collapsed), month calendar with click-to-filter, monthly breakdown, and a 7-day application rhythm bar chart. The left sidebar toggles the Tech Companies / SOEs tracks (data isolated per track via the track field in the Library tables) plus a Schedule view; a header badge shows online-sync status. Visual style: clean recruiting-tool look (light theme, blue-purple accents, rounded cards, colored status chips); all CSS/JS inline with no external dependencies; responsive for PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus 4 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "秋招投递记录工作台（双语自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "creator": { "name": "碎琼乱玉", "name_en": "碎琼乱玉", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-02T17:36:59+08:00", "updated_at": "2026-09-02T18:05:00+08:00", "quality_score": 97, "is_recommended": false, "recommended_order": 0 }, { "id": "qq-email-customer-followup-tracker", "title": "客户邮件跟进追踪表", "title_en": "Customer Email Follow-up Tracker", "subtitle": "按 UID 深读正文，跟进状态一条不漏", "subtitle_en": "Deep-read by UID, never lose track of a follow-up", "description": "跟合作方邮件往来最怕两件事：正文只扫了一眼摘要就漏了关键截止时间，以及过两天想找回来却发现「我到底回没回过这封」。这个案例演示了用 QQ邮箱 skill 收取最近 30 天邮件、筛出客户与合作方往来邮件，再按 UID 逐封读取完整正文，提炼核心诉求和截止时间，最后按「待我回复 / 等待对方 / 已完结」三种状态整理成一张跟进追踪表，每条都带邮件 UID、日期和下一步动作。适合：维护多个客户的外贸/BD 同学、跟进合作项目的项目经理、对接供应商的采购、要按周复盘对外沟通的团队负责人，以及所有邮件回复拖延症患者。", "description_en": "Two things make partner email threads painful: skimming only the preview and missing a key deadline, and later wondering \"did I ever reply to this?\" This case uses the QQ Mail skill to fetch the last 30 days of email, filter out customer and partner correspondence, deep-read each important message in full by UID, extract the core ask and deadline, and organize everything into a follow-up tracker with three statuses — \"needs my reply\", \"waiting on them\", \"closed\". Every entry carries the email UID, date, and next action. Great for: foreign-trade and BD folks juggling multiple clients, project managers chasing partnerships, procurement coordinating suppliers, team leads reviewing external communication weekly, and chronic email procrastinators.", "categories": [ "office" ], "tags": [ "邮件跟进", "客户管理", "待办追踪", "正文提取" ], "tags_en": [ "Email Follow-up", "Customer Management", "Task Tracking", "Body Extraction" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "我在跟几个合作方的邮件往来，经常漏回或者漏掉正文里的截止时间。请用 QQ邮箱 skill 帮我做一次跟进整理：\n\n1. 先收取最近 30 天的邮件；\n2. 从中筛出客户/合作方发来的往来邮件；\n3. 对需要我处理的关键邮件，按收信列表里的 UID 用 get-body 读取完整正文，提炼核心诉求、涉及金额或日期、截止时间；\n4. 按「待我回复 / 等待对方 / 已完结」三种状态，把每条跟进事项整理成追踪表，附上邮件 UID、日期、发件人和下一步动作。\n\n最终生成一份跟进追踪面板（单文件 HTML、明亮暖色系）。", "prompt_en": "I correspond with several partners by email and often miss replies or deadlines buried in message bodies. Use the QQ Mail skill to run a follow-up review for me:\n\n1. First fetch my emails from the last 30 days;\n2. Filter out the correspondence from customers/partners;\n3. For key emails that need my action, read the full body via get-body using the UID from the receive list, and extract the core ask, any amounts or dates involved, and deadlines;\n4. Organize every follow-up item into a tracker with three statuses — \"needs my reply\", \"waiting on them\", \"closed\" — with the email UID, date, sender, and next action.\n\nFinally, generate a follow-up tracking panel (single-file HTML, bright warm color scheme).", "skills": [ { "id": "qq-email", "name": "QQ邮箱", "description": "QQ邮箱收发邮件（IMAP/SMTP），支持发信、收信、查看正文", "skill_id": "skill_2053082758542479360" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T16:56:18+08:00", "updated_at": "2026-08-27T17:23:40+08:00", "artifacts": [ { "file": "output.html", "label": "客户邮件跟进追踪面板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "qq-email-inbox-digest", "title": "一周收件箱速览周报", "title_en": "Weekly Inbox Digest", "subtitle": "7 天邮件自动分类，重点事项一目了然", "subtitle_en": "Seven days of email, auto-categorized at a glance", "description": "一到周一就头大：邮箱里堆了整整一周没细看的邮件，逐封点开又费时，不看又怕漏掉重要的事。这个案例演示了用 QQ邮箱 skill 收取最近 7 天的邮件，自动按工作协作、通知公告、订阅推广、个人事务分类，标出需要回复处理的重点邮件，并给每封邮件提炼一段摘要，最后排成一份可读性拉满的杂志风周报。适合：邮件多到看不过来的打工人、每周要做 inbox 清零的效率党、需要向领导汇报本周往来邮件的项目经理、想快速了解邮箱动态的学生党，以及所有被邮件焦虑困扰的人。", "description_en": "Monday dread: a full week of unread email piled up in your inbox. Opening every message takes forever, but skipping them risks missing something important. This case uses the QQ Mail skill to fetch the last 7 days of email, automatically group it into work collaboration, notifications, subscriptions, and personal items, flag the messages that need a reply, add a short digest for each email, and lay it all out in a magazine-style weekly report. Great for: office workers drowning in email, inbox-zero enthusiasts, project managers reporting weekly correspondence, students keeping tabs on their mailbox, and anyone with email anxiety.", "categories": [ "office", "productivity" ], "tags": [ "邮件周报", "收件箱管理", "信息摘要", "效率办公" ], "tags_en": [ "Email Digest", "Inbox Management", "Summarization", "Productivity" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "每到周一，我最头疼的就是邮箱里堆了一周没细看的邮件。请用 QQ邮箱 skill 的收信能力，收取我最近 7 天的邮件（可用 --days 7），然后帮我做一份「收件箱一周速览」：\n\n1. 按类别分组：工作协作、通知公告、订阅推广、个人事务；\n2. 标出需要我回复或处理的重点邮件；\n3. 每封邮件给出主题、发件人、日期和一段两三句话的摘要；\n4. 结尾给一个本周收件箱小结（总数、各类占比、待处理清单）。\n\n最终汇总成一份可读性高的周报（单文件 HTML、明亮配色、报纸杂志风排版）。", "prompt_en": "Every Monday my biggest headache is a week of unexamined email sitting in the inbox. Use the QQ Mail skill's receiving capability to fetch my emails from the last 7 days (--days 7), then build a \"weekly inbox digest\" for me:\n\n1. Group emails by category: work collaboration, notifications, subscriptions, personal;\n2. Flag the important ones that need my reply or action;\n3. For each email, show subject, sender, date, and a two-to-three sentence digest;\n4. End with a weekly summary (total count, category breakdown, to-do list).\n\nCompile everything into a highly readable weekly report (single-file HTML, bright color scheme, newspaper-magazine style layout).", "skills": [ { "id": "qq-email", "name": "QQ邮箱", "description": "QQ邮箱收发邮件（IMAP/SMTP），支持发信、收信、查看正文", "skill_id": "skill_2053082758542479360" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T16:56:18+08:00", "updated_at": "2026-08-27T17:23:40+08:00", "artifacts": [ { "file": "output.html", "label": "收件箱一周速览周报" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "qq-email-interview-inviter", "title": "批量发送个性化面试邀约", "title_en": "Bulk Personalized Interview Invitations", "subtitle": "一批发送 5 位候选人，称呼和岗位各不相同", "subtitle_en": "Five candidates in one batch, each with their own greeting and role", "description": "招聘季最头疼的就是重复劳动：同样一封面试邀约，要挨个改称呼、改岗位、改时间，发完还得自己记谁发了谁没发。这个案例演示了把一份 5 人候选名单交给 QQ邮箱 skill，用 SMTP 逐个发送个性化面试邀约，并汇总成一份发送任务报告——名单、个性化变量、实际发送命令、每封邮件的发送状态一目了然。适合：负责校招/社招的 HR、需要批量发通知的行政同学、给学员发课程通知的教务老师、要给多个客户发会议邀约的销售，以及所有不想再复制粘贴改称呼的人。", "description_en": "Recruiting season is full of copy-paste drudgery: the same interview invitation, but you have to tweak the greeting, role, and time for every single candidate — and then keep track of who got it. This case hands a 5-person candidate list to the QQ Mail skill, which sends personalized invitations one by one via SMTP and wraps everything into a dispatch report: the roster, personalization variables, the actual send commands, and the delivery status of each email. Great for: HR running campus or lateral hiring, admins sending bulk notices, teachers notifying students, sales reps inviting multiple clients, and anyone tired of hand-editing greetings.", "categories": [ "office" ], "tags": [ "邮件发送", "批量处理", "招聘助手", "个性化" ], "tags_en": [ "Email Sending", "Batch Processing", "Recruiting Helper", "Personalization" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "我是 HR，本周要给 5 位通过初筛的候选人发面试邀约。请用 QQ邮箱 skill 的发信能力，把下面名单逐个发送个性化面试邀约邮件（每位候选人一封，称呼、岗位、面试时间不同，其余内容一致）：\n\n- 张明远，应聘产品经理，面试时间 周三 14:00\n- 李思雨，应聘运营专员，面试时间 周三 16:00\n- 王浩然，应聘前端工程师，面试时间 周四 10:00\n- 陈嘉怡，应聘数据分析师，面试时间 周四 15:00\n- 刘子墨，应聘产品经理，面试时间 周五 11:00\n\n邀约邮件包含：问候、岗位名称、面试时间（线上腾讯会议）、注意事项（提前 10 分钟入场）、回复确认提醒。全部发送完成后，汇总每封邮件的发送状态和使用的发送命令，生成一份本次批量发送的任务报告（单文件 HTML、明亮配色）。", "prompt_en": "I am an HR and need to send interview invitations to 5 candidates who passed the initial screening this week. Using the QQ Mail skill's sending capability, send each of them a personalized invitation email (one per candidate; greeting, role, and interview time differ, the rest is identical):\n\n- Zhang Mingyuan, Product Manager candidate, Wednesday 14:00\n- Li Siyu, Operations Specialist candidate, Wednesday 16:00\n- Wang Haoran, Frontend Engineer candidate, Thursday 10:00\n- Chen Jiayi, Data Analyst candidate, Thursday 15:00\n- Liu Zimo, Product Manager candidate, Friday 11:00\n\nEach invitation should include: a greeting, the role name, the interview time (online via Tencent Meeting), notes (join 10 minutes early), and a reminder to reply to confirm. After all emails are sent, summarize the delivery status of each email and the send commands used, and generate a dispatch report for this batch (single-file HTML, bright color scheme).", "skills": [ { "id": "qq-email", "name": "QQ邮箱", "description": "QQ邮箱收发邮件（IMAP/SMTP），支持发信、收信、查看正文", "skill_id": "skill_2053082758542479360" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T16:56:18+08:00", "updated_at": "2026-08-27T17:23:40+08:00", "artifacts": [ { "file": "output.html", "label": "批量发送任务报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "qqmusic-artist-profile", "title": "周杰伦音乐档案", "title_en": "Jay Chou Music Profile", "subtitle": "搜索聚合歌曲/专辑/MV/歌单，生成沉浸式艺人档案", "subtitle_en": "Aggregate songs, albums, MVs and playlists into an immersive artist profile", "description": "想完整回顾一位歌手的作品，歌曲、专辑、MV 却东一块西一块，翻起来太费劲。\n\n这个案例让 QQ音乐助手把搜索结果聚合成一页精致的艺人档案：歌手数据总览、热门歌曲 Top10、按发行时间排的专辑长廊、MV 热播榜、精选歌单推荐，全部带播放链接，像翻一本音乐杂志。\n\n适合：想系统回顾偶像作品的粉丝、做歌手专题的内容编辑、找演出或翻唱素材的音乐从业者。", "description_en": "Want to review an artist's complete work, but their songs, albums and MVs are scattered everywhere?\n\nThis case has the QQ Music assistant aggregate search results into one polished artist profile: an artist overview card, Top 10 hits, an album gallery sorted by release date, an MV hot list, and curated playlist picks — all with play links, like flipping through a music magazine.\n\nFor: fans revisiting their idol's discography, editors working on artist features, and musicians looking for performance or cover material.", "categories": [ "search", "creation" ], "tags": [ "QQ音乐", "周杰伦", "歌手档案", "专辑", "搜索" ], "tags_en": [ "QQ Music", "Jay Chou", "Artist Profile", "Albums", "Search" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "周杰伦沉浸式音乐档案页面" } ], "prompt": "用 QQ音乐助手搜索并整理歌手「周杰伦」的真实数据，生成沉浸式艺人档案单文件 HTML。依次调 /discover/search：type=5 查歌手档案（歌曲数、专辑数），type=0 搜热门歌曲，type=1 搜专辑（发行时间、播放量、分类），type=3 搜 MV（播放次数、发布时间），type=2 搜相关歌单；可再调 /playlists/detail 深挖 1 个高质量歌单的完整曲目。页面需包含：①歌手数据总览卡（歌曲数/专辑数）；②热门歌曲 Top10（带播放链接）；③专辑长廊（按发行时间排序，标注播放量）；④MV 热播榜（按播放次数排序）；⑤精选歌单推荐区。暗色唱片质感、排版精致有杂志感，单文件 HTML、资源全内联、可离线打开。所有数据必须真实搜索获取，禁止编造，页面底部注明查询时间。", "prompt_en": "Use the QQ Music assistant to search and organize real data for the artist \"Jay Chou\", then generate an immersive single-file HTML artist profile. Call /discover/search in sequence: type=5 for the artist profile (song count, album count), type=0 for popular songs, type=1 for albums (release date, play count, category), type=3 for MVs (play count, release date), and type=2 for related playlists; optionally call /playlists/detail to dig into one high-quality playlist's full track list. The page must include: 1) an artist overview card (song/album counts); 2) Top 10 popular songs (with play links); 3) an album gallery sorted by release date with play counts; 4) an MV hot list sorted by play count; 5) a curated playlist section. Dark vinyl-record texture, refined magazine-style layout, single-file HTML with all resources inlined, works offline. All data must come from real searches — no fabrication. Note the query time in the page footer.", "skills": [ { "id": "qqmusic", "name": "QQ音乐助手", "description": "QQ音乐官方智能助手，支持歌曲搜索、每日推荐、AI歌单、排行榜、听歌报告与AI音乐解读。", "skill_id": "skill_2070033533400236032" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "usage_count": 0, "created_at": "2026-07-24T17:41:00+08:00", "updated_at": "2026-07-27T12:35:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "qqmusic-charts-weekly-report", "title": "QQ音乐巅峰榜周报", "title_en": "QQ Music Top Charts Weekly Report", "subtitle": "一键汇总全站榜单，看清本周乐坛风向", "subtitle_en": "Aggregate every chart in one page and read this week's music trends at a glance", "description": "想知道最近什么歌最火，又懒得把飙升榜、热歌榜、新歌榜挨个翻一遍。\n\n这个案例让 QQ音乐助手实时拉取排行榜全部分组，再深挖收听人数最高的几个榜单的完整曲目，自动产出一份暗色数据周报：各榜单收听人数对比、重点榜单 Top10、跨榜热歌交叉分析，每首歌都带播放链接。\n\n适合：想快速掌握乐坛动态的音乐爱好者、做内容选题的自媒体人、需要热歌素材的电台和短视频创作者。", "description_en": "Want to know what's hot right now but too lazy to flip through every chart one by one?\n\nThis case has the QQ Music assistant pull all chart groups in real time, dig into the full track lists of the most-listened charts, and produce a dark-themed data report: listen-count comparison across charts, Top 10 of key charts, and cross-chart hit analysis — every song with a play link.\n\nFor: music lovers tracking trends, content creators picking topics, and radio or short-video makers looking for hot tracks.", "categories": [ "data", "search" ], "tags": [ "QQ音乐", "排行榜", "数据报告", "热歌榜", "可视化" ], "tags_en": [ "QQ Music", "Charts", "Data Report", "Top Songs", "Visualization" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "QQ音乐巅峰榜数据周报页面" } ], "prompt": "用 QQ音乐助手获取当前真实排行榜数据，做一份「QQ音乐巅峰榜数据周报」单文件 HTML 报告。先调 /charts 拿到全部榜单分组（巅峰榜、地区榜、特色榜等），记录每个榜单的收听人数与歌曲总数；再挑收听人数最高的 3-4 个榜单调 /charts/detail 拿完整曲目。页面需包含：①榜单分组总览与各榜单收听人数对比图；②重点榜单 Top10 歌曲排行（歌名+歌手+播放链接）；③跨榜热歌分析——统计同时登上多个榜单的歌曲及其出现次数；④数据更新时间与来源标注。暗色音乐氛围、图表用内联 SVG/Canvas 绘制、信息密度高，单文件 HTML、资源全内联、可离线打开。所有数据必须真实调用接口获取，禁止编造，页面底部注明数据抓取时间。", "prompt_en": "Use the QQ Music assistant to fetch current real chart data and build a \"QQ Music Top Charts Weekly Report\" as a single-file HTML page. First call /charts to get all chart groups (Top Charts, Regional Charts, Featured Charts, etc.), recording each chart's listener count and total tracks; then pick the 3-4 most-listened charts and call /charts/detail for their full track lists. The page must include: 1) an overview of chart groups with a listener-count comparison chart; 2) Top 10 rankings of key charts (song name + artist + play link); 3) cross-chart hit analysis — songs appearing on multiple charts with occurrence counts; 4) data update time and source attribution. Dark music atmosphere, charts drawn with inline SVG/Canvas, high information density, single-file HTML with all resources inlined, works offline. All data must be fetched from the real API — no fabrication. Note the data fetch time in the page footer.", "skills": [ { "id": "qqmusic", "name": "QQ音乐助手", "description": "QQ音乐官方智能助手，支持歌曲搜索、每日推荐、AI歌单、排行榜、听歌报告与AI音乐解读。", "skill_id": "skill_2070033533400236032" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "usage_count": 0, "created_at": "2026-07-24T17:41:00+08:00", "updated_at": "2026-07-27T12:35:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "qqmusic-listening-report", "title": "我的QQ音乐听歌画像", "title_en": "My QQ Music Listening Portrait", "subtitle": "听歌报告+AI解读，一页看懂你的音乐DNA", "subtitle_en": "Listening report plus AI interpretation — your music DNA on one page", "description": "听了那么久歌，想知道自己到底是什么风格的听众，光翻歌单可看不出来。\n\n这个案例让 QQ音乐助手拉取你的本周与本月听歌报告——日均时长、Top 歌手与歌曲、最爱时段、连续听歌天数、流派分布、小众占比、听歌城市、单曲循环之最，再调 AI 解读生成一段专属于你的音乐画像，汇总成一页温暖的个人听歌报告。\n\n适合：喜欢复盘听歌习惯的乐迷、想晒个性化报告的用户、对推荐算法好奇的 QQ 音乐用户。", "description_en": "After listening for so long, do you actually know what kind of listener you are? You can't tell just by scrolling your playlists.\n\nThis case has the QQ Music assistant pull your weekly and monthly listening reports — daily average time, top artists and songs, favorite hours, consecutive listening days, genre distribution, niche ratio, listening cities, and your most-looped track — then calls AI interpretation to write a personal music portrait, all wrapped into one warm listening report page.\n\nFor: music fans who love reviewing their habits, users who want to share a personalized report, and anyone curious about QQ Music's personalization.", "categories": [ "data" ], "tags": [ "QQ音乐", "听歌报告", "AI解读", "个人画像", "年度报告" ], "tags_en": [ "QQ Music", "Listening Report", "AI Interpretation", "Personal Portrait", "Yearly Report" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "个人听歌画像报告页面" } ], "prompt": "用 QQ音乐助手生成「我的听歌画像」个人报告单文件 HTML。依次调 /me/report：timeKey=w 拿本周报告（日均听歌时长、对比上周浮动、Top 歌手、Top 歌曲），timeKey=m 拿本月报告（Top 艺人、Top 歌曲、最爱时段、连续听歌天数、流派分布、小众/热门占比、尝鲜歌曲数、听歌城市、单曲循环之最、深夜单曲）；再调 /assistant/ai_interpretation（query 填「分析我的听歌风格，我是怎样的听众」）拿 AI 流式解读并整理成文。页面需包含：①本周速览数据卡；②月度 Top 歌手与 Top 歌曲榜（带播放链接）；③听歌习惯可视化（24 小时时段分布、流派占比、小众 vs 热门）；④趣味版块（循环之最、深夜单曲、听歌城市足迹）；⑤AI 音乐画像专区。温暖的年度报告风格、视觉精致有情感，单文件 HTML、资源全内联、可离线打开。所有数据必须真实调接口获取，禁止编造，注明统计周期。", "prompt_en": "Use the QQ Music assistant to generate a \"My Listening Portrait\" personal report as a single-file HTML page. Call /me/report in sequence: timeKey=w for this week's report (daily average listening time, week-over-week change, top artists, top songs), timeKey=m for this month's report (top artists, top songs, favorite hours, consecutive listening days, genre distribution, niche vs mainstream ratio, new song count, listening cities, most-looped song, late-night song); then call /assistant/ai_interpretation (query: \"分析我的听歌风格，我是怎样的听众\") to get the streaming AI interpretation and compile it into text. The page must include: 1) a weekly overview data card; 2) monthly top artists and top songs (with play links); 3) listening habit visualizations (24-hour distribution, genre share, niche vs mainstream); 4) fun sections (most-looped song, late-night song, listening city footprint); 5) an AI music portrait section. Warm annual-report style, refined and emotional visuals, single-file HTML with all resources inlined, works offline. All data must be fetched from the real API — no fabrication. Note the statistics period on the page.", "skills": [ { "id": "qqmusic", "name": "QQ音乐助手", "description": "QQ音乐官方智能助手，支持歌曲搜索、每日推荐、AI歌单、排行榜、听歌报告与AI音乐解读。", "skill_id": "skill_2070033533400236032" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "created_at": "2026-07-24T17:41:00+08:00", "updated_at": "2026-07-27T12:35:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "quadratic-function-graph-and-properties", "title": "二次函数的图像与性质", "title_en": "二次函数的图像与性质", "subtitle": "初中数学 · 二次函数的图像与性质", "subtitle_en": "初中数学 · 二次函数的图像与性质", "description": "初中数学《二次函数》互动课件，用 a、h、k 参数滑块实时观察抛物线的开口、顶点与平移。", "description_en": "初中数学《二次函数》互动课件，用 a、h、k 参数滑块实时观察抛物线的开口、顶点与平移。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "quadratic-function-graph-and-properties.html", "artifacts": [ { "file": "quadratic-function-graph-and-properties.html", "label": "二次函数的图像与性质｜互动教学网页" } ], "prompt": "帮我做一个初中数学《二次函数的图像与性质》的互动教学网页，要能用滑块调参数看抛物线变化。", "prompt_en": "帮我做一个初中数学《二次函数的图像与性质》的互动教学网页，要能用滑块调参数看抛物线变化。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "quarterly-sales-trend-comparison", "title": "季度成交趋势对比", "title_en": "Quarterly Sales Trend Comparison", "subtitle": "本季度与上季度同期成交趋势及驱动拆解", "subtitle_en": "A practical Weimob playbook for quarterly sales trend comparison", "description": "做季度成交趋势对比时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“季度成交趋势对比”展开：对比本季度截至昨日与上季度同期的 GMV、订单量和客单价，按周定位变化最明显的阶段，并判断增长是由订单量还是客单价驱动。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、同比环比分析运营人员、经营分析团队。", "description_en": "The hardest part of quarterly sales trend comparison is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates quarterly sales trend comparison through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Period-over-Period Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "季度分析", "GMV", "订单量", "客单价", "趋势对比" ], "tags_en": [ "Quarterly Sales Trend Comparison", "Business Diagnosis", "Period-over-Period Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "对比本季度截至昨天与上季度同期的GMV、订单量和客单价，按周定位增长或下滑最明显的时间段，并判断变化主要由订单量还是客单价驱动。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a quarterly sales trend comparison covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "rabbit-dream-handdraw", "title": "小兔追梦看海记 - 手绘动画", "title_en": "The Rabbit's Dream - Hand-drawn Animation", "subtitle": "把绘本故事变成温暖的手绘动画短片", "subtitle_en": "Turn a picture book story into a warm hand-drawn animation", "description": "想要把一篇绘本故事变成生动的手绘动画视频？这个案例展示了如何用 handdraw 手绘动画能力，将《小兔追梦看海记》的教学设计文档转化为一段 3 分钟的儿童绘本风格手绘动画。视频讲述小兔子怀揣看海梦想，历经河流、高山、荆棘等艰难险阻，凭借机智与坚持最终抵达大海的故事。适合教师制作教学动画、创作者将文字内容视觉化。", "description_en": "Want to turn a picture book story into a vivid hand-drawn animation? This case demonstrates how to use the handdraw skill to transform a teaching design document into a 3-minute children's picture book style animation. The video tells the story of a rabbit who dreams of seeing the ocean, overcomes rivers, mountains, and thorns, and finally reaches the sea through cleverness and persistence. Suitable for teachers creating educational animations and creators visualizing text content.", "tags": [ "手绘动画", "绘本故事", "教学视频", "儿童绘本" ], "tags_en": [ "hand-drawn animation", "picture book", "educational video", "children illustration" ], "prompt": "读取《小兔追梦看海记》文档内容，提炼故事情节，用手绘动画的方式生成一个3分钟的儿童绘本风格动画视频。故事分为三部分：兔子梦想的开始、追梦途中遇到的艰难险阻（过河、翻山、穿越荆棘）、最终梦想成真到达大海。选用儿童绘本画风，温暖明快的圆润造型。每页配一句简短旁白，画面只画看得见的具体事物。", "prompt_en": "Read the document 'The Rabbit's Dream of Seeing the Sea', extract the story plot, and generate a 3-minute children's picture book style hand-drawn animation video. The story has three parts: the beginning of the rabbit's dream, the difficulties encountered on the journey (crossing a river, climbing a mountain, passing through thorns), and finally achieving the dream of reaching the ocean. Use children's picture book style with warm, bright, rounded shapes. Each page has a short narration, and the visuals only depict concrete, visible objects.", "artifact_type": "video", "preview": "output.mp4", "external_url": "", "scenario": "", "scenario_use": "", "task_mode": "", "categories": [ "creation", "education" ], "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "output.mp4", "label": "手绘动画视频" } ], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "usage_count": 0, "created_at": "2026-07-14T21:20:00+08:00", "updated_at": "2026-07-14T21:20:00+08:00", "skills": [], "experts": [ { "id": "VibeknowHanddraw", "name": "vibeknow手绘视频讲解师", "description": "老师讲历史、家长讲绘本、医生讲康复、理财博主讲避坑，专业内容边画边讲成手绘科普视频，51种风格随选。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "89133bd8e85e", "name_en": "vibeknow hand-drawn explainer", "description_en": "Teachers explain history, parents read picture books, doctors explain rehab, finance bloggers warn of pitfalls — turn professional content into hand-drawn explainer videos, 51 styles to choose from.", "expert_id": "ex_kdKBSetkjUyA" } ], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "realtime-hotlist-dashboard", "title": "实时热搜榜数据看板", "title_en": "Real-time Trending Hotlist Dashboard", "subtitle": "一键抓取百度实时热搜，整理成清爽的排行榜看板", "subtitle_en": "Scrape Baidu's live trending list and turn it into a clean ranking dashboard in one shot", "description": "想快速掌握当下全网都在聊什么，却不想在一堆网页里手动复制粘贴？打开热搜页，条目一堆、热度值散落各处，整理起来又慢又乱。\n\n这个案例演示了：让 AI 用 QQ 浏览器自动化打开百度实时热搜榜，真实抓取前几十条热搜的排名、标题和热度指数，注明抓取时间和来源，然后自动整理成一个信息密度很高的单文件 HTML 数据看板——带排名和热度可视化，能离线打开、直接展示。\n\n适合：想快速做热点简报的运营、需要每日盯榜的自媒体、以及想看网页自动化怎么落地的开发者。", "description_en": "Want to know what the whole internet is talking about right now, without copy-pasting from a dozen web pages? Trending pages are crowded, heat scores are scattered, and organizing it all is slow and messy.\n\nThis case shows how to let AI use QQ Browser automation to open Baidu's real-time trending list, actually scrape the rank, title and heat index of dozens of entries, note the capture time and source, then automatically turn it into an information-dense single-file HTML dashboard — with ranking and heat visualization, ready to open offline and present directly.\n\nFor: operators who need quick hot-topic briefings, content creators who track trending lists daily, and developers curious about real-world web automation.", "categories": [ "search" ], "tags": [ "实时热点", "数据抓取", "可视化看板" ], "tags_en": [ "Trending", "Web Scraping", "Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "实时热搜榜数据看板" } ], "prompt": "用 QQ 浏览器自动化 skill 打开百度实时热搜榜（https://top.baidu.com/board?tab=realtime），真实抓取当前榜单前 15 到 20 条热搜，逐条提取排名、标题、热度指数以及页面上的简短描述，并记录数据来源 URL 和抓取时间。数据必须真实抓取，不要使用示意或编造的数据。\n\n然后把这些真实数据整理成一个信息密度高、耐看的单文件 HTML 排行榜看板，保存为 output.html：能一眼看出名次高低和热度强弱即可，风格你来把握。所有资源内联、可离线打开，页面内跳转用按钮加滚动而非锚点链接。", "prompt_en": "Use the QQ Browser automation skill to open Baidu's real-time trending list (https://top.baidu.com/board?tab=realtime) and actually scrape the current top 15 to 20 entries, extracting for each the rank, title, heat index and any short description shown, and record the source URL and capture time. The data must be genuinely scraped — do not use placeholder or fabricated data.\n\nThen organize the real data into an information-dense, readable single-file HTML ranking dashboard saved as output.html: it should make rank order and heat magnitude obvious at a glance; the visual style is up to you. Inline all resources so it opens offline, and use buttons with scrolling for in-page navigation rather than anchor links.", "skills": [ { "id": "qqbrowser-skill", "name": "QQ浏览器自动化", "description": "面向 AI 代理的浏览器自动化技能，可导航网页、填表、点击、截图、提取数据", "skill_id": "skill_2070018474844057600" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-07-27T03:32:36Z", "updated_at": "2026-07-27T07:18:04Z", "is_recommended": false, "recommended_order": 0 }, { "id": "reconciliation-demand-letter", "title": "月末对账 + 催款函，一次生成", "title_en": "Reconciliation Demand Letter", "subtitle": "对账单＋措辞得体催款函一次生成，收款不伤客情", "subtitle_en": "Reconciliation statement plus a tactful demand letter in one pass — collect without souring relationships", "description": "输入往来明细，自动生成往来对账单和措辞得体的催款函，超期账款一键提醒，收款不伤客情。", "description_en": "Enter the transaction details: it generates an account reconciliation statement and a tactfully worded demand letter, with one-click reminders for overdue payments — collections without damaging customer relationships.", "categories": [ "creation" ], "tags": [ "对账", "催款函", "应收账款", "内勤" ], "tags_en": [ "Reconciliation", "Demand Letter", "Accounts Receivable", "Back Office" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "对账单与催款函（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易内勤专家。请帮我生成某客户的往来对账单和催款函。\n\n【往来信息】\n- 客户：{如 江苏鼎盛建工}\n- 对账截止日：{如 2026-07-31}\n- 往来明细：{逐笔填写 日期/业务摘要/发货金额/已收款。例如：\n 07-05 螺纹Φ25 80吨 发货300,800 已收280,800}\n\n【要求】\n1. 生成往来对账单表：逐笔列出发货金额、已收款、应收余额，并合计出应收总额；\n2. 生成一份措辞得体的付款提醒函（催款函）：说明未结笔数、应收余额（附大写）、超期天数、限期付款日期、异议核对方式；\n3. 语气专业礼貌、不伤客情，符合中国B2B贸易催款习惯；查询【行业信息资产库】里这个行业的账期与催款口径。\n\n若无往来数据，请说明需要哪些字段并给示例。请生成对账单+催款函。", "skills": [ { "id": "smart-reconciliation", "name": "智能对账", "description": "TradeBuddy把内勤最费眼的“对数”交给 AI：票、货、款、单等多份数据放一起，按行业口径逐项比对，对得上的打勾、对不上的标出来、差多少一目了然，该预警的主动挑出来。不止财务对账，各类“多份数据", "skill_id": "skill_2090731715759472640" } ], "experts": [ { "id": "neiqin-expert", "name": "B2B贸易内勤专家", "description": "懂B2B贸易的内勤专家＝AI版录单员+档案室：识单、发票管理、智能对账、出表出单、盯订单进度。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_5Hu9b9KuHQwa" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "refund-order-value-risk-list", "title": "退款订单金额风险清单", "title_en": "Refund Order Value Risk List", "subtitle": "近30天退款订单金额与商品风险排查", "subtitle_en": "A practical Weimob playbook for refund order value risk list", "description": "做退款订单金额风险清单时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“退款订单金额风险清单”展开：汇总近30天退款订单，按退款金额、退款频次和商品集中度生成风险清单，标记需要优先核查的订单与商品，并给出处置建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、订单履约运营人员、经营分析团队。", "description_en": "The hardest part of refund order value risk list is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates refund order value risk list through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Order Fulfillment specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "退款订单", "退款金额", "风险清单", "商品集中度", "处置建议" ], "tags_en": [ "Refund Order Value Risk List", "Mall Operations", "Order Fulfillment", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "汇总近30天退款订单，按退款金额、退款频次和商品集中度生成风险清单，标记需要优先核查的订单与商品，并给出处置建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a refund order value risk list covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" }, { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "regulation-version-diff-reader-1", "title": "反垄断法 2007 vs 2022 修订对照解读", "title_en": "AML 2007 vs 2022 Revision Diff Report", "subtitle": "公平竞争审查、平台经济、罚款幅度三大重点 + 草案转正式稿对照", "subtitle_en": "Fair competition review, platform economy & fine escalations, with draft-to-final comparison", "description": "想搞清一部法规改了什么、为什么改、影响谁，最怕的就是新旧条文混在一起逐字比到眼瞎。这个案例演示了如何把《反垄断法》2007 年法、2021 年修正草案、2022 年修正正式稿三份文本交给 AI：先由法规数据库拉取全部历史版本与条文原文，再对相邻版本做逐词级 diff，把公平竞争审查制度入法、平台经济数据算法条款、罚款幅度大幅上调（最高 10 倍）等 15 处变更按『事实—规则—动因—影响—趋势』五层拆解，并聚合立法机关说明与黄勇、王先林、王晓晔等学者观点，最终输出一份带时间线、变更统计和双视图（并排/统一）对比的 HTML 报告。\n\n适合：企业法务与合规人员、反垄断领域律师、政策研究者、关注平台经济监管的投资人、需要做法规沿革梳理的法律学习者。", "description_en": "Trying to figure out what a regulation changed, why, and who it affects is painful when old and new provisions are jumbled together for line-by-line comparison. This case demonstrates handing the 2007 AML, the 2021 amendment draft, and the 2022 final revision to AI: historical versions and article texts are pulled from a regulation database, adjacent versions are diffed word-by-word, and 15 changes — from the codification of fair competition review and data/algorithm rules for platform economy to fine escalations of up to 10x — are analyzed through a five-layer framework (facts-rules-motivations-impact-trends), with official legislative statements and expert views from Huang Yong, Wang Xianlin, Wang Xiaoye aggregated into an HTML report featuring a timeline, change statistics, and side-by-side/unified diff views.\n\nFor: corporate legal & compliance teams, antitrust lawyers, policy researchers, investors tracking platform regulation, and law students working on regulatory evolution analysis.", "categories": [ "education", "search" ], "tags": [ "反垄断法", "法规版本对照", "公平竞争审查", "平台经济监管", "罚款幅度" ], "tags_en": [ "Antitrust Law", "regulation version diff", "fair competition review", "platform economy regulation", "penalty escalation" ], "artifact_type": "html", "preview": "report.html", "prompt": "请对照《中华人民共和国反垄断法》2007 年版本与 2022 年修订版本（2022.08.01 施行），并把 2021 年 10 月公布的《修正草案（征求意见稿）》一并纳入对比，重点看三处：①公平竞争审查制度（新增第五条）；②平台经济领域反垄断（新增第九条、第十九条，第二十二条数据算法条款）；③罚款幅度（第五十六、五十八、六十二、六十三条）。\n\n要求：1）条文原文必须从北大法宝 MCP 真实查询获取并注明版本与链接，禁止凭记忆；2）相邻版本做逐词级 diff，输出 git 风格双视图（并排/统一）HTML 报告；3）每条变更按『事实—规则—动因—影响—趋势』分层解读，至少写到动因层；4）用 WebSearch 检索全国人大审议说明、学者与律所观点并附署名、机构、日期、原文链接，标注官方说明/支持/谨慎立场，归纳共识与分歧；5）报告需包含版本时间线、变更统计（修改/新增/删除/条号迁移）、类型筛选器与深色模式；6）草案阶段的评论必须注明『正式稿可能已调整』，并单独标注草案与正式稿的文字差异；7）报告尾部附免责声明（不构成法律意见）。数据须来自真实查询并注明来源，禁止示意或模拟数据。", "prompt_en": "Compare the 2007 version of the PRC Anti-Monopoly Law with the 2022 revision (effective 2022-08-01), and include the October 2021 Amendment Draft (exposure draft) in the comparison. Focus on three areas: 1) the fair competition review regime (new Article 5); 2) platform-economy antitrust rules (new Articles 9 and 19, and the data/algorithm clause in Article 22); 3) penalty amounts (Articles 56, 58, 62, 63).\n\nRequirements: 1) Article texts must be retrieved in real time from the pkulaw MCP database, with version and link cited — no recall from memory; 2) produce word-level diffs between adjacent versions as a git-style HTML report with side-by-side/unified views; 3) analyze each change through a five-layer framework (facts–rules–motivations–impact–trends), reaching at least the motivation layer; 4) use WebSearch to gather NPC review statements and views from scholars and law firms, each with name, affiliation, date and source link, labelled as official/supportive/cautious, and summarize consensus and disputes; 5) the report must include a version timeline, change statistics (modified/added/deleted/renumbered), a type filter and dark-mode support; 6) comments on the draft stage must note \"the final version may have adjusted\", with explicit draft-vs-final wording differences; 7) end with a disclaimer (not legal advice). Data must come from real queries with sources cited; no illustrative or simulated data.", "skills": [ { "id": "regulation-version-diff-reader", "name": "法规历史版本对照解读", "description": "对法律、法规、规章及司法解释做历史版本对照解读，输出带时间线、变更统计与双视图对比的 HTML 报告。", "skill_id": "skill_2090718573269643264" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 99, "source": "official", "usage_count": 0, "created_at": "2026-08-21T15:35:00+08:00", "updated_at": "2026-08-21T15:36:00+08:00", "artifacts": [ { "file": "report.html", "label": "反垄断法修订对照解读 HTML 报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "regulation-version-diff-reader-2", "title": "《数据安全法》版本对照解读", "title_en": "Data Security Law Version Diff Analysis", "subtitle": "征求意见稿 vs 2021 正式稿：分类分级、重要数据义务、出境评估、罚则四大重点", "subtitle_en": "Draft vs 2021 final: data classification, important-data duties, cross-border assessment & penalties", "description": "法规改版最怕的就是『不知道改了啥、改了影响谁』——尤其像《数据安全法》这种基础性法律，草案到正式稿的每一处措辞变化都可能直接改写企业的合规义务。这个案例用北大法宝逐条拉取《数据安全法》2020 年征求意见稿与 2021 年正式稿条文，对分类分级、重要数据处理者义务、数据出境安全评估、罚则变化四类重点做逐词级 diff，叠加官方说明与专家观点（网信办、人大法工委、律所），最终产出一份带版本时间线、变更统计与双视图对比（并排/统一）的 HTML 报告，并附半导体制造企业的合规影响要点。\n\n适合：合规/法务人员做法规沿革研究；数据安全岗位评估新法义务落地；半导体等制造业企业梳理重要数据识别与出境合规；研究政策变化的分析师与咨询顾问。", "description_en": "Nothing is more annoying than a revised law where you cannot tell what actually changed and who is affected. This case pulls every article of the Data Security Law (2020 exposure draft vs 2021 final version) via PKULaw, runs a token-level diff on four focus areas - data classification, important-data processor duties, cross-border data security assessment and penalties - and layers on official explanations and expert commentary (CAC, NPC Legislative Affairs Commission, law firms). The result is a self-contained HTML report with a version timeline, change statistics and dual-view (side-by-side / unified) diffs, plus compliance impact notes for semiconductor manufacturers.\n\nFor: compliance and legal teams doing regulatory history research; data-security officers assessing new obligations; semiconductor and manufacturing firms identifying important data and export compliance; analysts and consultants tracking policy shifts.", "categories": [ "education", "search" ], "tags": [ "法规解读", "数据安全", "版本对比", "合规", "政策研究" ], "tags_en": [ "law analysis", "data security", "version diff", "compliance", "policy research" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.html", "prompt": "请对照《中华人民共和国数据安全法》征求意见稿和 2021 年正式稿（包括草案）做版本对照解读，重点看四类内容：① 数据分类分级制度；② 重要数据处理者义务；③ 数据出境安全评估；④ 罚则变化。\n\n我们是半导体制造企业，请在报告里顺带写合规影响要点。\n\n要求：\n1. 用北大法宝 MCP 检索法规全部历史版本并构建时间线（get_law_list → get_article 按版本全名取条）；北大法宝未收录的草案条文，以中国人大网官方公布文本为准并注明来源。\n2. 对两版对应条文做逐词级 diff，输出变更统计（修改/新增/删除条数、字符增删规模、条号迁移）。\n3. 每条变更按『事实—规则—动因—影响—趋势』五层写解读卡片，重大变更至少写到动因层，并注明『针对草案，正式稿可能已调整』。\n4. 用联网搜索聚合官方说明与专家观点，专家须实名并附来源链接，标注共识与分歧。\n5. 输出带版本时间线、变更统计、并排/统一双视图 diff 的 HTML 报告，资源全内联、可离线打开。\n6. 报告尾注『不构成法律意见』。", "prompt_en": "Compare the exposure draft (including the bill) of the Data Security Law of the People's Republic of China with its 2021 final version, focusing on four areas: 1) data classification and grading; 2) obligations of important-data processors; 3) cross-border data security assessment; 4) penalty changes.\n\nWe are a semiconductor manufacturer - please include compliance impact notes for us in the report.\n\nRequirements:\n1. Use the PKULaw MCP to retrieve all historical versions of the law and build a timeline (get_law_list -> get_article with the full versioned title); where PKULaw does not hold the draft, use the officially published text on npc.gov.cn and cite the source.\n2. Run a token-level diff on the corresponding articles of the two versions and output change statistics (modified/added/deleted counts, character churn, renumbered items).\n3. Write interpretation cards for every change following the fact-rule-motivation-impact-trend framework; reach at least the motivation layer for major changes and note 'based on the draft, the final version may have adjusted' where relevant.\n4. Aggregate official statements and expert commentary via web search; experts must be named with source links, and consensus vs disputes must be flagged.\n5. Output an HTML report with a version timeline, change statistics, and side-by-side/unified diff views; all assets must be inlined and the file must open offline.\n6. End the report with a 'not legal advice' disclaimer.", "skills": [ { "id": "regulation-version-diff-reader", "name": "法规历史版本对照解读", "description": "对法律、法规、规章及司法解释做历史版本对照解读，输出带时间线、变更统计与双视图对比的 HTML 报告", "skill_id": "skill_2090718573269643264" } ], "experts": [], "mcps": [ { "id": "pkulaw", "name": "北大法宝", "description": "法律检索与核验 MCP：法规版本清单、逐条取文、语义检索、名称识别" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-08-21T15:31:00+08:00", "updated_at": "2026-08-21T15:33:00+08:00", "artifacts": [ { "file": "report.html", "label": "report" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "regulation-version-diff-reader-3", "title": "公司法修订对照解读：2018修正 → 2023修订", "title_en": "Company Law Revision Comparison: 2018 Amendment → 2023 Revision", "subtitle": "注册资本五年实缴、股东失权与催缴、股权转让出资责任、治理结构四大主题逐条 diff", "subtitle_en": "Clause-by-clause diff on 5-year capital contribution, share forfeiture, transfer liability & governance", "description": "公司法改了，最怕搞不清新旧版本到底差在哪：注册资本认缴是不是改成实缴了？股东不按期出资怎么办？股权转出去了还要不要对没缴的出资负责？监事会被审计委员会取代了吗？本案例演示用「法规历史版本对照解读」能力，调取《中华人民共和国公司法》2018 年修正与 2023 年修订两个版本的官方文本，围绕注册资本认缴改实缴、股东失权与催缴、股权转让后的出资责任、董事会/监事会治理结构四大主题逐条做逐词级 diff，产出一份带版本时间线、变更统计、并排/统一双视图对比、分层解读与专家观点聚合的 HTML 报告。\n\n适合：需要做法规沿革研究的公司法务与合规人员、处理公司设立与股权交易业务的律师、关注注册资本实缴新政的企业创始人与财务负责人、研究公司治理与资本制度的学者与研究生。", "description_en": "When the Company Law is revised, the hardest part is figuring out exactly what changed between versions: Did the registered capital move from subscription to paid-in? What happens if a shareholder fails to contribute on time? After transferring equity, are you still liable for unpaid contributions? Is the supervisory board being replaced by the audit committee? This case demonstrates the Regulation Version Diff Reader capability: it pulls the official texts of the 2018 Amendment and 2023 Revision of the PRC Company Law and runs a token-level diff on four key themes — the shift from subscription to a 5-year paid-in capital rule, share forfeiture and capital calls, post-transfer contribution liability, and board/supervisory governance structure — producing an HTML report with a version timeline, change statistics, side-by-side & unified diff views, layered analysis, and expert opinions. \n\nFor: in-house counsel and compliance officers tracking regulatory evolution, lawyers handling company incorporation and equity transactions, founders and CFOs affected by the paid-in capital rules, and scholars researching corporate governance and capital systems.", "categories": [ "education", "search" ], "tags": [ "公司法", "法规解读", "版本对比", "注册资本", "公司治理" ], "tags_en": [ "Company Law", "Regulation Analysis", "Version Diff", "Registered Capital", "Corporate Governance" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "公司法修订对照解读报告（时间线 + 变更统计 + 双视图 diff）" } ], "prompt": "请对照《中华人民共和国公司法》2018 年修正与 2023 年修订两个版本，重点看以下四大主题：① 注册资本认缴改实缴（五年缴足、加速到期）；② 股东失权与催缴（董事核查催缴义务、失权程序与救济）；③ 股权转让后的出资责任（未届期转让、瑕疵转让的补充/连带责任）；④ 董事会/监事会治理结构（审计委员会替代监事会、小型公司机构简化、股东会与董事会职权重构）。\n\n要求：\n1. 条文与版本信息必须通过北大法宝检索获取真实官方文本并注明来源，禁止凭记忆或编造；\n2. 先在两版中分别定位相关条号（条号会迁移，不要假设新旧版同条号是同一内容），再逐条提取原文；\n3. 输出一份单文件 HTML 报告（资源全内联、可离线打开），包含：版本时间线、变更统计与字符增删规模、逐条 git 风格 diff（并排对照与统一视图可切换）、按『事实→规则→动因→影响→趋势』分层的解读卡片、官方说明与专家观点聚合（实名+来源链接）；\n4. 区分新设公司与存量公司的适用规则与时间表，报告末尾附免责声明。", "prompt_en": "Compare the 2018 Amendment and the 2023 Revision of the PRC Company Law, focusing on four themes: ① registered capital from subscription to paid-in (5-year payment deadline, accelerated maturity); ② shareholder forfeiture and capital calls (directors' duty to verify and call, forfeiture procedure and remedies); ③ post-transfer contribution liability (supplementary/joint liability for unexpired and defective transfers); ④ board/supervisory governance structure (audit committee replacing supervisory board, simplified structures for small companies, rebalancing of shareholders' meeting vs board powers).\n\nRequirements:\n1. Obtain all article texts and version info from real official sources via PKULaw search and cite them; never rely on memory or fabricate; 2. Locate the relevant article numbers in each version separately (numbers migrate between versions; never assume the same number means the same content), then extract the original text of each article; 3. Deliver a single-file HTML report (all resources inlined, openable offline) containing: a version timeline, change statistics with added/deleted character scale, per-article git-style diff (side-by-side and unified views switchable), analysis cards layered as Facts → Rules → Rationale → Impact → Trends, and aggregated official statements and expert opinions (with real names and source links); 4. Distinguish rules and timelines for newly established vs existing companies, and append a disclaimer at the end.", "skills": [ { "id": "regulation-version-diff-reader", "name": "法规历史版本对照解读", "description": "对法律、法规、规章及司法解释做历史版本对照解读：拉取全部历史版本，逐条提取相邻版本条文做逐词级 diff，分条归纳变更实质并聚合官方说明与专家观点，输出带时间线、变更统计与双视图对比的 HTML 报告", "skill_id": "skill_2090718573269643264" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-08-21T07:27:13Z", "updated_at": "2026-08-21T07:30:00Z", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-infographic-goujian-sword", "title": "越王勾践剑科普图解", "title_en": "Anatomical Diagram of the Sword of Goujian", "subtitle": "天下第一剑 · 一图拆解春秋青铜复合剑", "subtitle_en": "First Sword under Heaven — single-page breakdown of a Spring and Autumn bronze composite sword", "description": "每次想把一件国宝讲清楚，最头疼的就是：文字一大堆、图没有重点、读者划两下就走。这个案例演示了如何用 tanyuan-relic-infographic-expert 把春秋时期镇国之宝「越王勾践剑」拆成一张可展陈、可传播的单文件 HTML 长图：从尺寸/铸造/铭文/暗格花纹/科学检测，到历史地位与流传，配探元知识库来源标注，开箱即展。\n\n适合：博物馆社教老师做展陈小报、文博自媒体做公众号内容、配展板的微缩图、博物馆语音导览的视觉伴侣、教学课件中关于春秋青铜器单元的封面与解析页。", "description_en": "Explaining a national treasure in detail is painful: walls of text, no clear visuals, and readers scroll past in two seconds. This case shows how the tanyuan-relic-infographic-expert turns the Spring and Autumn treasure 「Sword of Goujian」 into a single-file HTML infographic ready for exhibition and sharing: from dimensions / metallurgy / inscription / dark-pattern inlay / scientific tests to historical provenance, every fact attributed to the TanYuan knowledge base.\n\nSuitable for: museum educators producing exhibition handouts, cultural-heritage content creators for WeChat public accounts, micro posters paired with exhibition panels, visual companions for museum audio guides, and cover pages or analysis sheets for school-courseware units on Spring and Autumn bronzeware.", "categories": [ "creation", "education" ], "tags": [ "文物图解", "春秋青铜", "镇馆之宝", "湖北省博物馆", "单文件 HTML" ], "tags_en": [ "relic infographic", "Spring and Autumn bronze", "museum treasure", "Hubei Provincial Museum", "single-file HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请调用 tanyuan-relic-infographic-expert 专家，按其定义的「文物图解长图」工作流，为「越王勾践剑」生成一张单文件 HTML 科普图解长图。\n\n要求：\n1. 数据真实：使用 tanyuan-wb-mcp-server 的 search_relics 与 search_movable_relics 真实检索（查询短语「越王勾践剑」）—— 本次会话已确认 MCP 可用，检索字段包含通长 55.7cm / 重量 0.875kg / 11 道同心圆（间隔 0.2mm）/ 鸟虫书铭文「越王鸠浅自作用剑」/ 蓝色琉璃 + 绿松石剑格 / 1965 年湖北江陵望山一号楚墓出土 / 一级文物·禁止出境 / 学界「硫化处理」与「富锡细晶层热扩散工艺」两种工艺解释并存 / 「嫁妆说 vs 战利品说」楚墓出土背景辩论 / 方壮猷主持释读（郭沫若「邵滑」说 vs 唐兰「鸠浅」说）。\n2. 文物本体图：探元检索无 images 字段时，按专家 SKILL.md 第 5.2 条走 Wikimedia Commons 真实照片（如 File:Sword of Goujian, Hubei Provincial Museum, 2015-04-06 01-edit.jpg，CC BY-SA 4.0，作者蘇伯軒），下载 base64 内联；aspect-ratio 等于真实图片比例（如 1200/251）。\n3. 视觉：使用简约版模板（默认款），浅色宣纸底，配色可切「经典蓝红 / 青绿·石青石绿 / 朱红·朱砂宫墙」三主题之一。\n4. 结构：标题区 → HERO 顶图（真实照片 base64 内联，比例=真实 aspect-ratio）→ 器物特征（4—6 条，逐处标 [探元文物库] 或 [探元知识库]）→ 工艺详解（含两种解释并存）→ 铭文释读过程（方壮猷→郭沫若→唐兰）→ 历史意义与楚墓出土背景（嫁妆说 vs 战利品说）→ 科学检测与研究（谭德睿 3000 倍显微镜）→ 教学辅助 SVG 示意（剑首同心圆 + 菱形暗格纹结构）→ 底部信息条 → 页脚来源标注。\n5. 硬约束：单文件 HTML、CSS/图片/SVG 全部内联、不得 https:// 外链、不得 href=\"#\" 空锚点、不得 emoji；所有数值与日期逐处标源。", "prompt_en": "Use the tanyuan-relic-infographic-expert to produce, by its defined single-page HTML infographic workflow, an infographic for the 「Sword of Goujian」 (Spring and Autumn period).\n\nRequirements:\n1. Real data: query dimensions, metallurgy, inscription, excavation context and historical provenance via tanyuan-wb-mcp-server (search_relics + search_movable_relics, query string「越王勾践剑」) — MCP confirmed reachable this session. Fields include: total length 55.7cm / weight 0.875kg / 11 concentric rings (0.2mm spacing) / bird-worm-script inscription「越王鸠浅自作用剑」/ blue glass + turquoise sword guard / 1965 Jiangling Wangshan No.1 Chu tomb / Grade-1 relic prohibited from overseas exhibition / two coexisting craft interpretations (sulfurization vs tin-rich fine-crystal layer hot-diffusion) / 「dowry vs war-trophy」 debate on Chu-tomb context / Fang Zhuangyou's reading session (Guo Moruo's「Shaohua」vs Tang Lan's「Jiuqian」).\n2. Relic photo: when TanYuan returns no `images` field, follow expert SKILL.md §5.2 to fetch from Wikimedia Commons (e.g. File:Sword of Goujian, Hubei Provincial Museum, 2015-04-06 01-edit.jpg, CC BY-SA 4.0, photographer 蘇伯軒); download + base64 inline; aspect-ratio equals actual image (e.g. 1200/251).\n3. Visual: «concise edition» template (default) on light xuan-paper; pick one of three themes.\n4. Structure: title block → HERO banner (real photo, base64 inlined, ratio = actual aspect-ratio) → physical features (4–6 items, every entry tagged [探元文物库] / [探元知识库]) → craft details (with two coexisting interpretations) → inscription reading history (Fang Zhuangyou → Guo Moruo → Tang Lan) → historical significance & Chu-tomb context (dowry vs war-trophy) → scientific detection & research (Tan Derui 3000x microscope) → teaching-auxiliary SVG schematic (sword pommel concentric rings + diamond lattice pattern structure) → bottom info bar → footer sources.\n5. Hard constraints: single HTML, all CSS/images/SVG inlined, no https:// links, no href=\"#\" anchors, no emoji; every number and date source-tagged.", "skills": [], "experts": [ { "id": "tanyuan-relic-infographic", "name": "文物图解专家", "description": "基于探元文物库+知识库+MCP检索的单文件文物图解专家，使用简约版/典藏版模板输出可展陈、可传播的 HTML 长图", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_OPJE3pWwnB1j" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:06:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "越王勾践剑科普图解" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "relic-lookup-bronze-protruding-mask", "title": "三星堆青铜纵目面具速查", "title_en": "Quick Lookup of the Sanxingdui Bronze Protruding-Eyes Mask", "subtitle": "商代古蜀神像 · 三星堆博物馆藏 · 一级文物", "subtitle_en": "Shang-dynasty Shu sacred image · Sanxingdui Museum · Grade-1 relic", "description": "想讲清楚那件「眼睛凸得像柱子、额头还嵌着一条龙」的青铜面具，常常还没开口就卡壳——尺寸多大、哪件出土、哪种解释更主流、铭文铭文它有没有，搜一晚上资料也凑不齐一栏对齐的字段。\n\n这个案例演示了如何用 tanyuan-relic-lookup-expert 文物速查助手，用户说一句「三星堆青铜纵目面具的基本信息」，专家就在一轮内调用探元文物库结构化检索、补一次知识库（造型/工艺/出土/学术争议），按档位 A 出卡：表格 - 核心特征 - 工艺与含义 - 出土与流传 - 配图 - 来源标注，每一行都能溯源到检索结果，绝不靠「大概」补位。\n\n适合：青铜器馆策展与社教老师、文博科普自媒体、博物馆语音导览撰稿、文史与艺术史专业学生、C 端文博爱好者做笔记与推文素材。", "description_en": "Trying to explain that 'column-shaped eyes and a dragon-shaped forehead' bronze mask often leaves you stuck before you even start — its dimensions, which one came from which pit, which theory is mainstream, even whether any inscription exists, are scattered across half a night's worth of searching and never align into one clean column.\n\nThis case shows how to use the tanyuan-relic-lookup-expert: when the user asks 'basic information about the Sanxingdui bronze protruding-eyes mask', the expert within one round calls the TanYuan relic structured search, supplements with a knowledge-base pass (shape / craft / excavation / academic debates), and emits a level-A card: table → key features → craft and meaning → excavation and provenance → image → source attribution, every row traceable to retrieval and never filled in by guesswork.\n\nBest for: bronze-ware museum curators and docents, cultural-heritage content creators, museum audio-guide writers, art-history and culture-history students, and serious cultural-relic enthusiasts taking notes or building social-media material.", "categories": [ "business" ], "tags": [ "文物速查", "三星堆", "青铜面具", "古蜀文明", "一级文物" ], "tags_en": [ "Relic Lookup", "Sanxingdui", "Bronze Mask", "Ancient Shu", "Grade-1 Relic" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "青铜纵目面具速查卡" } ], "prompt": "请按 tanyuan-relic-lookup-expert 文物速查助手的工作流，给出一张「青铜纵目面具」的速查卡。\n\n# 用户问题\n三星堆青铜纵目面具的基本信息\n\n# 检索步骤\n1. 先调一次 `search_relics`，query：「青铜纵目面具的基本信息，请返回文物名称、别名、年代、类型、类别、等级、馆藏机构、出土地、尺寸、封面、基本介绍和特征介绍」。\n2. 拿到结构化字段后，再用一次 `search_movable_relics`，query：「三星堆 青铜纵目面具」，目的是补足造型、工艺、出土与学术争议。\n3. 同名近名消歧：用户问的是「纵目面具」，若命中的是「青铜大面具」或「青铜戴冠纵目面具」等相近器物，必须列清年代 - 馆藏 - 出土地再请用户选定，不能蒙混。\n\n# 输出档位 A（单件精确命中 → 速查卡）\n字段渲染顺序：标题 → 别名（如有） → 表格 → 核心特征（3-5 条要点，每条一句） → 工艺与含义（可含学术争议并存说法） → 出土与流传 → 配图（cover 字段以 http/https 开头才内联，相对路径直接跳过）→ 底部来源标注。\n\n# 硬约束\n- 不臆造：尺寸 - 馆藏 - 出土地字段以检索结果为唯一来源，缺则跳过该行，不写「暂无 - 未知」。\n- 图片：cover 字段不以 http/https 开头则跳过图片行；保留原状不强行展示坏图。\n- 不复述检索过程：禁用「检索到 N 条」「调用数据库」等过程叙述与工具名 / requestId / 原始 JSON / SQL。\n- 全程中文，统一以「腾讯探元文物库 / 腾讯探元知识库」作为资料来源标识。\n\n# 用户输入\n三星堆青铜纵目面具的基本信息", "prompt_en": "Use the tanyuan-relic-lookup-expert workflow to produce a quick-lookup card for the 'Bronze Protruding-Eyes Mask'.\n\n# User question\nBasic information about the Sanxingdui bronze protruding-eyes mask\n\n# Retrieval steps\n1. Call `search_relics` once, query: \"Basic information about the Sanxingdui bronze protruding-eyes mask, please return relic name, alias, era, type, category, grade, holding museum, findspot, dimensions, cover, basic introduction and feature introduction\".\n2. After receiving structured fields, call `search_movable_relics` once with query \"Sanxingdui bronze protruding-eyes mask\" to supplement shape, craftsmanship, excavation and academic debate context.\n3. Same/near-name disambiguation: the user is asking for the 'protruding-eyes mask'; if results hit 'bronze large mask' or 'bronze crowned protruding-eyes mask' or other close-but-different pieces, you MUST list each with dynasty - museum - findspot and ask the user to pick one — never disguise one as the other.\n\n# Output level A (single-hit → factsheet card)\nField order: title → alias (if any) → table → 3-5 key features → craft and meaning (academic debates may coexist) → excavation and provenance → image (only inline if cover URL starts with http/https; relative paths skip the image line entirely) → bottom source attribution.\n\n# Hard constraints\n- No fabrication: dimensions, museum, findspot etc. must come from retrieval only — missing fields get skipped, never 'unknown / TBD'.\n- Images: cover URLs not starting with http/https skip the image line silently — never render a broken image.\n- No process narration: forbid 'retrieved N hits', 'calling database', tool names, requestId, raw JSON, SQL.\n- Default Chinese; cite 'Tencent TanYuan relic database' / 'Tencent TanYuan knowledge base' as the source.\n\n# User input\nBasic information about the Sanxingdui bronze protruding-eyes mask", "skills": [], "experts": [ { "id": "tanyuan-relic-lookup-expert", "name": "文物速查专家", "description": "依托腾讯探元文物库与知识库，报出文物名称即可速查年代、尺寸、出土地、馆藏机构与核心特征。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9f48a9c74dcd", "expert_id": "ex_gaZLEEnjaBQw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:05:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-lookup-jun-kiln-tripod-censer", "title": "钧窑天青釉贴塑花三足炉速查", "title_en": "Quick Lookup of the Jun-Kiln Sky-Blue Glazed Tripod Censer", "subtitle": "宋代钧瓷礼器 · 吉林省博物院藏 · 雨过天青云破处", "subtitle_en": "Song-dynasty Jun porcelain ritual vessel · Jilin Provincial Museum · 'sky-blue after rain' classic", "description": "提起「钧窑三足炉」，外行只知道天青色很好看，内行却常被一堆零散信息烦到——哪个馆藏、什么型号、贴塑纹饰怎么辨认、蚯蚓走泥纹是不是真的看得见，开口讲一分钟，回答就断线。\n\n这个案例演示了如何用 tanyuan-relic-lookup-expert 文物速查助手，用户问一句「宋代钧窑天青釉贴塑花三足炉的基本信息」，专家就在一轮内调探元文物库结构化检索，再补一次知识库（釉料配方 - 窑变工艺 - 贴塑装饰 - 用途场景），按档位 A 出卡：表格 - 核心特征 - 釉色工艺 - 用途与学术价值 - 馆藏来源，每一行都能溯源到检索结果，不靠「大概」「应该是」补位。\n\n适合：瓷器与钧窑方向研究者、博物馆社教与陈列文案、文物拍卖与鉴赏专栏作者、艺术院校宋瓷教学、文博文创产品开发者。", "description_en": "Talk of a 'Jun-kiln tripod censer' and laypeople only know 'sky-blue looks nice', while insiders are tripped up by scattered facts — which museum holds it, what model, how to recognize the applied-flower decoration, whether the 'earthworm-in-mud' crackle is really visible — answer breaks off within a minute.\n\nThis case shows how to use the tanyuan-relic-lookup-expert: when the user asks 'basic information about the Song-dynasty Jun-kiln sky-blue glazed applied-flower tripod censer', the expert within one round calls the TanYuan relic structured search, then supplements with a knowledge-base pass (glaze chemistry, kiln-transmutation technique, applied decoration, usage context), and emits a level-A card: table → key features → glaze and craft → use and academic value → museum source, every row traceable to retrieval and never filled in by guesswork.\n\nBest for: porcelain and Jun-kiln researchers, museum docent and exhibition copywriters, auction and connoisseurship column writers, art-school Song porcelain teachers, and cultural-product developers.", "categories": [ "business" ], "tags": [ "文物速查", "钧窑", "宋代瓷器", "天青釉", "三足炉" ], "tags_en": [ "Relic Lookup", "Jun Kiln", "Song Porcelain", "Sky-Blue Glaze", "Tripod Censer" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "钧窑天青釉贴塑花三足炉速查卡" } ], "prompt": "请按 tanyuan-relic-lookup-expert 文物速查助手的工作流，给出一张「钧窑天青釉贴塑花三足炉」的速查卡。\n\n# 用户问题\n宋代钧窑天青釉贴塑花三足炉的基本信息\n\n# 检索步骤\n1. 先调一次 `search_relics`，query：「钧窑天青釉贴塑花三足炉的基本信息，请返回文物名称、别名、年代、类型、类别、等级、馆藏机构、出土地、尺寸、封面、基本介绍和特征介绍」。\n2. 拿到结构化字段后，再用一次 `search_movable_relics`，query：「钧窑 三足炉 釉色」，目的补足釉料 - 窑变 - 贴塑工艺与用途。\n3. 同名近名消歧：钧窑系三足炉有多个型号（天青釉贴塑花 / 紫斑 / 双耳等），命中多件时按年代 - 馆藏 - 尺寸 / 等级区分，请用户选定。\n\n# 输出档位 A（单件精确命中 → 速查卡）\n字段渲染顺序：标题 → 别名（如有） → 表格 → 核心特征（3-5 条要点，每条一句） → 釉色与工艺（贴塑纹饰 + 釉料 / 窑变） → 用途与学术价值 → 配图（cover 字段以 http/https 开头才内联，相对路径直接跳过）→ 底部来源标注。\n\n# 硬约束\n- 不臆造：尺寸 - 馆藏 - 出土地字段以检索结果为唯一来源，缺则跳过该行，不写「暂无 - 未知 - 大概」。\n- 图片：cover 字段不以 http/https 开头则跳过图片行，保留原状不强行展示坏图。\n- 不复述检索过程：禁用「检索到 N 条」「调用数据库」等过程叙述与工具名 / requestId / 原始 JSON / SQL。\n- 全程中文，统一以「腾讯探元文物库 / 腾讯探元知识库」作为资料来源标识。\n\n# 用户输入\n宋代钧窑天青釉贴塑花三足炉的基本信息", "prompt_en": "Use the tanyuan-relic-lookup-expert workflow to produce a quick-lookup card for the 'Jun-kiln Sky-Blue Glazed Applied-Flower Tripod Censer'.\n\n# User question\nBasic information about the Song-dynasty Jun-kiln sky-blue glazed applied-flower tripod censer\n\n# Retrieval steps\n1. Call `search_relics` once, query: \"Basic information about the Jun-kiln sky-blue glazed applied-flower tripod censer, please return relic name, alias, era, type, category, grade, holding museum, findspot, dimensions, cover, basic introduction and feature introduction\".\n2. After receiving structured fields, call `search_movable_relics` once with query \"Jun kiln tripod censer glaze color\" to supplement glaze chemistry, kiln-transmutation technique and usage context.\n3. Same/near-name disambiguation: many Jun-kiln tripod censers exist (sky-blue flower-applied / purple-spot / double-eared …). For multiple hits, list each with dynasty - museum - dimensions / grade and ask the user to pick.\n\n# Output level A (single-hit → factsheet card)\nField order: title → alias (if any) → table → 3-5 key features → glaze and craft (applied ornament + glaze chemistry / kiln-transmutation) → use and academic value → image (only inline if cover URL starts with http/https; relative paths skip the image line entirely) → bottom source attribution.\n\n# Hard constraints\n- No fabrication: dimensions, museum, findspot etc. must come from retrieval only — missing fields get skipped, never 'unknown / TBD'.\n- Images: cover URLs not starting with http/https skip the image line silently — never render a broken image.\n- No process narration: forbid 'retrieved N hits', 'calling database', tool names, requestId, raw JSON, SQL.\n- Default Chinese; cite 'Tencent TanYuan relic database' / 'Tencent TanYuan knowledge base' as the source.\n\n# User input\nBasic information about the Song-dynasty Jun-kiln sky-blue glazed applied-flower tripod censer", "skills": [], "experts": [ { "id": "tanyuan-relic-lookup-expert", "name": "文物速查专家", "description": "依托腾讯探元文物库与知识库，报出文物名称即可速查年代、尺寸、出土地、馆藏机构与核心特征。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9f48a9c74dcd", "expert_id": "ex_gaZLEEnjaBQw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:05:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-lookup-qingming-scroll", "title": "清明上河图速查", "title_en": "Quick Lookup: Along the River During the Qingming Festival", "subtitle": "北宋张择端存世孤本 · 知识库佐证 + 通行权威著录", "subtitle_en": "Zhang Zeduan's sole surviving scroll · knowledge-base evidence with canonical records", "description": "查一件大众名物，最怕两种结果：一是硬凑出一张全是「未知 / 暂无」的半张卡，二是查不到就直接把用户劝退。\n\n这个案例演示的是 tanyuan-relic-lookup-expert 的第三条路：用户问「清明上河图」，结构化文物库确实没有独立条目——但知识库里有扎实的可核验记载。助手不硬凑卡片、也不生造字段：先以更短 query 重试结构化检索，确认未收录后转向知识库检索，把能确认的事实稳稳定定写出来——作者张择端（存世孤本）、年代北宋末年、绢本设色长卷（通行著录约 24.8 × 528.7 cm）、故宫博物院藏；再给出构图、内容、价值三条核心看点，以及卷首「插柳」民俗、明清仇英仿本等知识库佐证。\n\n对无法从结构化字段确认的内容（精确尺寸、馆藏编号），案例在末尾明确标注「依据探元知识库文献与通行权威著录整理，若需精确到具体馆藏著录可进一步核查」——既交付可用信息，也讲清信息边界。\n\n走法对应专家的降级路径：结构化库未收录 → 知识库佐证 + 权威著录。这条路径常被新手忽略，遇冷就直接劝退用户；资深用法是把它当作一次换路径的检索——用户更欣赏一个「能讲清讲了多少、讲不清也讲清为什么」的助手。\n\n适合：博物馆社教老师讲案头查不到的大众名物、文物拍卖与媒体撰稿人快速核实、文博科普视频作者做脚本底稿、文博文创内容创作者。", "description_en": "Looking up a household-name artwork usually ends in one of two bad outcomes: a half-empty card padded with \"unknown / N/A\", or an outright dismissal when the first query misses.\n\nThis case demonstrates a third path. Asked about Along the River During the Qingming Festival, the structured relic database returns no standalone record — yet the knowledge base holds solid, verifiable material. The assistant neither pads the card nor invents fields: it retries the structured search with a shorter query, confirms the gap, then turns to the knowledge base and states what can be confirmed — Zhang Zeduan (sole surviving work), late Northern Song, handscroll in ink and colour on silk (canonically c. 24.8 x 528.7 cm), Palace Museum, Beijing — followed by three key observations on composition, content and value, plus knowledge-base evidence such as the willow-planting custom at the scroll's opening and the later Ming copies by Qiu Ying.\n\nWhere structured fields cannot confirm a value (exact dimensions, accession number), the case states so explicitly at the end: sourced from the knowledge base and canonical records, with further verification recommended for collection-specific data. It delivers usable information while being honest about its limits.\n\nThe workflow follows the expert's fallback path: structured miss, then knowledge-base evidence plus canonical records. Novices often abandon the lookup here; experienced users treat it as a route change.\n\nBest for: museum educators handling famous objects absent from the catalogue, auction and media writers fact-checking quickly, heritage video authors drafting scripts, and cultural-creative content producers.", "categories": [ "business" ], "tags": [ "文物速查", "清明上河图", "传世名画", "档位C降级", "稳妥作答" ], "tags_en": [ "Relic Lookup", "Qingming Scroll", "Chinese Masterpiece", "Level-C Graceful Degradation", "Faithful Answer" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "清明上河图速查" } ], "prompt": "请按 tanyuan-relic-lookup-expert 文物速查助手的工作流，回应用户对「清明上河图」的速查请求。\n\n# 用户问题\n清明上河图的基本信息\n\n# 检索步骤（请严格按顺序执行并尝试）\n1. 先调一次 `search_relics`，query：「清明上河图的基本信息，请返回文物名称、年代、馆藏机构、尺寸、封面」，预期可能命中为空。\n2. 若第一步无命中，**不展示任何错误**，再以更短 query 试一次：「清明上河图 / Along the River During the Qingming Festival」。\n3. 仍无命中时，调一次 `search_movable_relics`，query：「清明上河图」，从知识库中搜集可引用的相关片段（如「北宋张择端所画，现藏北京故宫」「中华传世之瑰宝」等）。\n4. 多轮仍无结构化字段时，**走档位 C**——见下方输出规则。\n\n# 输出档位 C（探元库未收录 → 稳妥作答，不硬凑卡片）\n模板：\n> 关于「<用户所问>」，探元文物库中暂时没有收录到对应的藏品记录。根据公开资料：\n>\n> <稳妥、准确的说明，可引用知识库检索片段>。\n>\n> 如果你能补充它的<年代 / 出土地 / 馆藏机构>等线索，我可以再帮你更准确地定位一次。\n>\n> 来源：腾讯探元知识库 + 公开资料\n\n# 硬约束\n- 不臆造：未在检索中确认的字段（具体尺寸、具体馆藏号、保护级别、入藏年份等）**不要编**。\n- 不复述检索过程：禁用「检索为空」「接口未命中」「数据库没收录」等技术性失败措辞，改为「探元文物库中暂时没有收录到对应的藏品记录」这一统一表达。\n- 不发散：清明上河图不是本专家速查范畴的典型对象，不要长篇展开历史背景，只给 2-4 句稳妥、可溯源的说明。\n- 全程中文，统一以「腾讯探元知识库 + 公开资料」作为资料来源标识。\n\n# 用户输入\n清明上河图的基本信息", "prompt_en": "Use the tanyuan-relic-lookup-expert workflow to respond to the user's lookup request for 'Along the River During the Qingming Festival'.\n\n# User question\nBasic information about Along the River During the Qingming Festival\n\n# Retrieval steps (execute in order)\n1. Call `search_relics` once, query: \"Basic information about Along the River During the Qingming Festival, please return relic name, era, holding museum, dimensions, cover\". The hit rate may be empty.\n2. If step 1 misses, do NOT show any error — retry with a shorter query: \"Along the River During the Qingming Festival\".\n3. If still empty, call `search_movable_relics` once with query \"Along the River During the Qingming Festival\" to collect quotable fragments (e.g., 'painted by Zhang Zeduan in the Northern Song, now in the Palace Museum, Beijing', 'a Chinese treasure').\n4. After multiple rounds still no structured fields, take **Level C** — see output rule below.\n\n# Output Level C (not in TanYuan library → graceful answer, no forced card)\nTemplate:\n> About 「\n> .\n>\n> If you can supplement , I can help you locate it more precisely.\n>\n> Source: Tencent TanYuan knowledge base + public sources\n\n# Hard constraints\n- No fabrication: fields unconfirmed in retrieval (specific dimensions, specific accession number, protection grade, accession year, etc.) MUST NOT be invented.\n- No process narration: forbid technical failure phrases like 'retrieval empty', 'endpoint missed', 'database has no entry'; use only the unified expression 'the TanYuan relic database has not yet recorded a matching entry'.\n- Stay in scope: Along the River During the Qingming Festival is not the typical object of this expert — do not deliver long historical background; only 2-4 steady, source-traceable sentences.\n- Default Chinese; cite 'Tencent TanYuan knowledge base + public sources' as the source.\n\n# User input\nBasic information about Along the River During the Qingming Festival", "skills": [], "experts": [ { "id": "tanyuan-relic-lookup-expert", "name": "文物速查专家", "description": "依托腾讯探元文物库与知识库，报出文物名称即可速查年代、尺寸、出土地、馆藏机构与核心特征。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9f48a9c74dcd", "expert_id": "ex_gaZLEEnjaBQw" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:05:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "report-analysis", "title": "报告解读", "title_en": "Report Interpretation", "subtitle": "看不懂的报告有人陪你看：指标逐项解读 + 综合研判 + 下一步建议 + 对症医院医生推荐，一页讲清", "subtitle_en": "Reports you can't read, walked through with you: item-by-item interpretation, overall assessment, next steps, and matching hospitals & doctors in one page", "description": "医疗报告是写给专业人士看的，可拿到它的却是普通人——满纸术语像一道门槛，把人挡在“自己身体状况”的门外。严不严重、怎么办、找谁看，这几个最朴素的问题，恰恰最难找到答案。\n\n以一份乳腺病理报告为例：把报告交给微小医，它会陪你逐项读懂病理诊断、免疫组化指标和淋巴结情况，给出综合研判和下一步该怎么做的建议，还帮你推荐杭州真实的三甲医院和对症医生——生成一页图文解读，存图发给家人，一起商量着来。", "description_en": "Medical reports are written for professionals, yet they land in the hands of ordinary people — the wall of jargon becomes a barrier between you and your own health. The simplest questions — how serious is it, what do I do, whom do I see — are often the hardest to answer.\n\nTake a breast pathology report as an example: hand it to Weixiaoyi, and it walks you through the diagnosis, immunohistochemistry markers and lymph node status one by one, gives an overall assessment with clear next steps, and recommends real top-tier hospitals and the right doctors in Hangzhou — all in a one-page interpretation you can save and share with family, so you can decide together.", "categories": [ "business" ], "tags": [ "报告解读", "病理报告", "就医推荐", "微医", "HTML生成" ], "tags_en": [ "Report Interpretation", "Pathology Report", "Care Recommendation", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "report-analysis.html", "artifacts": [ { "file": "report-analysis.html", "label": "病理报告解读与就医推荐 HTML 页面" } ], "prompt": "帮我解读这份报告，以及针对报告结果有什么推荐的医院和医生？将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203）。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果。", "prompt_en": "Interpret this report for me, and based on the results, recommend suitable hospitals and doctors. Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for sharing; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203). HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation.", "skills": [], "experts": [ { "id": "we-health-expert", "name": "微小医", "description": "微医官方 AI 健康管理专家：报告解读、就医推荐、健康与用药咨询", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_iLQ9gn0VnEOh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-20T16:06:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "research-ai-coding-business-model", "title": "代码的新经济：AI 编程工具商业模式研究", "title_en": "The New Economics of Code: A Study of AI Coding Tool Business Models", "subtitle": "Cursor / GitHub Copilot / Windsurf 的市场、变现与竞争深度拆解", "subtitle_en": "A deep dive into the market, monetization and competition of Cursor / GitHub Copilot / Windsurf", "description": "AI 编程工具已成为开发者的主流基础设施，但它们究竟如何赚钱、竞争格局走向何方？本案由「风向标」趋势研究专家运用桌面研究与多源数据三角验证方法，系统拆解了 AI 代码助手赛道：从 85% 的开发者采用率、约 80 亿美元市场规模，到 GitHub Copilot、Cursor、Windsurf 三强的定位差异，再到订阅/席位/用量/模型差价的四层变现栈与单位经济挑战。交付一份风投 deck 风格的单文件 HTML 深度研究报告，含核心读数、市场增长曲线、竞争象限、变现拆解与趋势展望。适合产品经理、投资人、创业者与关注 AI 商业化的从业者。", "description_en": "AI coding tools have become mainstream developer infrastructure, but how do they actually make money and where is competition heading? Using desk research and multi-source triangulation, the TrendResearcher expert dissects the AI code assistant market: from 85% developer adoption and an ~$8B market, to the positioning of GitHub Copilot, Cursor and Windsurf, and the four-layer monetization stack (subscription / seats / usage / model margin) and unit-economics challenges. Delivers a VC-deck-style single-file HTML deep-research report with key metrics, a market growth curve, a competitive quadrant, monetization breakdown and outlook. Ideal for product managers, investors, founders and anyone tracking AI commercialization.", "categories": [ "search" ], "tags": [ "深度研究", "AI编程工具", "商业模式", "市场分析" ], "tags_en": [ "Deep Research", "AI Coding Tools", "Business Model", "Market Analysis" ], "artifact_type": "html", "prompt": "帮我做一份关于 AI 编程工具赛道商业模式的深度研究报告，做成单文件 HTML，风投/产品分析 deck 风格（明亮现代 SaaS 商业风、靛蓝+翠绿强调、卡片化）。要覆盖：执行摘要核心结论、核心读数（开发者采用率、GitHub Copilot 用户与付费、Cursor 估值、市场规模）、市场增长曲线、竞争象限与玩家对比、变现模式拆解（订阅/企业席位/API用量/模型差价）、单位经济与挑战、SWOT 和趋势展望，关键数据要标注来源、页脚附完整来源清单，核心数据（采用率、Copilot/Cursor 用户与营收、市场规模等）须联网搜索真实公开来源并标注，仅在缺乏公开数据时对推算部分注明为估算。", "prompt_en": "Create a deep-research report on the business models of the AI coding tools market as a single-file HTML in a VC / product-analysis deck style (bright modern SaaS look, indigo + emerald accents, card-based). Cover: an executive-summary conclusion, key metrics (developer adoption, GitHub Copilot users & paid, Cursor valuation, market size), a market growth curve, a competitive quadrant and player comparison, a monetization breakdown (subscription / enterprise seats / API usage / model margin), unit economics and challenges, SWOT and an outlook. Cite sources for key data, include a full source list in the footer, and label estimated figures as real (web-searched).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "AI 编程工具商业模式深度研究报告" } ], "skills": [], "experts": [ { "id": "TrendResearcher", "name": "风向标", "description": "持续追踪行业和技术趋势，为产品战略提供前瞻性洞察", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e98dc4dfd783", "name_en": "Zara", "description_en": "Continuously tracks industry and technology trends to provide forward-looking insights for product strategy", "expert_id": "ex_DdCsaoq4AtcO" } ], "mcps": [], "scenario": "深度研究", "scenario_id": 103, "scenario_en": "Deep Research", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T11:00:00Z", "updated_at": "2026-06-22T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-cheetah-conservation", "title": "奔跑在消失的疆域：猎豹保护现状研究", "title_en": "Running Out of Range: A Study on Cheetah Conservation", "subtitle": "种群、威胁与跨境保护行动的深度研究报告", "subtitle_en": "A deep-research report on populations, threats, and transboundary conservation", "description": "全球最快的陆地动物正退出它曾经的疆域：野外仅存约 6,500 只、现存范围不足历史的 10%。本案由「深研研」专家以综合深度研究方法论，基于 IUCN 红色名录与 WWF（2024）真实数据，系统梳理猎豹的种群现状、历史下降趋势、三大威胁（人兽冲突/非法贸易/栖息地丧失）与跨境保护行动，交付一份图文并茂、来源可追溯的单文件 HTML 深度研究报告。适合野生动物保护研究者、科普内容创作者、环保 NGO 与关注生物多样性的读者。", "description_en": "The fastest land animal is running out of range: only ~6,500 remain in the wild, across just 10% of its historical territory. Using a comprehensive deep-research methodology and real data from the IUCN Red List and WWF (2024), the expert 'Iris' synthesizes cheetah population status, the century-long decline, three major threats (human-wildlife conflict, illegal trade, habitat loss) and transboundary conservation action into a single-file, source-traceable HTML report. Ideal for wildlife researchers, science communicators, environmental NGOs, and biodiversity-minded readers.", "categories": [ "search" ], "tags": [ "深度研究", "野生动物保护", "猎豹", "生物多样性" ], "tags_en": [ "Deep Research", "Wildlife Conservation", "Cheetah", "Biodiversity" ], "artifact_type": "html", "prompt": "我想要一份关于猎豹（Acinonyx jubatus）保护现状的深度研究报告，做成精美的单文件 HTML（CSS/JS 全内联，可离线打开）。请基于 IUCN 红色名录和 WWF 2024 的真实数据，包含：hero 大图标题、执行摘要、核心数据速览（野外约 6500 只 / IUCN 易危 / 仅存历史范围 10% / 亚洲亚种极危）、一张历史种群下降趋势的 SVG 折线图、三大威胁章节（人兽冲突、盗猎与非法贸易、栖息地丧失与碎片化）、历史 vs 现存范围对比、种群与繁育章节、保护行动与展望，以及页脚的完整来源清单。视觉走自然保护科考/野生动物纪录片风格，大地暖色调，关键数据正文标注来源。", "prompt_en": "I want a deep-research report on the conservation status of the cheetah (Acinonyx jubatus), built as a polished single-file HTML (all CSS/JS inlined, works offline). Base it on real data from the IUCN Red List and WWF 2024, and include: a hero image with title, an executive summary, a key-stats overview (~6,500 in the wild / IUCN Vulnerable / only 10% of historical range / Asiatic subspecies Critically Endangered), an SVG line chart of the historical population decline, a three-threats section (human-wildlife conflict, poaching & illegal trade, habitat loss & fragmentation), a historical vs. current range comparison, a population & breeding section, conservation actions & outlook, and a full source list in the footer. Use a wildlife-documentary / field-research visual style with earthy warm tones, citing sources inline for key figures.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "猎豹保护现状深度研究报告" } ], "skills": [], "experts": [ { "id": "DeepResearchExpert", "name": "深研研", "description": "综合性深度研究专家，支持多源信息检索、事实验证、知识发现和结构化报告生成，含微信公众号文章搜索能力", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "54a931baf096", "name_en": "Iris", "description_en": "Comprehensive deep research expert supporting multi-source retrieval, fact verification, knowledge discovery, and structured report generation", "expert_id": "ex_mMbwwmFA9n9P" } ], "mcps": [], "scenario": "深度研究", "scenario_id": 103, "scenario_en": "Deep Research", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T11:00:00Z", "updated_at": "2026-06-22T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-frontier-llm-benchmark", "title": "三强争锋：顶尖大模型能力全景评测", "title_en": "Clash of the Titans: A Full-Spectrum Benchmark of Frontier LLMs", "subtitle": "GPT-5 vs Claude 4 Opus vs Gemini 2.5 Pro 深度对比研究", "subtitle_en": "An in-depth comparison of GPT-5, Claude 4 Opus, and Gemini 2.5 Pro", "description": "面对 GPT-5、Claude 4 Opus、Gemini 2.5 Pro 三大旗舰模型，团队该如何按场景选型？本案由「深研研」深度研究专家整合 OpenAI、Anthropic、Google 官方数据与 AIbase 第三方评测，围绕编程、数学、多模态、长上下文与定价五大维度做事实交叉验证，产出一份暗色基准跑分台风格的单文件 HTML 深度研究报告：含执行摘要、关键发现卡、多基准分组柱状图、五维能力雷达、定价与长上下文对比及场景选型建议表，正文逐项标注来源、页脚附完整来源清单与研究局限。适合需要做模型选型决策的研发负责人、架构师与技术决策者。", "description_en": "How should teams choose between GPT-5, Claude 4 Opus, and Gemini 2.5 Pro for different scenarios? The DeepResearchExpert integrates official data from OpenAI, Anthropic, and Google with third-party AIbase evaluations, cross-verifying facts across coding, math, multimodal, long-context, and pricing dimensions. The deliverable is a single-file HTML deep-research report in a dark benchmark-lab style: executive summary, key-finding cards, grouped benchmark bar charts, a five-dimension capability radar, pricing and long-context comparisons, and a scenario-based recommendation table, with inline citations and a full source list. Ideal for R&D leads, architects, and technical decision-makers making model-selection decisions.", "categories": [ "search" ], "tags": [ "深度研究", "大模型评测", "模型选型", "基准对比" ], "tags_en": [ "Deep Research", "LLM Benchmark", "Model Selection", "Benchmark Comparison" ], "artifact_type": "html", "prompt": "帮我做一份顶尖大模型能力对比的深度研究报告，对象是 GPT-5、Claude 4 Opus、Gemini 2.5 Pro。要用真实基准数据（编程 SWE-bench、数学 AIME、多模态 MMMU、长上下文窗口、定价），正文关键数据要标注来源，页脚列完整来源清单。风格走暗色科技评测实验室/基准跑分台的硬核质感。页面要有执行摘要、关键发现卡、多基准分组柱状图、多维能力雷达图、定价对比、长上下文对比、场景选型建议表。最终交付一个所有 CSS/JS 内联的单文件 HTML，界面中文。", "prompt_en": "Create a deep-research report comparing frontier LLMs — GPT-5, Claude 4 Opus, and Gemini 2.5 Pro — using real benchmark data (coding SWE-bench, math AIME, multimodal MMMU, context window, pricing). Cite sources inline for key figures and list a full source bibliography in the footer. Use a hardcore dark tech evaluation-lab / benchmark-scoreboard aesthetic. Include an executive summary, key-finding cards, grouped benchmark bar charts, a multi-dimension capability radar, pricing comparison, long-context comparison, and a scenario-based recommendation table. Deliver a single HTML file with all CSS/JS inlined, Chinese UI.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "顶尖大模型能力全景评测报告" } ], "skills": [], "experts": [ { "id": "DeepResearchExpert", "name": "深研研", "description": "综合性深度研究专家，支持多源信息检索、事实验证、知识发现和结构化报告生成，含微信公众号文章搜索能力", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "54a931baf096", "name_en": "Iris", "description_en": "Comprehensive deep research expert supporting multi-source retrieval, fact verification, knowledge discovery, and structured report generation", "expert_id": "ex_mMbwwmFA9n9P" } ], "mcps": [], "scenario": "深度研究", "scenario_id": 103, "scenario_en": "Deep Research", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T11:00:00Z", "updated_at": "2026-06-22T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-global-ev-landscape", "title": "驶向电动：全球电动汽车市场格局研究", "title_en": "Driving Electric: A Deep Research on the Global EV Market Landscape", "subtitle": "2024 回顾、2025/2030 展望与中欧美三大市场的结构性分化", "subtitle_en": "2024 Review, 2025/2030 Outlook, and the Structural Divergence of China, Europe and the US", "description": "全球电动化进程已跨过 20% 渗透率临界线，但增长并不均质——中国领跑、新兴市场起飞、美国掉队。本案由「风向标」专家运用趋势研究方法论，以 IEA《Global EV Outlook 2025》权威数据为底座，对 2024 年市场格局做了系统复盘，并对 2025/2030 进行情景展望，重点拆解中/欧/美三大市场的结构性分化、电池降价催生的平价拐点、新兴市场第二曲线、充电基建及石油替代效应。交付一份单文件 HTML 深度研究报告，含双轴趋势图、区域分组柱状对比图与完整来源清单。适合产品战略、投资研究、出海规划及行业分析人员参考。", "description_en": "Global EV adoption has crossed the 20% penetration threshold, yet growth is far from uniform — China leads, emerging markets take off, and the US falls behind. Using the Trend Researcher methodology and grounded in the IEA Global EV Outlook 2025, this case systematically reviews the 2024 landscape and projects 2025/2030 scenarios, dissecting the structural divergence among China, Europe and the US, the price-parity tipping point driven by falling battery costs, the second growth curve in emerging markets, charging infrastructure, and oil-displacement effects. Delivers a single-file HTML deep-research report with a dual-axis trend chart, grouped regional bar comparison, and a full source list — ideal for product strategy, investment research and market-entry planning.", "categories": [ "search" ], "tags": [ "深度研究", "电动汽车", "市场格局", "趋势展望" ], "tags_en": [ "Deep Research", "Electric Vehicles", "Market Landscape", "Trend Outlook" ], "artifact_type": "html", "prompt": "帮我做一份『全球电动汽车市场格局』的深度研究报告，单文件 HTML，清洁能源/绿色科技数据简报风格（明亮清新、电光青绿主色）。要基于 IEA Global EV Outlook 2025 的真实数据：2024 全球销量 1700 万辆+、占比首破 20%、同比 25%；中/欧/美三区域占比分化（中国近一半→2030 约 80%，欧洲 20%→58%，美国约 11%→20%）；新兴市场增长、电池降价与平价拐点、充电基建、2030 展望与每日替代 500 万桶石油。要有 executive summary、核心读数、全球销量与占比双轴趋势图（SVG，历史到 2030，标注里程碑）、中欧美三区域分组柱状对比图、风险分析和页脚完整来源清单。图表坐标轴要容纳最大值、刻度比例一致，无 emoji，界面中文。", "prompt_en": "Create a single-file HTML deep-research report on the global EV market landscape, in a clean-energy / green-tech data-briefing style (bright, fresh, electric teal-green palette). Ground it in real IEA Global EV Outlook 2025 data: 2024 global sales 17M+, share crossing 20% for the first time, +25% YoY; the divergence among China/Europe/US (China ~half rising to ~80% by 2030, Europe 20%→58%, US ~11%→20%); emerging-market growth, falling battery prices and the price-parity tipping point, charging infrastructure, and the 2030 outlook with over 5 million barrels/day of oil displacement. Include an executive summary, headline readouts, a dual-axis trend chart of global sales and share (SVG, history through 2030 with milestone annotations), a grouped bar comparison of China/Europe/US, risk analysis, and a complete source list in the footer. Chart axes must accommodate maximums with consistent scales; no emoji; Chinese UI.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "全球电动汽车市场格局深度研究报告" } ], "skills": [], "experts": [ { "id": "TrendResearcher", "name": "风向标", "description": "持续追踪行业和技术趋势，为产品战略提供前瞻性洞察", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e98dc4dfd783", "name_en": "Zara", "description_en": "Continuously tracks industry and technology trends to provide forward-looking insights for product strategy", "expert_id": "ex_DdCsaoq4AtcO" } ], "mcps": [], "scenario": "深度研究", "scenario_id": 103, "scenario_en": "Deep Research", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T11:00:00Z", "updated_at": "2026-06-22T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-gold-price-drivers", "title": "黄金的高光之年：2025 金价牛市深度复盘", "title_en": "Gold's Banner Year: A Deep Dive into the 2025 Bull Run", "subtitle": "从突破 3000 到逼近 4600 美元——降息、去美元化与避险三重共振", "subtitle_en": "From $3,000 to nearly $4,600 — the triple resonance of rate cuts, de-dollarization, and safe-haven demand", "description": "2025 年国际金价走出 1979 年以来最强年度行情，全年涨约 65%、年内逾 50 次刷新历史高位。本案由「严研行」行业研究专家运用行业全景与驱动因素拆解方法论，对这轮历史性牛市进行系统复盘：梳理台阶式价格走势、拆解降息/去美元化/避险三大驱动、剖析 5002 吨总需求的结构构成与央行连续三年超千吨购金，并给出 2026 年展望与下行风险边界。交付为单文件 HTML 深度研究报告，配真实数据引用与完整来源清单。适合需要快速建立金价逻辑框架的研究员、基金经理与财经内容创作者。", "description_en": "In 2025 international gold prices delivered their strongest annual rally since 1979 — up about 65% with over 50 record highs. Using the Market Researcher methodology of sector landscape and driver decomposition, this case systematically reviews the historic bull run: the staircase price path, the three drivers (rate cuts, de-dollarization, safe-haven demand), the composition of 5,002 tonnes of total demand, three straight years of 1,000+ tonne central-bank buying, plus a 2026 outlook and downside-risk boundaries. Delivered as a single-file HTML deep-research report with cited real data and a full source list. For analysts, PMs, and finance content creators who need a fast, rigorous framework on gold.", "categories": [ "search" ], "tags": [ "深度研究", "黄金市场", "大宗商品", "金价驱动" ], "tags_en": [ "Deep Research", "Gold Market", "Commodities", "Price Drivers" ], "artifact_type": "html", "prompt": "帮我做一份关于 2025 年国际金价牛市的深度研究报告，要做成精美的单文件 HTML（CSS/JS 全部内联，可直接打开）。视觉走高端财经研报、黑金奢华质感。内容要包含：hero 大图标题、执行摘要、核心读数（全年涨约65%、盘中峰值约4549.96美元/盎司、全球需求5002吨、央行连续三年净购超千吨）、金价台阶式走势图（标注3月破3000、10月上4000、年底逼近4600这三个关键点位）、降息/去美元化/避险三大驱动因素章节、需求结构构成图（投资/央行/金饰/工业合计100%）、央行购金趋势对比柱状图、展望与风险，以及页脚完整数据来源清单（WGC、新华社、Bullionvault）。正文关键数据要标注来源。注意中国习惯涨红跌绿。", "prompt_en": "Create a deep-research report on the 2025 international gold bull run as a polished single-file HTML (all CSS/JS inlined, opens directly in a browser). Use a high-end financial research, black-and-gold luxury aesthetic. Include: a hero image with title, an executive summary, core readings (about +65% for the year, intraday peak near $4,549.96/oz, global demand of 5,002 tonnes, central banks net-buying 1,000+ tonnes for three straight years), a staircase price chart marking the three key levels (breaking $3,000 in March, $4,000 in October, nearing $4,600 at year-end), a section on the three drivers (rate cuts / de-dollarization / safe-haven demand), a demand-composition chart (investment / central banks / jewelry / industrial summing to 100%), a central-bank purchase trend bar chart, an outlook and risks section, and a complete source list in the footer (WGC, Xinhua, Bullionvault). Cite sources next to key figures in the body. Follow the Chinese convention where up moves are shown in red.", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "2025 国际金价牛市深度研究报告" } ], "skills": [], "experts": [ { "id": "FsiMarketResearcher", "name": "严研行", "description": "面向分析师与基金经理的行业研究分析师,产出行业全景、竞争格局、可比公司估值表与主题选股清单等研究交付物。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c5487ce4b84a", "name_en": "Market Researcher", "description_en": "Industry research analyst for analysts and fund managers, delivering sector landscapes, competitive analysis, comparable valuation tables, and thematic stock lists", "expert_id": "ex_Q89Vl704GHHa" } ], "mcps": [], "scenario": "深度研究", "scenario_id": 103, "scenario_en": "Deep Research", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-06-22T11:00:00Z", "updated_at": "2026-06-22T11:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "residential-lease-review", "title": "住宅租赁合同风险审查", "title_en": "Residential Lease Risk Review", "subtitle": "AI智能识别主体、标的与履约条款中的隐藏风险", "subtitle_en": "AI-powered detection of hidden risks in parties, subject matter, and performance clauses", "description": "住宅租赁是生活中最常见的合同类型之一，但往往因为“只是租个房”而忽略关键条款。本案例演示腾讯电子签资深合同法务专家如何对一份标准住宅租赁合同进行全方位风险扫描，自动识别主体信息缺失、标的物描述模糊、支付含税不明等高频风险点，并给出专业修改建议。\n\n适合：房东、房产中介、租房平台运营者、法务合规人员。", "description_en": "Residential leases are among life's most common contracts, yet critical clauses are often overlooked. This case demonstrates how Tencent eSign's Contract Legal Expert performs comprehensive risk scanning on a standard residential lease — automatically identifying hidden risks like incomplete party information, vague property descriptions, and ambiguous tax-inclusive pricing — with professional modification advice.\n\nSuitable for: landlords, property agents, rental platform operators, compliance officers.", "categories": [ "business" ], "tags": [ "主体信息", "租赁标的", "解约返还" ], "tags_en": [ "Party Information", "Lease Subject", "Termination & Return" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请对我提供的《住宅房屋租赁合同》进行严格尺度的合同风险审查。重点检查出租方和承租方的主体资格是否完整有效、租赁标的物描述是否明确（含面积、附属设施）、租金及押金是否区分含税价、提前解约条件是否清晰、违约金设置是否合理、争议解决方式是否明确可执行。请输出高/中/低分级的风险清单及每条修改建议。", "prompt_en": "Please perform a strict-scope risk review on my Residential Lease Agreement. Focus on party qualification completeness, property description clarity (area & facilities), rent/deposit tax treatment, early-termination conditions, reasonableness of penalty clauses, and enforceability of dispute resolution. Output a tiered high/mid/low risk list with modification advice per item.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T16:56:11+08:00", "updated_at": "2026-08-13T18:08:23+08:00", "artifacts": [ { "file": "output.html", "label": "基于 V1.3 真实审查任务的互动审查页面" }, { "file": "住宅租赁合同_完整审查报告_V1.6.html", "label": "可视化商务版完整审查报告（HTML）" }, { "file": "住宅租赁合同_完整审查报告_V1.3.md", "label": "完整审查报告数据版（Markdown）" }, { "file": "住宅房屋租赁合同_审查批注_V1.3.docx", "label": "腾讯电子签官方导出的 V1.3 审查批注合同" }, { "file": "住宅房屋租赁合同_审查摘要_V1.3.xlsx", "label": "腾讯电子签官方导出的 V1.3 审查摘要 Excel" }, { "file": "真实审查结果_V1.3.json", "label": "腾讯电子签 V1.3 审查任务原始返回数据" }, { "file": "住宅房屋租赁合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "resource-list-copy", "title": "每天发货源，一键生成资源单文案", "title_en": "Resource List Copy", "subtitle": "微信群/朋友圈/一对一跟单三版文案，复制即发", "subtitle_en": "Three versions of source-list copy — WeChat group, Moments and 1-on-1 — ready to paste", "description": "输入库存货源和价格，自动生成微信群、朋友圈、一对一跟单三版资源单文案，带行情话术，复制即发。", "description_en": "Enter your stock and prices: it generates three versions of source-list copy for WeChat groups, Moments and one-on-one follow-ups, complete with market-quote talking points — copy and send.", "categories": [ "creation" ], "tags": [ "资源单", "货源文案", "微信营销", "B2B贸易" ], "tags_en": [ "Resource List", "Source Copy", "WeChat Marketing", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "资源单文案三连版（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易销售专家。请帮我生成今天要发的资源单文案。\n\n【今日货源】\n{逐行填写：仓库/品类/规格/价格/现货量。例如：\n- 南京库 螺纹钢 HRB400E Φ20，3760元/吨，现货80吨\n- 南京库 高线 Φ8，3850元/吨，少量}\n\n【要求】\n1. 调用【找钢行情MCP】核对当日价格与涨跌，查询【行业信息资产库】里这个行业发货源的话术口径；\n2. 生成3个版本：①微信群群发简洁版（规格价格清单+互动引导）②朋友圈行情带货版（结合今日涨跌造紧迫感）③一对一跟单版（针对具体客户催单）；\n3. 语气口语化、接地气、符合钢贸圈微信发货习惯，带emoji，复制即可发。\n\n请生成3版资源单文案。", "skills": [ { "id": "resource-list-posting", "name": "发资源单", "description": "TradeBuddy把“我有一批货”这件事做扎实：随手一说、拍张照或甩个 Excel，就整理成一张规范货盘——既能直接复制去发、又是一条系统认得的标准货品记录，还顺手告诉你现在有没有人正在找这批货。", "skill_id": "skill_2091154547653066752" } ], "experts": [ { "id": "xiaoshou-expert", "name": "B2B贸易销售专家", "description": "懂B2B贸易的销售专家：发货盘、写文案、出报价、盯商机、经营客户、看业绩、谈价成单，助你多成单。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_42GwnCqPUv7a" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "restful-api-design", "title": "RESTful API 设计与代码生成", "title_en": "RESTful API Design and Code Generation", "subtitle": "一句 Prompt 生成符合规范的完整后端 API 代码", "subtitle_en": "Generate complete, standards-compliant backend API code from a single prompt", "description": "手动设计 RESTful API 需要兼顾路由命名规范、HTTP 状态码、请求校验、统一响应格式、鉴权中间件等大量细节，遗漏任何环节都可能导致接口不一致或安全隐患。本案例演示如何通过 API Design Skill，一次性为电商平台订单管理模块生成完整的 Express + TypeScript 实现代码，涵盖 CRUD 接口设计、Zod 参数校验、JWT 鉴权、分页排序筛选、统一错误处理及完整的类型定义。适合后端开发者、全栈工程师及希望快速搭建标准化 API 服务的技术团队。", "description_en": "Designing RESTful APIs manually requires careful attention to route naming conventions, HTTP status codes, request validation, unified response formats, and authentication middleware -- missing any of these can lead to inconsistent interfaces or security vulnerabilities. This case demonstrates how to use the API Design Skill to generate a complete Express + TypeScript implementation for an e-commerce order management module in one shot, covering CRUD endpoint design, Zod parameter validation, JWT authentication, pagination/sorting/filtering, unified error handling, and comprehensive type definitions. Ideal for backend developers, full-stack engineers, and teams looking to quickly build standardized API services.", "tags": [ "REST API", "Express", "TypeScript", "Node.js", "后端开发" ], "tags_en": [ "REST API", "Express", "TypeScript", "Node.js", "Backend Development" ], "artifact_type": "code", "categories": [ "dev" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.ts", "prompt": "请为电商平台订单管理模块设计完整的 RESTful API 并生成 Express + TypeScript 实现。要求：1) RESTful CRUD 接口（GET/POST/PUT/DELETE）；2) 分页、筛选、排序参数；3) 统一响应格式和错误处理；4) 请求校验（Zod）；5) JWT 鉴权中间件；6) 完整 TypeScript 类型定义。\n\n代码注释和用户可见字符串用简体中文，标识符保留英文。", "prompt_en": "Please design a complete RESTful API for an e-commerce platform's order management module and generate an Express + TypeScript implementation. Requirements: 1) RESTful CRUD endpoints (GET/POST/PUT/DELETE); 2) pagination, filtering, and sorting parameters; 3) a unified response format and error handling; 4) request validation (Zod); 5) JWT authentication middleware; 6) complete TypeScript type definitions.\n\nCode comments and user-visible strings should be in Simplified Chinese; keep identifiers in English.", "skills": [ { "id": "api-design", "name": "API 设计", "description": "RESTful API 接口设计与文档生成", "slug": "api-design" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "source": "official", "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.ts", "label": "RESTful API 设计" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "resume-bilingual-from-zero", "title": "从 0 生成中英双语简历：应届生 30 分钟搞定 master 版本", "title_en": "Bilingual Resume from Zero: 30-Minute Master Version for New Grads", "subtitle": "粘贴个人信息 → AI 一键生成中英双语 master 简历，英文版是 native rewrite 不是机翻", "subtitle_en": "Paste your info → AI generates bilingual master resume in one shot. English version is native rewrite, not machine translation", "description": "应届生第一次写英文简历，机翻得满屏 Chinglish？\n本案例演示用 resume-assistant 的 generate 模式，从 0 给李思源（北邮 26 届，字节火山引擎 + 商汤实习）生成中英双语 master 简历：\n1. 中文版采用紧凑表格 + bullet 混排版式，重点数据加粗，1.5 页 ATS 友好\n2. 英文版是 native rewrite（非翻译）：「独立负责」译为 \"Independent owner\"，「单卡日均节省」译为 \"saved per card per day\"，符合英文招聘官阅读习惯\n3. 文化转换：奖学金等级、竞赛排名格式都按英文招聘惯例改写（如 ICPC \"Silver Medal\"）\n4. 量化数据全部保留，单位换算（¥18 万 → ¥180K）符合英语习惯\n\n适用人群：第一次写简历的应届生 / 需要中英双语简历投海外或外企的求职者 / 嫌自己英文写作 \"AI 味\" 太重的同学。", "description_en": "First time writing an English resume and ending up with translation-flavored Chinglish?\nThis case shows resume-assistant's generate mode creating Li Siyuan's (BUPT 2026, ByteDance + SenseTime intern) bilingual master resume from scratch:\n1. Chinese version: compact tables + bullets, key metrics bolded, 1.5-page ATS-friendly layout\n2. English version is a NATIVE REWRITE (not translation): '独立负责' becomes 'Independent owner', '单卡日均节省' becomes 'saved per card per day' — matches how English recruiters actually read\n3. Cultural adaptation: scholarship tiers, competition rank formats, etc. rewritten per English recruiting norms (e.g., ICPC 'Silver Medal')\n4. All quantitative data preserved with unit conversion (¥18万 → ¥180K) per English convention\n\nBest for: New grads writing their first resume / candidates needing bilingual resumes for overseas or foreign companies / anyone whose English writing feels too 'AI-flavored'.", "tags": [ "AI简历", "中英双语", "应届生", "Generate模式", "Native Rewrite" ], "tags_en": [ "AI Resume", "Bilingual", "New Grad", "Generate Mode", "Native Rewrite" ], "prompt": "## 任务\n\n用 resume-assistant 的 generate 模式，从 0 生成一份中英双语 master 简历。\n\n**Preflight 预设**：\n- 模板：混合（表格 + bullet 混排）\n- 长度：自动（约 1.5 页，ATS 友好）\n- 语言：中英双版（独立重写）\n\n## 候选人原始信息（粘贴你的个人信息）\n...\n\n## 输出要求\n\n### 中文版\n- 紧凑表格 + bullet 混排，关键量化数据加粗\n- 按经历密度排序模块，标题副栏给出明确求职方向\n\n### 英文版\n- Native rewrite，符合英文招聘官阅读习惯\n- 数字单位英语化、奖项用国际通用格式\n- bullet 用强动词开头，避免翻译腔和 AI 味", "prompt_en": "## Task\n\nUse resume-assistant's generate mode to create a bilingual Chinese-English master resume from scratch.\n\n**Preflight presets**:\n- Template: Hybrid (mixed tables + bullets)\n- Length: Auto (about 1.5 pages, ATS-friendly)\n- Language: Bilingual Chinese-English (independently rewritten)\n\n## Candidate's raw information (paste your own details)\n...\n\n## Output requirements\n\n### Chinese version\n- Compact tables + bullets, key quantitative metrics bolded\n- Order modules by experience density; state a clear target direction in the header\n\n### English version\n- Native rewrite, matching how English recruiters read\n- Localize number units; use internationally recognized formats for awards\n- Start bullets with strong action verbs; avoid translation-flavored phrasing and an AI feel", "artifact_type": "md", "categories": [ "business", "office", "productivity" ], "cover_image": "cover.png", "preview": "resume-zh.md", "artifacts": [ { "file": "resume-zh.md", "label": "中文简历", "label_en": "Chinese Resume" }, { "file": "resume-en.md", "label": "英文简历（Native Rewrite）", "label_en": "English Resume (Native Rewrite)" } ], "skills": [ { "id": "resume-assistant", "name": "AI 简历助手", "description": "面向中国求职者的可追溯 JD 定制简历助手：从 master 简历派生多版本，按 JD 重写并标注三维度防幻觉证据，输出中英双版 + 战略附录 + ATS 友好 PDF。", "skill_id": "skill_2057443453823152128" } ], "experts": [], "quality_score": 96, "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 100, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "resume-jd-match-score", "title": "JD 匹配度打分：工程型简历去投研究岗能拿多少分？", "title_en": "JD Match Score: How Many Points Can an Engineering Resume Get for a Research Role?", "subtitle": "用 score 模式给简历↔JD 做 5 维度量化评估，输出可执行的修改建议", "subtitle_en": "Score mode quantifies resume-to-JD fit across 5 dimensions and outputs actionable improvements", "description": "海投每家都被秒拒？不是简历差，而是不知道差在哪。\n本案例演示用 resume-assistant 的 score 模式，给李思源（典型工程型候选人）的简历对腾讯 AI Lab 算法研究员 JD 做精细打分：\n1. 5 维度量化评分（硬技能 / 经历相关性 / 学术产出 / ATS / 加分项），总分 71/100\n2. 16 个核心关键词逐条命中分析（工程类 90% vs 研究类 33%）\n3. 找出结构性偏差（候选人=工程型 vs 岗位=研究型）并给出改投建议\n4. TOP 3 修改建议按 ROI 排序，明确区分「改写已有内容」 vs 「需要补做的事」\n5. 跨岗位匹配度对比表，告诉你该把精力放在哪类岗位\n\n适用人群：投了 30+ 简历都没回应、不知道自己卡在哪一环、需要把每份简历的 ROI 算清楚的求职者。", "description_en": "Sending 30+ resumes with zero replies? It's not your resume — it's that you don't know where you're losing points.\nThis case shows how resume-assistant's score mode gives Li Siyuan's resume (a typical engineering profile) a precise scoring against Tencent AI Lab Researcher JD:\n1. 5-dimension quantified scoring (hard skills / experience relevance / academic output / ATS / extras), total 71/100\n2. 16 core keywords matched line-by-line (engineering hits 90% vs research hits 33%)\n3. Identify structural mismatch (engineering candidate vs research role) and recommend better-fit roles\n4. Top 3 fixes sorted by ROI, clearly separating 'rewrite existing content' vs 'things you need to actually do'\n5. Cross-role matching scoreboard showing where to focus your effort\n\nBest for: Job seekers with 30+ unanswered applications who need to figure out exactly where they're losing points and optimize ROI per application.", "tags": [ "AI简历", "JD打分", "Score模式", "求职策略", "应届生" ], "tags_en": [ "AI Resume", "JD Score", "Score Mode", "Job Strategy", "New Grad" ], "prompt": "## 任务\n\n用 resume-assistant 的 score 模式，给一份简历对目标 JD 做匹配度打分。这里以「工程型候选人 → 腾讯 AI Lab 算法研究员」为例。\n\n## 候选人 master 简历摘要（粘贴你的简历要点）\n...\n\n## 目标 JD（粘贴目标岗位 JD）\n...\n\n## 评分要求（5 维度，共 100 分）\n1. 硬技能匹配（30）\n2. 经历相关性（25）\n3. 学术产出（20）\n4. ATS 关键词覆盖率（15）\n5. 加分项（10）\n\n## 输出要求\n1. 总分 + 评级 + 一句话定位\n2. 5 维度分数表\n3. 关键词命中明细表\n4. 每维度得分项与扣分项\n5. TOP 3 修改建议（按 ROI 排序，区分「改写已有内容」与「需要补做的事」）\n6. 跨岗位匹配度对比表\n\n打分基于简历已声明的事实。", "prompt_en": "## Task\n\nUse resume-assistant's score mode to score a resume against a target JD for fit. Here we use \"engineering-type candidate → Tencent AI Lab Algorithm Researcher\" as an example.\n\n## Candidate master resume summary (paste your resume highlights)\n...\n\n## Target JD (paste the target role's JD)\n...\n\n## Scoring requirements (5 dimensions, 100 points total)\n1. Hard skill match (30)\n2. Experience relevance (25)\n3. Academic output (20)\n4. ATS keyword coverage (15)\n5. Extras (10)\n\n## Output requirements\n1. Total score + rating + one-line positioning\n2. 5-dimension score table\n3. Keyword hit detail table\n4. Scoring and deduction items per dimension\n5. Top 3 fixes (sorted by ROI, separating \"rewrite existing content\" vs \"things you need to actually do\")\n6. Cross-role matching comparison table\n\nAll scoring is based on facts stated in the resume.", "artifact_type": "md", "categories": [ "business", "office", "productivity" ], "cover_image": "cover.png", "preview": "score-report.md", "artifacts": [ { "file": "score-report.md", "label": "JD 匹配度打分报告", "label_en": "JD Match Score Report" } ], "skills": [ { "id": "resume-assistant", "name": "AI 简历助手", "description": "面向中国求职者的可追溯 JD 定制简历助手：从 master 简历派生多版本，按 JD 重写并标注三维度防幻觉证据，输出中英双版 + 战略附录 + ATS 友好 PDF。", "skill_id": "skill_2057443453823152128" } ], "experts": [], "quality_score": 88, "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 100, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "resume-tailor-jd-newgrad", "title": "校招简历适配 JD：把通用简历改成快手 AI Infra 校招版", "title_en": "JD Tailor for New Grads: Adapt General Resume to Kuaishou AI Infra Campus Role", "subtitle": "输入应届生通用简历 + 目标 JD → AI 重排经历优先级、重写 bullet 命中关键词、生成定制化版本 + 投递战略附录", "subtitle_en": "Input general resume + target JD → AI reprioritizes experience, rewrites bullets to hit keywords, generates tailored version + strategic appendix", "description": "应届生秋招撒网投简历，每家公司都用同一份会被秒拒？\n本案例演示用 resume-assistant 的 tailor 模式，把李思源（北邮 26 届，字节火山引擎 + 商汤实习）的通用简历定制成快手 AI Infra 校招版，并配套生成投递战略附录：\n1. JD 关键词级匹配（覆盖率 62% → 84%，10 个核心关键词命中 9 个）\n2. 经历重排（vLLM 推理优化提到第一条，K8s Operator 紧随其后）\n3. Bullet 重写命中 JD（显式写出 PagedAttention、Continuous Batching、Spot Instance）\n4. 三维度 Provenance 防幻觉审计（事实 / 语气 / 技能 三层全部通过，零编造）\n5. 战略附录（v0.3.2 默认必出）：4 个优势、Gap 分析、3 道高频面试题 + STAR 扩写、Cover Letter 三种 hook，所有内容可追溯到简历或 JD\n\n适用人群：26/27 届应届生，有过 1-2 段实习，需要按目标公司 JD 做定制化简历，并希望直接拿到面试准备建议的同学。", "description_en": "Are you a new grad sending the same resume everywhere and getting auto-rejected?\nThis case shows how resume-assistant's tailor mode adapts Li Siyuan's general resume (BUPT 2026 grad, ByteDance Volcano Engine + SenseTime intern) into a Kuaishou AI Infra campus version, plus a strategic interview-prep appendix:\n1. JD keyword coverage 62% → 84%, hitting 9 of 10 core keywords\n2. Experience reordered (vLLM inference optimization first, K8s Operator next)\n3. Bullets rewritten to embed JD keywords explicitly (PagedAttention, Continuous Batching, Spot Instance)\n4. 3-dimensional provenance audit (factual / tonal / skill — all passed, zero fabrication)\n5. Strategic appendix (v0.3.2 default): 4 strengths, Gap analysis, 3 likely interview questions with STAR expansions, 3 cover-letter hooks — all traceable to resume or JD\n\nBest for: 2026/2027 new grads with 1-2 internships needing JD-tailored resumes plus actionable interview prep.", "tags": [ "AI简历", "校招", "JD定制", "AI Infra", "应届生" ], "tags_en": [ "AI Resume", "Campus Recruit", "JD Tailor", "AI Infra", "New Grad" ], "prompt": "## 任务\n\n使用 resume-assistant 的 tailor 模式，把一份应届生通用简历按目标 JD 定制成针对性版本。\n\n## 候选人通用简历\n（粘贴你的 master 简历）：\n...\n\n## 目标 JD\n（粘贴目标岗位 JD 与核心关键词）：\n...\n\n## 定制要求\n- 按 JD 重排经历优先级，把最相关的经历提到最前\n- 在 bullet 中自然嵌入 JD 核心关键词\n- 所有量化数据原样保留，不编造、不夸大\n- 给出 JD 关键词覆盖率统计\n\n## 输出\n- `tailored-resume-zh.md`：定制化中文简历\n- `strategic-appendix.md`：投递战略附录", "prompt_en": "## Task\n\nUse resume-assistant's tailor mode to adapt a new-grad general resume into a targeted version for a specific JD.\n\n## Candidate general resume\n(Paste your master resume):\n...\n\n## Target JD\n(Paste the target role's JD and its core keywords):\n...\n\n## Tailoring requirements\n- Reorder experiences by relevance to the JD, moving the most relevant to the top\n- Naturally embed the JD's core keywords into the bullets\n- Keep all quantitative data exactly as-is — no fabrication, no exaggeration\n- Provide JD keyword coverage statistics\n\n## Output\n- `tailored-resume-zh.md`: the tailored Chinese resume\n- `strategic-appendix.md`: submission-strategy appendix", "artifact_type": "md", "categories": [ "business", "office", "productivity" ], "cover_image": "cover.png", "preview": "tailored-resume-zh.md", "artifacts": [ { "file": "tailored-resume-zh.md", "label": "定制化中文简历", "label_en": "Tailored Chinese Resume" }, { "file": "strategic-appendix.md", "label": "投递战略附录（面试准备 + Cover Letter Hooks）", "label_en": "Strategic Appendix (Interview Prep + Cover Letter Hooks)" }, { "file": "result.json", "label": "JD 匹配 + Provenance 审计结果", "label_en": "JD Matching + Provenance Audit Results" } ], "skills": [ { "id": "resume-assistant", "name": "AI 简历助手", "description": "面向中国求职者的可追溯 JD 定制简历助手：从 master 简历派生多版本，按 JD 重写并标注三维度防幻觉证据，输出中英双版 + 战略附录 + ATS 友好 PDF。", "skill_id": "skill_2057443453823152128" } ], "experts": [], "quality_score": 97, "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 100, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "retail-business-management-workbench", "title": "零售行业经营管理工作台", "title_en": "Retail Business Management Workbench", "subtitle": "左侧可收起侧栏 + 营销/商品/仓配/数仓多模块切换 + 实时读取资料库数据", "subtitle_en": "Collapsible sidebar + multi-module switching (Marketing/Products/Warehouse/Data) + real-time database rendering", "description": "零售门店每天面对营销排期、SKU 上百件、仓配成本、需求管理等多条业务线，页面在多个系统里反复切换，难以形成统一的经营视图。\n\n本案例把营销计划、商品资料、仓配费用、数仓需求四个典型零售管理场景合并到一个单页面工作台：左侧可收起侧栏切换模块，每个模块通过资料库在线数据库实时渲染数据，所有数据可双轨持久化。\n\n适合：零售门店运营、连锁品牌总部、烘焙/餐饮/生鲜等本地化零售行业负责人，及需要把多张报表合并到统一视图的运营团队。", "description_en": "A single-page retail business management workbench that merges marketing plans, product data, warehouse distribution costs and data warehouse requirements into one view: a collapsible sidebar with multi-module switching, all modules rendered in real time from the online database. Suitable for retail store operations, chain brand HQ and local retail teams.", "categories": [ "workspace" ], "tags": [ "零售管理", "经营工作台", "多模块切换", "可收起侧栏", "资料库联动" ], "tags_en": [ "Retail Management", "Business Workbench", "Multi-module", "Collapsible Sidebar", "Database-driven" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "零售经营管理工作台" } ], "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/KZVubEzFfcZXPTYPJxxAC3?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个零售行业经营管理工作台单页面 HTML，左侧可收起侧栏（带边缘悬浮手柄，可点击收起态 logo 再次展开，悬停图标弹 tooltip），包含四个模块切换：营销计划看板、商品资料看板、仓配费用分析、数仓需求管理。每个模块通过资料库在线数据库实时读取并渲染（使用 window.__SMART_PAGE__.database SDK，databaseId 以字符串字面量硬编码），烘焙主题视觉（暖米色 / 烘焙红 / 圆体字），CSS/JS 无外部依赖。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the database capability to build a retail business management workbench single-page HTML. Left side has a collapsible sidebar (with edge floating handle, click the collapsed logo to expand again, hover icons show tooltips), including four module switches: Marketing Plan Board, Product Data Board, Warehouse Distribution Cost Analysis, Data Warehouse Requirement Management. Each module reads and renders data in real time from the online database (using the window.__SMART_PAGE__.database SDK, databaseId hardcoded as a string literal), bakery theme visuals (warm beige / bakery red / rounded fonts), and no external dependencies.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "孙北北面包店", "name_en": "Sun Beibei Bakery", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-20T18:56:00+08:00", "updated_at": "2026-08-20T18:56:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "rfm-vip-retention-plan", "title": "RFM核心VIP维系方案", "title_en": "RFM Core VIP Retention Plan", "subtitle": "核心 VIP 客户分层、权益与跟进话术", "subtitle_en": "A practical Weimob playbook for rfm core vip retention plan", "description": "做RFM核心VIP维系方案时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“RFM核心VIP维系方案”展开：通过最近消费时间、消费频次和消费金额识别核心 VIP 客户，基于可获取的客户价值信息制定分层维系策略、建议权益和跟进话术。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、客户分层洞察运营人员、经营分析团队。", "description_en": "The hardest part of rfm core vip retention plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates rfm core vip retention plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Customer Segmentation specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "RFM模型", "VIP客户", "客户价值", "维系策略", "跟进话术" ], "tags_en": [ "RFM Core VIP Retention Plan", "Customer Operations", "Customer Segmentation", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "用RFM模型帮我找出核心VIP客户，按最近消费时间、消费频次和消费金额说明入选依据，制定分层维系策略、建议权益和可直接使用的跟进话术；只输出方案，不创建任务。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a rfm core vip retention plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "risk-inspection-report", "title": "经营有没有雷？一份风险巡检报告帮你排查", "title_en": "Risk Inspection Report", "subtitle": "应收/供应商/资金库存/行情四维扫描，分级预警", "subtitle_en": "Four-dimension scan — receivables, suppliers, cash & stock, market — with graded alerts", "description": "自动巡检客户应收、供应商、资金库存、行情四大风险维度，分级预警并给出应对建议。", "description_en": "It automatically inspects four risk dimensions — customer receivables, suppliers, cash and inventory, and market conditions — with graded early warnings and response recommendations.", "categories": [ "data" ], "tags": [ "风险巡检", "应收风控", "供应商风险", "B2B贸易" ], "tags_en": [ "Risk Inspection", "Receivables Risk", "Supplier Risk", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "经营风险巡检报告（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易运营专家。请帮我做一次经营风险巡检。\n\n【数据来源】\n- 我已绑定业务系统/上传了经营数据：请读取应收、供应商、库存、客户等数据；\n- 我暂时没有数据：说明需要哪些字段并给示例报告。\n\n【要求】\n1. 从4个维度巡检：①客户/应收风险（超期账款、客户集中度）②供应商风险（命中【找钢黑名单】的重点标注）③资金/库存风险（周转、压库/断货）④行情风险（调用【找钢行情MCP】判断价格波动）；\n2. 查询【行业信息资产库】里这个行业的风控口径（如超期天数阈值、集中度警戒线），对每项风险分高/中/低三级；\n3. 每项风险给出具体应对建议。\n\n若数据源未就绪或返回不足，请以通用经营风控方法论为主产出并说明。请生成完整风险巡检报告。", "skills": [ { "id": "risk-inspection", "name": "风险巡检", "description": "TradeBuddy把“手上一批合作方，定期扫一遍看谁出问题”的风控诉求，做成批量、高风险置顶、来源可追溯的风险清单。与供应商背调同源，区别在背调深挖一家、巡检批量扫存量。", "skill_id": "skill_2091156854531039232" } ], "experts": [ { "id": "yunying-expert", "name": "B2B贸易运营专家", "description": "懂B2B贸易的运营专家：面向老板与管理员，做生意体检、看团队人效、批量扫风险、沉淀企业知识。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_D7992fbSsXNx" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "robot-conference-event-overview", "title": "机器人大会事件速览", "title_en": "Robot Conference Event Overview", "subtitle": "", "subtitle_en": "", "description": "无需逐篇刷长文，快速解锁热门事件核心资讯！efunds 热点梳理，自动检索易方达公众号公开素材，生成事实型行业事件速览，观点原文引用，内容有据可查。适合赛道投资者、财经资讯爱好者、运营研究人员等。", "description_en": "Tired of reading a dozen long WeChat posts just to understand what a major industry conference (like the World Robot Conference) actually covered and what signals it sent? This case shows how to use E Fund hotspot combing to automatically retrieve real public articles from the two official accounts \"E Fund\" and \"E Fund Index\", and structure them into a plain-text briefing using the three-part format: Event Overview → Reference Materials → Fixed Risk Disclosure — facts only, market viewpoints quoted verbatim, nothing fabricated.\n\nTarget audience: investors following robotics / AI / advanced-manufacturing themes, research or operations staff who need quick industry-event briefs, and everyday readers who want AI to digest financial news efficiently.", "categories": [ "finance" ], "tags": [ "机器人大会", "热点梳理", "事件速览", "易方达", "财经资讯" ], "tags_en": [ "Robot Conference", "Hotspot Combing", "Event Overview", "E Fund", "Financial News" ], "artifact_type": "md", "preview": "robot-conference-event-overview.md", "artifacts": [ { "file": "robot-conference-event-overview.md", "label": "机器人大会事件速览（纯文字）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请使用 efunds 热点梳理能力检索「易方达基金」「易方达指数通」公众号公开文章，梳理世界机器人大会最新资讯，输出纯文字三段式结果：\n1. 事件概述：2‑3 句客观事实，不加观点、不预测后市；\n2. 参考资料：引用 1‑2 篇原文观点（100‑200 字，带引号 + 标题 + 公众号 + 时间 + 链接），无资料则如实说明；\n3. 末尾添加固定风险提示。\n全部信息取自真实检索结果，禁止编造。", "prompt_en": "Use the E Fund hotspot combing capability to search the two official accounts \"E Fund\" and \"E Fund Index\" for real public articles, and comb through the latest news about the World Robot Conference. Output plain text (no HTML) in the three-part format: ① Event Overview — 2-3 sentences of objective facts, no viewpoint, no market forecast; ② Reference Materials — 1-2 articles, market viewpoints quoted verbatim (100-200 chars, in quotes, with source title + account + time + link), state honestly if no material; ③ Fixed risk disclosure at the end. All information must come from real search results, nothing fabricated.", "skills": [ { "id": "efunds-hotspot-combing", "name": "易方达热点梳理", "description": "基于 efunds 检索易方达系公众号，对财经 / 市场 / 政策热点做资讯与事实梳理", "skill_id": "skill_2092287909491437568" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-28T14:38:37+08:00", "updated_at": "2026-08-31T17:12:13+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "saas-onboarding-flow", "title": "SaaS 产品 Onboarding 引导流程", "title_en": "SaaS Product Onboarding Flow", "subtitle": "多步骤向导 + 动效过渡，提升用户激活率", "subtitle_en": "Multi-step wizard with animations to boost user activation", "description": "许多 SaaS 产品在用户注册后面临激活率低的困境——用户不知道从哪里开始、核心功能难以被发现、首次体验缺乏引导感。本案例演示了一个 Superhuman/Raycast 风格的 5 步 Onboarding 引导流程，包含角色选择、偏好配置、快捷键教学等关键环节，通过精致的暗色 UI 与流畅的过渡动效，让用户在 30 秒内完成初始化并感受到产品的高级质感。适合 SaaS 产品经理、增长团队和前端开发者参考，可直接复用到自有产品的新用户激活流程中。", "description_en": "Many SaaS products struggle with low activation rates after sign-up — users don't know where to start, core features remain undiscovered, and the first experience lacks guided flow. This case demonstrates a Superhuman/Raycast-style 5-step onboarding wizard featuring role selection, preference configuration, and keyboard shortcut tutorials, all wrapped in a polished dark UI with smooth transition animations that help users complete setup in under 30 seconds while experiencing premium product quality. Ideal for SaaS product managers, growth teams, and frontend developers looking to replicate a high-converting onboarding experience in their own products.", "categories": [ "business", "creation" ], "tags": [ "用户引导", "SaaS", "动效设计", "转化率", "产品体验" ], "tags_en": [ "Onboarding", "SaaS", "Animation", "Conversion", "UX" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Onboarding 引导页面" } ], "prompt": "请生成一个 Superhuman / Raycast 风格的 SaaS 产品 Onboarding 引导页面（单文件 HTML，内联无外部依赖），暗色主题、高端克制、键盘优先的设计美学，带步骤切换动画和顶部进度条。\n\n5 个步骤：\n1. 欢迎页：产品 Logo + 品牌标语 + 「开始体验」\n2. 角色选择：设计师/开发者/产品经理/运营 四张卡片，选中有反馈\n3. 偏好设置：主题色选择 + 通知频率\n4. 快捷键教学：可视化键盘展示 3-4 个核心快捷键\n5. 完成页：庆祝动画 + 「进入工作台」\n\n支持前进/后退。", "prompt_en": "Please create a Superhuman / Raycast-style SaaS product onboarding page (single-file HTML, fully inline with no external dependencies), with a dark theme, a high-end, restrained, keyboard-first design aesthetic, with step-transition animations and a progress bar at the top.\n\n5 steps:\n1. Welcome page: product logo + brand tagline + \"Get Started\"\n2. Role selection: four cards — designer / developer / product manager / operations, with feedback on selection\n3. Preferences: theme color selection + notification frequency\n4. Keyboard-shortcut tutorial: a visual keyboard showing 3-4 core shortcuts\n5. Completion page: a celebration animation + \"Enter Workspace\"\n\nSupport going forward/backward.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 9, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-05-03T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": true, "recommended_order": 18 }, { "id": "saas-ops-dashboard", "title": "SaaS 运营数据 Dashboard 单页应用", "title_en": "SaaS Operations Analytics Dashboard (SPA)", "subtitle": "用 React + Recharts 做一个深色主题的运营数据看板，KPI、趋势、渠道、留存、热力一屏看全，产出可离线运行的单文件源码", "subtitle_en": "A dark-themed SaaS analytics dashboard with React + Recharts — KPIs, trends, channels, retention and heatmap on one screen, delivered as a single self-contained HTML file", "description": "想给自己的 SaaS 产品做一个能在周会上直接投屏的运营数据看板，却卡在「图表怎么搭、组件怎么拆、数据怎么编得自洽」上？这个案例让 EdgeOne Makers 开发专家团真刀真枪帮你把它做出来。\n\n演示里，主理人编排前端成员用 Vite + React + Tailwind CSS 写出一整屏深色主题的运营看板：顶部 Hero 展示产品名与数据周期；DAU / MAU / 付费转化率 / MRR 四张关键指标卡带环比箭头；7 日活跃用户趋势面积图、用户来源渠道占比环形图、5 功能 × 7 天的功能使用热力矩阵、次日/3日/7日/30日留存漏斗条形图，以及四张含指标异动描述与建议动作的运营洞察卡片。组件按 KPICard、AreaChart、DonutChart、Heatmap、FunnelBar、InsightCard 清晰拆分，图表用 Recharts，合成数据各指标间逻辑自洽（MAU 142.8K、黏性 36.7%、付费转化率 4.6% 对应 MRR ¥56.5万、留存逐级递减），不依赖任何外部图片资源。最终产物是一个零外部依赖、可离线渲染的单文件 HTML Dashboard：双击即可在浏览器打开，也可以直接 `edgeone makers deploy` 一键部署上线。\n\n适合：要做数据看板/管理后台的前端同学、需要给产品搭运营仪表盘的独立开发者、想学 Recharts 图表组件拆分的工程师、做 SaaS / 增长分析演示的产品与运营。", "description_en": "Want a SaaS operations dashboard you can throw straight onto the big screen in a weekly review, but get stuck on \"how do I build the charts, split the components, and fabricate self-consistent data\"? This case has the EdgeOne Makers dev team build the whole thing for real.\n\nIn the demo, the team lead orchestrates a frontend specialist to build a full-screen dark-themed dashboard with Vite + React + Tailwind CSS: a top Hero showing the product name and data period; four KPI cards for DAU / MAU / paid conversion / MRR with MoM arrows; a 7-day active-user area chart, a traffic-channel donut chart, a 5-feature × 7-day usage heatmap, a day-1/3/7/30 retention funnel bar chart, and four operational insight cards with anomaly descriptions and suggested actions. Components are cleanly split into KPICard, AreaChart, DonutChart, Heatmap, FunnelBar and InsightCard, charts use Recharts, and the synthetic data is internally consistent (MAU 142.8K, 36.7% stickiness, 4.6% paid conversion mapping to ¥565K MRR, monotonically decreasing retention), with zero external image dependencies. The final deliverable is a single self-contained HTML dashboard with no external dependencies that renders fully offline: just double-click to open it in a browser, or run edgeone makers deploy to ship it live.\n\nGreat for: frontend devs building dashboards/admin panels, indie hackers who need an ops dashboard for their product, engineers learning Recharts component breakdown, and product/growth folks preparing SaaS analytics demos.", "categories": [ "dev", "data", "agent" ], "tags": [ "数据看板", "React", "Recharts", "运营分析", "单文件应用" ], "tags_en": [ "Dashboard", "React", "Recharts", "Ops Analytics", "Single-file App" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "运营数据 Dashboard 单文件应用" } ], "prompt": "用 Vite + React 帮我做一个 SaaS 产品运营数据 Dashboard 单页应用，深色主题。\n\n页面包含：\n- 顶部 Hero 展示产品名称与当前数据周期；\n- DAU / MAU / 付费转化率 / MRR 四张关键指标卡（含环比箭头）；\n- 一个 7 日活跃用户趋势面积图；\n- 一个用户来源渠道占比环形图（直接 / 搜索 / 社交 / 广告 4 段）；\n- 一个功能使用热力矩阵（5 个核心功能 × 7 天，色阶表示调用量）；\n- 一组用户留存漏斗条形图（次日 / 3 日 / 7 日 / 30 日）；\n- 底部放 4 张运营洞察卡片（含指标异动描述与建议动作）。\n\n组件拆分清晰（KPICard、AreaChart、DonutChart、Heatmap、FunnelBar、InsightCard），数据用合成示意数据但各指标间逻辑自洽，图表用 Recharts 或 ECharts，样式用 Tailwind CSS，不依赖外部图片资源，界面文字用简体中文。\n\n开发请走 EdgeOne Makers 开发专家团：由主理人编排前端成员实现页面与图表组件，完成后主理人本地预览验证。最终产物希望是一个零外部依赖、可离线打开的单文件 HTML，方便我直接双击查看或后续一键部署。", "prompt_en": "Build me a SaaS operations analytics dashboard as a single-page app with Vite + React, dark theme.\n\nThe page should include:\n- A top Hero showing the product name and current data period;\n- Four KPI cards for DAU / MAU / paid conversion rate / MRR (with month-over-month arrows);\n- A 7-day active-user trend area chart;\n- A traffic-channel donut chart (Direct / Search / Social / Ads, 4 segments);\n- A feature-usage heatmap (5 core features × 7 days, color scale for call volume);\n- A user-retention funnel bar chart (day-1 / day-3 / day-7 / day-30);\n- Four operational insight cards at the bottom (each with an anomaly description and a suggested action).\n\nKeep components cleanly split (KPICard, AreaChart, DonutChart, Heatmap, FunnelBar, InsightCard), use synthetic placeholder data that is internally consistent across metrics, draw charts with Recharts or ECharts, style with Tailwind CSS, do not rely on external image resources, and use Simplified Chinese for the UI text.\n\nFor development, use the EdgeOne Makers dev team: the lead orchestrates a frontend specialist to implement the page and chart components, then previews it locally. I'd like the final deliverable to be a single self-contained HTML file with zero external dependencies that I can open offline by double-clicking, or deploy later in one shot.", "scenario": "", "scenario_use": "", "task_mode": "", "skills": [], "experts": [ { "id": "edgeone-makers-team-lead", "expert_id": "ex_EDEph9cW7RjS", "name": "Makers 开发专家团", "description": "在 EdgeOne Makers 上编排前端、后端、AI Agent 成员完成 Web 全栈开发与一键部署的专家团", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "e20cf15557d4" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-06-25T16:40:00+08:00", "updated_at": "2026-06-25T16:40:00+08:00", "source": "official", "is_recommended": false, "recommended_order": 0 }, { "id": "sale-contract", "title": "办公设备买卖合同", "title_en": "Office Equipment Sales Contract", "subtitle": "把采购数量、价款、验收和质保一次写清", "subtitle_en": "Turn quantities, payment, acceptance and warranty into a clear contract.", "description": "采购设备时，数量、验收和售后往往散落在聊天记录里，后续最容易产生争议。\n\n本案例演示腾讯电子签合同法务专家如何将 50 台笔记本电脑采购需求整理为可编辑的《办公设备买卖合同》，并突出价款、交付、验收、质保与违约责任。\n\n适合：采购负责人、行政团队、供应商管理人员、初创企业经营者。", "description_en": "Equipment purchases often scatter quantities, acceptance rules and after-sales commitments across chats, creating avoidable disputes.\n\nThis case demonstrates how the Tencent eSign Contract Legal Expert turns a 50-laptop procurement brief into an editable Office Equipment Sales Contract, highlighting payment, delivery, acceptance, warranty and liability.\n\nSuitable for: procurement leads, administrative teams, supplier managers and startup operators.", "categories": [ "business" ], "tags": [ "买卖合同", "设备采购", "验收", "质保" ], "tags_en": [ "Sales Contract", "Equipment Procurement", "Acceptance", "Warranty" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请起草一份办公设备买卖合同。甲方为[甲方名称]，乙方为[乙方名称]；采购[设备名称]共[数量]台，单价[单价]元，含税总价[总价]元。约定[预付款比例]预付款、[尾款条件]后支付尾款；明确交付期限、验收标准、质保期限、售后责任、违约金、赔偿上限及争议解决。请输出可直接签署的中文 Word 合同，条款层级清晰，重点保障采购方的验收与售后权益。", "prompt_en": "Please draft an office equipment sales contract. Party A is [甲方名称] and Party B is [乙方名称]. Party A purchases [设备名称], [数量] units, at [单价] RMB per unit, for a tax-inclusive total of [总价] RMB. Specify a [预付款比例] advance payment and final payment after [尾款条件]. Clearly define the delivery period, acceptance criteria, warranty, after-sales responsibilities, liquidated damages, liability cap and dispute resolution. Produce a signable Chinese Word contract with clear clause hierarchy and strong protection for the purchaser's acceptance and after-sales rights.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-11T21:04:09+08:00", "updated_at": "2026-08-12T13:37:31+08:00", "artifacts": [ { "file": "output.html", "label": "V3.0 互动合同起草页面" }, { "file": "办公设备买卖合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "sales-associate-growth-opportunity-map", "title": "导购成长机会地图", "title_en": "Sales Associate Growth Opportunity Map", "subtitle": "导购个体画像、标杆定位与辅导方向", "subtitle_en": "A practical Weimob playbook for sales associate growth opportunity map", "description": "做导购成长机会地图时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“导购成长机会地图”展开：分析每位导购的业务优势、成长短板和辅导机会，识别标杆与需要支持的人员，让团队业绩提升更有方向。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购业绩管理运营人员、经营分析团队。", "description_en": "The hardest part of sales associate growth opportunity map is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates sales associate growth opportunity map through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Performance Management specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "导购画像", "标杆定位", "业绩辅导", "成长机会", "团队提升" ], "tags_en": [ "Sales Associate Growth Opportunity Map", "Store and Sales Associate Operations", "Sales Associate Performance Management", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我看看每位导购更擅长什么、卡在哪里，以及谁适合做标杆、谁需要什么样的辅导，让团队业绩提升更有方向。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a sales associate growth opportunity map covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "guide-management", "name": "导购业绩管理", "description": "查询导购交易与业绩明细，支持按业绩类型、订单来源、客户、商品和订单信息进行筛选。", "skill_id": "skill_2087079228047712256" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-associate-performance-ranking", "title": "导购业绩排行榜", "title_en": "Sales Associate Performance Ranking", "subtitle": "导购月度业绩排行与标杆方法拆解", "subtitle_en": "A practical Weimob playbook for sales associate performance ranking", "description": "做导购业绩排行榜时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“导购业绩排行榜”展开：统计各导购本月销售额、成交单量和客单价，生成排行榜并提炼TOP导购值得复制的成交方法。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购业绩管理运营人员、经营分析团队。", "description_en": "The hardest part of sales associate performance ranking is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates sales associate performance ranking through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Performance Management specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "导购排行", "销售额", "成交单量", "客单价", "优秀经验" ], "tags_en": [ "Sales Associate Performance Ranking", "Store and Sales Associate Operations", "Sales Associate Performance Management", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "统计本月截至昨天各导购的销售额、成交单量和客单价，生成业绩排行榜，并提炼TOP导购值得复制的成交方法。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a sales associate performance ranking covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "guide-management", "name": "导购业绩管理", "description": "查询导购交易与业绩明细，支持按业绩类型、订单来源、客户、商品和订单信息进行筛选。", "skill_id": "skill_2087079228047712256" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-associate-tiered-coaching", "title": "导购业绩分层辅导", "title_en": "Tiered Sales Associate Coaching", "subtitle": "导购能力分层与差异化辅导计划", "subtitle_en": "A practical Weimob playbook for tiered sales associate coaching", "description": "做导购业绩分层辅导时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“导购业绩分层辅导”展开：根据各导购本月净业绩、成交单量和客单价进行分层，识别高频成交型、高客单型、均衡型和待提升型人员，并制定差异化辅导计划。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购业绩管理运营人员、经营分析团队。", "description_en": "The hardest part of tiered sales associate coaching is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates tiered sales associate coaching through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Performance Management specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "导购分层", "净业绩", "成交单量", "客单价", "辅导计划" ], "tags_en": [ "Tiered Sales Associate Coaching", "Store and Sales Associate Operations", "Sales Associate Performance Management", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "根据本月截至昨天各导购的净业绩、成交单量和客单价进行分层，识别高频成交型、高客单型、均衡型和待提升型人员，并为各层制定具体辅导计划。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a tiered sales associate coaching covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "guide-management", "name": "导购业绩管理", "description": "查询导购交易与业绩明细，支持按业绩类型、订单来源、客户、商品和订单信息进行筛选。", "skill_id": "skill_2087079228047712256" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-battle-ai-finance-asset-pack", "title": "AI财税销售资产包", "title_en": "AI Finance Sales Asset Pack", "subtitle": "一页纸、客户案例、异议FAQ与ROI计算器的完整销售组合", "subtitle_en": "A complete sales package with one-pager, customer stories, objection FAQ, and ROI calculator", "description": "销售 AI 财税产品最怕的不是没有材料，而是材料彼此脱节：第一次沟通只有产品介绍，深入推进时没有案例，客户提出异议时没有统一口径，谈 ROI 时又只能现场口算。这个案例演示了销售作战团队如何为 AI 财税 SaaS 产出一整套可直接使用的销售资产：一页纸产品介绍、3 个跨行业成功客户案例、20 条常见异议 FAQ，以及一个可交互的 HTML ROI 计算器。最终交付一个适合销售、售前和管理层共同使用的 Markdown + HTML 资产包。\n\n适合：销售总监、AE、售前顾问、行业顾问、市场团队", "description_en": "The hardest part of selling AI finance software is not a lack of materials—it's when the materials do not connect: product intro for the first call, but no customer stories for deeper conversations; no unified answers for objections; and no structured way to discuss ROI beyond rough estimates. This case shows how the Sales Battle Team builds a complete sales asset pack for an AI finance & tax SaaS product: a one-page product sheet, three cross-industry customer success stories, 20 objection-handling FAQs, and an interactive HTML ROI calculator. The final deliverable is a Markdown + HTML package that sales, pre-sales, and leadership teams can all use.\n\nBest for: Sales Directors, Account Executives, Solution Consultants, Industry Advisors, Marketing Teams", "categories": [ "business" ], "tags": [ "销售资产包", "AI财税SaaS", "客户案例", "异议处理", "ROI计算器" ], "tags_en": [ "sales asset pack", "AI finance SaaS", "customer stories", "objection handling", "ROI calculator" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "one-pager.html", "artifacts": [ { "file": "one-pager.html", "label": "一页纸产品介绍 HTML" }, { "file": "roi-calculator.html", "label": "交互式 ROI 计算器 HTML" }, { "file": "output.md", "label": "销售资产包 Markdown 总包" } ], "prompt": "请为销售团队打造可直接使用的销售资产包，围绕“AI 财税 SaaS 产品”：\n1. 一页纸产品介绍（PDF/HTML 打印友好）\n2. 3 个成功客户案例（不同行业，500-2000 字/篇）\n3. HTML 版交互式 ROI 计算器（用户输入企业规模、当前财税成本，得到使用本产品的预期节省）\n4. 销售常见异议 FAQ 20 条\n\n全部以 Markdown + HTML 打包交付。", "prompt_en": "Please build a ready-to-use sales asset pack for a sales team, centered on an \"AI Fintech/Tax SaaS product\":\n1. A one-page product overview (PDF/HTML, print-friendly)\n2. 3 customer success stories (different industries, 500-2000 words each)\n3. An HTML interactive ROI calculator (the user enters their company size and current finance/tax costs, and gets the expected savings from using this product)\n4. 20 common sales objection FAQs\n\nDeliver everything packaged as Markdown + HTML.", "skills": [], "experts": [ { "id": "SalesBattleTeam", "name": "销售作战团队", "description": "销售攻坚体系，客户研究锁定目标、外联策略提升触达、竞品情报预警风险、销售预测优化资源分配", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "ff06833f0e30", "name_en": "Sales Battle Team", "description_en": "Sales effectiveness system: account research identifies targets, outreach strategy improves reach, competitive intel warns risks, pipeline forecasting optimizes resource allocation", "expert_id": "ex_ATh4MnfFCKxR" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:55:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "sanxingdui-central-plains-bronze-correlation", "title": "三星堆面具与中原青铜器关联对比", "title_en": "Sanxingdui Bronze Mask vs Central Plains Bronze Vessel Correlation", "subtitle": "同源异流:商代晚期古蜀与中原青铜器的工艺、造型与文化比较", "subtitle_en": "Same Origin, Different Streams: A comparison of craft, form, and culture between late Shang Shu (Sanxingdui) and Central Plains bronzes", "description": "研究古蜀文明最头疼的,就是三星堆那些神性夸张的青铜面具跟前母戊鼎这种规整庄严的中原礼器到底有什么关系?研究文物又怕资料零散、工艺断章、文化关系缺乏交叉核验。\n\n这个案例演示了如何借助腾讯探元文物库与知识库的并行检索,主代理亲自做三路 WebSearch 多源核验,覆盖工艺技法、造型特征、材质成分、文化内涵与历史交流脉络五个维度,最终交付包含配图与对比雷达图的 Markdown/HTML 双格式对比报告。\n\n适合:商代考古与青铜器研究者、博物馆策展与社教人员、中华文明探源工程与三星堆研究爱好者、艺术史与古代工艺美术专业学生、文物爱好者与高端文旅讲解员。", "description_en": "Researching the ancient Shu civilization usually stalls on one big question: how do the otherworldly Sanxingdui bronze masks actually relate to canonical Central Plains ritual bronzes like the Houmuwu Ding? Traditional digs into late Shang bronze work scatter sources, fragment craft notes, and rarely cross-check cultural-exchange evidence. This case demonstrates how to pair the Tencent Tanyuan relic and knowledge bases for parallel structured retrieval, then run three concurrent main-agent WebSearch passes for cross-verification across five dimensions—craftsmanship technique, form characteristics, material composition, cultural meaning, and historical-cultural exchange patterns—delivering a cross-checked comparison report with embedded images and a contrast radar chart in both Markdown and HTML.", "categories": [ "business" ], "tags": [ "三星堆", "中原青铜器", "商代考古", "后母戊鼎", "中华文明探源" ], "tags_en": [ "Sanxingdui", "Central Plains Bronze", "Shang Archaeology", "Houmuwu Ding", "Chinese Civilization Origins" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含对比雷达图)" } ], "prompt": "请按「腾讯探元学术研究-关联研究」专家的工作流,关联对比「三星堆青铜面具」与「中原青铜礼器」在工艺、造型、文化层面的关系。\n\n研究范围(覆盖五维):\n- 工艺技法:陶范法/分铸-嵌铸/铸后补铸、合金配比(铜-锡-铅)、空心设计、装饰浮雕工艺\n- 造型特征:人像vs兽面、纵目/凸眼柱/夔龙冠 vs 饕餮纹/夔龙纹/扉棱/云雷纹\n- 材质成分:铜-锡-铅三元合金,有否添磷、空心铸件、四足热裂处理\n- 文化内涵:神性夸张+祭祀崇拜(古蜀)vs 王权礼制+祖先沟通(商)\n- 交流脉络:丝绸之路前的早期跨区域交流史,长江中下游文化间接影响\n\n交付要求(强制):\n1. 探元两库并行(知识库语义+文物数据库结构化) + 主代理 2-3 次 WebSearch 多源核验(对象 A/对象 B/文化背景)\n2. 配图 ≥ 2 张,优先 Wikimedia / 官方博物馆公开图源;图表 ≥ 1 张(雷达图,SVG 渲染)\n3. 报告:对比结构(对象 A 概述 → 对象 B 概述 → 逐维度对比 → 文化影响与交流 → 结论 → 证据与局限 → 参考来源 → 图片与图表来源)\n4. 仅写「对比结论」,不含行动建议;参考文献只列网络公开来源;图表与图注嵌入对应章节;主要结论同步到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Correlation Research' expert workflow to comparatively research the relationship between 'Sanxingdui bronze masks' and 'Central Plains bronze ritual vessels' across craft, form, and culture.\n\nResearch scope (five dimensions):\n- Craftsmanship: piece-mold casting, after-cast patching, alloy recipes (Cu-Sn-Pb), hollow casting, decorative relief\n- Form: anthropomorphic vs zoomorphic faces, protruding eyes/peak-dragon crown vs taotie/kui dragon/flange/cloud-thunder patterns\n- Material: Cu-Sn-Pb ternary alloy, optional phosphorus addition, hollow leg casting, four-foot thermal-crack mitigation\n- Cultural meaning: spiritual-transcendent & sacrificial worship (Shu) vs royal-ritual & ancestral communication (Shang)\n- Exchange: pre-Silk-Road trans-regional exchange, indirect Lower Yangtze influence\n\nDelivery requirements (mandatory):\n1. Parallel Tanyuan dual bases (knowledge base semantics + relic database structural) + 2-3 main-agent WebSearch passes (Object A / Object B / cultural context)\n2. ≥ 2 images (Wikimedia / official museum public sources preferred); ≥ 1 chart (radar SVG)\n3. Report structure: contrastive (Object A summary → Object B summary → per-dimension comparison → cultural exchange → conclusion → limits → references → image/chart sources)\n4. Conclusions only, no recommendations; references limited to public web sources; charts/figures embedded in corresponding sections; key conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-correlation", "name": "文博关联研究专家", "description": "基于探元文物库与知识库并行采集,由主代理直接联网多源交叉核验,交付四格式跨对象/跨域对比研究报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "75f1347b2045", "expert_id": "ex_pEyI3eIkynB5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:11:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-position-recommend", "title": "职位推荐：本科会计能报哪些岗", "title_en": "Position Match: Jobs an Accounting Grad Can Apply For", "subtitle": "按报考资格实时筛出此刻还能报名的编制岗位", "subtitle_en": "Live-filter public-sector positions still open, matched to your eligibility", "description": "翻职位表是考编最耗人的环节：几千行 Excel，学历专业逐列比对，比对完报名往往已截止。\n\n本案例输入「普通本科 · 工商管理类（会计学）· 想要会计岗 · 只看报名中」，从全国真实职位库筛出 102 个符合资格且在报名期内的岗位，按推荐值取前 10，每条带单位、地区、薪资、招录人数与来源公告。\n\n适合：备考公务员/事业编的应届生、跨省报考的考生、帮学生择岗的就业老师。", "description_en": "Scanning position spreadsheets is the most draining part of exam prep: thousands of rows, degree and major compared column by column, and registration often closes before you finish.\n\nThis case takes \"regular bachelor's, Business Administration (Accounting), wants accounting posts, open-for-application only\" and filters the live nationwide database down to 102 eligible open positions, returning the top 10 by recommendation score with unit, region, salary, headcount and source announcement.\n\nFor: fresh graduates preparing for civil-service exams, cross-province applicants, and career advisors.", "categories": [ "search" ], "tags": [ "公考", "选岗", "职位推荐", "会计" ], "tags_en": [ "civil service", "job matching", "position recommendation", "accounting" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "报考资格匹配岗位清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我是普通本科学历，专业属于工商管理类（会计学），想找会计相关的编制岗位，只要现在还能报名的。请用真实职位库实时筛选：1) 只保留我够格报考且状态为「报名中」的岗位；2) 按推荐值取前 10 条，列出岗位、单位、地区、考试类型、招录人数、薪资标签、来源公告；3) 数据必须来自实时查询，不要示例或估算；4) 最后提示报考前要核对的事项。", "prompt_en": "I hold a regular bachelor's degree in Business Administration (Accounting) and want accounting-related public-sector positions still open for application. Query the live position database: 1) keep only positions I'm eligible for with status \"open\"; 2) return the top 10 by recommendation score with position, unit, region, exam type, headcount, salary tags and source announcement; 3) data must come from a live query, no samples or estimates; 4) end with what to verify before applying.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "school-route-safety", "title": "《上学路上》交通安全闯关教学", "title_en": "《上学路上》交通安全闯关教学", "subtitle": "小学一年级 · 《上学路上》交通安全闯关教学", "subtitle_en": "小学一年级 · 《上学路上》交通安全闯关教学", "description": "一年级《上学路上》交通安全互动课件，用闯关游戏教交通标志、红绿灯与斑马线通行。", "description_en": "一年级《上学路上》交通安全互动课件，用闯关游戏教交通标志、红绿灯与斑马线通行。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "school-route-safety.html", "artifacts": [ { "file": "school-route-safety.html", "label": "《上学路上》交通安全闯关教学｜互动教学网页" } ], "prompt": "帮我做一个小学一年级《上学路上》交通安全的互动教学网页，要设计成卡通闯关游戏。", "prompt_en": "帮我做一个小学一年级《上学路上》交通安全的互动教学网页，要设计成卡通闯关游戏。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "skills": [], "experts": [ { "id": "scrna-seq-expert", "name": "单细胞转录组分析专家", "description": "scRNA-seq 单细胞转录组标准分析全流程（云端 WDL + 本地 scanpy/Seurat）", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_GIBKBuib5zoX" } ], "mcps": [], "id": "scrna-cell-communication", "title": "PBMC 免疫通讯图谱：LIANA 真实数据分析", "title_en": "PBMC Immune Communication Atlas: LIANA Real-Data Analysis", "subtitle": "pbmc68k_reduced · 10 类免疫细胞 · consensus 资源 · 抗原呈递主轴", "subtitle_en": "pbmc68k_reduced · 10 immune cell types · consensus resource · antigen presentation axis", "description": "拿到一份 PBMC 10 类免疫细胞混合数据（10X v1，scanpy 内置 `pbmc68k_reduced`，FACS 排序真值），想知道这些细胞之间在「喊话」什么——谁是主要信号发送者，谁是接收者，抗原呈递轴怎么工作？本案例用 LIANA 框架（Python 版的 CellChat + NicheNet 综合体）在 omnipath consensus 资源上跑出 1288 个配体-受体互作：覆盖 6 阶段（数据加载 / 互作全景 / sender × receiver 热图 / 出度入度 / MHC-II 通路聚焦 / 配体表达点图）。最直观的发现是树突状细胞 HLA-DR → CD4 T 是这条数据集里最强的抗原呈递轴，出度 Dendritic 31、入度 CD4 Naive 16——这与 FACS 排序中 DC 富集、CD4 T 是其天然配对完全吻合。所有数字与图表来自真实 LIANA 执行，可复现。适合：做组织微环境通讯分析的研究者、想替代 CellChat 用 Python 单栈做 LR 配对分析、想理解「互作图谱怎么从单细胞里来」的初学者。", "description_en": "You have a 10-class PBMC immune cell dataset (10X v1, bundled in scanpy as `pbmc68k_reduced`, with FACS-sorted ground truth) and want to know what these cells are saying to each other — who are the main signal senders, who receives, and how does the antigen presentation axis work? This case uses the LIANA framework (a Python combination of CellChat + NicheNet) on the omnipath consensus resource and produces 1288 ligand-receptor interactions across 6 stages: data load / global interaction landscape / sender × receiver heatmap / out-degree and in-degree / MHC-II pathway focus / ligand expression dotplot. The most striking finding: dendritic-cell HLA-DR → CD4 T is the strongest antigen presentation axis in this dataset, with Dendritic out-degree 31 and CD4 Naive in-degree 16 — matching the FACS-sorted cell type abundances. All numbers and figures from real LIANA execution. Suitable for tissue microenvironment communication analysis, switching from CellChat R to a Python single-stack LR analysis, and learners who want to understand how interaction maps emerge from single-cell data.", "categories": [ "data", "education" ], "tags": [ "LIANA", "pbmc68k", "MHC-II", "细胞通讯", "consensus" ], "tags_en": [ "LIANA", "pbmc68k", "MHC-II", "cell communication", "consensus" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "PBMC 免疫通讯图谱（自包含 HTML）" } ], "prompt": "你是一位资深单细胞免疫学分析专家，熟练使用 LIANA 框架（Python 版 CellChat + NicheNet 综合体）。请严格按下面指令一次性产出最终文档，不要询问补充问题。\n\n# 任务\n构建一份名为《PBMC 免疫通讯图谱：LIANA 真实数据分析》的中文 HTML 文档，作为可复跑、可对照的免疫通讯分析演示。\n\n# 输入（真实数据，scanpy 内置，可一键下载）\n使用 `sc.datasets.pbmc68k_reduced()`：700 细胞 × 765 HVGs，已带 10 个 FACS 排序真值（`bulk_labels`：Dendritic / CD14+ Monocyte / CD19+ B / CD4+/CD25 T Reg / CD8+ Cytotoxic T / CD8+/CD45RA+ Naive Cytotoxic / CD56+ NK / CD4+/CD45RO+ Memory / CD34+ / CD4+/CD45RA+/CD25- Naive T）。将 `bulk_labels` 简化为中文标签（去复杂 +/- 符号），赋给 `cell_type` 列。X 用 `.raw.X`（log1p 标准化后的表达）。\n\n# 输出\n单文件 HTML：output.html，自包含（7 张结果图全部 base64 data URI 内嵌、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航）。\n\n# 6 阶段结构（每节固定格式）\n每节五段式：\n1. 简短引言（2-3 句，目标 + 关键决策）\n2. Python 代码（scanpy + liana）：可直接运行 + 关键维度变化\n3. R 等效提示（CellChat）：1-3 行核心函数路径\n4. 核心结果数值表\n5. 结果图 base64 内嵌 + 图注\n\n# 6 阶段必须覆盖\n1. 加载 + bulk_labels → 简化 cell_type 标签\n2. LIANA rank_aggregate(consensus 资源, expr_prop=0.1, n_perms=100)，过滤小群体（min 5 cells/侧）\n3. Top 20 配体-受体互作（magnitude_rank 升序）\n4. 显著互作（mag_rank<0.05）sender × receiver 热图\n5. 每类细胞出度/入度（谁是发送者、谁是接收者）\n6. MHC-II 通路聚焦：HLA-D 系列 → CD4 配对，APC 源 × CD4 目标 lr_means 累加热图 + 配体表达点图（10 个 HLA-D 家族基因）\n\n# 必须嵌入的 7 张图\n1. 10 类细胞构成条形图\n2. Top 20 配体-受体互作水平条形图\n3. sender × receiver 互作数热图\n4. 出度/入度分组条形图\n5. MHC-II lr_means 累加热图（APC → CD4）\n6. Top 互作配体在各细胞类型中的表达点图\n7. HLA-D 家族 10 个基因在各细胞类型中的表达点图\n\n# 关键发现（必须出现）\n- MHC-II → CD4 T 是主轴\n- 树突状细胞是最强抗原呈递者（HLA-D 全家族高表达）\n- CD4 Naive / CD4 Mem 是主要接收者\n- 提供 LIANA vs CellChat 对比表", "prompt_en": "You are a senior single-cell immunology expert proficient in the LIANA framework (Python combination of CellChat + NicheNet). Produce the final document in one pass without asking follow-up questions.\n\n# Task\nBuild a Chinese HTML document titled 'PBMC Immune Communication Atlas: LIANA Real-Data Analysis' as a reproducible, comparable immune communication analysis demo.\n\n# Input (real data, bundled in scanpy)\nUse `sc.datasets.pbmc68k_reduced()`: 700 cells × 765 HVGs with 10 FACS-sorted labels in `bulk_labels` (Dendritic / CD14+ Monocyte / CD19+ B / CD4+/CD25 T Reg / CD8+ Cytotoxic T / CD8+/CD45RA+ Naive Cytotoxic / CD56+ NK / CD4+/CD45RO+ Memory / CD34+ / CD4+/CD45RA+/CD25- Naive T). Map bulk_labels to simplified Chinese labels (strip complex +/- symbols) and assign to `cell_type` column. Use `.raw.X` as X (log1p-normalized expression).\n\n# Output\nSingle-file HTML: output.html, self-contained (7 figures embedded as base64 data URIs, no external resources, content visible by default, no hash anchors).\n\n# 6-stage structure (fixed per-section 5-part format)\n1. Brief intro (2-3 sentences)\n2. Python code (scanpy + liana)\n3. R equivalent (CellChat)\n4. Numeric results table\n5. Base64 figure + caption\n\n# Stages to cover\n1. Load + bulk_labels to simplified cell_type\n2. LIANA rank_aggregate(consensus resource, expr_prop=0.1, n_perms=100), filter small groups (min 5 cells/side)\n3. Top 20 ligand-receptor interactions (ascending by magnitude_rank)\n4. Significant interactions (mag_rank<0.05) sender × receiver heatmap\n5. Per-celltype out-degree/in-degree (who sends, who receives)\n6. MHC-II pathway focus: HLA-D series → CD4, APC source × CD4 target lr_means sum heatmap + ligand expression dotplot (10 HLA-D family genes)\n\n# Required 7 figures\n1. 10-celltype composition bar chart\n2. Top 20 LR interactions horizontal bar chart\n3. sender × receiver interaction count heatmap\n4. Out-degree/in-degree grouped bar chart\n5. MHC-II lr_means sum heatmap (APC → CD4)\n6. Top-interaction ligands expression dotplot across cell types\n7. HLA-D family 10-gene expression dotplot across cell types\n\n# Required key findings\n- MHC-II → CD4 T is the main axis\n- Dendritic cells are the strongest antigen presenters (high expression across all HLA-D family)\n- CD4 Naive / CD4 Mem are the main receivers\n- Include a LIANA vs CellChat comparison table", "cover_image": "cover.png", "cover_mode": "landscape", "created_at": "2026-08-29T17:20:00+08:00", "updated_at": "2026-08-29T17:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scrna-seq-expert", "title": "scRNA-seq 9 阶段实战手册", "title_en": "scRNA-seq 9-Stage Practical Handbook", "subtitle": "从原始矩阵到细胞通讯，Python(scanpy) + R(Seurat) 双栈对照", "subtitle_en": "From raw matrix to cell communication, with Python (scanpy) and R (Seurat) side by side", "description": "想学单细胞分析，教程要么只有零散代码、要么只给结论不给中间数字，跟着跑一遍还是不知道每一步该看什么。这个案例生成一份 9 阶段完整实战手册：数据读入、QC、标准化+HVG+PCA、聚类+UMAP、细胞注释、批次整合（Harmony）、差异分析、拟时序、细胞通讯（CellChat），每节固定「引言 + Python 代码 + R 等效提示 + 关键结果数值表 + 结果图」五段式，11 张结果图全部内嵌，单文件自包含、复制到任何浏览器即可查看。适合：单细胞入门研究生、生信培训讲师、搭建分析 SOP 的团队、方案验证与选型对比的分析师。", "description_en": "Learning single-cell analysis usually means scattered code snippets or conclusions without intermediate numbers — you finish a tutorial still unsure what to check at each step. This case generates a complete 9-stage practical handbook: data loading, QC, normalization + HVG + PCA, clustering + UMAP, cell annotation, batch integration (Harmony), differential analysis, pseudotime, and cell communication (CellChat). Each section follows a fixed five-part format: intro, Python code, R equivalent hints, key numeric result tables, and a figure. All 11 figures are embedded — one self-contained file you can open in any browser. Suitable for: single-cell beginners, bioinformatics trainers, teams building analysis SOPs, and analysts comparing methods.", "categories": [ "data", "education" ], "tags": [ "scRNA-seq", "scanpy", "Harmony", "拟时序", "细胞通讯" ], "tags_en": [ "scRNA-seq", "scanpy", "Harmony", "Pseudotime", "Cell Communication" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "scRNA-seq 9 阶段实战手册（自包含 HTML）" } ], "prompt": "你是一位资深单细胞转录组分析专家，习惯在 Python（scanpy）与 R（Seurat）两栈之间无缝切换。请严格按下面指令一次性产出最终文档，不要询问补充问题。\n\n# 任务\n构建一份名为《scRNA-seq 9 阶段实战手册》的中文 HTML 文档，作为可复跑、可对照、可讲解的完整流水线演示。\n\n# 输入（合成数据，可复现）\n请用代码生成一份合成 PBMC 数据，模拟下面规格：\n- 3,000 cells × ~10,000 expressed genes（用稀疏高斯过程模拟 8 个细胞类型的真实 marker 表达模式）\n- 2 个技术批次（b1/b2，约 6:4 比例），批次 2 在 5 个基因上有 ~15% 的均值偏移\n- 加入合理的线粒体基因比例分布（大多数细胞 <10%，少量到 30%）\n- 加入细胞周期相关基因作为可选评分依据\n\n# 输出\n单文件 HTML：output.html，要求自包含（所有 CSS 内联、11 张结果图全部用 base64 data URI 内嵌、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航），单文件可移植、可离线打开。\n\n# 9 阶段结构（每节固定格式）\n每节必须按以下五段式呈现，缺一不可：\n\n## 阶段 N · <阶段名>\n1. 简短引言（2-3 句，说明目标 + 关键决策点）\n2. Python 代码（scanpy）：可直接运行的代码块，每步加行内注释；代码后给出关键维度变化（如过滤后 cell 数、cluster 数、HVG 数、PC 数等）\n3. 等价 R 提示（Seurat/其他）：给出对应 R 版本的 1-3 行核心函数路径（如 Seurat::NormalizeData()），不要复写完整 R 代码，仅用注释标识等效步骤\n4. 核心结果（数值表）：用表格写本阶段拿到的结果数值，包括关键维度变化（cells/genes/HVG/clusters）、选用参数（min_genes、n_top_genes、n_pcs、resolution 等）、关键阈值（线粒体比例 cutoff、doublet score 阈值等）\n5. 结果图：用 matplotlib 生成 PNG，base64 内嵌到 HTML；下方写图注说明怎么读、主要发现\n\n# 9 个阶段必须覆盖\n1. 数据读入 —— 合成数据生成 + AnnData/Seurat 对象构造 + 关键检查\n2. 质控 (QC) —— 计算 nFeature/nCount/pct.mt，过滤前后数值 + QC violin 图\n3. 标准化 + HVG + PCA —— normalize、log1p、HVG 数 + PCA elbow 图\n4. 聚类 + UMAP —— Leiden/Louvain、resolution 探索 + UMAP-cluster 着色\n5. 自动 + 手工注释 —— 用 CellTypist 风格规则/已知 marker 列表分配 8 个 cell types + Marker 点图 + z-score 热图\n6. 批次整合 —— Harmony 整合前后对比（计算一个混合度指标，如 kBET 或 iLISI）+ 整合前后 UMAP 对比\n7. 差异分析 (DE) —— 选一个 cluster vs rest，输出 top markers + 火山图\n8. 拟时序 —— Monocle3/PAGA 选根节点 + Pseudotime UMAP\n9. 细胞通讯 —— CellChat 跑全部配体-受体对，统计每个 cell type 的入度/出度 + CellChat chord/network 图\n\n# 必须嵌入的 11 张图\n1. QC 指标分布 violin（nFeature, nCount, pct.mt 三联图）\n2. PCA elbow（解释方差 vs PC 编号）\n3. UMAP 按 cluster 着色\n4. UMAP 按 cell type 注释着色\n5. Marker dotplot（gene × cluster，size=pct, color=mean）\n6. Marker heatmap（top 5 marker per cluster 的 z-score 热图）\n7. 批次整合前后 UMAP 对比（左 raw，右 Harmony）\n8. 差异基因火山图（log2FC vs -log10 p-value，关键基因高亮）\n9. 拟时序 UMAP（按伪时间着色 + 根节点标记）\n10. CellChat 出度/入度条形图\n11. CellChat 网络图（点 + 边，或 chord diagram）\n\n# 风格与排版\n- 中文输出；每张代码块前写一句「这一步在做什么」的指引\n- 表格里写真实能拿到的结果数字，而不是模板填空\n- 每张图注 1-2 句，告诉读者怎么读、主要发现\n- 末尾必须有「前置条件清单」：通讯需先有可靠注释、聚类分辨率探索、整合需 batch 列等\n\n# 不要做的事\n- 不留 TODO 或占位符；不输出多轮对话引导\n- 不留任何外链引用，所有图必须 base64 内嵌\n- 不使用 emoji 作为章节符号；不写「完整代码见附录」等偷懒话\n\n# 完成判定\n完成后回报：1) 文件路径 2) 文件大小 3) 嵌入图数量 4) 9 个阶段小节数。", "prompt_en": "You are a senior single-cell transcriptomics expert fluent in both Python (scanpy) and R (Seurat). Follow the instructions below to produce the final document in one pass, without asking follow-up questions.\n\n# Task\nBuild a Chinese-language HTML document titled \"scRNA-seq 9-Stage Practical Handbook\" — a reproducible, cross-referenced, presentable full-pipeline demo.\n\n# Input (synthetic data, reproducible)\nGenerate a synthetic PBMC dataset in code with these specs:\n- 3,000 cells × ~10,000 expressed genes (sparse Gaussian process simulating realistic marker expression of 8 cell types)\n- 2 technical batches (b1/b2, ~6:4 ratio); batch 2 has a ~15% mean shift on 5 genes\n- A reasonable mitochondrial gene ratio distribution (most cells <10%, a few up to 30%)\n- Cell-cycle genes included as optional scoring basis\n\n# Output\nSingle-file HTML: output.html — self-contained (all CSS inlined, all 11 result figures embedded as base64 data URIs, no external resource references, content visible by default, no hash-anchor navigation), portable and openable offline.\n\n# 9-stage structure (fixed per-section format, all five parts required)\n\n## Stage N · \n1. Brief intro (2-3 sentences: goal + key decision points)\n2. Python code (scanpy): runnable code blocks with inline comments; report key dimension changes after the code (post-filter cell count, cluster count, HVG count, PC count, etc.)\n3. Equivalent R hints (Seurat/others): 1-3 lines of core function paths (e.g. Seurat::NormalizeData()); do not rewrite full R code, just annotate equivalent steps\n4. Core results (numeric table): real numbers from this stage — key dimension changes (cells/genes/HVG/clusters), chosen parameters (min_genes, n_top_genes, n_pcs, resolution, etc.), key thresholds (mitochondrial ratio cutoff, doublet score threshold, etc.)\n5. Result figure: matplotlib PNG embedded as base64; caption below explaining how to read it and the main finding\n\n# The 9 stages must cover\n1. Data loading — synthetic data generation + AnnData/Seurat object construction + key checks\n2. QC — compute nFeature/nCount/pct.mt, pre/post-filter numbers + QC violin plot\n3. Normalization + HVG + PCA — normalize, log1p, HVG count + PCA elbow plot\n4. Clustering + UMAP — Leiden/Louvain, resolution exploration + UMAP colored by cluster\n5. Auto + manual annotation — CellTypist-style rules/known marker lists to assign 8 cell types + marker dotplot + z-score heatmap\n6. Batch integration — Harmony before/after comparison (compute a mixing metric such as kBET or iLISI) + before/after UMAP comparison\n7. Differential analysis (DE) — one cluster vs rest, top markers + volcano plot\n8. Pseudotime — Monocle3/PAGA root selection + pseudotime UMAP\n9. Cell communication — CellChat over all ligand-receptor pairs, per-cell-type in/out degree + CellChat chord/network figure\n\n# The 11 required figures\n1. QC metric violins (nFeature, nCount, pct.mt triplet) 2. PCA elbow (variance explained vs PC number) 3. UMAP colored by cluster 4. UMAP colored by cell-type annotation 5. Marker dotplot (gene × cluster, size=pct, color=mean) 6. Marker heatmap (z-score of top 5 markers per cluster) 7. Before/after batch-integration UMAP (left raw, right Harmony) 8. DEG volcano plot (log2FC vs -log10 p-value, key genes highlighted) 9. Pseudotime UMAP (colored by pseudotime + root marked) 10. CellChat out/in degree bar chart 11. CellChat network figure (nodes + edges, or chord diagram)\n\n# Style and layout\n- Chinese output; one line of \"what this step does\" before each code block\n- Tables must contain real obtainable result numbers, not template placeholders\n- Each figure caption 1-2 sentences: how to read it, main finding\n- The document must end with a \"prerequisites checklist\": communication requires reliable annotation, clustering resolution exploration, integration requires a batch column, etc.\n\n# Do not\n- Leave TODOs or placeholders; produce multi-turn dialogue prompts\n- Leave any external link references; all figures must be base64-embedded\n- Use emoji as section symbols; write lazy lines like \"full code in appendix\"\n\n# Completion criteria\nWhen done, report: 1) file path 2) file size 3) number of embedded figures 4) number of the 9 stage sections.", "skills": [], "experts": [ { "id": "scrna-seq-expert", "name": "单细胞转录组分析专家", "description": "scRNA-seq 单细胞转录组标准分析全流程（云端 WDL + 本地 scanpy/Seurat）", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_GIBKBuib5zoX" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-27T22:03:55+08:00", "updated_at": "2026-08-30T11:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "source": "official", "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "skills": [], "experts": [ { "id": "scrna-seq-expert", "name": "单细胞转录组分析专家", "description": "scRNA-seq 单细胞转录组标准分析全流程（云端 WDL + 本地 scanpy/Seurat）", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_GIBKBuib5zoX" } ], "mcps": [], "id": "scrna-trajectory-velocity", "title": "paul15 骨髓分化轨迹与 RNA 速率实战", "title_en": "paul15 Myeloid Differentiation Trajectory & RNA Velocity", "subtitle": "真实数据 PAGA · DPT 伪时间 · 合成 spliced/unspliced · scVelo 速率", "subtitle_en": "Real data PAGA · DPT pseudotime · synthetic spliced/unspliced · scVelo velocity", "description": "手头有一份真实的骨髓单细胞数据（Paul 等 2015 造血分化，scanpy 内置 `paul15`，2730 细胞 × 3451 基因，FACS 排序真值），想回答：分化顺序与分叉点是什么？每个谱系（Erythroid / Granulocyte / Monocyte / Megakaryocyte）的速率驱动基因？scVelo 速率方向与 PAGA/DPT 是否一致？本案例用 PAGA 粗粒度图谱揭示簇 7（MEP 区）作为中枢节点向 4-5 个分支辐射；用 DPT 扩散伪时间给每个细胞一个 0-1 的时间排序；用基于 DPT 合成的 spliced/unspliced（unspliced 沿 DPT 前移 0.05 dt）跑 scVelo stochastic 模型，速率流场从 UMAP 中心向外辐射——与 PAGA 拓扑完全吻合；按谱系 top 8 速率基因展示 Ccl5、Id2、S100A8/A9 等。**注意**：paul15 原始数据无 RNA 剪接计数，本案例为演示 velocity 流程基于 DPT 合成演示数据，真实分析应先用 velocyto/STARsolo 从 BAM 重建。6 张图全部内嵌。适合：做分化/发育轨迹的研究者、PAGA 与 DPT 想从入门到上手、学习 scVelo 速度分析的初学者。", "description_en": "You have a real myeloid single-cell dataset (Paul et al. 2015, bundled in scanpy as `paul15`, 2730 cells × 3451 genes, FACS-sorted ground truth) and want to know: what's the differentiation order and branching point? What are the velocity-driven genes for each lineage (Erythroid / Granulocyte / Monocyte / Megakaryocyte)? Do scVelo velocity directions agree with PAGA/DPT? This case uses PAGA coarse-grained graphs to reveal cluster 7 (MEP region) as a central hub radiating to 4-5 branches, computes DPT diffusion pseudotime for every cell, runs scVelo stochastic on DPT-synthesized spliced/unspliced (unspliced shifted forward 0.05 dt) yielding a velocity stream that radiates from the UMAP center — perfectly matching the PAGA topology — and displays top-8 velocity genes per lineage (Ccl5, Id2, S100A8/A9, etc). **Note**: paul15 has no splice counts; this case synthesizes demo data based on DPT to demonstrate velocity flow — for real analysis, generate spliced/unspliced from velocyto/STARsolo first. All 6 figures embedded. Suitable for differentiation/development researchers, PAGA/DPT learners, scVelo beginners.", "categories": [ "data", "education" ], "tags": [ "paul15", "PAGA", "DPT", "scVelo", "RNA 速率", "轨迹推断" ], "tags_en": [ "paul15", "PAGA", "DPT", "scVelo", "RNA velocity", "trajectory" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "paul15 轨迹与速率实战（自包含 HTML）" } ], "prompt": "你是一位资深单细胞轨迹与 RNA 速率分析专家，熟练 PAGA / DPT / scVelo 全套工具链。请严格按下面指令一次性产出最终文档，不要询问补充问题。\n\n# 任务\n构建一份名为《paul15 骨髓分化轨迹与 RNA 速率实战》的中文 HTML 文档，作为可复跑、可对照的分化轨迹演示。\n\n# 输入（真实数据，scanpy 内置，可一键加载）\n使用 `sc.datasets.paul15()`：2730 细胞 × 3451 基因，原始 counts，带 19 个 FACS 排序真值 `paul15_clusters`（如 `1Ery`、`7MEP`、`9GMP`）。**不直接用这 19 个原簇**——重做 Leiden（resolution=1.0 → 12 个新簇），并把 19 个原簇按 lineage 缩写归为 5 大谱系大类 + Other（Erythroid / Granulocyte / Monocyte / Megakaryocyte / Dendritic 加上未匹配的 `Ba / CD4+/CD8+ T / NK / ILC / pDC` 等归入 Other）。\n\n# 输出\n单文件 HTML：output.html，自包含（6 张图全部 base64 data URI 内嵌、无外部资源、内容默认可见、不使用 hash 锚点）。\n\n# 6 阶段结构（每节五段式：引言 + Python 代码 + R 等效 + 数值表 + 图注）\n\n# 6 阶段必须覆盖\n1. 加载 paul15 + 简化 lineage 标签（5 谱系 + Other） + 标准化 + HVG(1500) + PCA(50)\n2. neighbors + UMAP + Leiden(1.0) + 谱系多数投票映射 + 双联图（leiden/lineage）\n3. PAGA：`sc.tl.paga(groups='leiden')` + `sc.pl.paga(layout='fr', threshold=0.05)`，展示簇 7 作为中枢节点\n4. diffmap(15) + 选 Erythroid 区根节点 + `sc.tl.dpt(n_dcs=10)` + DPT UMAP\n5. 与 DPT 相关 top 24 基因（正负各 12）→ 50 个 DPT 分箱的 z-score 热图\n6. **基于 DPT 合成 spliced/unspliced**：spliced = 沿 DPT 移动平均平滑 + 噪声；unspliced = spliced 沿 DPT 前移 0.05 dt + 噪声。`scv.pp.filter_and_normalize` → `sc.pp.neighbors` → `scv.pp.moments` → `scv.tl.velocity(stochastic)` → `scv.tl.velocity_graph` → `scv.pl.velocity_embedding_stream`\n7. 按谱系 top 8 速率基因水平条形图（4 谱系 × 8 基因）\n\n# 必须嵌入的 6 张图\n1. UMAP 双联（leiden / lineage）\n2. PAGA 粗粒度连通性图（force-directed layout）\n3. DPT UMAP（紫到黄） + 红圈根节点\n4. 24 基因 z-score 热图（含 r 值标注）\n5. scVelo 速率流场 stream plot（按谱系着色）\n6. 4 谱系 × top 8 速率基因水平条形图\n\n# 关键说明（必须出现）\n- paul15 原始数据无 RNA 剪接计数，本案例合成 spliced/unspliced 用于演示 velocity 流程\n- 真实分析应先用 velocyto/STARsolo 从 BAM 重建剪接计数后从 .loom 导入\n- 提供 PAGA vs Monocle3 vs Slingshot 对比表", "prompt_en": "You are a senior single-cell trajectory and RNA velocity analysis expert proficient in the PAGA / DPT / scVelo toolchain. Produce the final document in one pass without asking follow-up questions.\n\n# Task\nBuild a Chinese HTML document titled 'paul15 Myeloid Differentiation Trajectory & RNA Velocity' as a reproducible, comparable differentiation trajectory demo.\n\n# Input (real data, bundled in scanpy)\nUse `sc.datasets.paul15()`: 2730 cells × 3451 genes, raw counts, with 19 FACS-sorted labels in `paul15_clusters` (e.g. `1Ery`, `7MEP`, `9GMP`). **Do not use the 19 original clusters directly** — re-run Leiden (resolution=1.0 → 12 new clusters) and map the 19 original clusters by lineage abbreviation to 6 broad categories (Erythroid / Granulocyte / Monocyte / Megakaryocyte / Dendritic / Lymphoid).\n\n# Output\nSingle-file HTML: output.html, self-contained (6 figures embedded as base64 data URIs, no external resources, content visible by default, no hash anchors).\n\n# 6-stage structure (fixed per-section 5-part format)\n\n# Stages to cover\n1. Load paul15 + simplified lineage labels + normalize + HVG(1500) + PCA(50)\n2. neighbors + UMAP + Leiden(1.0) + lineage majority vote mapping + double-plot (leiden/lineage)\n3. PAGA: `sc.tl.paga(groups='leiden')` + `sc.pl.paga(layout='fr', threshold=0.05)`, show cluster 7 as central hub\n4. diffmap(15) + select Erythroid-region root cell + `sc.tl.dpt(n_dcs=10)` + DPT UMAP\n5. Top 24 DPT-correlated genes (12 positive + 12 negative) → 50 DPT bins z-score heatmap\n6. **Synthesize spliced/unspliced from DPT**: spliced = DPT moving-average smooth + noise; unspliced = spliced shifted forward 0.05 dt in DPT + noise. `scv.pp.filter_and_normalize` → `sc.pp.neighbors` → `scv.pp.moments` → `scv.tl.velocity(stochastic)` → `scv.tl.velocity_graph` → `scv.pl.velocity_embedding_stream`\n7. Top 8 velocity genes per lineage horizontal bar chart (4 lineages × 8 genes)\n\n# Required 6 figures\n1. UMAP double-plot (leiden / lineage)\n2. PAGA coarse-grained connectivity graph (force-directed layout)\n3. DPT UMAP (purple to yellow) + red circle root cell\n4. 24-gene z-score heatmap (with r values annotated)\n5. scVelo velocity stream plot (colored by lineage)\n6. 4 lineages × top 8 velocity gene horizontal bar chart\n\n# Required key notes\n- paul15 has no RNA splice counts; this case synthesizes spliced/unspliced for velocity demo only\n- For real analysis, generate spliced/unspliced from velocyto/STARsolo first then import from .loom\n- Include a PAGA vs Monocle3 vs Slingshot comparison table", "cover_image": "cover.png", "cover_mode": "landscape", "created_at": "2026-08-29T17:20:00+08:00", "updated_at": "2026-08-29T17:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sector-hotspot-tactical-power-equipment", "title": "电力设备及新能源 · 轻量战术研判", "title_en": "Power Equipment & New Energy — Tactical Research", "subtitle": "A股板块四段式拆解：逻辑归因 → 梯队研判 → 操作预案 → 路径推演", "subtitle_en": "A-share sector four-phase breakdown: Logic Attribution → Echelon Analysis → Action Plan → Path Projection", "description": "还在纠结板块轮动节奏？以「电力设备及新能源」为例，用 sector-hotspot-analysis 技能的轻量战术研判模式，走完逻辑归因→梯队研判→操作预案→路径推演的完整四段式分析链路。基于真实行情数据（2026-08-04），结合「十五五」新型电力系统规划催化事件，识别领涨龙头、补涨标的与掉队品种，给出条件化的仓位取向与多条路径推演。适合需要快速拆解题材并形成可执行操作框架的投资者。", "description_en": "Struggling with sector rotation timing? Using Power Equipment & New Energy as an example, this case walks through the full four-phase tactical analysis pipeline of the sector-hotspot-analysis skill: Logic Attribution → Echelon Analysis → Action Plan → Path Projection. Based on real market data (2026-08-04) and the '15th Five-Year Plan' new power system catalyst, it identifies leading stocks, catch-up plays, and laggards, with conditional position orientation and multi-path projections. Ideal for investors needing rapid sector breakdown with actionable frameworks.", "categories": [ "finance" ], "tags": [ "板块轮动", "电力设备", "新能源", "战术研判", "A股" ], "tags_en": [ "sector rotation", "power equipment", "new energy", "tactical research", "A-share" ], "artifact_type": "html", "preview": "tactical-analysis-power-equipment.html", "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请做 A 股「轻量战术研判」：题材/板块聚焦 电力设备及新能源，输出四段式 Markdown（逻辑归因 → 梯队研判 → 操作预案 → 路径推演）。数据必须走可信 skill（westock-data / neodata-financial-search / wb-finance-skill），不要编造涨跌幅。报告含市场日期、板块代码、领涨/补涨/协同/掉队标的的行情数据、条件化操作预案、多路径推演及免责声明。\n\nSuccess Criteria: 所有输出基于真实查询数据，四段式结构完整，每段有明确结论和条件化判断；防守观察位给出具体价位；路径推演含概率估计和触发条件。", "prompt_en": "Perform an A-share 'Lightweight Tactical Research': sector/theme focused on Power Equipment & New Energy, output four-phase Markdown (Logic Attribution → Echelon Analysis → Action Plan → Path Projection). Data must come from trusted skills (westock-data / neodata-financial-search / wb-finance-skill), do not fabricate price changes. Report includes market date, sector code, market data for leading/catch-up/synergy/laggard stocks, conditional action plan, multi-path projections, and disclaimer.\n\nSuccess Criteria: All output based on real queried data, complete four-phase structure, each phase with clear conclusions and conditional judgments; defensive observation levels include specific price points; path projections include probability estimates and trigger conditions.", "skills": [ { "id": "sector-hotspot-analysis", "name": "跨市场板块轮动与热点分析", "name_en": "Cross-Market Sector Rotation & Hotspot Analysis", "description": "跨市场板块轮动与热点分析，覆盖资金流向、板块强弱与战术级研究结论", "skill_id": "skill_2085355360696082432" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "tactical-analysis-power-equipment.html", "label": "战术研判报告（HTML）" }, { "file": "tactical-analysis-power-equipment.md", "label": "战术研判报告（Markdown）" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "usage_count": 0, "created_at": "2026-08-04T04:01:06.000Z", "updated_at": "2026-08-04T04:01:06.000Z", "is_recommended": false, "recommended_order": 0 }, { "id": "sector-news-weekly-dashboard", "title": "半导体板块要闻看板", "title_en": "Semiconductor Sector News Dashboard", "subtitle": "板块资讯按主题自动归类，晨会前扫一眼心里有数", "subtitle_en": "Sector news auto-grouped by theme—one glance before the morning meeting", "description": "看板块新闻最头疼的是信息太散：政策、公司、产业链混在一起，捋不出主线。这个案例演示了如何让 AI 多角度检索半导体板块的真实报道，按主题归类成一块要闻看板，顶部统计概览、下部分区卡片，每条都标注来源和日期，晨会前扫一眼就心里有数。\n\n适合：关注行业轮动的投资者、行业研究员、需要板块情报的投研和产品同学。", "description_en": "The hardest part of following sector news is how scattered it is—policy, company updates, and supply-chain stories all mixed together with no clear thread. This case shows how AI retrieves real semiconductor sector reports from multiple angles and organizes them into a themed news dashboard: stats overview on top, themed card sections below, every item labeled with source and date. One quick scan before the morning meeting and you know exactly what is going on.\n\nBest for: investors tracking sector rotation, industry researchers, and investment research or product teams who need sector intelligence.", "categories": [ "finance", "search" ], "tags": [ "半导体", "板块要闻", "数据看板", "行业研究" ], "tags_en": [ "Semiconductor", "Sector News", "Dashboard", "Industry Research" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "半导体板块要闻看板" } ], "prompt": "帮我盘点半导体板块近期的要闻。请用平安证券资讯查询 skill 从多个角度真实检索（如「半导体板块 最新报道」「芯片 行业政策」「晶圆代工 动态」「存储芯片 行情」等），把召回资讯按主题归类（如行业政策、公司动态、产业链、市场行情），每个主题内按日期由新到旧排列，注明来源与日期。最终产出一份单文件 HTML 板块要闻看板：顶部为板块概览与要闻统计，下方按主题分区展示资讯卡片，页脚注明数据来源（平安证券资讯知识库）与「本内容仅为资讯整理，不构成投资建议」。视觉方向：数据看板风格、统计卡片加分区布局、适合投研晨会快速浏览；所有资源内联、可离线打开。", "prompt_en": "Help me round up recent important news in the semiconductor sector. Use the Ping An News Search skill to run real queries from multiple angles (e.g. \"semiconductor sector latest reports\", \"chip industry policy\", \"foundry updates\", \"memory chip market\"), group the retrieved news by theme (e.g. industry policy, company updates, supply chain, market trends), sort each theme by date (newest first), and label each item with source and date. Produce a single-file HTML sector news dashboard: an overview with news statistics at the top, themed sections of news cards below, and a footer stating the data source (Ping An Securities news knowledge base) and \"This content is for informational purposes only and does not constitute investment advice.\" Visual direction: dashboard style with stat cards and sectioned layout, suitable for a quick scan before research morning meetings; all resources inlined so the file works offline.", "skills": [ { "id": "news-search", "name": "平安证券资讯查询", "description": "按股票、公司、行业、ETF、概念或市场主题检索相关新闻、快讯和财经报道。", "skill_id": "skill_2079451206686916608" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-24T17:48:03+08:00", "updated_at": "2026-07-27T12:49:23+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "segmented-private-traffic-content-calendar", "title": "分层私域内容排期", "title_en": "Segmented Private-Traffic Content Calendar", "subtitle": "四类客群的7天私域内容运营方案", "subtitle_en": "A practical Weimob playbook for segmented private-traffic content calendar", "description": "做分层私域内容排期时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“分层私域内容排期”展开：面向新客、活跃会员、沉睡会员和高价值客户，分别生成企微一对一话术、社群内容、商品推荐和7天内容排期；只输出方案，不执行群发。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、私域触达运营运营人员、经营分析团队。", "description_en": "The hardest part of segmented private-traffic content calendar is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates segmented private-traffic content calendar through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Private-Traffic Outreach specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "私域分层", "企微话术", "社群内容", "商品推荐", "七天排期" ], "tags_en": [ "Segmented Private-Traffic Content Calendar", "Customer Operations", "Private-Traffic Outreach", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "针对新客、活跃会员、沉睡会员和高价值客户，分别生成企微一对一触达话术、社群内容和商品推荐文案，并输出一套7天内容排期；只生成方案和文案，不执行群发。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a segmented private-traffic content calendar covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" }, { "id": "content-creation", "name": "内容创作与分发", "description": "创作营销文案、公众号图文、小红书和朋友圈内容；可按渠道生成素材并发布或保存入库。", "skill_id": "skill_2087078523417223168" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "selfmedia-creator-workbench", "title": "诘忍生的自媒体创作台", "title_en": "Self-Media Creator Workbench", "subtitle": "灵感、选题、发布计划，一个页面管完整个内容流水线", "subtitle_en": "Ideas, topics, and publishing plans — your whole content pipeline on one page", "description": "做自媒体最头疼的不是写不出，而是灵感随手记在各处、选题想到就忘、发布节奏全凭感觉，时间一长账号就断更了。\n\n这个案例演示了一个「自媒体创作台」：打开页面先看「今日」区，今天要发什么、哪些选题逾期了一目了然；灵感库把刷到的好内容随手存档、按标签筛选；选题库用看板管理「待写→写作中→待发布→已发布」全流程，优先级和平台清清楚楚；发布页按日历日期排期，月底还有本月发布统计。数据存在资料库数据表里，增删改实时写回，手机电脑打开都是同一份。\n\n适合：想稳定更新的自媒体创作者、一边读博/上班一边做账号的斜杠青年、需要管理多平台发布节奏的运营、想给自己搭一个轻量内容管理系统的效率党。", "description_en": "The hardest part of running a self-media account isn't writing — it's that ideas get jotted down everywhere, topics are forgotten as soon as they pop up, and the publishing rhythm runs on vibes, until the account quietly goes silent.\n\nThis case shows a self-media creator workbench: open the page and the Today section tells you what to publish today and which topics are overdue; the idea library archives great content you stumble across, filterable by tags; the topic board manages the full pipeline from to-write → writing → ready → published, with priority and platform on every card; the publishing tab schedules by date and wraps up the month with publishing stats. Data lives in a Library structured table — every add, edit, or delete writes back in real time, so desktop and mobile always show the same thing.\n\nBest for: creators who want to publish consistently, side-hustlers running an account alongside a PhD or a day job, operators managing multi-platform publishing cadences, and anyone who wants a lightweight personal CMS.", "categories": [ "workspace" ], "tags": [ "自媒体工作台", "选题管理", "灵感收集", "发布计划", "资料库数据表" ], "tags_en": [ "Self-Media Workbench", "Topic Management", "Idea Collection", "Publishing Schedule", "Library Tables" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "自媒体创作台单文件（接资料库数据表，内置中英切换与示例数据兜底）" } ], "prompt": "请使用资料库的能力帮我一键做同款（https://www.workbuddy.cn/space/d/IrLW26DvPgoxxZ5wU6m0Hg），需要实现数据在线存储和双向同步。优先走一键做同款的链路，若没有能力则按以下方式搭建：\n\n帮我搭一个「自媒体创作台」。\n\n四个模块：\n1. 今日：聚合展示今天要发布的内容和逾期的选题，支持一键标记已发布，可折叠\n2. 灵感库：随手记录刷到的选题灵感（标题、描述、标签、来源链接），支持按标签筛选，灵感可一键转成选题\n3. 选题库：看板管理选题全流程（待写→写作中→待发布→已发布），每张卡片带优先级（P0/P1/P2）、目标平台（B站/小红书/抖音/YouTube/公众号）、计划日期和内容方向\n4. 发布：按计划日期排列发布节奏，附本月发布统计\n\n数据存储：用资料库能力自建一张数据表（标题、类型、状态、描述、标签、来源、优先级、平台、计划日期、内容方向、创建日期），页面实时读写这张表（增删改直接写回），离线时用本地缓存兜底；再把工作台 HTML 部署到资料库，交付一个在线链接。\n\n界面：低饱和治愈系配色、卡片式布局、圆角柔和，信息密度适中，PC/手机双适配（移动端底部 Tab 导航）；零外链，图标用内联 SVG 不用 emoji。\n\n交付：资料库在线工作台——在线 HTML 页面 + 1 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://www.workbuddy.cn/space/d/IrLW26DvPgoxxZ5wU6m0Hg, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build me a self-media creator workbench. Four modules:\n1. Today: aggregates what should be published today and which topics are overdue, with one-tap mark-as-published, collapsible\n2. Idea library: quickly archive content ideas you come across (title, description, tags, source link), filterable by tag; any idea can be converted into a topic with one tap\n3. Topic board: a kanban managing the full topic pipeline (to-write → writing → ready → published); every card carries priority (P0/P1/P2), target platform (Bilibili/RED/Douyin/YouTube/WeChat official account), planned date, and content direction\n4. Publishing: arranges the publishing cadence by planned date, with monthly publishing stats Data storage: use the Library to self-create one structured table (title, type, status, description, tags, source, priority, platform, planned date, content direction, created date); the page reads and writes this table in real time (adds/edits/deletes write straight back), with a local-cache fallback when offline; then deploy the workbench HTML to the Library and deliver an online link. Interface: low-saturation healing color palette, card-based layout, soft rounded corners, moderate information density, responsive on desktop and mobile (bottom tab navigation on mobile); with zero external links, icons as inline SVG (no emoji).\n\nDeliverable: an online Library workbench — an online HTML page plus 1 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929" } ], "mcps": [], "creator": { "name": "诘忍生", "name_en": "诘忍生", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "community", "usage_count": 0, "created_at": "2026-08-14T20:53:29+08:00", "updated_at": "2026-08-19T23:35:51+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "semiconductor-outlook", "title": "半导体行业怎么看？", "title_en": "How to Read the Semiconductor Industry?", "subtitle": "一文读懂半导体｜热点·行情·产业", "subtitle_en": "Read the semiconductor sector in one article: hotspots, market, industry", "description": "想搞懂半导体板块，热点、行情、观点、产业链四个维度要一个个单独查，数据东拼西凑，还总担心来源不权威。\n\n这个案例把易方达四大官方数据能力串成一条流水线，聚焦半导体，产出一份信息密度高、全部真实可溯源的单页 HTML：从近期热点到六只核心指数走势，再到基金经理观点和产业链结构，一次看完。\n\n适合：想快速了解某个行业的投资者。", "description_en": "Understanding the semiconductor sector requires checking hotspots, market, views and supply chain separately, with scattered data and uncertain sources. This case chains four official E Fund data capabilities into one pipeline, focused on the semiconductor topic, delivered as one dense, fully traceable single-page HTML: from recent hotspots to the 3-month trend of six core semiconductor indexes, then to manager views and supply chain structure, all in one read.\n\nIdeal for: investors who want a quick overview of an industry.", "categories": [ "finance" ], "tags": [ "半导体", "芯片", "热点概览", "基金经理观点", "产业链" ], "tags_en": [ "semiconductor", "chip", "hotspots", "manager views", "supply chain" ], "artifact_type": "html", "preview": "output.html", "prompt": "生成单文件 HTML《半导体行业怎么看？》。所有数据必须通过易方达官方 Skill 真实查询、注明来源与日期，禁止编造/示意/模拟。四部分顺序：① 热点梳理（efunds-hotspot-combing 事件分支）② 行情走势（index-query，六只核心半导体指数近 3 月折线对比）③ 相关观点（efunds-manager-views）④ 产业链结构（efunds-topic-overview）。E-Fund VI 风格，涨红跌绿，CSS/JS 全内联，顶部「易方达基金」徽章+数据日期，页脚带数据日期与风险提示。", "prompt_en": "Generate a single-file HTML report \"How to Read the Semiconductor Industry?\". All data must be obtained through real queries using the official E Fund skills, with source and date noted - no fabrication/placeholder/simulated data. Four sections in order: 1. Hotspot combing (efunds-hotspot-combing, event branch) 2. Market trends (index-query, 3-month line comparison of six core semiconductor indexes) 3. Related views (efunds-manager-views) 4. Supply chain structure (efunds-topic-overview). E-Fund VI style, gains red / losses green, all CSS/JS inlined, top badge \"E Fund\" with the data date, footer carries the data date and a risk disclaimer.", "skills": [ { "id": "efunds-hotspot-combing", "name": "易方达热点梳理", "description": "易方达基金「热点梳理」总入口：内置 3 类场景（事件热点梳理 / 近期热点速览 / 热点周报），全部基于 efunds-mcp 检索「易方达基金」「易方达指数通」两个公众号的公开文章；AI 仅做资讯与事实梳理、不做观点总结，涉及市场观点的内容一律引用原文。", "skill_id": "skill_2092287909491437568" }, { "id": "efunds-topic-overview", "name": "易方达主题概念速览", "description": "当用户提到某个行业、投资策略、资产类别、热门主题等，想获取相关信息、辅助理解概念时使用。基于易方达基金各官方微信公众号已发布文章，自动围绕关键词梳理整合相关资讯，输出结构化内容（包含概念阐释、市场动态、相关表现等），帮助用户快速了解某一主题的相关内容。", "skill_id": "skill_2092171456670191616" }, { "id": "efunds-manager-views", "name": "易方达基金经理观点检索", "description": "当用户想了解最新投研观点或某位基金经理时使用。基于易方达基金官网和各官方微信公众号已发布文章，自动检索并整合基金经理们公开发表的投资观点，根据用户关注维度输出观点汇总、基金经理档案等结构化内容，帮助用户获取专业资讯。", "skill_id": "skill_2089896329561710592" }, { "id": "index-query", "name": "指数查询", "description": "查询证券指数的基础信息、行情、成分股、行业分布、估值、收益及相关基金等数据", "skill_id": "skill_2089730438115336192" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-25T09:20:56+08:00", "updated_at": "2026-08-31T17:25:42+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "output.html", "label": "完整分析报告（单文件 HTML）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "seo-keyword-research", "title": "SEO 关键词调研报告", "title_en": "SEO Keyword Research Report", "subtitle": "输入目标关键词，AI 产出完整的搜索量、竞争度、长尾词与选题策略报告", "subtitle_en": "Input target keyword, AI delivers a complete report on search volume, competition, long-tail keywords and content strategy", "description": "做跨境电商独立站，想靠内容获取自然流量，可光是关键词调研就让人头大——搜索量在哪查、竞争度怎么判断、长尾词怎么挖、选题怎么定？手动整理一次至少半天，还不一定准。\n\n这个案例演示了输入目标关键词「跨境电商独立站」，AI 一站式产出完整的 SEO 关键词调研报告：核心关键词搜索量估算与竞争度分析、20 个长尾关键词挖掘（含搜索意图分类）、关键词难度评级（高/中/低）、10 个基于关键词的文章标题建议、竞品排名前 5 页面分析摘要、关键词布局策略建议。报告 3000 字以上，结构化 Markdown 格式，直接指导后续内容生产。\n\n适合：跨境电商运营、独立站站长、SEO 专员、内容营销团队、数字营销从业者。", "description_en": "Running a cross-border e-commerce DTC site and want organic traffic through content, but keyword research alone is overwhelming — where to check search volume, how to judge competition, how to mine long-tail keywords, how to pick topics? Manual research takes at least half a day and may still be inaccurate.\n\nThis case demonstrates inputting the target keyword 'cross-border e-commerce DTC site' and having AI produce a complete SEO keyword research report: core keyword search volume estimation and competition analysis, 20 long-tail keywords with search intent classification, keyword difficulty ratings (high/medium/low), 10 keyword-based article title suggestions, top-5 competitor page analysis summary, and keyword placement strategy recommendations. The report is 3000+ words in structured Markdown, directly guiding subsequent content production.\n\nIdeal for: Cross-border e-commerce operators, DTC site owners, SEO specialists, content marketing teams, digital marketing professionals.", "categories": [ "search", "business" ], "tags": [ "SEO", "关键词调研", "跨境电商", "长尾关键词", "竞争度分析" ], "tags_en": [ "SEO", "Keyword Research", "Cross-border E-commerce", "Long-tail Keywords", "Competition Analysis" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "SEO关键词调研报告" } ], "prompt": "帮我做一份 SEO 关键词调研报告，目标关键词是「跨境电商独立站」。要求包含：核心关键词搜索量估算与竞争度分析、20 个长尾关键词挖掘（含搜索意图分类）、关键词难度评级（高/中/低）、内容选题建议（10 个基于关键词的文章标题）、竞品排名前 5 的页面分析摘要、关键词布局策略建议。输出为结构化 Markdown 报告，3000 字以上。", "prompt_en": "Help me produce an SEO keyword research report, with the target keyword being \"cross-border e-commerce independent stores (DTC sites).\" Requirements: estimate the search volume and analyze the competitiveness of the core keyword, mine 20 long-tail keywords (with search-intent classification), rate keyword difficulty (high/medium/low), provide content topic suggestions (10 article titles based on keywords), a summary analysis of the top 5 ranking competitor pages, and keyword layout strategy recommendations. Output a structured Markdown report, 3,000+ words.", "skills": [ { "id": "seo-ops", "name": "seo-ops", "description": "SEO 运营自动化，内容攻击简报、GSC 优化与趋势侦察", "description_en": "SEO operations: content attack briefs, GSC optimization, and trend scouting", "skill_id": "skill_2053083137368264704" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "series-and-parallel-circuits", "title": "串并联电路规律与等效电阻", "title_en": "串并联电路规律与等效电阻", "subtitle": "中学物理 · 串并联电路规律与等效电阻", "subtitle_en": "中学物理 · 串并联电路规律与等效电阻", "description": "中学物理《串并联电路》互动课件，用可调电压电阻的虚拟电表，归纳串并联规律并化简等效电阻。", "description_en": "中学物理《串并联电路》互动课件，用可调电压电阻的虚拟电表，归纳串并联规律并化简等效电阻。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "series-and-parallel-circuits.html", "artifacts": [ { "file": "series-and-parallel-circuits.html", "label": "串并联电路规律与等效电阻｜互动教学网页" } ], "prompt": "帮我做一个中学物理《串并联电路》的互动教学网页，要有能调参数的虚拟电路和电表。", "prompt_en": "帮我做一个中学物理《串并联电路》的互动教学网页，要有能调参数的虚拟电路和电表。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "service-contract", "title": "企业管理咨询服务合同", "title_en": "Management Consulting Services Contract", "subtitle": "把服务范围、成果验收与知识产权落成条款", "subtitle_en": "Convert service scope, deliverable acceptance and IP ownership into enforceable clauses.", "description": "咨询项目最怕服务范围说不清、成果难验收、报告归属不明确。\n\n本案例将组织效能诊断与经营策略研讨需求，生成可编辑的《企业管理咨询服务合同》，把服务边界、交付物、付款节点、保密与成果权属可视化呈现。\n\n适合：咨询采购方、专业服务机构、企业战略团队、项目负责人。", "description_en": "Consulting projects often run into trouble when scope, acceptance rules and report ownership are vague.\n\nThis case turns an organizational effectiveness and strategy workshop brief into an editable Management Consulting Services Contract, visualizing scope, deliverables, payment milestones, confidentiality and ownership.\n\nSuitable for: consulting buyers, professional service firms, corporate strategy teams and project owners.", "categories": [ "business" ], "tags": [ "服务合同", "咨询项目", "成果验收", "知识产权" ], "tags_en": [ "Service Contract", "Consulting", "Deliverable Acceptance", "Intellectual Property" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请起草一份企业管理咨询服务合同。甲方委托乙方提供[咨询主题]服务，服务内容包括[服务范围]，交付物为[报告/方案/研讨纪要]，服务期[起止日期]，含税服务费[金额]元。请明确成果验收方式、付款节点、双方配合义务、保密期限、成果知识产权归属、乙方责任上限、违约责任及争议解决。请输出可直接签署的中文 Word 合同。", "prompt_en": "Please draft a management consulting services contract. Party A engages Party B for [咨询主题] services, including [服务范围], with deliverables of [报告/方案/研讨纪要], for [起止日期] and a tax-inclusive service fee of [金额] RMB. Clearly specify the acceptance method, payment milestones, each party's cooperation obligations, confidentiality term, IP ownership, Party B's liability cap, breach liability and dispute resolution. Produce a signable Chinese Word contract.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-11T21:04:09+08:00", "updated_at": "2026-08-12T13:37:31+08:00", "artifacts": [ { "file": "output.html", "label": "V3.0 互动合同起草页面" }, { "file": "企业管理咨询服务合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "silk-road-tang-byzantine-correlation", "title": "唐代金银器与拜占庭工艺关联研究", "title_en": "Tang Gold-Silver Ware vs Byzantine Craftsmanship", "subtitle": "何家村窖藏 vs 东罗马鎏金银盘:锤鍱与鎏金在丝绸之路两端的两种命运", "subtitle_en": "Hejiacun Hoard vs East Roman Gilt-Silver Plate: two destinies of hammering and gilding along the Silk Road", "description": "唐代那么繁盛的金银器跟地中海那边的拜占庭有什么关系?是不是简单照搬?研究丝路金银器又怕断章:粟特、波斯萨珊、拜占庭各有贡献,常被混淆。\n\n这个案例演示了如何借助腾讯探元文物库与知识库的并行检索,主代理亲自做三路 WebSearch 多源核验(对象 A 唐代金银器、对象 B 拜占庭鎏金银盘、丝绸之路双向交流脉络),覆盖工艺技法、造型特征、材质成分、文化内涵与历史交流五个维度,最终交付包含配图与对比雷达图的 Markdown/HTML 双格式对比报告。\n\n适合:丝绸之路与中外文化交流研究者、唐代金银器与博物馆学专业者、拜占庭艺术史研究者、中西物质文明交流史爱好者、文物爱好者与高端文旅讲解员、艺术史与考古专业学生。", "description_en": "How does the glorious Tang gold-silver ware actually relate to Byzantine Mediterranean craft? Was it a simple copy or a complex dialogue? Studying Silk Road metalwork easily fragments sources, conflating Sogdian, Sasanian Persian, and Byzantine contributions into a single 'Western' stream. This case demonstrates how to pair Tencent Tanyuan relic and knowledge bases for parallel retrieval, plus three concurrent main-agent WebSearch passes (Object A Tang metalwork, Object B Byzantine gilt plate, bidirectional Silk Road exchange), covering five dimensions—craft, form, material, cultural meaning, and historical exchange—delivering a cross-checked comparison report with embedded images and a contrast radar chart in Markdown and HTML formats.", "categories": [ "business" ], "tags": [ "唐代金银器", "拜占庭", "东罗马鎏金银盘", "丝绸之路", "何家村窖藏" ], "tags_en": [ "Tang Gold-Silver", "Byzantium", "East Roman Gilt-Silver Plate", "Silk Road", "Hejiacun Hoard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含对比雷达图)" } ], "prompt": "请按「腾讯探元学术研究-关联研究」专家的工作流,关联对比「唐代金银器(以何家村窖藏鎏金舞马衔杯纹银壶为代表)」与「拜占庭鎏金银器(以甘肃省博物馆藏东罗马鎏金银盘为代表)」在工艺与文化层面的关系。\n\n研究范围(覆盖五维):\n- 工艺技法:锤揲(捶揲)+錾刻+鎏金基础工艺组合;萨珊凸纹装饰 vs 拜占庭高浮雕传统;粟特带把杯、波斯萨珊立鸟纹/联珠纹、拜占庭高足杯对唐代影响\n- 造型特征:扁圆皮囊壶/八棱带把杯/高足杯 vs 圆形高浮雕大盘/圈足/卷唇/斜弧壁\n- 材质成分:薄胎银+局部鎏金 vs 高纯度银胎+整体鎏金;盘心高浮雕模压镶嵌\n- 文化内涵:皇室祭祀/舞马祝寿/装饰奢华 vs 高等级收藏/外交礼品/贸易货币\n- 交流脉络:丝绸之路北道要隘(靖远)、萨珊-粟特-拜占庭-唐的多向互动、双向输出(唐→日本正仓院)\n\n交付要求(强制):\n1. 探元两库并行 + 主代理 2-3 次 WebSearch 多源核验\n2. 配图 ≥ 2 张(鎏金舞马衔杯纹银壶 + 东罗马鎏金银盘),公开图源(故宫敦煌特展/陕历博/陕西传媒网);图表 ≥ 1 张(雷达图 SVG)\n3. 报告结构:对比结构(对象 A 概述 → 对象 B 概述 → 逐维度对比 → 丝路双向交流 → 结论 → 证据与局限 → 参考来源 → 图片与图表来源)\n4. 仅写\"对比结论\",不含建议;参考文献只列网络公开来源;主要结论同步到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Correlation Research' expert workflow to comparatively research the relationship between 'Tang gold-silver ware (Hejiacun hoard gilded silver ewer with dancing horse as representative)' and 'Byzantine gilt-silver ware (East Roman gilt-silver plate at Gansu Provincial Museum as representative)' across craft and culture.\n\nResearch scope (five dimensions):\n- Craft: hammering-repoussé + chasing + gilding; Sasanian relief vs Byzantine high-relief traditions; Sogdian handled cups, Sasanian standing-bird/pearl-roundel motifs, Byzantine stemmed cups and their impact on Tang metalwork\n- Form: flat-round leather-bag ewer / eight-lobed handled cup / stemmed cup vs round high-relief plate / ring foot / rolled rim / slanted wall\n- Material: thin silver body + partial gilding vs high-purity silver + overall gilding; center-disk repoussé inlay\n- Cultural meaning: royal ritual / dancing-horse birthday banquet / decorative luxury vs high-grade collection / diplomatic gift / trade currency\n- Exchange: Northern Silk Road corridor (Jingyuan ferry), Sasanian-Sogdian-Byzantine-Tang multi-directional interaction, two-way transmission (Tang→Shōsōin Japan)\n\nDelivery requirements (mandatory):\n1. Parallel Tanyuan dual bases + 2-3 main-agent WebSearch passes for cross-source verification\n2. ≥ 2 images (gilded dancing-horse ewer + Byzantine gilt plate), public sources (Palace Museum Dunhuang Special Exhibition / Shaanxi History Museum / Shaanxi Media); ≥ 1 chart (radar SVG)\n3. Report structure: contrastive (Object A summary → Object B summary → per-dimension → Silk Road exchange → conclusion → limits → references → image/chart sources)\n4. Conclusions only, no recommendations; references limited to public web sources; key conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-correlation", "name": "文博关联研究专家", "description": "基于探元文物库与知识库并行采集,由主代理直接联网多源交叉核验,交付四格式跨对象/跨域对比研究报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "75f1347b2045", "expert_id": "ex_pEyI3eIkynB5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:11:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sku-inventory-turnover-diagnosis", "title": "SKU库存周转诊断", "title_en": "SKU Inventory Turnover Diagnosis", "subtitle": "SKU 周转天数、可售天数与库存风险诊断", "subtitle_en": "A practical Weimob playbook for sku inventory turnover diagnosis", "description": "做SKU库存周转诊断时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“SKU库存周转诊断”展开：结合各 SKU 近 30 天销量和当前库存，计算周转天数及预计可售天数，同时识别积压和潜在断货风险。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、库存分析运营人员、经营分析团队。", "description_en": "The hardest part of sku inventory turnover diagnosis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates sku inventory turnover diagnosis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Inventory Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "库存周转", "可售天数", "库存积压", "断货风险", "SKU分析" ], "tags_en": [ "SKU Inventory Turnover Diagnosis", "Mall Operations", "Inventory Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我结合各SKU近30天销量和当前库存，计算库存周转天数及预计可售天数，识别周转过慢或可能断货的商品，按风险排序并给出库存优化建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a sku inventory turnover diagnosis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sku-stockout-risk-check", "title": "缺货风险SKU排查", "title_en": "SKU Stockout Risk Check", "subtitle": "零低库存 SKU 缺货风险与可售天数评估", "subtitle_en": "A practical Weimob playbook for sku stockout risk check", "description": "做缺货风险SKU排查时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“缺货风险SKU排查”展开：找出当前零库存和低库存 SKU，结合近 30 天销量评估缺货风险、预计可售时间和紧急处理顺序。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、库存分析运营人员、经营分析团队。", "description_en": "The hardest part of sku stockout risk check is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates sku stockout risk check through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Inventory Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "零库存", "低库存", "缺货风险", "可售天数", "库存预警" ], "tags_en": [ "SKU Stockout Risk Check", "Mall Operations", "Inventory Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我列出当前零库存和低库存的SKU，结合近30天销量按缺货风险排序，估算剩余可售天数，并给出紧急处理建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a sku stockout risk check covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "slow-moving-inventory-clearance-plan", "title": "滞销库存梯度清仓", "title_en": "Tiered Slow-Moving Inventory Clearance Plan", "subtitle": "未动销库存分阶段清仓与渠道分发方案", "subtitle_en": "A practical Weimob playbook for tiered slow-moving inventory clearance plan", "description": "做滞销库存梯度清仓时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“滞销库存梯度清仓”展开：针对连续 90 天未动销库存设计分阶段降价、搭配销售和渠道分发方案，降低库存占用。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、促销活动运营人员、经营分析团队。", "description_en": "The hardest part of tiered slow-moving inventory clearance plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates tiered slow-moving inventory clearance plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Promotion Operations specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "梯度清仓", "滞销库存", "分阶段降价", "搭配销售", "渠道分发" ], "tags_en": [ "Tiered Slow-Moving Inventory Clearance Plan", "Mall Operations", "Promotion Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "针对连续90天未动销的库存，帮我制定梯度清仓方案，包括分阶段降价节奏、搭配销售策略和渠道分发建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a tiered slow-moving inventory clearance plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "smart-page-sync-weekly", "title": "AI Page 项目双周报 · PREP", "title_en": "AI Page Bi-weekly Report · PREP", "subtitle": "用 PREP 观点回环把双周报做成 5 屏给完结论的可交互 HTML", "subtitle_en": "Turn a bi-weekly report into a 5-screen interactive HTML using PREP", "description": "写周报最难的不是写——是「怎么让 Leader 30 秒就读完」。\n\n这个案例基于腾讯文档「智能页面」（Smart Page）技能，用 PREP 观点回环（Point-Reason-Example-Point）+ 腾讯蔚蓝皮肤，把一份 AI Page 项目双周报做成单文件可交互 HTML。首屏左右分栏：左 P 大字结论（本期评价 + 下期重点），右 R 三条原因卡片。中段 4 个关键数字滚动到视口自动从 0 累加，下面接 5 条核心成果时间线 + 3 条外部声音引述。再往下是下一步行动卡片，owner / ETA 一目了然。最后红/黄卡片明确标出 3 条风险（含 mitigation） + 3 条求助清单（含 urgency）。整体不超过 5 屏，PREP 的灵魂就是短。\n\n适合：\n- 每周/双周需要写周报、月报的产品经理 / 项目经理 / Tech Lead\n- 跨部门协作汇报、轻量同步周会、月度复盘\n- 受够 PPT 翻 5 页才看到结论、想要首屏给结论的同学\n- 希望让汇报有可读性、有数据、有求助、有跟进项的人", "description_en": "The hardest part of writing a weekly report is not writing — it is getting your leader to read it in 30 seconds.\n\nThis case is built on Tencent Docs Smart Page. Using the PREP loop (Point-Reason-Example-Point) + Tencent Blue skin, an AI Page bi-weekly report is assembled into a single-file interactive HTML. The first fold splits left/right: a big-typography Point on the left (this period's verdict + next focus), three Reason cards on the right. Mid-section: 4 key metrics that count up from 0 on scroll, followed by a 5-item achievement timeline and 3 external voice quotes. Then a Next-Steps card grid showing owner & ETA. Finally red/amber cards explicitly marking 3 risks (with mitigations) and 3 asks (with urgency). The whole report stays under 5 screens — brevity is the soul of PREP.\n\nBest for:\n- PMs, project managers, and tech leads writing weekly / bi-weekly / monthly reports\n- Cross-team sync, lightweight weekly stand-ups, monthly retrospectives\n- People sick of flipping 5 PPT pages before seeing the conclusion\n- Anyone who wants their reports to be readable, data-backed, with explicit asks and follow-ups", "categories": [ "office" ], "tags": [ "双周报", "PREP", "可交互 HTML", "腾讯蔚蓝", "项目同步" ], "tags_en": [ "Bi-weekly Report", "PREP", "Interactive HTML", "Tencent Blue", "Project Sync" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI Page 项目双周报（可交互 HTML）" } ], "prompt": "请使用「腾讯文档智能页面 · Smart Page」技能，帮我生成一份项目双周报的可交互 HTML 汇报网页。\n\n参数：\n- 场景：sync（周期同步 / 双周报）\n- 叙事范式：prep（PREP 观点回环：Point-Reason-Example-Point）\n- 皮肤：tencent-blue（腾讯蔚蓝，司内汇报默认）\n- 受众：直属 Leader + 协作团队\n\n内容框架（5 屏 PREP 回环，每屏自成结论）：\n1. P · 一句话结论：本期整体评价 + 下期重点\n2. R · 1-3 条原因\n3. E · 例证：4 个关键数字 + 4-5 条核心成果（含 owner 与状态）+ 1-3 条客户/协作方原话\n4. 下一步：3 条行动，含 owner 与 ETA\n5. P · 重申结论 + 风险与求助清单\n\n第一步：（提供本期项目进展、关键数据与成果）。\n...\n\n输出要求：\n- 单文件 HTML，CSS / JS / 字体 / 图片全部内联\n- 首屏即给结论，不折叠隐藏\n- 关键数字恰好 4 个，整体不超过 5 屏\n- 具备滚动揭示、数字 CountUp、Sticky 章节编号、进度条，并支持 prefers-reduced-motion 降级\n- 最后发布到腾讯文档，返回可在线访问的链接", "prompt_en": "Please use the \"Tencent Docs Smart Page\" skill to generate an interactive HTML report page for a project bi-weekly report.\n\nParameters:\n- Scenario: sync (periodic sync / bi-weekly report)\n- Narrative: prep (the PREP loop: Point-Reason-Example-Point)\n- Skin: tencent-blue (Tencent azure, default for internal reports)\n- Audience: direct leader + collaborating teams\n\nContent framework (5 screens of PREP loop, each self-contained with a conclusion):\n1. P · One-sentence conclusion: this period's overall assessment + next focus\n2. R · 1-3 reasons\n3. E · Evidence: 4 key numbers + 4-5 core achievements (with owner and status) + 1-3 verbatim quotes from customers/collaborators\n4. Next steps: 3 actions with owner and ETA\n5. P · Restate conclusion + risks and asks\n\nStep 1: (provide this period's project progress, key metrics, and achievements).\n...\n\nOutput requirements:\n- Single-file HTML, with all CSS / JS / fonts / images inlined\n- The conclusion must appear on the first screen, not hidden behind a collapse\n- Exactly 4 key numbers, no more than 5 screens overall\n- Include scroll reveal, number CountUp, sticky section numbering, a progress bar, and a prefers-reduced-motion fallback\n- Finally publish to Tencent Docs and return an online-accessible link", "skills": [ { "id": "smart-page", "name": "腾讯文档智能页面", "description": "下一代可交互在线汇报网页生成工具，覆盖汇报 / 述职 / 周报 / 分享 / 项目复盘 / 数据看板等场景，生成即获永久在线链接", "skill_id": "skill_2053082927661453312" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-05-07T23:15:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "sorting-algorithms-visualization-comparison", "title": "冒泡/快速/归并排序可视化对比", "title_en": "冒泡/快速/归并排序可视化对比", "subtitle": "计算机科学 · 冒泡/快速/归并排序可视化对比", "subtitle_en": "计算机科学 · 冒泡/快速/归并排序可视化对比", "description": "排序算法可视化课件，用单步动画演示冒泡、快排、归并，并提供同数据同台对决与复杂度对比。", "description_en": "排序算法可视化课件，用单步动画演示冒泡、快排、归并，并提供同数据同台对决与复杂度对比。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "sorting-algorithms-visualization-comparison.html", "artifacts": [ { "file": "sorting-algorithms-visualization-comparison.html", "label": "冒泡/快速/归并排序可视化对比｜互动教学网页" } ], "prompt": "帮我做一个常见排序算法可视化对比的互动教学网页，要能单步演示冒泡、快排和归并的执行过程。", "prompt_en": "帮我做一个常见排序算法可视化对比的互动教学网页，要能单步演示冒泡、快排和归并的执行过程。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sourcing-comparison-report", "title": "采购找货，一份比价报告帮你选对供应商", "title_en": "Sourcing Comparison Report", "subtitle": "多方比价、核到货价、查风险，附采购建议", "subtitle_en": "Multi-source price comparison, landed-cost check and supplier risk, with a buying recommendation", "description": "输入采购品类和数量，自动多方比价、核到货价、查供应商风险，输出带采购建议的寻源比价报告。", "description_en": "Enter the category and quantity to source: it compares quotes from multiple suppliers, verifies landed costs and checks supplier risk, then outputs a sourcing comparison report with a purchasing recommendation.", "categories": [ "data" ], "tags": [ "采购寻源", "比价", "供应商风控", "贸易流通" ], "tags_en": [ "Sourcing", "Price Comparison", "Supplier Risk", "Trade & Distribution" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "寻源比价报告（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易采购专家。请为我做一次采购寻源比价，并输出一份可决策的比价报告。\n\n【采购需求】\n- 品类/规格：{如 螺纹钢 HRB400E Φ20}\n- 数量：{如 200吨}\n- 交货地：{如 南京市江宁区}\n- 期望到货时间：{如 3~5天内}\n\n【报告要求】\n1. 调用【找钢行情MCP】取该品类在交货地的真实成交价区间作为市场参考；\n2. 列出至少4~5家供应商比价：含税单价、运费、到货价、账期、交货周期；\n3. 对每家做风控标注——命中【找钢黑名单】的标为高风险；\n4. 给出采购建议：首选/备选各1家并说明理由（不能只看最低价，要综合到货价、账期、交货、风控）；\n5. 查询【行业信息资产库】里这个品类的议价要点，给出可争取的议价空间；\n6. 输出为一页结构清晰的比价报告，含比价表 + 采购建议。\n\n若行情或风控数据源未就绪，请以通用采购方法论为主产出并说明。请生成完整报告。", "skills": [ { "id": "network-wide-sourcing-compare", "name": "全网找货比价", "description": "TradeBuddy把 B2B 贸易「一个料号打一圈全网、聚成一张比价清单」的行业习惯，封装成一句话就能出结果的能力。", "skill_id": "skill_2091152134004711424" }, { "id": "supplier-background-check", "name": "供应商背调", "description": "TradeBuddy把“合作前先把对方查一圈”的风控习惯，做成一句话出结果的供应商背景调查与风险核查——工商、涉诉、舆情、黑名单一并核验，给倾向性判断而非绝对结论。", "skill_id": "skill_2091152618275139584" } ], "experts": [ { "id": "caigou-expert", "name": "B2B贸易采购专家", "description": "懂B2B贸易的采购专家：解析需求、看真实行情、全网找货比价、供应商背调、议价话术，帮你把好采购关。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6kRSFmkAwD8M" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "steel-trade-knowledge", "title": "新人入行必看：钢铁贸易知识速览", "title_en": "Steel Trade Knowledge", "subtitle": "钢铁建材贸易速答、术语表与避坑提醒", "subtitle_en": "Quick answers, glossary and pitfall alerts for steel & building-material trade", "description": "输入你想了解的主题，自动生成钢铁/建材贸易的常见问题速答、术语表和避坑提醒，新人快速上手。", "description_en": "Enter the topic you want to learn: it generates quick FAQ answers, a glossary and pitfall reminders for steel and building-material trade, helping newcomers get up to speed fast.", "categories": [ "creation" ], "tags": [ "行业知识", "新人培训", "术语", "贸易流通" ], "tags_en": [ "Industry Knowledge", "Onboarding", "Glossary", "Trade & Distribution" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "钢铁贸易知识速览（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易运营专家。请帮我整理一份行业知识速览，用于{用途，如 新业务员入行培训}。\n\n【主题】\n{如 钢铁贸易新人必懂；或指定某方向：行情怎么看/术语/账期风险等}\n\n【要求】\n1. 常见问题速答：挑3~5个新人最常问的问题（如牌号含义、报价口径区别、怎么看行情），用大白话解答；\n2. 必备术语表：列出该行业高频术语（如材质单、质保书、背靠背、承兑、锁价等），查询【行业信息资产库】里这个行业的术语口径，可对任一术语展开解释；\n3. 避坑提醒：列出新人易踩的坑（报价口径、材质质保、赊销资信、锁价风险等）。\n\n内容要准确、接地气、符合中国钢铁/建材贸易实际。请生成完整知识速览。", "skills": [ { "id": "knowledge-butler", "name": "知识管家", "description": "TradeBuddy把企业散落的制度、规则、手册、话术、问答，做成“一句话就能问、答案有出处”的内部知识库助手。答案一律基于检索到的企业资料，检索不到就说没有，绝不用通用常识冒充企业规定。", "skill_id": "skill_2090731525359386624" } ], "experts": [ { "id": "yunying-expert", "name": "B2B贸易运营专家", "description": "懂B2B贸易的运营专家：面向老板与管理员，做生意体检、看团队人效、批量扫风险、沉淀企业知识。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_D7992fbSsXNx" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "stereo-seq-expert", "title": "Stereo-seq 空间转录组分析手册", "title_en": "Stereo-seq Spatial Transcriptomics Handbook", "subtitle": "SAW 上游投递到 3D 重建，7 大场景全流程代码", "subtitle_en": "From SAW upstream submission to 3D reconstruction — full code for 7 scenarios", "description": "做空间转录组分析最头疼的是流程割裂：上游投递在 DCS 平台、下游分析换 Stereopy、细胞分割又得学 Spateo，网上没有一份把 7 个场景串起来的中文手册。这个案例生成一份覆盖 SAW-ST-V8 上游主分析、Stereopy 标准下游、多样本整合（Harmony）、Spateo 细胞分割、连续切片 3D 重建、CCI 细胞通讯、SCC 结构数字化 7 大场景的完整手册：每个场景都有目标说明、可运行代码、产出文件清单、结果图和解读要点，11 张图全部内嵌，单文件自包含可直接当团队培训材料。适合：Stereo-seq 平台用户、空间组学方向研究生、搭建空间转录组 SOP 的团队、DCS 云平台任务投递者。", "description_en": "Spatial transcriptomics workflows are painfully fragmented: upstream submission on the DCS platform, downstream analysis in Stereopy, cell segmentation in yet another tool (Spateo) — and no single Chinese handbook ties the seven scenarios together. This case generates a complete handbook covering SAW-ST-V8 upstream analysis, standard Stereopy downstream, multi-sample integration (Harmony), Spateo cell segmentation, serial-section 3D reconstruction, CCI cell communication, and SCC structural digitization. Every scenario includes goal, runnable code, output file list, a result figure, and interpretation points — 11 figures all embedded in one self-contained file ready for team training. Suitable for: Stereo-seq platform users, spatial-omics graduate students, teams building spatial transcriptomics SOPs, and DCS cloud task submitters.", "categories": [ "data", "education" ], "tags": [ "Stereo-seq", "空间转录组", "Stereopy", "细胞分割", "细胞通讯" ], "tags_en": [ "Stereo-seq", "Spatial Transcriptomics", "Stereopy", "Cell Segmentation", "Cell Communication" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Stereo-seq 空间转录组 7 场景分析手册（自包含 HTML）" } ], "prompt": "你是 Stereo-seq 空间转录组分析专家。请生成一份以分析结果为核心的自包含 Playbook 手册：\n\n要求：\n1. 覆盖 7 个场景：SAW-ST-V8 上游主分析、Stereopy 标准下游（GEF → 聚类注释）、多样本整合（MSData + Harmony）、Spateo 细胞分割、连续切片 3D 重建、CCI 细胞通讯、SCC 结构数字化。\n2. 每个场景必须包含：分析目标、完整可运行的 Python/dcs-cli 代码（带注释）、产出文件清单（含文件名和用途）、代表性结果图、关键解读要点。\n3. 用 matplotlib 预先生成 11 张代表性结果图（QC 小提琴、空间聚类散点、UMAP、marker 热图、批次混色对比、时间序列空间演变、细胞分割三联图、3D 重建渲染、CCI 网络圆圈图、SCC 层梯度、SAW 流水线流程图）。\n4. 将所有图片以 base64 data URI 内联嵌入，产出单文件自包含报告（无外链）。\n5. 末尾附避坑清单和图表文件索引。\n\n约束：\n- 输出单文件 HTML（output.html）：所有 CSS 内联、图片 base64 内嵌、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航；所有代码基于此前 9 个 skill 实测验证的真实 API 编写，不要虚构函数签名。\n- 中文版，代码标识符保持英文生态惯例。\n- 文档可直接作为团队培训手册使用。", "prompt_en": "You are a Stereo-seq spatial transcriptomics analysis expert. Generate a result-oriented, self-contained playbook handbook:\n\nRequirements:\n1. Cover 7 scenarios: SAW-ST-V8 upstream analysis, standard Stereopy downstream (GEF → clustering and annotation), multi-sample integration (MSData + Harmony), Spateo cell segmentation, serial-section 3D reconstruction, CCI cell communication, and SCC structural digitization.\n2. Each scenario must include: analysis goal, complete runnable Python/dcs-cli code (commented), output file list (filenames and purposes), a representative result figure, and key interpretation points.\n3. Pre-generate 11 representative result figures with matplotlib (QC violin, spatial clustering scatter, UMAP, marker heatmap, batch mixing comparison, time-series spatial evolution, cell segmentation triplet, 3D reconstruction render, CCI network circle plot, SCC layer gradient, SAW pipeline diagram).\n4. Inline all figures as base64 data URIs and produce a single-file self-contained report (no external links).\n5. Append a pitfall checklist and a figure file index at the end.\n\nConstraints:\n- Output a single-file HTML (output.html): all CSS inlined, figures base64-embedded, no external resource references, content visible by default, no hash-anchor navigation; all code must be based on real APIs verified by the previous 9 skills — do not fabricate function signatures.\n- Chinese language; code identifiers follow English ecosystem conventions.\n- The document must be usable directly as team training material.", "skills": [], "experts": [ { "id": "stereo-seq-expert", "name": "Stereo-seq空间转录组分析专家", "description": "Stereo-seq 空间转录组全流程分析（SAW 投递 / Stereopy 下游 / Spateo 分割·对齐·3D），内嵌 saw-st-v8 / dcs-cli / saw-st-v8-makeref / saw-st-v8-realign / saw-st-v8-reanalyze-lasso / saw-st-v8-gef2gem / stereopy-docs / spateo / stereo-seq-technology-principles 共 9 个 skill", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_b2vdl3DaOp7A" } ], "mcps": [ { "id": "dcs-cloud", "name": "DCS Cloud", "description": "华大 DCS 云平台生物信息任务投递与调度" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-27T22:03:55+08:00", "updated_at": "2026-08-30T11:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "stock-deep-dive-card", "title": "个股基本面深挖卡片", "title_en": "Stock Deep-Dive Card", "subtitle": "以贵州茅台 600519.SH 为例", "subtitle_en": "Example: Kweichow Moutai 600519.SH", "description": "想看懂一只股票，往往要在行情软件、财报、研报、资金流之间反复切换，数据散、口径乱，还容易凭记忆记错数字。\n\n本案例用 NeoData 金融搜索服务，对贵州茅台(600519.SH)真实查询实时行情、估值分位、三年财报、今日资金流向、机构评级与近期事件，一次性汇成一张信息密度高、排版专业的单文件 HTML「个股深挖卡片」：结论先行，行情/估值/基本面/资金/机构/事件分区呈现，红涨绿跌，内置中英双语，资源全内联可离线打开。\n\n适合：想快速吃透一只股票的个人投资者；需要产出个股速览卡片的研究员/分析师；做投研自动化的开发者；关注 A 股龙头的财经内容创作者；想学习金融数据查询与可视化的学习者。", "description_en": "Understanding a single stock usually means jumping between the quote app, filings, research notes and money flow — data is scattered, definitions vary, and it's easy to misremember numbers.\n\nThis case uses NeoData Financial Search to run real queries on Kweichow Moutai (600519.SH) — live quote, valuation percentiles, three-year financials, today's capital flow, analyst ratings and recent events — and compiles them into one high-density, professionally laid-out single-file HTML deep-dive card: verdict first, with quote/valuation/fundamentals/flow/analysts/events sections, red-up-green-down, built-in bilingual text, all resources inlined for offline use.\n\nGood for: retail investors who want to grasp a stock fast; researchers/analysts producing single-stock snapshot cards; developers automating research workflows; finance content creators covering A-share leaders; learners studying financial data querying and visualization.", "categories": [ "finance", "data", "search" ], "tags": [ "个股分析", "金融数据", "基本面", "可视化卡片", "贵州茅台" ], "tags_en": [ "Stock Analysis", "Financial Data", "Fundamentals", "Visual Card", "Kweichow Moutai" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "prompt": "用 NeoData 金融搜索服务，帮我为贵州茅台(600519.SH)做一张个股基本面深挖卡片，产物是一个单文件 HTML。\n\n请务必先通过 NeoData 官方脚本真实查询以下数据，全部取最新值，不要凭记忆或编造，也不要混用其他数据源：\n1. 实时行情：最新价、涨跌幅、开盘/最高/最低、成交量与成交额、换手率、总市值；\n2. 估值：PE、PB、股息率，以及历史分位；\n3. 最新财报关键指标：营业收入、归母净利润及同比增速、毛利率、净利率、ROE、每股收益；\n4. 今日主力资金流向与净流入；\n5. 近期券商评级与盈利预测（目标价、买入/增持占比、未来几年 EPS 与增速预测）；\n6. 近期重大事件：分红送配、股份回购、提价等。\n\n把每条查询回包里的关键数字与结论提炼出来，标注数据时间，然后汇总成一张信息密度高、排版专业的深挖卡片。\n\n结构建议：结论先行 → 行情快照 → 估值水位 → 基本面财报表 → 资金流向 → 机构观点/评级 → 事件追踪 → 复现查询清单 → 口径与风险提示。\n\n样式方向：浅色专业金融风，红涨绿跌（中国习惯），信息密度高、排版整洁；做成单文件 HTML，所有资源内联、可离线打开。", "prompt_en": "Use NeoData Financial Search to build a fundamental deep-dive card for Kweichow Moutai (600519.SH), delivered as a single-file HTML.\n\nFirst run real queries via the official NeoData script for all of the following, always taking the latest values — do not rely on memory or fabricate, and do not mix in other data sources:\n1. Live quote: last price, change %, open/high/low, volume and turnover, turnover rate, market cap;\n2. Valuation: PE, PB, dividend yield, plus historical percentiles;\n3. Latest financials: revenue, net profit attributable and YoY growth, gross margin, net margin, ROE, EPS;\n4. Today's main-force capital flow and net inflow;\n5. Recent broker ratings and forecasts (target price, buy/overweight share, EPS and growth forecasts for coming years);\n6. Recent major events: dividends, buybacks, price hikes, etc.\n\nExtract the key figures and conclusions from each query response, mark the data timestamps, then compile them into a high-density, professionally laid-out deep-dive card.\n\nSuggested structure: verdict first → quote snapshot → valuation level → fundamentals table → capital flow → analyst views/ratings → event tracker → reproduction query list → scope & risk notes.\n\nStyle direction: light professional finance look, red-up-green-down (China convention), high information density and clean layout; make it a single-file HTML with all resources inlined and openable offline.", "skills": [ { "id": "neodata-financial-search", "skill_id": "skill_2053082432761950208", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据" } ], "experts": [], "mcps": [], "artifacts": [ { "file": "output.html", "label": "贵州茅台个股深挖卡片" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-22T03:01:47Z", "updated_at": "2026-07-22T03:58:02Z", "is_recommended": false, "recommended_order": 0 }, { "id": "stock-news-flash-brief", "title": "澜起科技动态追踪简报", "title_en": "Montage Tech News Tracking Brief", "subtitle": "多角度检索热门个股真实资讯，时间线简报即查即得", "subtitle_en": "Multi-angle retrieval of real news on a hot stock, delivered as a timeline brief", "description": "想追一只热门股的消息，最烦的就是东一条西一条，看完这条忘了那条。这个案例演示了如何让 AI 从多个角度检索澜起科技的真实资讯，把同一事件的多方报道归纳在一起，整理成一份按时间线排列的深色快讯风简报，每条都标注来源和日期。\n\n适合：盯盘的个人投资者、需要快速了解个股动态的研究助理、做晨会准备的投研同学。", "description_en": "Tracking news on a hot stock is frustrating—bits and pieces scattered everywhere, forgotten as soon as you read them. This case shows how AI retrieves real news about Montage Technology from multiple angles, groups reports of the same event together, and organizes everything into a dark, flash-style timeline brief with source and date labels on every item.\n\nBest for: individual investors watching the market, research assistants who need quick stock updates, and investment researchers preparing for morning meetings.", "categories": [ "finance", "search" ], "tags": [ "个股追踪", "资讯简报", "时间线", "快讯" ], "tags_en": [ "Stock Tracking", "News Brief", "Timeline", "News Flash" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "澜起科技动态追踪简报" } ], "prompt": "帮我追踪澜起科技（688008）的最新动态。请用平安证券资讯查询 skill 从多个角度真实检索（如「澜起科技 最新消息」「澜起科技 业绩」「澜起科技 股价异动 原因」等），把召回资讯按日期由新到旧整理，同一事件的多条报道归纳在一起，注明每条的来源与日期。最终产出一份单文件 HTML 个股资讯简报：顶部为核心动态概览，主体按时间线排列资讯条目，配事件归纳板块，页脚注明数据来源（平安证券资讯知识库）与「本内容仅为资讯整理，不构成投资建议」。视觉方向：深色金融快讯质感、时间线排版、信息密度高、适合快速扫读；所有资源内联、可离线打开。", "prompt_en": "Help me track the latest developments of Montage Technology (688008). Use the Ping An News Search skill to run real queries from multiple angles (e.g. \"Montage Tech latest news\", \"Montage Tech earnings\", \"Montage Tech price movement reasons\"), organize the retrieved news by date (newest first), group multiple reports of the same event together, and label each item with its source and date. Produce a single-file HTML stock news brief: a highlights overview at the top, a timeline of news items as the main body, an event-grouping section, and a footer stating the data source (Ping An Securities news knowledge base) and \"This content is for informational purposes only and does not constitute investment advice.\" Visual direction: dark financial news-flash aesthetic, timeline layout, high information density, easy to scan; all resources inlined so the file works offline.", "skills": [ { "id": "news-search", "name": "平安证券资讯查询", "description": "按股票、公司、行业、ETF、概念或市场主题检索相关新闻、快讯和财经报道。", "skill_id": "skill_2079451206686916608" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-24T17:46:45+08:00", "updated_at": "2026-07-27T12:49:23+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "stock-roundtable-ai-vs-energy", "title": "AI vs 新能源：今年哪个方向更好？", "title_en": "AI vs. New Energy: Which Sector to Back This Year?", "subtitle": "趋势派、估值派、信号派三方实时数据圆桌，给你一个真正有分歧的答案", "subtitle_en": "Trend, valuation, and signal traders debate live with real-time data — genuine disagreement included", "description": "每次做方向选择都想找人问问，但身边的人要么不懂要么只会说「分散投资」。这个案例演示了炒股搭子圆桌专家如何处理「AI vs 新能源」这道年度选择题：先查实时行情（AI板块 PE 66.7倍/3年分位89.8%，新能源 PE 48.8倍），再从产业趋势、估值定价、资金信号三个维度展开真实分歧——星仔认为 AI 景气度碾压是今年第一顺位，文仔算出储能 PEG 0.8 比 AI 性价比更高，洲仔则提示4月22日地缘窗口前任何方向都要打折扣。最终不给单一结论，而是三条操作路线图供你按自己偏好选择。\n\n适合：在 AI 和新能源之间纠结选择的投资者、想听到真实分歧而非统一口径的股民、需要量化依据支撑方向判断的散户。", "description_en": "When it's time to pick a sector, you want real debate — not vague \"diversify\" advice. This case shows how the Stock Roundtable expert handles the classic AI vs. New Energy dilemma: starting with live data (AI sector PE 66.7x at 89.8% 3-year percentile; New Energy PE 48.8x), then surfacing genuine disagreements across three lenses. Xingzai sees AI momentum as the year's clear leader; Wenzai calculates that energy storage's PEG of 0.8 beats AI on value; Zhouzai flags the April 22 geopolitical window as the biggest variable overriding any sector call. The output isn't a single verdict — it's three decision paths mapped to your personal priorities.\n\nIdeal for: investors torn between AI and New Energy, stock pickers who want genuine disagreement instead of consensus takes, and retail traders who need quantitative anchors before committing to a direction.", "categories": [ "agent" ], "tags": [ "A股方向选择", "AI板块", "新能源", "多空分歧", "实时数据" ], "tags_en": [ "A-share sector picks", "AI sector", "new energy", "bull-bear debate", "live data" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI vs 新能源圆桌讨论分析报告" } ], "prompt": "我在 AI 和新能源之间做方向选择，请帮我组织一场圆桌讨论。\n\n背景：投资期限约 6-12 个月；关注 A 股 AI 算力产业链 vs 新能源；纠结哪个方向风险回报比更好。\n\n请按炒股搭子圆桌流程：\n1. 先查今日实时行情（大盘 + 两个板块的 PE 与涨幅）\n2. 查宏观背景，判断有无系统性风险\n3. 安排星仔（趋势派）、文仔（估值派）、洲仔（信号派）并行分析后汇总\n4. 明确点出成员间的核心分歧\n5. 给出三条差异化操作路线图：激进看多 AI / 保守配置新能源 / 观望等待信号\n\n要求：每个数字都要有实时数据出处，不要「综合来看」式的空泛结论；三条路线之间差异要足够大，让我感受到真实分歧。", "prompt_en": "I'm choosing a direction between AI and new energy — please help me organize a roundtable discussion.\n\nBackground: investment horizon about 6-12 months; sectors of interest are the A-share AI computing-power supply chain vs. new energy; I'm unsure which offers the better risk/reward.\n\nPlease follow the stock-buddy roundtable process:\n1. First look up today's live market data (the broad index + PE and change for both sectors)\n2. Check the macro backdrop and judge whether there's any systemic risk\n3. Bring in Xingzai (trend school), Wenzai (valuation school), and Zhouzai (signal school) to analyze in parallel, then summarize\n4. Clearly point out the core disagreements among members\n5. Provide three differentiated action roadmaps: aggressively bullish on AI / conservative new-energy focus / wait-and-watch for signals\n\nRequirements: every number must cite a live data source; no hand-wavy \"all things considered\" conclusions; the three roadmaps must differ enough that I can feel the real disagreement.", "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "全球多市场金融数据搜索服务，自然语言查询股票、基金、宏观、外汇、大宗商品等全品类金融数据", "description_en": "Natural language financial data search across global markets: stocks, funds, macro, forex & commodities", "skill_id": "skill_2053082432761950208" } ], "experts": [ { "id": "StockPartnerTeam", "name": "腾讯自选股金融分析专家团", "description": "蒸馏了六位真实的炒股大神的实战经验，涵盖产业策略、信号捕捉、估值评估、抄底布局、基本面研究、短线交易六大视角，基于真实数据各抒己见，提供多角度分析参考。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "3acdcc5c61b9", "name_en": "Tencent WeStock Financial Analysis Team", "description_en": "Distilled real-world experience of six veteran stock investors, covering industry strategy, signal hunting, valuation analysis, contrarian entry, fundamentals research, and short-term trading — multi-angle analysis references grounded in real-time data.", "expert_id": "ex_cRUYbcfeBpop" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-16T17:30:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "stock-screening-tencent-buy", "title": "PE<20+ROE>15%选股 & 腾讯能买吗", "title_en": "PE<20 + ROE>15% Stock Screening & Should You Buy Tencent", "subtitle": "基本面、估值、逆向三派同场，从筛选到建仓一条龙", "subtitle_en": "Fundamentals, valuation, and contrarian experts debate — from screening to position building", "description": "想用PE和ROE筛好公司，但筛出来不知道谁更值得买？腾讯PE 18倍看着不贵，但又怕高位站岗。这个案例演示了炒股搭子圆桌如何处理这类先筛再评的问题：先按PE<20+ROE>15%+市值>500亿筛出美的、福耀、格力等标的，再让钊仔逐个验证基本面（格力PE 6.4是低估还是陷阱？），文仔横向对比PEG（福耀0.65 vs 腾讯1.1谁更值？），坤仔给腾讯的支撑位分批建仓计划（514→476→414→359，每个价位都有K线依据）。最终三条路线图：重仓腾讯的/选性价比的/等趋势确认的。\n\n适合：想用财务指标筛股但不会进一步评判的投资者、纠结腾讯该不该买的港股股民、需要分批建仓具体计划的人。", "description_en": "You can screen stocks with PE and ROE, but then what? Tencent looks cheap at 18x PE, but is it a trap? This case shows how the Stock Roundtable handles the classic screen-then-evaluate workflow: first filter for PE<20 + ROE>15% + market cap>CNY 50B to find Midea, Fuyao, Gree, etc., then have Zhaozai verify fundamentals (Gree at 6.4x PE — undervalued or value trap?), Wenzai compare PEG across picks (Fuyao 0.65 vs Tencent 1.1), and Kunzai lay out a support-driven scaling plan for Tencent (514→476→414→359 HKD, each level anchored to chart support). Three decision paths: go heavy on Tencent / pick best value / wait for trend confirmation.\n\nIdeal for: investors who screen by financials but can't judge what to buy, Hong Kong stock pickers torn on Tencent, and anyone needing a concrete scaling plan with price levels.", "categories": [ "agent" ], "tags": [ "选股筛选", "PE ROE", "腾讯", "分批建仓", "性价比对比" ], "tags_en": [ "stock screening", "PE ROE", "Tencent", "scaling in", "value comparison" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "output.html", "label": "PE<20+ROE>15%选股与腾讯分析" } ], "preview": "output.html", "prompt": "帮我筛选 PE 低于 20、ROE 大于 15%、市值超 500 亿的股票（A股+港股），然后重点分析腾讯现在能不能买。\n\n腾讯关注点：当前约 18 倍 PE、22% ROE，想知道是否偏贵、与筛出的其他公司性价比如何、若买怎么分批建仓。\n\n请安排三位成员并行分析：\n1. 钊仔（基本面派）：验证每家公司基本面，重点拆解腾讯财报、业务与管理层\n2. 文仔（估值派）：横向对比 PEG / PE 分位，给出便宜到贵的排序与核心+卫星仓位建议\n3. 坤仔（逆向派）：给腾讯一份支撑位驱动的分批建仓计划\n\n要求：每个数字都要有实时数据出处，结论具体可执行。", "prompt_en": "Help me screen for stocks with PE below 20, ROE above 15%, and market cap over CNY 50 billion (A-shares + Hong Kong stocks), then focus on whether Tencent can be bought now.\n\nTencent focus: currently about 18x PE and 22% ROE; I want to know whether it's expensive, how its value-for-money compares to the other screened companies, and how to scale into a position if buying.\n\nPlease arrange three members to analyze in parallel:\n1. Zhaozai (fundamentals school): verify each company's fundamentals, with a deep breakdown of Tencent's financials, business, and management\n2. Wenzai (valuation school): compare PEG / PE percentiles, give a cheapest-to-most-expensive ranking, and recommend core + satellite position sizing\n3. Kunzai (contrarian school): give Tencent a support-level-driven scaling-in plan\n\nRequirement: every number must have a real-time data source, and conclusions must be specific and actionable.", "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "全球多市场金融数据搜索服务，自然语言查询股票、基金、宏观、外汇、大宗商品等全品类金融数据", "description_en": "Natural language financial data search across global markets: stocks, funds, macro, forex & commodities", "skill_id": "skill_2053082432761950208" } ], "experts": [ { "id": "StockPartnerTeam", "name": "腾讯自选股金融分析专家团", "description": "蒸馏了六位真实的炒股大神的实战经验，涵盖产业策略、信号捕捉、估值评估、抄底布局、基本面研究、短线交易六大视角，基于真实数据各抒己见，提供多角度分析参考。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "3acdcc5c61b9", "name_en": "Tencent WeStock Financial Analysis Team", "description_en": "Distilled real-world experience of six veteran stock investors, covering industry strategy, signal hunting, valuation analysis, contrarian entry, fundamentals research, and short-term trading — multi-angle analysis references grounded in real-time data.", "expert_id": "ex_cRUYbcfeBpop" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 85, "source": "official", "usage_count": 0, "created_at": "2026-04-16T18:30:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "stock-up-advice", "title": "该不该囤货？一份备货建议帮你决策", "title_en": "Stock-Up Advice", "subtitle": "行情×库存销量合并测算，该不该囤一目了然", "subtitle_en": "Live market quotes meet your inventory and sales data — stock up or not, at a glance", "description": "输入品类和当前库存，结合实时行情与行业研判，测算补货量、给出分批建仓与风险控制建议。", "description_en": "Enter the category and current stock: combined with real-time market quotes and industry analysis, it estimates replenishment volumes and gives staged stock-building and risk-control advice.", "categories": [ "data" ], "tags": [ "囤货备货", "行情研判", "库存决策", "贸易流通" ], "tags_en": [ "Stock-Up Decision", "Market Analysis", "Inventory Planning", "Trade & Distribution" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "囤货备货建议报告（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易采购专家。请帮我判断某品类当前是否值得囤货，并给出备货建议。\n\n【品类与库存】\n- 品类/规格：{如 螺纹钢 HRB400E}\n- 区域：{如 华东}\n- 当前库存：{如 180吨}\n- 日均出货：{如 20吨/天}\n\n【要求】\n1. 调用【找钢行情MCP】取该品类当前均价及近一周/一月涨跌，查询【行业信息资产库】里这个行业的行情研判口径（淡旺季、限产、库存去化、下游需求），判断价格趋势；\n2. 测算：可用天数、安全库存、建议补货量；\n3. 给出备货建议：是否补货、补多少、是否分批建仓、锁价还是观望、资金占用比例控制；\n4. 给出风险提示与止盈/止损参考。\n\n若行情数据源未就绪或返回不足，请以通用采购与库存管理方法论为主产出并说明。请生成完整备货建议。", "skills": [ { "id": "stock-up-advice", "name": "囤货备货建议", "description": "TradeBuddy把“这个价位要不要囤、备多少”的纠结，做成“行情 + 自己库存销量”摆到一起、依据清楚的备货参考。只摆依据供判断，不替你做决定。", "skill_id": "skill_2091152387889201152" }, { "id": "market-quote-check", "name": "看行情", "description": "TradeBuddy把“随时问一句现在什么价、涨还是跌”的诉求，做成取真实行情、口径清楚、来源可追溯的行情速览。采购和销售共用一套。价格一律来自真实行情数据，绝不编价。", "skill_id": "skill_2090729375350423552" } ], "experts": [ { "id": "caigou-expert", "name": "B2B贸易采购专家", "description": "懂B2B贸易的采购专家：解析需求、看真实行情、全网找货比价、供应商背调、议价话术，帮你把好采购关。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6kRSFmkAwD8M" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "store-product-mix-comparison", "title": "门店商品结构差异分析", "title_en": "Store Product Mix Comparison", "subtitle": "多门店热销商品与成交结构差异洞察", "subtitle_en": "A practical Weimob playbook for store product mix comparison", "description": "做门店商品结构差异分析时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“门店商品结构差异分析”展开：对比各门店近30天的热销商品、商品类别、GMV和客单价结构，识别门店之间的商品经营差异，并提出选品和陈列建议。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、门店业绩分析运营人员、经营分析团队。", "description_en": "The hardest part of store product mix comparison is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates store product mix comparison through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Store Performance Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "门店商品", "商品结构", "热销商品", "选品建议", "陈列建议" ], "tags_en": [ "Store Product Mix Comparison", "Store and Sales Associate Operations", "Store Performance Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "对比各门店近30天的热销商品、商品类别、GMV和客单价结构，识别门店之间的商品经营差异，并提出选品和陈列建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a store product mix comparison covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "merchant-store", "name": "商户与门店查询", "description": "查询可访问商户、门店组织节点、店铺经营概览和待办事项，帮助快速了解当前店铺状态。", "skill_id": "skill_2087080985104273408" }, { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "store-sales-after-sales-health-check", "title": "门店销售售后健康诊断", "title_en": "Store Sales and After-Sales Health Check", "subtitle": "门店销售与售后综合健康评估", "subtitle_en": "A practical Weimob playbook for store sales and after-sales health check", "description": "做门店销售售后健康诊断时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“门店销售售后健康诊断”展开：综合各门店本月销售额、订单量、客单价和售后表现评估经营健康度，筛选成交偏弱或售后风险突出的门店。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、门店业绩分析运营人员、经营分析团队。", "description_en": "The hardest part of store sales and after-sales health check is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates store sales and after-sales health check through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Store Performance Analysis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "门店健康度", "销售表现", "订单量", "客单价", "售后风险" ], "tags_en": [ "Store Sales and After-Sales Health Check", "Store and Sales Associate Operations", "Store Performance Analysis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "综合各门店本月截至昨天的销售额、订单量、客单价和售后表现，评估门店经营健康度，列出成交偏弱或售后风险突出的门店，并给出优先干预建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a store sales and after-sales health check covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "merchant-store", "name": "商户与门店查询", "description": "查询可访问商户、门店组织节点、店铺经营概览和待办事项，帮助快速了解当前店铺状态。", "skill_id": "skill_2087080985104273408" }, { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "student-side-hustle-radar", "title": "大学生搞钱情报站", "title_en": "Student Side-Hustle Radar", "subtitle": "筛选、比较与启动第一份副业", "subtitle_en": "Discover, compare, and start your first side hustle", "description": "想找兼职，却被碎片信息、收入不透明和不知道从哪一步开始困住？这个工作台把匿名演示情报、12 个常见方向、个人匹配和从 0 到 1 的起跑计划放到同一处：先看机会，再比较方向，填一次自己的时间、目标和技能，最后把要做的事拆成可勾选的步骤。适合正在找兼职、实习或线上副业的大学生，也适合想把第一笔收入的尝试做得更有节奏的人。", "description_en": "Looking for a side hustle can mean fragmented openings, unclear pay, and no idea where to begin. This workbench brings anonymized demo opportunities, 12 common tracks, a personal match flow, and zero-to-one starter plans into one place: inspect opportunities, compare tracks, enter your time, goals, and skills once, then turn the work into checkable steps. It is for university students seeking part-time work, internships, or online side hustles — and for anyone who wants to make a first-income experiment more deliberate.", "tags": [ "大学生兼职", "副业筛选", "方向匹配", "起跑计划", "进度追踪" ], "tags_en": [ "Student jobs", "Side-hustle filters", "Track matching", "Starter plans", "Progress tracking" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "大学生搞钱情报站（双语自包含演示）" } ], "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "prompt": "请使用资料库的能力帮我一键做同款 https://workbuddy.link/p/pZcGgLKyjHkHMvNbF5CKI6，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「大学生搞钱情报站」 工作台，面向正在找兼职、实习和线上副业的大学生。页面数据全部存资料库数据表，运行时读写；不要展示本地存储、备份、恢复、导入或导出功能。\n\n五个模块：\n1. 首页控制台：今天要处理、已收录方向、适合我的方向、正在尝试、本周任务和关联资料表；\n2. 兼职雷达：机会情报卡与方向库同页展示，支持关键词、类型、收入、时间投入、远程程度等筛选，能把一条情报收录为自建方向，也能新增、删除、横向比较方向；\n3. 帮我选：收集专业、可投入时间、月收入目标、通勤、技能和偏好，按收入、时间自由、远程、成长、简历价值等维度推荐方向；\n4. 起跑器：每个方向都有从 0 到 1 的可展开步骤、完成状态、截止日、第一笔收入记录，以及技能、作品集和渠道准备清单；\n5. 我的进度：汇总进行中方向、下一步、最近截止和第一笔收入状态。\n\n数据结构使用五张资料库表：兼职情报、我的方向、我的尝试、我的情况、关联资料表。每张表需有与页面读写字段一一对应的 schema；情报和方向支持新增、更新、删除，尝试和个人情况支持持续更新。内置匿名演示数据：约 29 条机会情报、12 个方向、2 条起跑记录和 1 份报价参考表。\n\n视觉使用明亮、紧凑的学生效率工作台：深蓝紫顶部、浅色卡片、清晰的任务优先级与状态色。所有 CSS/JS/演示数据内联，不依赖外部 CDN。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the library capability to clone this workbench in one click: https://workbuddy.link/p/pZcGgLKyjHkHMvNbF5CKI6, with online data storage and two-way sync. Prefer the one-click clone path; if unavailable, build as follows: Build a 'Student Side-Hustle Radar' as a workbench for university students seeking part-time work, internships, and online side hustles. Store all page data in library databases and read/write it at runtime. Do not present local storage, backup, restore, import, or export as product features. Five modules:\n1. Home dashboard: Today’s Focus, saved tracks, best fits, in-progress tracks, this week’s tasks, and a connected data table;\n2. Opportunity Radar: show opportunity cards and a track library together, with filters for keyword, type, income, time commitment, remote suitability, and more. A user can save an opportunity as a personal track, add or delete tracks, and compare tracks side by side;\n3. Find My Fit: collect major, available time, monthly income target, commute, skills, and preferences; recommend tracks by income, flexibility, remote suitability, growth, and resume value;\n4. Starter: each track has expandable zero-to-one steps, completion state, due date, first-income record, and preparation checklists for skills, portfolio, and channels;\n5. My Progress: summarize active tracks, next steps, nearest due dates, and first-income status. Use five library databases: Opportunity Feed, My Tracks, My Attempts, My Profile, and Connected Data Tables. Each schema must map directly to the page’s read/write fields; opportunities and tracks support create, update, and delete, while attempts and profile support ongoing updates. Include anonymized demo data: about 29 opportunities, 12 tracks, 2 starter records, and 1 rate-reference table. Visual direction: a bright, compact student productivity workbench with a deep blue-purple header, light cards, and clear task-priority and state colors. All CSS, JS, and demo data are inline with no external CDN dependency.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "creator": { "name": "复旦奕昂", "name_en": "Fudan Yi'ang", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 91, "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "skills": [], "created_at": "2026-08-31T15:45:00+08:00", "updated_at": "2026-08-31T15:45:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "summer-heat-prevention-video", "title": "夏季防暑降温科普手绘动画", "title_en": "Summer Heat Prevention Handdrawn Animation", "subtitle": "3分钟手绘动画讲解中暑预防与急救知识", "subtitle_en": "3-minute handdrawn animation explaining heat stroke prevention and first aid", "description": "夏季高温频发中暑事故，很多人缺乏预防和急救知识。本案例用 vibeknow 手绘动画师专家，将一份防暑科普文稿转化为 3 分钟暖心扁平风格的手绘动画短片，涵盖中暑症状分级、五招防暑方法、移降补送急救口诀和重点人群注意事项，适合社区科普宣传和校园健康教育。适用于需要将文档/文稿快速转化为手绘动画视频的场景。", "description_en": "Heat stroke incidents spike every summer, yet many lack prevention and first-aid knowledge. This case uses the vibeknow handdraw animator expert to transform a heat prevention document into a 3-minute warm flat-style handdrawn animation covering symptom levels, five prevention tips, the Move-Cool-Hydrate-Send first-aid protocol, and key population guidance. Ideal for community health education and school outreach. Useful when you need to quickly turn documents into handdrawn animation videos.", "prompt": "请根据这份夏季防暑降温科普文稿，制作一个3分钟的手绘动画视频。内容涵盖：什么是中暑及原因、中暑三级症状、五项预防方法、移降补送急救口诀、重点人群注意事项和常见误区。风格选择暖心扁平插画。", "prompt_en": "Create a 3-minute handdrawn animation video based on this summer heat prevention document. Cover: what heat stroke is and its causes, three symptom levels, five prevention methods, the Move-Cool-Hydrate-Send first-aid protocol, key population precautions, and common misconceptions. Use the warm flat illustration style.", "tags": [ "手绘动画", "健康科普", "视频生成", "文档转视频" ], "tags_en": [ "handdrawn animation", "health education", "video generation", "document to video" ], "categories": [ "creation", "education" ], "scenario": "", "scenario_use": "", "task_mode": "设计创意", "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "手绘动画视频" } ], "skills": [], "experts": [ { "id": "VibeknowHanddraw", "name": "vibeknow手绘视频讲解师", "description": "老师讲历史、家长讲绘本、医生讲康复、理财博主讲避坑，专业内容边画边讲成手绘科普视频，51种风格随选。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "89133bd8e85e", "name_en": "vibeknow hand-drawn explainer", "description_en": "Teachers explain history, parents read picture books, doctors explain rehab, finance bloggers warn of pitfalls — turn professional content into hand-drawn explainer videos, 51 styles to choose from.", "expert_id": "ex_kdKBSetkjUyA" } ], "mcps": [], "source": "official", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "usage_count": 0, "created_at": "2026-07-15T11:23:00+08:00", "updated_at": "2026-07-15T11:59:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "supplier-background-check", "title": "合作前先查一查：供应商背调报告", "title_en": "Supplier Background Check", "subtitle": "工商资质、司法失信一键核查，评级+合作建议", "subtitle_en": "Business registration, litigation and default records checked in one pass, with rating and advice", "description": "输入供应商名称，自动核查工商资质、司法与失信风险，出具评级和合作建议，避免踩坑。", "description_en": "Enter a supplier's name: it verifies business registration, litigation and dishonesty risks, then issues a rating and cooperation advice to help you avoid bad partners.", "categories": [ "data" ], "tags": [ "供应商背调", "尽职调查", "风控", "采购" ], "tags_en": [ "Supplier Background Check", "Due Diligence", "Risk Control", "Procurement" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "供应商背调报告（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易采购专家。请帮我对一家供应商做背景调查。\n\n【被调查方】\n- 供应商名称：{如 南京中储贸易有限公司}\n- 拟采购品类：{如 螺纹钢}\n\n【要求】\n1. 工商与资质：调用【企查查MCP】核查成立时间、注册资本、经营状态、经营范围是否与采购品类匹配；\n2. 风险核查：是否命中【找钢黑名单】、司法诉讼/被执行、经营异常/行政处罚，逐项列出并标注结果；\n3. 综合评级：给出低/中/高风险评级；\n4. 合作建议：查询【行业信息资产库】里这个行业的供应商风控口径，给出预付比例、材质质保要求、单笔额度、提额条件等具体建议。\n\n若某项数据源未就绪或查不到，请标注「未获取」，不要臆造。请生成完整背调报告。", "skills": [ { "id": "supplier-background-check", "name": "供应商背调", "description": "TradeBuddy把“合作前先把对方查一圈”的风控习惯，做成一句话出结果的供应商背景调查与风险核查——工商、涉诉、舆情、黑名单一并核验，给倾向性判断而非绝对结论。", "skill_id": "skill_2091152618275139584" } ], "experts": [ { "id": "caigou-expert", "name": "B2B贸易采购专家", "description": "懂B2B贸易的采购专家：解析需求、看真实行情、全网找货比价、供应商背调、议价话术，帮你把好采购关。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_6kRSFmkAwD8M" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "svg-animated-dashboard", "title": "SVG 动画数据看板", "title_en": "SVG Animated Data Dashboard", "subtitle": "纯 SVG 实现的带入场动画的数据分析仪表盘", "subtitle_en": "Pure SVG data analytics dashboard with animated chart entries", "description": "不依赖任何图表库，如何用纯 SVG 打造专业数据看板？本案例展示使用原生 SVG 绘制环形图、柱状图、折线图和迷你图，配合 CSS 动画实现入场效果，4 个 KPI 卡片悬停交互。展示无依赖、高性能的数据可视化方案。适合追求轻量化的前端开发者和数据产品设计师。", "description_en": "How to build a professional data dashboard with pure SVG and zero dependencies? This case demonstrates donut charts, bar charts, line charts, and sparklines — all in native SVG with CSS entry animations and interactive KPI cards.", "categories": [ "data", "dev" ], "tags": [ "SVG图表", "数据看板", "零依赖", "CSS动画" ], "tags_en": [ "SVG Charts", "Data Dashboard", "Zero Dependencies", "CSS Animation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "SVG 动画数据看板页面" } ], "prompt": "请创建一个纯 SVG 实现的动画数据看板页面（单文件 HTML，内联无外部依赖），专业蓝/青/绿配色、图表带入场动画、响应式。\n\n图表全部用 SVG 绘制：动画环形图（各区域收入）、柱状图（增长动画）、折线图（绘制动画）、KPI 卡片内嵌迷你图。\n顶部 4 个 KPI：Revenue $2.4M(+12%)、Users 48.5K(+8%)、Conversion 3.2%(-0.4%)、NPS 72(+5)。\n含侧边导航 + 日期范围选择器（纯视觉）。", "prompt_en": "Create a pure-SVG animated data dashboard page (single-file HTML, inlined with no external dependencies), with a professional blue/cyan/green color scheme, charts with entry animations, and responsive layout.\n\nAll charts drawn with SVG: animated donut chart (revenue by region), bar chart (growth animation), line chart (drawing animation), KPI cards with embedded sparklines.\nTop 4 KPIs: Revenue $2.4M (+12%), Users 48.5K (+8%), Conversion 3.2% (-0.4%), NPS 72 (+5).\nInclude side navigation + a date range picker (purely visual).", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "symptom-abpain", "title": "肚子痛转到右下腹，别硬扛", "title_en": "Pain Moved to Lower Right — Do Not Wait", "subtitle": "多轮问诊识别急腹症信号，直指急诊外科", "subtitle_en": "Multi-turn interview flags acute abdomen, points to ER surgery", "description": "肚脐周围痛了半天，后来固定到右下腹，按着更疼，还有点低烧。是肠胃炎还是更急的？\n\n这个案例演示把「腹痛」交给症状自查：它多轮追问疼痛转移、压痛反跳痛、伴随症状、排便排气、预警信号，识别出「转移性右下腹痛 + 固定压痛 + 低热」高度提示急性阑尾炎，直接给出紧急度 🔴 高、建议立即急诊外科，并提醒「别吃止痛药掩盖腹痛、先禁食水」。\n\n适合：急腹症早识别、夜间分诊、互联网医院导诊、学校/工地医务室。", "description_en": "Periumbilical pain for half a day, then fixed to the lower right, worse on pressure, low fever. Gastritis or something urgent?\n\nThis case shows handing \"abdominal pain\" to symptom self-check: it asks about pain migration, rebound tenderness, accompaniments, bowel sounds, red flags, recognizes \"migratory right-lower pain + fixed tenderness + low fever\" strongly suggesting acute appendicitis, gives urgency 🔴 high, advises immediate ER surgery, and warns \"do not take painkillers that mask it, NPO first\".\n\nGood for: early acute-abdomen detection, night triage, tele-health routing, school/site clinics.", "categories": [ "education", "productivity" ], "tags": [ "症状自查", "腹痛", "急腹症", "阑尾炎", "分诊" ], "tags_en": [ "self-check", "abdominal pain", "acute abdomen", "appendicitis", "triage" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "从昨晚开始肚脐周围绞痛，今天早上疼到了右下腹，一按更疼，还有点低烧不太想吃东西。这要紧吗？要不要马上去医院？", "prompt_en": "Since last night I have had cramping pain around my navel; this morning it moved to the lower right abdomen, worse when pressed, with a low fever and poor appetite. Is this serious? Should I go to the hospital immediately?", "skills": [ { "id": "tencent-diagnosis", "name": "腾讯健康-AI就医助手-症状自查", "description": "AI 症状自查：对接 ADP LLMDiagnosisHealth API，多轮追问收集症状后输出结构化健康分析与就诊指引", "slug": "tencent-diagnosis" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "肚子痛转到右下腹，别硬扛" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "symptom-feedback", "title": "症状反馈", "title_en": "Symptom Feedback", "subtitle": "身体不舒服别硬扛：应对建议 + 原因自查 + 危险信号提醒 + 科室与医院医生推荐，一页看懂", "subtitle_en": "Don't tough it out when you feel unwell: relief tips, likely-cause self-check, red-flag warnings, and department & doctor picks in one page", "description": "几乎每个人都有过这样的时刻：身体突然不舒服，说不上大病却又不踏实——要不要紧？扛一扛还是去趟医院？真要去，又不知从哪一科挂起。这个“拿不准”的空档，往往最熬人。\n\n以「头痛」为例：把症状描述告诉微小医，它会给你贴心的应对建议、常见原因自查和危险信号提醒，告诉你该挂什么科，还推荐真实的医院和专家——一页看懂，心里有了底，再安心去看病。", "description_en": "Almost everyone has had this moment: something feels off with your body — not quite an emergency, but not nothing either. Should you tough it out or see a doctor? And if you go, where do you even start? That stretch of uncertainty is often the hardest part.\n\nTake a headache as an example: describe your symptom to Weixiaoyi, and it offers caring relief tips, likely-cause self-checks and red-flag warnings, tells you which department to register with, and recommends real hospitals and specialists — one page to read, and you can head to the doctor with peace of mind.", "categories": [ "business" ], "tags": [ "症状反馈", "健康咨询", "就医指南", "微医", "HTML生成" ], "tags_en": [ "Symptom Feedback", "Health Consultation", "Care Guide", "WeDoctor", "HTML Generation" ], "artifact_type": "html", "preview": "headache-guide.html", "artifacts": [ { "file": "headache-guide.html", "label": "头痛健康指南 HTML 页面" } ], "prompt": "头痛难受怎么办，可能是为啥，该挂什么号，有什么推荐的医院和医生？将结果生成html进行展示。html 功能：1. 如有跳转链接，要同时支持点击跳转及弹窗扫描二维码访问；2. 固定导航栏，根据页面尺寸显示在顶部或侧边，需高亮当前模块，点击切换定位页面至对应模块；3. 可将页面内容保存为本地图片，图片中的信息元素显隐、排版需根据分享场景做一下优化；4. 数据列表限制最多5条；5.头部信息标题、副标题文本需体现该页面为全面的健康服务，显示品牌名称“微小医”及logo（https://kano.guahao.com/zj8979871203）。html ui设计风格：1. \"医疗蓝扁平卡片流\"：纯白大圆角卡片，以 0.5px 细线代替阴影分隔，全部交互强调收敛到单一品牌蓝 #3f86ff 及其渐变/透明度变体，通过 PingFang SC 三档字重与四级灰阶建立信息层级，主次按钮（渐变实心 vs 浅蓝描边）形成清晰的操作引导优先级；2. 统一配内联的svg图标；3. 可交互按钮鼠标悬停时增加动效；4. 元素布局根据页面尺寸调整，以提高展示效果。", "prompt_en": "What should I do about a bad headache, what might be causing it, which department should I register with, and are there any recommended hospitals and doctors? Present the results as an HTML page. HTML features: 1. Any external link should support both direct click-through and a popup with a scannable QR code; 2. A fixed navigation bar that sits on top or on the side depending on page width, highlighting the current section and scrolling to the corresponding section on click; 3. The page content can be saved as a local image, with element visibility and layout optimized for sharing; 4. Data lists are limited to 5 items at most; 5. The header title and subtitle should convey that this page is a comprehensive health service, showing the brand name \"Weixiaoyi\" and its logo (https://kano.guahao.com/zj8979871203). HTML UI style: 1. \"Medical-blue flat card flow\": pure white cards with large border radius, 0.5px hairlines instead of shadows for separation, all interactive emphasis confined to a single brand blue #3f86ff and its gradient/opacity variants, information hierarchy built on three font weights of PingFang SC and four grayscale levels, with primary vs. secondary buttons (solid gradient vs. light-blue outline) forming a clear action priority; 2. Consistent inline SVG icons; 3. Hover animations on all interactive buttons; 4. Element layout adapts to page size for better presentation.", "skills": [ { "id": "we-health-skill", "name": "微医健康管家", "description": "微医AI健康管家（官方Skill），覆盖报告解读、就医推荐、用药咨询、健康咨询等全部医疗健康场景", "slug": "we-health-skill" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-19T17:15:00+08:00", "updated_at": "2026-08-24T18:30:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "symptom-headache", "title": "头痛两周，AI 帮你理清楚", "title_en": "Two Weeks of Headache, Sorted by AI", "subtitle": "多轮问诊，给出可能病因与就诊科室", "subtitle_en": "Multi-turn interview — likely causes and which department", "description": "右边太阳穴一跳一跳地疼了两星期，有时恶心。百度完越查越怕。\n\n这个案例演示把「反复头痛」交给症状自查：它像医生一样多轮追问——部位性质、频率、伴随症状、诱因、预警信号，逐步收窄范围，最后输出结构化健康分析报告：可能病因（偏头痛倾向高）、建议就诊科室（神经内科）、注意事项与家庭处理，并明确标注不替代诊断。\n\n适合：轻症自查分诊、体检中心预问诊、互联网医院导诊、企业医务室。", "description_en": "A throbbing pain at the right temple for two weeks, sometimes nauseous. Baidu only makes it scarier.\n\nThis case shows handing \"recurrent headache\" to symptom self-check: like a doctor it asks across turns — location/nature, frequency, accompaniments, triggers, red flags — narrowing the range, then outputs a structured report: likely causes (migraine high), department to visit (neurology), cautions and home care, clearly not a diagnosis.\n\nGood for: mild-symptom triage, pre-visit at checkup centers, tele-health routing, corporate clinic.", "categories": [ "education", "productivity" ], "tags": [ "症状自查", "头痛", "问诊", "就诊科室", "健康科普" ], "tags_en": [ "self-check", "headache", "interview", "department", "health literacy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "反复头痛两周了，右边太阳穴一跳一跳地疼，有时候恶心，熬夜之后更明显。这可能是什么问题，要不要去医院、挂哪个科？", "prompt_en": "I've had recurring headaches for two weeks — throbbing pain at my right temple, sometimes nausea, worse after staying up late. What could this be, should I see a doctor, and which department?", "skills": [ { "id": "tencent-diagnosis", "name": "腾讯健康-AI就医助手-症状自查", "description": "AI 症状自查：对接 ADP LLMDiagnosisHealth API，多轮追问收集症状后输出结构化健康分析与就诊指引", "slug": "tencent-diagnosis" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "头痛两周，AI 帮你理清楚" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "symptom-skinrash", "title": "一身红疹发痒，先别乱抹", "title_en": "Itchy Red Rash — Do Not Guess", "subtitle": "多轮问诊区分过敏与皮炎，给护理与预警", "subtitle_en": "Multi-turn interview tells allergy vs dermatitis, care and red flags", "description": "手臂胸前起一片红疹，发痒，昨天吃了海鲜、换了新护肤品。是过敏还是别的？\n\n这个案例演示把「皮疹」交给症状自查：它多轮追问形态分布、痒痛、近期接触、全身症状、过敏预警（喉头紧/呼吸困难），判断「风团样 + 瘙痒 + 海鲜/新护肤品诱因」倾向荨麻疹（过敏），建议皮肤科，并强调若出现喉咙发紧、呼吸困难须立即急诊。\n\n适合：皮肤问题自查、母婴/敏感肌护理、互联网医院导诊、药店皮肤药推荐前置。", "description_en": "A patch of red rash on arms and chest, itchy, after seafood and a new skincare yesterday. Allergy or something else?\n\nThis case shows handing \"rash\" to symptom self-check: it asks form/distribution, itch vs pain, recent exposure, systemic symptoms, allergy red flags (throat tightness/dyspnea), judges \"wheal-like + itch + seafood/new product\" leans urticaria (allergy), advises dermatology, and stresses immediate ER if throat tightens or breathing trouble.\n\nGood for: skin self-check, baby/sensitive-skin care, tele-health routing, pre-step for pharmacy skin meds.", "categories": [ "education", "productivity" ], "tags": [ "症状自查", "皮疹", "过敏", "荨麻疹", "分诊" ], "tags_en": [ "self-check", "rash", "allergy", "urticaria", "triage" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "手臂和胸前起了一片红疹，特别痒，昨天吃了海鲜还换了新护肤品。这是过敏吗，要怎么处理、需要看医生吗？", "prompt_en": "A patch of itchy red rash appeared on my arms and chest. Yesterday I ate seafood and switched to a new skincare product. Is this an allergy, how should I handle it, and do I need to see a doctor?", "skills": [ { "id": "tencent-diagnosis", "name": "腾讯健康-AI就医助手-症状自查", "description": "AI 症状自查：对接 ADP LLMDiagnosisHealth API，多轮追问收集症状后输出结构化健康分析与就诊指引", "slug": "tencent-diagnosis" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-22T17:14:53Z", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "一身红疹发痒，先别乱抹" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "tang-nara-buddhist-correlation", "title": "唐代与日本奈良佛像艺术风格对比", "title_en": "Tang vs Nara Buddhist Sculpture Artistic Style Comparison", "subtitle": "盛唐彩塑与奈良干漆夹纻:跨海造像艺术的本真传承与本土再创造", "subtitle_en": "High-Tang painted sculpture vs Nara dry-lacquer: authentic transmission and indigenous re-creation of Buddhist art across the East China Sea", "description": "讲起唐代佛像大家都能想象敦煌彩塑,但跟日本奈良佛像到底什么关系?是简单照搬还是有更深的本地化再创造?对比两座佛像又怕断章取义、漏掉干漆夹纻等关键工艺差异。\n\n这个案例演示了如何借助腾讯探元文物库与知识库的并行检索,主代理亲自做三路 WebSearch 多源核验(对象 A 唐代造像、对象 B 奈良佛像、中日佛教艺术交流脉络),覆盖工艺技法、造型特征、材质成分、文化内涵与历史交流五个维度,最终交付包含配图与对比雷达图的 Markdown/HTML 双格式对比报告。\n\n适合:佛教艺术史研究者、敦煌学与丝绸之路文化研究者、唐代社会与文化爱好者、奈良时代东亚文化交流研究者、博物馆策展与文博教育人员、艺术院校佛教美术专业学生。", "description_en": "Talking about Tang Buddhist sculpture most people picture Dunhuang painted statues, but how do they actually relate to Nara-period Japanese works? Are Nara sculptures simple knockoffs or do they carry deeper indigenous re-creation? Comparing two iconic traditions often fragments sources, misses key craft differences like dry-lacquer (干漆夹纻) hollow-body technique, and overlooks the cross-cultural exchange network. This case shows how to use Tencent Tanyuan relic and knowledge bases for parallel retrieval, plus three concurrent main-agent WebSearch passes (Object A Tang sculpture, Object B Nara sculpture, Sino-Japanese Buddhist art exchange), covering five dimensions—craftsmanship, form, material, cultural meaning, and historical exchange—delivering a cross-checked comparison report with embedded images and a contrast radar chart in Markdown and HTML formats.", "categories": [ "business" ], "tags": [ "唐代彩塑", "奈良时代", "鉴真东渡", "干漆夹纻", "佛教艺术" ], "tags_en": [ "Tang Painted Sculpture", "Nara Period", "Jianzhen Voyage", "Dry Lacquer", "Buddhist Art" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告(含对比雷达图)" } ], "prompt": "请按「腾讯探元学术研究-关联研究」专家的工作流,关联对比「唐代佛像(以敦煌莫高窟盛唐彩塑为代表)」与「日本奈良佛像(以唐招提寺鉴真和尚坐像为代表)」在艺术风格层面的关系。\n\n研究范围(覆盖五维):\n- 工艺技法:泥塑/木骨泥塑 vs 干漆夹纻(活脱干漆)/木雕,鎏金、敷彩、贴金箔等装饰工艺\n- 造型特征:面相丰满/菩萨世俗化 vs 写实肖像/精神化塑造,衣纹处理、身躯比例\n- 材质成分:木-苇草-泥-矿物颜料 vs 苎麻-生漆,以及两地原料来源\n- 文化内涵:世俗化圆满/丰腴神性 vs 写实人格/肖像化精神,以及中日审美差异\n- 交流脉络:遣唐使学问僧(最澄、空海)、鉴真东渡(公元754年)、中日文化输出的双向性\n\n交付要求(强制):\n1. 探元两库并行 + 主代理 2-3 次 WebSearch 多源核验\n2. 配图 ≥ 2 张(敦煌莫高窟代表彩塑 + 唐招提寺鉴真像),优先 Wikimedia / 公开博物馆图源;图表 ≥ 1 张(雷达图 SVG)\n3. 报告结构:对比结构(对象 A 概述 → 对象 B 概述 → 逐维度对比 → 中日交流脉络 → 结论 → 证据与局限 → 参考来源 → 图片与图表来源)\n4. 仅写「对比结论」,不含建议;参考文献只列网络公开来源;主要结论同步到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Correlation Research' expert workflow to comparatively research the artistic-style relationship between 'Tang Buddhist sculpture (Dunhuang Mogao high-Tang painted sculpture as representative)' and 'Nara-period Japanese Buddhist sculpture (Tōshōdai-ji Jianzhen portrait as representative)'.\n\nResearch scope (five dimensions):\n- Craft: clay sculpture / wood-boned clay vs dry-lacquer (活脱干漆) / wood carving; gilding, polychromy, gold-leaf application\n- Form: full-faced / secularized bodhisattvas vs realistic portrait / spiritualized expression; drapery handling, body proportions\n- Material: wood-reed-clay-mineral pigments vs ramie-raw lacquer, with raw material sources on both sides\n- Cultural meaning: secularized fullness / corporeal divinity vs realistic persona / portrait spirituality, plus Sino-Japanese aesthetic divergence\n- Exchange: Kentōshi monk-students (Saichō, Kūkai), Jianzhen's eastward voyage (754 CE), bidirectional nature of cultural transmission\n\nDelivery requirements (mandatory):\n1. Parallel Tanyuan dual bases + 2-3 main-agent WebSearch passes for cross-source verification\n2. ≥ 2 images (Dunhuang Mogao + Tōshōdai-ji Jianzhen), Wikimedia / public museum sources preferred; ≥ 1 chart (radar SVG)\n3. Report structure: contrastive (Object A summary → Object B summary → per-dimension → Sino-Japanese exchange → conclusion → limits → references → image/chart sources)\n4. Conclusions only, no recommendations; references limited to public web sources; key conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research-correlation", "name": "文博关联研究专家", "description": "基于探元文物库与知识库并行采集,由主代理直接联网多源交叉核验,交付四格式跨对象/跨域对比研究报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "75f1347b2045", "expert_id": "ex_pEyI3eIkynB5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:11:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tang-sancai-horse-postcard", "title": "唐三彩马文创 · 明信片", "title_en": "Tang Sancai Horse Cultural Product · Postcard", "subtitle": "文物研究档案风的 A6 印刷明信片", "subtitle_en": "An A6 print postcard in relic-study archive style", "description": "给文创品做明信片，既想有研究档案的雅致，又怕堆一堆泛化国风元素显得俗。\n\n案例以文创马鞍包为主体，从包本身提取研究性辅助图形（马纹线描、扣件放大、缝线轨迹），暖米白手工纸背景 + 雾青绿半透明色块，仅少量英文「OBJECT STUDY / 01」，输出安静雅致的 A6 竖版文物研究档案风明信片。\n\n适合：文创品牌营销、平面/编辑设计师、博物馆文创团队、精品礼品开发方", "description_en": "This case features a saddle-bag creative product, deriving study graphics from the bag itself (bridle line drawing, buckle close-up, stitch tracks), on warm ivory handmade paper with misty teal translucent blocks and minimal English 'OBJECT STUDY / 01' — a quiet, refined A6 relic-study postcard.\n\nGood for: Cultural brand marketers, Graphic & editorial designers, Museum creative teams, Premium gift developers", "categories": [ "creation" ], "tags": [ "文创明信片", "A6印刷", "档案设计", "皮具编辑" ], "tags_en": [ "Cultural Postcard", "A6 Print", "Archive Design", "Leather Editorial" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请基于我上传的文创品图，设计一张 A6 竖版印刷明信片，比例 111:154，风格为「当代文物研究档案 × 克制古画印刷 × 精品皮具编辑设计」。\n主体必须完整保留上传包的真实身份特征：马鞍包轮廓、绿色马鞍形翻盖、翻盖上的立体马纹、浅黄皮革包身、棕色皮革包边与肩带、绿色褶皱前袋、金属扣环、车线、皮革纹理和原有比例。不要改造包的结构，不要替换材质，不要把它变成真实古董或普通时装广告。\n构图采用“宽而圆的现代文创品研究式排版”：包作为前景主视觉，放在画面中下部，约占画面 50%–60%，保持完整清晰、真实皮革质感和综合色彩；上方、左右与底部留出充足呼吸空间。背景为暖米白手工纸，带非常细腻的纸纤维、轻微旧印刷颗粒和低对比灰褐线条。\n从包本身提取研究性辅助图形，而非加入泛化国风背景：将翻盖上的马纹转化为左上角的细线描马匹图；将金属扣、车线、褶皱皮革口袋和马鞍弧线转化为右上局部圆形放大图、底部局部结构线稿、虚线缝线轨迹和弧形构造线。所有辅助图形必须来源于包的可见结构，像产品档案或设计研究稿，保持很浅的灰褐色，不抢主体。\n色彩体系：暖米白纸色、低饱和灰褐线描、雾青绿与浅灰蓝的大面积半透明几何色块；保留包本身的绿色、浅黄、棕色与黄铜色。仅在角落加入一个很小的砖红色印章点睛，面积不超过画面 2%。整体安静、雅致、现代，不要高饱和。\n排版只使用极少量英文信息：左侧小号高对比衬线体 “OBJECT STUDY” 与编号 “01”；若文字无法清晰呈现，宁可省略文字。不要加入中文标题、品牌名、年代、地点或任何未经提供的历史信息。\n避免：山水、宫殿、古建筑、人物、市集、花卉、龙凤、祥云、灯笼、旅游纪念品风、金箔、厚重国潮、复杂边框、镜面产品摄影、3D 渲染感、过重噪点、水印、乱码。最终输出为清晰的 A6 竖版 PNG。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-promo-expert", "name": "文创宣传视觉设计专家", "description": "将定稿文创设计转化为海报、明信片、收藏小卡、易拉宝等宣传物料", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c6229d08eb8d", "expert_id": "ex_zaYhWoDF5zrp" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-bronze-taotie-origin", "title": "青铜器兽面纹的起源与精神象征", "title_en": "Origin and Symbolism of the Bronze Beast-Face Motif", "subtitle": "从史前图像→二里头铜牌→商周礼器，追踪兽面纹（旧称「饕餮纹」）的演化线索与多重解释", "subtitle_en": "Trace the 3000-year evolution from Liangzhu jades to Shang-Zhou ritual bronzes, and decode its 'theocracy-to-kingship' symbolism", "description": "博物馆里青铜鼎上的两只大眼睛常常让人发怵——很多人想知道那到底是什么、有没有具体含义。\n\n关于名称：正文与封面所说的「饕餮纹」是沿用已久的旧称，「兽面纹」是现代考古与博物馆展陈中更常用的称呼，两者指的是同一类以中轴展开、突出双目的正面兽首纹样；案例正文已专门说明这一称谓差异。\n\n这个案例演示了腾讯探元文博知识问答专家如何回答历史溯源问题——它从史前兽面图像、二里头镶嵌铜牌讲到商周青铜礼器，同时明确区分“可观察的图像联系”和“尚未证实的直接传承”。全文以镶嵌绿松石兽面纹铜牌饰、商代兽面纹铜斝两件真实文物配图，分层呈现动物综合、祭祀信仰与权力表达等解释，并用正文 [1][2][3][4] 对应文末来源。\n\n适合：博物馆与考古专业学生；青铜器收藏与研究者；做博物馆深度讲解的志愿者；对纹饰学、纹样发展史感兴趣的设计与艺术史学习者；做'5·18 国际博物馆日'、'文化遗产日'主题推文的公众号编辑。", "description_en": "The two huge eyes on a Shang dynasty bronze ding can be unnerving — many visitors want to know what's behind the motif.\n\nThis case shows how the Tencent TanYuan cultural-heritage Q&A expert answers a 'historical origin' question — it traces the full evolutionary chain of the beast-face motif (formerly called 'taotie'): from Liangzhu jades (ca. 3300-2300 BCE), to Erlitou pottery of the Xia, to Shang dynasty bronze ritual vessels, to the Western Zhou. Each stage's visual features and symbolic meaning are spelled out, and the answer makes clear that the motif is an idealised composite of multiple animals (ox, tiger, dragon variants), and that its use shifted from natural-spirit worship to a symbol of kingship. Two real artefacts anchor the narrative: the inlaid turquoise beast-face bronze plaque from Erlitou (Yanshi), and the Shang dynasty 'beast-face li' at the Anhui Museum.\n\nBest for: Museum studies and archaeology students; bronze collectors and researchers; volunteer docents who do in-depth tours; design and art-history students interested in ornamental patterns; WeChat editors preparing 'International Museum Day (18 May)' or 'Cultural Heritage Day' themed posts.", "categories": [ "business", "education" ], "tags": [ "兽面纹", "饕餮纹", "青铜器", "商周礼器", "历史溯源" ], "tags_en": [ "beast-face motif", "taotie pattern", "bronze ritual vessel", "Shang Zhou bronzes", "motif origin" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整问答报告（起源链+实物对照+精神象征+来源）" } ], "prompt": "请回答：青铜器上的兽面纹（旧称饕餮纹）到底是什么？它最早出现在什么时候？后来为什么会变成商周礼器的'主流纹饰'？\n\n要求：\n1. **结论先行**：一两句话讲清——兽面纹是商周青铜礼器的重要纹饰，可能综合多种动物特征；其含义存在多种解释，不能把“政治符号”或“良渚直接起源”写成唯一结论\n2. **完整溯源链**：分四段展开并明确证据强弱：\n - **史前背景**：良渚等文化的兽面图像可作为参照，但与商周纹饰的直接传承仍有争议\n - **过渡期**：二里头文化镶嵌绿松石兽面纹铜牌饰显示兽面图像已与青铜、镶嵌工艺及礼仪场景发生联系\n - **鼎盛期**：商代兽面纹成为青铜礼器的重要主纹，并与地纹、器形共同组织\n - **转折期**：西周以后逐步简化、几何化，并转化为后世装饰母题\n3. **两种主流学术观点**：明确写出'多源综合说（主流）'和'自然神崇拜说'两条解释路线，承认学界至今无完全定论\n4. **精神象征三层**：早期=氏族图腾/自然神；商代=政教合一视觉统治工具；西周以后=秩序化、神秘感退场\n5. **实物对照**：至少使用 2 个已核实的真实锚点，其中包括二里头镶嵌绿松石兽面纹铜牌饰和一件商代兽面纹青铜礼器\n6. **名字溯源**：解释“饕餮纹”和“兽面纹”两种名称为何并见，避免写成学界已完全统一\n7. **学界争议和待考**：明确点出图像来源、具体动物原型和精神象征均未形成唯一结论，不添加与主题无关的铸造争议\n8. **图文与引用**：至少加入 2 张真实文物图；正文只用 [1][2][3] 标注事实或观点，不要反复写“来源于某知识库”；文末统一列明题名、机构、完整 URL 与图片出处\n9. **数据要可溯源**：每个时间点都对应可核验的考古时间框架，不确定处明确写“尚有争议”\n\n风格：能给一个对青铜器一无所知的人讲清楚 3000 年的演化路径；不要用过程叙述词；末尾给一条'下次类似问题'作为延伸钩子。", "prompt_en": "Please answer: What exactly is the beast-face motif (formerly 'taotie') on Chinese bronzes? When does it first appear? Why did it become the dominant motif on Shang-Zhou ritual vessels?\n\nRequirements:\n1. **Conclusion first**: Lead with the three verifications — it is an idealised composite of multiple animals (essentially a political symbol); it is not a real headless monster; its origins can be traced back to Liangzhu jades.\n2. **Full origin chain**: Four clearly delineated stages, each anchored to widely accepted archaeological dates:\n - **Proto-motif phase**: Liangzhu jades (ca. 3300-2300 BCE) + Longshan-culture eggshell black pottery / stone moulds — technical rehearsal.\n - **Transition phase**: Xia-dynasty Erlitou (Taosi, ca. 1900 BCE) + the inlaid turquoise beast-face bronze plaque from Erlitou (Yanshi) — the earliest bronze instance.\n - **Apex phase**: Shang — from 'independent / connected beast-face' two types → mid-Shang two-tier patterns → late Yinxu three-tier patterns.\n - **Turning point**: Western Zhou onward — simpler, more ordered, less terrifying.\n3. **Two mainstream scholarly views**: Spell out clearly 'multi-animal composite (mainstream)' and 'natural-spirit worship' — and acknowledge the lack of a complete consensus.\n4. **Three symbolic layers**: Early = clan totem / natural-spirit worship; Shang = theocratic-political visual tool; Western Zhou onward = rationalised, mystery recedes.\n5. **Real artefact anchors**: At least three — the inlaid turquoise beast-face bronze plaque from Erlitou (Yanshi); the Shang dynasty 'beast-face li' at the Anhui Museum (from the Funan Yueyahe discovery); the 'beast-face bronze yan' at the Anhui Huizhou History Museum.\n6. **Name etymology**: Why the ancients called it 'taotie', why modern archaeology renamed it 'beast-face', and the horn-shape sub-types (outward-curled, inward-curled, zigzag-horned, double-dragon-horned, sheep-horned, ox-horned, variant).\n7. **Open questions**: Make clear that 'whether the motif depicts any real animal' is still debated; for details without consensus (e.g. 'cast-in ear vs. one-piece casting'), explicitly say further research is needed.\n8. **Sources**: 'Tencent TanYuan knowledge base' by topic + 'Public sources' with URLs; different views should be separately attributed.\n9. **Traceable data**: Every date should match an archaeological consensus timeframe (e.g. 'Liangzhu culture, ca. 3300-2300 BCE').\n\nStyle: Make a 3000-year evolutionary path understandable to someone who knows nothing about bronzes. No process narration. End with one 'next time, try asking' hook.", "skills": [], "experts": [ { "id": "tanyuan-heritage-qa-expert", "name": "文博百科专家", "description": "依托腾讯探元文博知识库，从工艺原理与材质分析解答文博疑问，结论先行并标注权威资料来源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "97d0c0800a1f", "expert_id": "ex_qSpUhcNlQwWc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:00:29+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-kintsugi-craft", "title": "金缮工艺到底是什么——与锔瓷有什么不同", "title_en": "What Kintsugi Really Is — and How It Differs from Juci Riveting", "subtitle": "瓷器的'碎裂伤疤'怎么变成'金色艺术'——原理、工序、与锔瓷的本质区别", "subtitle_en": "How a broken porcelain scar becomes gold-traced art — principles, process, and why kintsugi differs from 'juci' (riveting)", "description": "看到一件用金色线条'勾'出裂痕的瓷器，很多人只觉得漂亮，却说不出它的工序与美学。\n\n本案例的主线是对比：封面取「金缮与锔瓷 · 有何不同」这一角度，正文也按此展开——先拆解金缮拼接、补缺、描线、饰金的完整流程，再与锔瓷（钻孔加金属锔钉的机械连接）逐项对照，并区分传统金缮与现代「金缮风格」修补。\n\n这个案例演示了腾讯探元文博知识问答专家如何回答传统工艺问题——它会拆解拼接、补缺、描线、饰金的完整流程，再与锔瓷对比，并区分传统金缮和现代'金缮风格'修补。全文选用东京国立博物馆与 Cooper Hewitt 馆藏实物配图，用正文 [1][2][3] 对应文末来源。\n\n适合：博物馆讲解员与非遗传承研究者；陶艺与漆艺爱好者；瓷器收藏家与工艺品买家；对'残缺美'感兴趣的设计师、文创从业者；想为自己破损的瓷器找修复方案的实用用户。", "description_en": "Many people see a porcelain bowl traced with gold along its crack and only register 'pretty', without knowing the technique behind it.\n\nThis case shows how the Tencent TanYuan cultural-heritage Q&A expert answers a 'traditional craft' question — first explaining the two key steps (lacquer bonding → gold application), then contrasting kintsugi with 'juci' (riveting-repair), and finally tracing the aesthetic back to the Japanese wabi-sabi philosophy. The full answer references a real artefact (a Southern Song 'black-body water dropper' at the Longquan Celadon Museum that has been kintsugi-repaired) and closes with a pragmatic checklist for anyone considering commissioning a kintsugi repair — to avoid being misled by cheap 'machine kintsugi'.\n\nBest for: Museum docents and intangible-heritage researchers; ceramics and lacquerware enthusiasts; porcelain collectors and craft buyers; designers / cultural-product creators interested in 'the beauty of imperfection'; end-users looking for a trustworthy way to repair a broken heirloom piece.", "categories": [ "business", "education" ], "tags": [ "金缮", "锔瓷", "侘寂", "非遗工艺", "瓷器修复" ], "tags_en": [ "kintsugi", "juci riveting", "wabi-sabi", "intangible craft heritage", "porcelain repair" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整问答报告（工序+对照+美学溯源+实用提醒+来源）" } ], "prompt": "请回答：金缮工艺到底是什么？它与锔瓷工艺有什么本质区别？\n\n要求：\n1. **结论先行**：第一段一两句话讲清'金缮是用大漆粘合、贴金装饰的瓷器修补工艺'，并强调它的本质不是'修好瓷器'而是'让残缺可视化为美'\n2. **两道关键工序**：把粘合（大漆常温固化）和装饰（描线+贴金）讲透，说明大漆固化的环境要求\n3. **与锔瓷对比**：用表格或对照段落讲清两种工艺的核心差异（锔瓷=金属钉加固实用性、金缮=漆粘+贴金艺术性）。明确'现代修复中两者常常结合使用'\n4. **美学溯源**：从日本侘寂（wabi-sabi）理念讲起——侘=朴素质朴、寂=时间痕迹——说明金缮为何重视残缺\n5. **现代延伸**：讲清金缮已从'修复工艺'演变为'当代陶瓷创作语言'，举例说明\n6. **真实文物对照**：使用权属清晰的博物馆馆藏金缮实物，不得编造馆藏或修复记录\n7. **实用提醒**：天然漆固化条件、技艺门槛、修复材料与修复师甄别建议\n8. **图文与引用**：至少加入 2 张真实实物图；正文只用 [1][2][3] 标注对应事实，不要反复写“来源于某知识库”；文末统一列明题名、机构、完整 URL 与图片授权\n9. **面向人群**：文博爱好者、文创设计师、瓷器收藏者、收藏破损瓷器的家庭用户都能读懂\n\n风格：不要用'我检索了两次''系统返回'这类过程语；专业术语第一次出现时用一句话解释（生漆、侘寂）；末尾给一条'下次类似问题'作为延伸钩子。", "prompt_en": "Please answer: What is kintsugi, and how does it fundamentally differ from 'juci' (riveting-repair)?\n\nRequirements:\n1. **Conclusion first**: Open with a one- or two-sentence verdict — kintsugi bonds the pieces back with urushi lacquer and traces the crack with gold; its essence is not 'fixing the porcelain' but 'making the imperfection visible as art'.\n2. **The two key steps**: Explain bonding (urushi cures at room temperature; the curing chamber) and decoration (lacquer line tracing + gold application). Include the environmental requirements for urushi curing.\n3. **Kintsugi vs juci**: Use a table or parallel paragraph to spell out the core difference — juci = metal-rivet reinforcement for utility; kintsugi = lacquer bond + gold trace for artistry. Make clear the two are often combined in modern repair.\n4. **Aesthetic origin**: Trace back to Japanese wabi-sabi — wabi = quiet simplicity; sabi = the patina of time — and explain why kintsugi values imperfection.\n5. **Modern extension**: Note that kintsugi has evolved from a 'repair technique' into a 'contemporary ceramic creative language', with an example.\n6. **Real artefact reference**: Anchor on a real object — the Southern Song 'black-body water dropper' at the Longquan Celadon Museum, which was kintsugi-repaired.\n7. **Pragmatic checklist**: Curing-environment requirements, skill thresholds, advice for choosing a real kintsugi artisan vs cheap 'machine kintsugi'.\n8. **Sources**: Separately label 'Tencent TanYuan knowledge base' (topic only) and 'Public sources' (must carry clickable URLs).\n9. **Audience**: Accessible to museum enthusiasts, cultural-product designers, porcelain collectors, and households with a broken heirloom piece.\n\nStyle: No process narration like 'I retrieved twice' or 'the system returned'. Explain every technical term the first time it appears (urushi, wabi-sabi). End with one 'next time, try asking' hook.", "skills": [], "experts": [ { "id": "tanyuan-heritage-qa-expert", "name": "文博百科专家", "description": "依托腾讯探元文博知识库，从工艺原理与材质分析解答文博疑问，结论先行并标注权威资料来源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "97d0c0800a1f", "expert_id": "ex_qSpUhcNlQwWc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:00:29+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-poster-chenghua-chicken-cup", "title": "成化斗彩鸡缸杯典藏明信片", "title_en": "Chenghua Doucai Chicken Cup Collector Postcard", "subtitle": "回纹描边、竖排题名、朱文印章、团花底纹——一盏明代御窑的典藏明信片", "subtitle_en": "Fret-pattern border, vertical title, vermilion seals, medallion ground — a Ming imperial-kiln collectible postcard", "description": "这个案例按既定设计规范制作成化斗彩鸡缸杯典藏明信片：外层双回纹边框与四角回纹角花，右侧竖排标题与年代小竖框，红色朱文印章、竖排英文名与竖排馆藏年代信息分列两侧，底部为馆藏实拍杯体。\n\n关于命名：产物与封面统一为「典藏明信片 / COLLECTOR POSTCARD」，由腾讯探元典藏海报专家产出，与「典藏海报」同属一套典藏视觉，区别只在成品用途与尺寸。\n\n执行要点：文物主体只用故宫博物院馆藏实拍，仅做背景分离与等比缩放，器形、斗彩纹饰与子母鸡雏均保持原貌；生图工具用于生成宣纸质感与团花、祥云底纹，不参与文物绘制；所有中文由字体排版并逐字核对，避免生成错字。\n\n适合：故宫与瓷器主题社教文创、明清瓷器爱好者、典藏类自媒体、文博文创产品经理、设计系文化创意方向学生。", "description_en": "A collectible postcard produced to the established design system: double fret border with corner ornaments, vertical title with a boxed reign mark, vermilion seals, vertical English name and collection details, and an unretouched museum photograph of the cup. Image generation is used only for the paper texture and medallion ground; all Chinese text is typeset and proofread.", "categories": [ "creation", "business" ], "tags": [ "文物明信片", "成化斗彩", "典藏视觉", "故宫博物院", "博物馆文创" ], "tags_en": [ "Heritage Postcard", "Chenghua Doucai", "Collectible Visual", "Palace Museum", "Museum Creative" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "成化斗彩鸡缸杯典藏明信片" } ], "prompt": "你是「腾讯探元典藏海报专家」，按既定设计规范制作「成化斗彩鸡缸杯」竖版典藏明信片。\n\n## 一、事实检索\n并行检索探元文物库与知识库，核对名称、年代、馆藏、尺寸（高 3.4cm、口径 8.3cm、足径 3.7cm）、斗彩工艺与子母鸡雏纹饰。缺失或冲突的事实标注待确认，不补写。\n\n## 二、文物主体\n1. 只使用探元 cover 或博物馆馆藏实拍图，允许背景分离、等比缩放与版面定位。\n2. 禁止把文物交给生图模型重绘、补绘、换纹饰或改器形；杯口、腹部、圈足与鸡雏纹饰必须与原图一致。\n3. 无可靠实物图时停止制作并说明缺少图源。\n\n## 三、必须调用生图工具\n调用生图工具生成非文物层：暖米黄做旧宣纸质感，右上角淡团花，左侧淡祥云，中下部保持干净留白，画面内不得出现任何文字、印章、边框与器物。禁止用纯色或 CSS 渐变代替生图背景。\n\n## 四、版式规范（固定结构）\n- 外框：双线回纹边框 + 四角回纹角花。\n- 右上：竖排馆藏与年代信息（故宫博物院藏 / 明成化年间 1465—1487）+ 朱文印章。\n- 中部偏右：竖排大字标题「成化斗彩鸡缸杯」。\n- 标题左侧：小竖框年代「明成化」、朱文印章「御窑」、竖排斜体英文名。\n- 标题右侧：竖排「典藏明信片」+ 竖排英文 COLLECTIBLE POSTCARD。\n- 下部：馆藏实拍杯体居中。\n- 底部：尺寸信息与「文物主体为馆藏实拍，未作形态重绘」。\n\n## 五、质检\n逐项比对实物原图（器形、口沿、圈足、鸡雏与花石纹饰），逐字核对标题、印章、英文与落款；确认生图背景中无文字与器物。\n\n## 六、交付\n渲染为 1024×1536 PNG，嵌入单文件 output.html，提供下载与打印按钮，无外链、无 emoji、无锚点。", "prompt_en": "Produce the Chenghua Doucai Chicken Cup collectible postcard to the fixed design system: double fret border with corner ornaments, vertical title, boxed reign mark, vermilion seals, vertical English name, vertical collection details, centered museum photograph, and a footer with dimensions plus the no-redraw notice. Verify all facts first. The artifact layer must be an unretouched museum photograph; you must invoke an image-generation tool for the paper texture and medallion ground only, and that generated layer must contain no text, seals, borders or objects. Typeset and proofread every Chinese character, render 1024×1536, and package a single offline output.html.", "skills": [], "experts": [ { "id": "tanyuan-heritage-poster", "name": "藏品视觉专家", "description": "把国宝做成竖版典藏海报或明信片，含文物主视觉、意境文案与馆藏落款，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d6fe16acd91e", "expert_id": "ex_9Wv6m472f1Ns" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:01:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-poster-goujian-sword", "title": "越王勾践剑典藏海报", "title_en": "Sword of Goujian: Collector Poster", "subtitle": "青铜千古、越地风华——一柄春秋名剑的水墨典藏视觉", "subtitle_en": "Bronze through the ages, the splendor of Yue — an ink-wash collector visual for a Spring and Autumn masterpiece", "description": "想把一件国宝做成一张典藏海报，关键不只是放大文物，而是让文物特征、字体、意境和信息层级共同形成完整的视觉叙事。\n\n本案例采用用户指定的「越王勾践剑」成品作为最终产物：以水墨山川、古建与江面为意境背景，突出剑身菱形暗格、青铜质感和馆藏信息，书法题字、诗句与篆印共同完成典藏海报表达。专家工作流已新增硬性要求：必须调用生图工具生成主视觉，不能用纯 HTML/CSS 拼一张海报；生图前先锁定真实文物特征，生成后逐项比对；若文物主体难以准确风格化，则转为特展海报、馆藏名片、展览预告等视觉类目，用展览主题、日期与馆藏信息丰富画面，而不是杜撰器物。\n\n适合：博物馆文创与新媒体运营、展览宣传与海报设计者、文博科普内容编辑、历史与美育教师、文博爱好者。", "description_en": "A collector poster that combines an AI-generated ink-wash hero visual with calligraphy, poetic copy, collection facts and a seal; the same visual can also be printed at postcard size. The workflow must invoke image generation for the main visual, constrain the artifact's verified features before generation, and validate them afterward. When close-up stylization cannot preserve the artifact, the expert should switch to an exhibition poster, collection card or event visual rather than inventing a relic.", "categories": [ "creation", "business" ], "tags": [ "典藏海报", "越王勾践剑", "特展海报", "水墨视觉", "博物馆文创" ], "tags_en": [ "Collector Poster", "Sword of Goujian", "Exhibition Poster", "Ink-Wash Visual", "Museum Creative" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "越王勾践剑文物明信片（单文件 HTML）" } ], "prompt": "请以「越王勾践剑」为主题，制作一张可下载、打印或分享的文物明信片/海报，输出单文件 HTML。\n\n## 一、先锁定真实文物特征\n并行调用探元文物库与知识库，至少核对：春秋晚期、湖北江陵望山一号墓出土、湖北省博物馆藏、通长约 55.6cm、短柄、剑首小型圆盘且内铸 11 道同心圆、剑茎两道凸箍、剑格正面蓝色琉璃/背面绿松石、剑身满饰黑色菱形暗格、八字鸟篆铭文仅位于剑身正面靠近剑格处。信息冲突或缺失时标注待确认，不得自行补写。\n\n## 二、必须调用生图工具\n1. 必须实际调用 ImageGen 或同等生图工具生成主视觉底图，禁止仅靠 HTML/CSS/SVG 拼接一张页面冒充生图结果。\n2. 优先使用真实馆藏图作为参考图，设置高参考图保真度；如当前生图工具不支持参考图，不得声称实现了像素级还原，必须在成品说明中标注风格化示意。\n3. 生成内容包含完整文物主体与水墨意境背景；HTML 只负责承载最终生图成品及必要文字，不得用网页元素拼造剑身、纹饰或背景。\n\n## 三、生图提示词约束\n竖版文博典藏视觉，暖白宣纸底，淡青灰水墨山川、江面、古建剪影与飞鸟，大面积留白。中央为完整越王勾践剑正面立姿。严格遵守以下结构：短柄，柄长只占全剑约 15%；剑首是比例克制的小型圆盘，不得画成权杖或大圆帽；剑茎仅两道凸箍；剑格小巧并带蓝色琉璃与青绿色镶嵌；剑身宽阔扁平、近格处较宽、向尖端缓慢收窄；全剑保持笔直；剑身连续分布黑色菱形暗格；鸟篆铭文只在靠近剑格的小区域，不得铺满剑身；青铜原色与青绿色包浆，不得变成金色、银色、黑色武侠剑，不增加剑穗、挂饰或幻想构件。风格化只作用于纸张、背景、光影与色调，不得改变器形、结构、比例、纹饰和铭文位置。生成底图中不要文字、印章、日期和 Logo。\n\n## 四、生成后硬性核验\n把生成图与真实文物图并排检查：短柄比例、剑首、两道凸箍、小型剑格、菱形暗格、铭文位置必须全部通过。任一项不一致立即重新生图，最多 3 次；3 次仍失败时，不得交付失真文物，改做「特展海报 / 馆藏名片 / 展览预告 / 研究成果视觉」：保留真实文物图，将生成能力用于环境背景，并增加展览主题、日期、地点、单元标题等信息丰富画面。\n\n## 五、排版与交付\n参考东方立轴和用户示例：书法主标题「越王勾践剑」，副题「青铜千古·越地风华」，诗句「吴戈越剑 湛卢遗风 / 千年不锈 寒光依旧」，落款「湖北省博物馆藏 / 春秋晚期青铜名剑」。文字使用可商用书法字体或字体转 SVG 路径，并内联到 HTML；逐字检查，不让图片模型生成中文。最终必须是完整的生图主视觉，不是网页临时拼贴；单文件 HTML 内联字体与图片，无外链、无 emoji、无锚点。", "prompt_en": "Create a vertical Sword of Goujian heritage postcard. Verify all artifact facts first. You MUST invoke an image-generation tool for the hero visual; HTML may only package the finished visual and typography, not fabricate the artifact or background with page elements. Constrain the generated sword to the verified short hilt, small disc pommel with concentric circles, two hilt rings, compact inlaid guard, broad tapering blade, continuous dark rhombic pattern, and a small bird-script inscription area near the guard. Compare every generated result with a real collection image and retry up to three times. If accurate stylization fails, switch to an exhibition poster, collection card or event visual using the real artifact image and AI-generated environmental background, adding exhibition theme, date and venue instead of inventing the relic.", "skills": [], "experts": [ { "id": "tanyuan-heritage-poster", "name": "腾讯探元", "description": "把国宝做成竖版典藏海报或明信片，含文物主视觉、意境文案与馆藏落款，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d6fe16acd91e", "expert_id": "ex_9Wv6m472f1Ns" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:01:00+08:00", "updated_at": "2026-08-31T17:22:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-poster-houmuwu-ding", "title": "后母戊鼎典藏海报", "title_en": "Houmuwu Ding Dark-Gold Hero Poster", "subtitle": "国宝重器标签、烫金标题、双侧竖排信息、饕餮拓片底纹——商代青铜重器典藏视觉", "subtitle_en": "National-treasure tag, gilded title, flanking vertical captions, taotie rubbing ground — a Shang bronze collector visual", "description": "这个案例按既定设计规范制作后母戊鼎典藏海报：左上角「国宝重器」方框标签，顶部分隔式导语与回纹纹饰，中央烫金大字标题配双层副题，两侧竖排文物属性与馆藏信息，中央朱印，下部为馆藏实拍方鼎。\n\n执行要点：文物主体只用中国国家博物馆馆藏实拍，仅做抠底与等比缩放，双耳、长方形深腹、四柱足与器身纹饰均保持原貌；生图工具用于生成暖白纸底与饕餮拓片底纹，不参与文物绘制；形制表述已修正为通耳高 133cm、双耳四足。\n\n适合：青铜重器主题社教文创、博物馆纪录片受众、商周青铜爱好者、文博文创产品经理、设计系文化创意方向学生。", "description_en": "A collector poster produced to the established design system: a boxed national-treasure tag, a segmented top lead line with fret ornament, a gilded title with two-tier subtitles, flanking vertical captions, a central seal, and an unretouched National Museum of China photograph of the ding. Image generation supplies only the paper and taotie rubbing ground.", "categories": [ "creation", "business" ], "tags": [ "典藏海报", "商代青铜", "后母戊鼎", "国家博物馆", "博物馆文创" ], "tags_en": [ "Collector Poster", "Shang Bronze", "Houmuwu Ding", "National Museum of China", "Museum Creative" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "后母戊鼎典藏海报" } ], "prompt": "你是「腾讯探元典藏海报专家」，按既定设计规范制作「后母戊鼎」竖版典藏海报。\n\n## 一、事实检索\n并行检索探元文物库与知识库，核对馆藏、出土信息、通耳高 133cm、口长 110cm、口宽 79cm、重 832.84kg、双耳、长方形深腹、四柱足与器身纹饰。形制为双耳四足，不得写成四耳；事实冲突时标注待确认。\n\n## 二、文物主体\n1. 只使用探元 cover 或中国国家博物馆馆藏实拍图，允许抠底、等比缩放与版面定位。\n2. 禁止把鼎交给生图模型重绘、补绘、增删耳足、替换纹饰或再造铭文。\n3. 内壁铭文在正面视角不可见时，不得画到器身外部。\n4. 无可靠实物图时停止制作并说明缺少图源。\n\n## 三、必须调用生图工具\n调用生图工具生成非文物层：暖白到浅米色纸底，中部一枚极淡的商代饕餮/回纹圆形拓片印记，上下保持大面积留白，画面内不得出现任何文字、印章、边框与立体器物。禁止用纯色或 CSS 渐变代替生图背景。\n\n## 四、版式规范（固定结构）\n- 外框：双线回纹边框 + 四角回纹角花。\n- 左上：方框标签「国宝重器」。\n- 顶部：分隔式导语「中 / 国 / 青 / 铜 / 文 / 明 / 瑰 / 宝」+ 左右细线 + 中央回纹纹饰。\n- 中央：烫金渐变大字标题「后母戊鼎」。\n- 标题下：主副题「商代青铜重器」（两侧细线）与「青铜之巅，礼制之威」。\n- 两侧：左竖排文物属性、右竖排文物级别与馆藏。\n- 中央朱印一枚。\n- 下部：馆藏实拍方鼎居中，不遮挡两侧竖排文字。\n- 底部：尺寸重量信息与「文物主体为馆藏实拍，未作形态重绘」。\n\n## 五、质检\n逐项比对实物原图（双耳、深腹、四柱足、器身纹饰、比例），逐字核对标题、副题、标签、印章与落款；确认生图背景中无文字与器物，且文物未遮挡侧栏文字。\n\n## 六、交付\n渲染为 1024×1536 PNG，嵌入单文件 output.html，提供下载与打印按钮，无外链、无 emoji、无锚点。", "prompt_en": "Produce the Houmuwu Ding collector poster to the fixed design system: fret border with corner ornaments, boxed national-treasure tag, segmented top lead line with fret ornament, gilded title, two-tier subtitles with hairlines, flanking vertical captions, central seal, centered museum photograph that must not overlap the side captions, and a footer with dimensions plus the no-redraw notice. Verify the facts first: 133 cm overall height, 110 × 79 cm mouth, 832.84 kg, two ears and four columnar legs. The artifact layer must be an unretouched museum photograph; invoke an image-generation tool only for the paper and faint taotie rubbing ground, which must contain no text, seals, borders or solid objects. Render 1024×1536 and package a single offline output.html.", "skills": [], "experts": [ { "id": "tanyuan-heritage-poster", "name": "藏品视觉专家", "description": "把国宝做成竖版典藏海报或明信片，含文物主视觉、意境文案与馆藏落款，事实源自探元知识库。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d6fe16acd91e", "expert_id": "ex_9Wv6m472f1Ns" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-14T18:01:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tanyuan-sancai-vs-porcelain", "title": "唐三彩为什么不是瓷器", "title_en": "Why Tang Sancai Is Not Porcelain", "subtitle": "从胎料、烧成温度、胎体状态三条门槛，把'是陶还是瓷'讲透", "subtitle_en": "Walk through the three thresholds (clay, firing temperature, body state) to settle the 'pottery or porcelain?' question", "description": "一提到唐三彩很多人下意识就说'它是瓷器'，但博物馆的展签明明写的是'唐三彩'三个字。\n\n这个案例演示了腾讯探元文博知识问答专家如何回答这类'X 为什么不是 Y'的概念辨析题——它会从胎料与配方、烧成温度、胎体状态三个相互关联的维度展开，最终明确归类到'低温铅釉陶'。全文以两件真实唐三彩实物配图说明釉色与露胎细节，并用正文 [1][2][3] 对应文末来源，避免来源名称打断阅读。\n\n适合：博物馆讲解员与志愿者；博物馆爱好者、教师与科普编辑；做文博短视频/公众号需要一段准确科普脚本的创作者；文物拍卖与工艺品行业做内容鉴别的人员。", "description_en": "Most people instinctively call Tang sancai 'porcelain', but museum labels never say that.\n\nThis case shows how the Tencent TanYuan cultural-heritage Q&A expert answers a 'why is X not Y?' question — it first lays out the three objective thresholds for calling something 'porcelain' (clay, firing temperature, body state), then compares Tang sancai against each one, and finally places it cleanly as 'low-fire lead-glazed earthenware'. The full answer is backed by real retrieval data, real numbers (800°C vs 1200°C+), a real artefact (Tang sancai camel at the Nanjing Museum), and a clean-up of the misconception that 'three colours = three colours'.\n\nBest for: Museum docents and volunteers; museum enthusiasts, teachers and popular-science editors; short-video / WeChat creators who need an accurate explainer script; auction and craft-industry content reviewers.", "categories": [ "business", "education" ], "tags": [ "唐三彩", "陶与瓷辨析", "低温铅釉", "概念辨析", "工艺原理" ], "tags_en": [ "Tang Sancai", "pottery vs porcelain", "low-fire lead glaze", "concept disambiguation", "craft principles" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "完整问答报告（结论+分节论证+常见误解+来源）" } ], "prompt": "请回答：唐三彩为什么不是瓷器？\n\n要求：\n1. **结论先行**：第一段直接给出判断（是低温铅釉陶，不是瓷器），一两句话讲清\n2. **立标准 + 逐条比对**：从胎料与配方、烧成温度、胎体状态三个相互关联的维度解释陶与瓷，再用唐三彩逐条比对；不要把材料、温度、吸水率写成彼此独立的绝对门槛\n3. **引用真实数据**：必须基于检索到的资料给出具体数字（800℃ 左右的釉烧 vs 1200℃+ 的瓷器烧成），不要凭印象编造\n4. **澄清常见误解**：点破'三彩不是只有三种颜色'、'三彩的彩与瓷器无关'两点\n5. **区分事实与综合判断**：对学界有定论的（如'瓷器判定标准'）直接讲，对'唐三彩的釉烧温度区间'这类来自不同文献的不同说法标注'综合判断'\n6. **图文与引用**：至少加入 2 张与正文紧密对应的真实文物图；正文只用 [1][2][3] 标注事实，不要反复写“来源于某知识库”；文末统一列明题名、机构、完整 URL 与图片出处\n7. **温度数据要有锚点**：给出约 800℃ 的低温釉烧参照，但不要用未经来源支持的类比\n\n风格：口语化、解释链条完整、对一个完全不懂陶瓷工艺的人也能讲懂；不出现'检索''知识库''查询'等过程叙述词；末尾给出 1 条'下次类似问题可以这样发问'作为延伸引导。", "prompt_en": "Please answer: Why is Tang sancai not considered porcelain?\n\nRequirements:\n1. **Conclusion first**: Open with a one- or two-sentence verdict — it is low-fire lead-glazed earthenware, not porcelain.\n2. **Standards before comparison**: First lay out the three objective thresholds for 'porcelain' — (a) body uses porcelain stone / kaolin instead of ordinary clay; (b) firing temperature high enough to vitrify the body; (c) body is vitrified, dense, very low water absorption with a high-fire glaze firmly fused to it. Then check Tang sancai against each threshold: its body material, its firing temperature, whether the body vitrified, water absorption, and whether lead glaze is a high-fire or low-fire glaze.\n3. **Real data only**: Cite concrete numbers backed by retrieved sources (around 800°C glaze-firing vs 1200°C+ for porcelain). Don't invent figures.\n4. **Debunk common misconceptions**: Clarify that 'three colours' refers to the polychrome family (lead-glaze palette dominated by yellow / green / white), not literally three colours, and that the word 'cai' has nothing to do with porcelain.\n5. **Distinguish established fact from synthesis**: Established consensus (e.g. the porcelain criteria) can be stated directly; for figures that differ across sources (e.g. the exact glaze-firing temperature range of Tang sancai) mark as 'synthesis'.\n6. **Always end with sources**: Two buckets — 'Tencent TanYuan knowledge base' (named by topic only) and 'Public sources' (every citation must carry a clickable URL, otherwise don't claim it as public).\n7. **Anchor unfamiliar numbers**: e.g. 'around 800°C, well below the melting point of steel' so a lay reader can follow.\n\nStyle: colloquial, full cause-and-effect chain, accessible to a reader with zero ceramics background. Never use process narration like 'searching', 'querying the knowledge base', 'retrieved twice'. End with one 'next time, try asking this related question' hook.", "skills": [], "experts": [ { "id": "tanyuan-heritage-qa-expert", "name": "文博百科专家", "description": "依托腾讯探元文博知识库，从工艺原理与材质分析解答文博疑问，结论先行并标注权威资料来源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "97d0c0800a1f", "expert_id": "ex_qSpUhcNlQwWc" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T19:00:29+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tapd-bug-quality-report", "title": "TAPD 缺陷质量分析报告", "title_en": "TAPD Bug Quality Analysis Report", "subtitle": "按严重程度、模块、处理人聚合缺陷，一键出质量趋势报告", "subtitle_en": "Aggregate bugs by severity, module and owner into a quality trend report", "description": "迭代快结束了，缺陷散落在 TAPD 里，谁也说不清这个版本质量到底稳不稳、哪个模块最坑、还有多少高危没闭环——每次做质量复盘都要手动导表格拉数据，费时又容易漏。\n\n这个案例帮你把一个项目或迭代里的全部缺陷自动拉下来，按严重程度、模块、处理人、状态四个维度聚合，再算出缺陷密度、重开率、平均修复时长、遗留缺陷等关键指标，最后生成一份暗色科技感的单文件 HTML 报告：概览卡片、严重程度环形图、近 6 周新增/修复趋势、高发模块 TOP 排行、代表性缺陷明细、质量结论与风险提示一应俱全，双开即看。\n\n适合测试负责人做版本质量评估、研发 Leader 定位高发模块、项目经理做迭代复盘汇报，也适合想给老板一份体面质量报告的你。", "description_en": "The sprint is wrapping up, bugs are scattered across TAPD, and nobody can say for sure whether this release is solid, which module is the biggest pain, or how many critical defects are still open. Every quality retro means manually exporting sheets and crunching numbers by hand.\n\nThis case pulls every bug from a project or iteration, aggregates them across four dimensions — severity, module, owner and status — then computes key metrics like defect density, reopen rate, average fix time and remaining defects. The output is a dark, tech-styled single-file HTML report: overview cards, a severity donut chart, a 6-week new-vs-fixed trend, a top hotspot-module ranking, a representative bug table, plus a quality conclusion and risk section — all viewable in one double-click.\n\nGreat for QA leads assessing release quality, dev leads pinpointing hotspot modules, PMs running sprint retros, and anyone who needs a presentable quality report for their boss.", "categories": [ "dev", "data" ], "tags": [ "TAPD", "缺陷管理", "质量分析", "测试", "数据可视化" ], "tags_en": [ "TAPD", "Bug Tracking", "Quality Analysis", "QA", "Data Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "缺陷质量分析报告" } ], "prompt": "请用 TAPD skill（tapd-openapi / tapd-cli）生成一份「缺陷质量分析报告」单文件 HTML。数据来源：优先用 TAPD 官方接口拉真实数据——调用 bug list / count / fields 拉取 workspaceid={{WORKSPACE_ID}} 下指定迭代内全部缺陷；若无权限或取不到，则用演示数据代替并明确标注「演示数据」。聚合分析：按严重程度、状态、模块、处理人分别聚合，统计近 6 周新增 vs 修复趋势，并计算缺陷密度、重开率、平均修复时长、遗留数、修复率等关键指标，各维度数字须自洽、加总一致。交付：输出单文件 HTML，资源全部内联、可离线双击打开。", "prompt_en": "Use the TAPD skill (tapd-openapi / tapd-cli) to generate a single-file HTML \"Bug Quality Analysis Report\". Data source: prefer real data via TAPD official APIs — call bug list / count / fields to fetch all bugs under workspaceid={{WORKSPACE_ID}} within a specified iteration; if access is unavailable or data cannot be retrieved, use demo data instead and clearly label it as \"demo data\". Aggregation: aggregate by severity, status, module and assignee, compute the new-vs-fixed trend over the last 6 weeks, and calculate key metrics such as defect density, reopen rate, average fix time, open count and fix rate — all numbers must be internally consistent and add up. Delivery: output a single-file HTML with all resources inlined, openable offline by double-click.", "skills": [ { "id": "tapd-openapi", "name": "TAPD", "description": "TAPD 项目管理平台操作（需求、缺陷、任务、迭代、Wiki）", "skill_id": "skill_2053085043504316416" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-27T11:42:00+08:00", "updated_at": "2026-07-27T11:42:00+08:00", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "tapd-iteration-health-dashboard", "title": "TAPD 迭代健康度看板", "title_en": "TAPD Iteration Health Dashboard", "subtitle": "一键汇总迭代进度、燃尽、需求健康度与成员负载", "subtitle_en": "One board for iteration progress, burndown, story health and member load", "description": "迭代到了中后期，进度到底行不行？需求卡在哪、谁过载了、还有几个缺陷没清，光靠翻 TAPD 列表根本看不清，开站会全靠嘴对。\n\n这个案例用 TAPD skill 拉取当前进行中的迭代及其下的需求、任务、缺陷，一次性汇总成「迭代健康度看板」：迭代概览、燃尽图（理想线 vs 实际）、需求状态分布、受阻需求风险、成员负载、缺陷压力、多迭代趋势对比，最后给一个综合健康度评级和行动建议。产物是暗色科技感的单文件 HTML，资源全内联、可离线打开。\n\n适合项目经理、Scrum Master、研发 TL 在每日站会/迭代复盘时快速看清健康度、及早发现延期与阻塞风险。", "description_en": "Mid-to-late in a sprint, is progress actually on track? Which stories are stuck, who is overloaded, how many bugs are still open — scrolling through TAPD lists tells you none of it, and standups turn into guesswork.\n\nThis case uses the TAPD skill to pull the current in-progress iteration with its stories, tasks and bugs, and rolls them up into one Iteration Health Dashboard: iteration overview, burndown (ideal vs actual), story-status distribution, blocked-story risks, member load, defect pressure, multi-iteration trend comparison, plus an overall health rating with action items. The output is a dark tech-styled single-file HTML with all resources inlined, openable offline.\n\nGreat for PMs, Scrum Masters and dev leads to grasp sprint health at a glance during daily standups and retros, catching delays and blockers early.", "categories": [ "dev", "office" ], "tags": [ "TAPD", "敏捷开发", "迭代管理", "燃尽图", "项目看板" ], "tags_en": [ "TAPD", "Agile", "Iteration Management", "Burndown", "Project Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "迭代健康度看板" } ], "prompt": "用 TAPD skill（tapd-openapi / tapd-cli）拉取当前进行中的迭代及其下的需求、任务、缺陷，workspaceid 用占位符 {{WORKSPACE_ID}}，运行前替换为你的项目 id；无权限时改用演示数据并标注「演示数据」。把数据汇总成「迭代健康度看板」，含迭代概览、燃尽图（理想线 vs 实际剩余工作量）、需求状态分布、成员负载与过载识别、受阻风险、缺陷压力及综合健康度评级与建议，燃尽/完成率/工作量等数字需自洽。最终输出单文件 HTML，资源全内联、可离线打开。", "prompt_en": "Use the TAPD skill (tapd-openapi / tapd-cli) to pull the current in-progress iteration with its stories, tasks and bugs; use placeholder {{WORKSPACE_ID}} for workspaceid and replace it with your project id before running. If you lack access, use demo data clearly labeled as \"demo data\" on the page. Roll it up into an Iteration Health Dashboard covering the iteration overview, burndown (ideal vs actual remaining effort), story-status distribution, member load with overload detection, blocked-story risks, defect pressure, plus an overall health rating with action items; keep the burndown, completion-rate and effort numbers internally consistent. Output a single-file HTML with all resources inlined, openable offline.", "skills": [ { "id": "tapd-openapi", "name": "TAPD", "description": "TAPD 项目管理平台操作（需求、缺陷、任务、迭代、Wiki）", "skill_id": "skill_2053085043504316416" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-27T11:42:00+08:00", "updated_at": "2026-07-27T11:42:00+08:00", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "tapd-story-review-board", "title": "TAPD 需求评审看板一键生成", "title_en": "One-Click TAPD Story Review Board", "subtitle": "把一个迭代的需求与变更历史整理成可视化评审材料", "subtitle_en": "Turn an iteration's stories and change history into a visual review board", "description": "开评审会前，产品经理经常要在 TAPD 里挨个点开需求，手动整理优先级、状态、验收标准和这段时间改了什么，一个迭代十几条需求折腾大半天还容易漏掉变更风险。\n\n这个案例用 TAPD skill 拉取指定迭代下的全部需求（story list）和需求变更历史（story changes），自动整理成一份需求评审看板：顶部是需求总数和各状态概览，中间用环形图 / 条形图展示按优先级、状态、模块的分布，下面是完整需求明细表，还单独把近期发生优先级 / 范围变更的需求拎出来做「变更风险区」，最后附评审 checklist 和结论区。产物是单文件 HTML，暗色科技感，可离线打开。\n\n适合产品经理、项目经理、迭代负责人在需求评审会前快速准备材料，也适合团队 leader 一眼掌握迭代内需求全貌。", "description_en": "Before a review meeting, product managers often have to open each story in TAPD one by one, manually collecting priorities, statuses, acceptance criteria and what changed recently. A dozen stories in an iteration can take half a day, and change risks are easily missed.\n\nThis case uses the TAPD skill to pull all stories (story list) and story change history under a given iteration, then automatically compiles a review board: an overview of total story count and status breakdown at the top, donut / bar charts showing distribution by priority, status and module, a full story detail table below, plus a dedicated \"change risk\" panel that surfaces stories with recent priority or scope changes, and finally a review checklist and conclusion section. The output is a single-file HTML with a dark tech aesthetic that opens offline.\n\nIdeal for product managers, project managers and iteration owners preparing materials before a review meeting, and for team leads who want a full picture of the iteration's requirements at a glance.", "categories": [ "dev", "office" ], "tags": [ "TAPD", "需求管理", "项目管理", "评审", "看板" ], "tags_en": [ "TAPD", "Requirement Management", "Project Management", "Review", "Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "需求评审看板" } ], "prompt": "请用 TAPD skill 通过官方接口，拉取指定迭代下的全部需求（story list）及其变更历史（story changes），我的 workspace id 是 {{WORKSPACE_ID}}；若暂无 TAPD 访问权限，可用一批虚构演示数据生成，但须在页面明确标注为演示数据、不暗示真实。\n\n请把这些需求整理成一份需求评审看板，包含：概览（项目、迭代、评审日期、需求总数与各状态数量）、按优先级 / 状态 / 模块的分布、需求明细（ID / 标题 / 模块 / 优先级 / 状态 / 负责人 / 工作量 / 验收标准）、近期发生优先级或范围变更的变更风险区、以及评审 checklist 与结论。\n\n输出为单文件 HTML，可离线直接打开。", "prompt_en": "Please use the TAPD skill to pull, via the official API, all stories (story list) and their change history (story changes) under a given iteration. My workspace id is {{WORKSPACE_ID}}; if I don't currently have TAPD access, you may generate with a set of fictional demo data, but it must be clearly labeled as demo data on the page and not imply it is real.\n\nOrganize these stories into a story review board covering: an overview (project, iteration, review date, total story count and per-status counts), distribution by priority / status / module, story details (ID / title / module / priority / status / owner / effort / acceptance criteria), a change-risk zone for stories with recent priority or scope changes, and a review checklist with the verdict.\n\nOutput as a single-file HTML that opens offline directly.", "skills": [ { "id": "tapd-openapi", "name": "TAPD", "description": "TAPD 项目管理平台操作（需求、缺陷、任务、迭代、Wiki）", "skill_id": "skill_2053085043504316416" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-07-27T11:42:00+08:00", "updated_at": "2026-07-27T11:42:00+08:00", "cover_image_en": "cover-en.png", "preview_en": "output.html", "is_recommended": false, "recommended_order": 0 }, { "id": "taxue-aroma-delivery-review", "title": "投放复盘报告一键生成", "title_en": "One-Click Ad Campaign Retrospective Report", "subtitle": "投完一波广告，把后台数据变成一份能下载打印的复盘文档", "subtitle_en": "Turn raw campaign data into a downloadable, print-ready retrospective", "description": "广告投完了，钱花了、单出了，但「这波到底打得怎么样、下次该改哪里」往往说不清——数据散在后台，看得懂的人才看得懂，想给老板或客户交一份说明也无从下手。\n\n这个案例把一整套投手的复盘方法装进一次对话：给出渠道、周期、花了多少钱、换来多少结果和当初的目标，就能得到一份完整报告。顶部先用大白话讲清「总账 / 做对了什么 / 还能更好 / 下次这样投」四段结论，往下接六节数据文档——表现总览与成交漏斗、逐日走势折线图（标出爬坡·高峰·衰退三段与成本拐点）、阶段对比与效率缺口测算、素材分析（结构·排行·建议）、五维星级评价、下一波建议与目标对照。术语全部翻成人话，好坏只跟自己的目标比，没有的数据如实标注不估算。\n\n适合：投完一波想认真总结的广告主、需要给客户交复盘文档的代理商投手、刚接手投放要快速看懂账户表现的运营新人。", "description_en": "When a campaign wraps up, the money is spent and the orders are in — but \"how did we actually do, and what should change next time?\" usually goes unanswered. The data sits scattered in a dashboard that only specialists can read, and putting together an explanation for a boss or client feels like starting from scratch.\n\nThis case packs a full media buyer's retrospective method into a single conversation. Provide the channel, date range, spend, results, and your original target, and you get a complete report. It opens with four plain-language conclusions — the bottom line, what worked, what could be better, and how to run it next time — followed by six sections of data: a performance overview with conversion funnel, a daily trend line chart (marking the ramp-up, peak, and decline phases plus the cost inflection point), phase-by-phase comparison with efficiency gap analysis, creative analysis (structure, ranking, recommendations), a five-dimension star rating, and next-round recommendations with target benchmarks. Jargon is translated into everyday language, performance is judged only against your own goals, and missing data is labeled as such rather than estimated.\n\nGood for: advertisers who want a real post-campaign review, agency buyers who owe clients a retrospective document, and marketers new to ad operations who need to understand account performance fast.", "categories": [ "business" ], "tags": [ "广告", "投放复盘", "数据分析", "报告生成" ], "tags_en": [ "Advertising", "Campaign Retrospective", "Data Analysis", "Report Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "踏雪香薰投放复盘报告" } ], "prompt": "投完一波广告，帮我出一份完整的、能下载打印的复盘报告（HTML）。\n\n【基本情况】\n- 品牌 / 产品：{品牌或店铺名}，{产品和客单价，如「东方香氛香薰蜡烛，168-588 元」}\n- 渠道：{腾讯广告朋友圈+视频号 / 抖音巨量 / 小红书 / 千川 / 快手 / 百度}\n- 周期：{起始日期} 至 {结束日期}，已结案\n- 考核：主要要 {订单 / 线索 / 涨粉}，当初想控制在 {一个多少钱} 以内\n\n【数据】（没有的直接写「无」，不用估算）\n- 总花费 {金额}｜结果数 {订单或线索数}｜成交额 {GMV}\n- 曝光 / 点击：{曝光量} / {点击量}\n- 逐日明细：{每天的日期、消耗、结果数}\n- 素材明细：{每条的名称、视频还是图片、卖点、消耗、点击率、结果数}\n- 中间调整：{第几天加过预算 / 换过素材 / 改过出价}\n\n【产出】\n顶部先用大白话给四段结论（这波的总账 / 做对了什么 / 还能更好的地方 /【下次这样投】行动清单），下面接六节数据文档：01 表现总览 + 成交漏斗｜02 逐日走势折线图（标出爬坡·高峰·衰退三段和成本拐点）｜03 阶段对比 + 效率缺口测算｜04 素材分析（结构 / 排行 / 建议三部分）｜05 五维星级综合评价｜06 下一波建议 + 目标对照。\n\n写到 outputs/ 并打开预览。", "prompt_en": "My ad campaign has wrapped up. Please produce a complete, downloadable and print-ready retrospective report (HTML).\n\n[Basics]\n- Brand / product: {brand or shop name}, {product and price range, e.g. \"Eastern-aesthetic scented candles, 168-588 RMB\"}\n- Channel: {Tencent Ads WeChat Moments + Channels / Douyin Ocean Engine / Xiaohongshu / Qianchuan / Kuaishou / Baidu}\n- Date range: {start date} to {end date}, campaign closed\n- Target: primarily {orders / leads / followers}, originally aiming to stay under {target cost per result}\n\n[Data] (write \"none\" for anything you don't have — do not estimate)\n- Total spend {amount} | Results {orders or leads} | Revenue {GMV}\n- Impressions / clicks: {impressions} / {clicks}\n- Daily breakdown: {date, spend, results for each day}\n- Creative breakdown: {name, video or image, selling point, spend, CTR, results for each}\n- Mid-flight changes: {which day you raised budget / swapped creatives / adjusted bids}\n\n[Deliverable]\nOpen with four plain-language conclusions (the bottom line / what worked / what could be better / a numbered \"run it this way next time\" action list), then six data sections: 01 Performance overview + conversion funnel | 02 Daily trend line chart (marking ramp-up, peak and decline phases plus the cost inflection point) | 03 Phase comparison + efficiency gap analysis | 04 Creative analysis (structure / ranking / recommendations) | 05 Five-dimension star rating | 06 Next-round recommendations + target benchmarks.\n\nWrite it to outputs/ and open a preview.", "skills": [], "experts": [ { "id": "ad-delivery-allrounder", "name": "艾投·金手指AI广告投手", "description": "用大白话帮普通人把广告从建到复盘全程管到尾。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "01193abea1bc", "expert_id": "ex_EYovtHW9hlEA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T08:00:00Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "taxue-aroma-material-strategy", "title": "香薰新品朋友圈素材策略", "title_en": "WeChat Moments Creative Strategy for a Scented Candle Launch", "subtitle": "8 万预算 30 天，把「我要推广」拆成可开工的素材清单", "subtitle_en": "Turn an 80K budget over 30 days into a ready-to-execute creative checklist", "description": "第一次投广告最卡的不是钱，是不知道该做多少条素材、做哪几类、每条配什么文案。凭感觉做十几张图投出去，跑不动也不知道问题在哪。\n\n这个案例把一句「我要推广香薰产品踏雪，预算 8 万投 30 天，希望获客成本 100 元以内」，直接算成一份可开工的素材策略：按日预算倒推出广告数与素材数，推荐适配的素材类型，逐条给出尺寸、文案和可直接出图的生图提示词，最后合成一份离线可看、可打印的单文件 HTML 手册。\n\n适合：第一次投效果广告的广告主、一人公司/独立品牌主理人、需要快速出素材 brief 的运营与设计。", "description_en": "The hardest part of a first ad campaign isn't the budget — it's not knowing how many creatives to make, which formats to use, or what copy goes with each one. Ship a dozen images on instinct and, when they underperform, you have no idea why.\n\nThis case turns a single sentence — \"I want to promote my scented candle brand Taxue, 80,000 budget over 30 days, target CPA under 100\" — into a ready-to-execute creative strategy. It back-calculates ad and creative counts from the daily budget, recommends the creative formats that fit, then specifies dimensions, copy, and ready-to-use image-generation prompts for every single item, delivered as a single-file offline HTML manual.\n\nGood for: advertisers running performance ads for the first time, one-person companies and indie brand founders, and marketers or designers who need a creative brief fast.", "categories": [ "business" ], "tags": [ "广告", "营销", "推广", "素材策略" ], "tags_en": [ "Advertising", "Marketing", "Promotion", "Creative Strategy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "踏雪香薰素材策略与执行手册" } ], "prompt": "我想推广我的香薰产品\"踏雪\"，主打东方美学与情绪疗愈，材料选用天然手工大豆蜡与蜂蜡混合配方，搭配木质烛芯，瓶身采用冰川质感玻璃或微粗糙感素烧陶瓷，兼具空间摆件属性与收藏价值。我的预算是80000元，广告想投放30天，目标是吸引人直接下单购买，每个下单的新用户，获客成本希望在100元以内，想主投朋友圈渠道，请帮我产出素材策略。", "prompt_en": "I want to promote my scented candle product \"Taxue\" (Treading Snow), centered on Eastern aesthetics and emotional healing. It's made from a natural handcrafted blend of soy wax and beeswax with a wooden wick; the vessel uses glacier-textured glass or slightly coarse unglazed ceramic, serving as both a decorative object and a collectible. My budget is 80,000 RMB over 30 days of ad delivery. The goal is to drive direct purchases, with a target customer acquisition cost under 100 RMB per new paying user. I want to focus on WeChat Moments as the main channel. Please produce a creative material strategy for me.", "skills": [], "experts": [ { "id": "material-manager-suce", "name": "苏策·金手指素材经理", "description": "手把手带你做出可以直接投放的效果广告素材。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c59eafdb16dd", "expert_id": "ex_Y6SygSNKVerC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T06:44:16Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tcm-guizhi-formula-breakdown", "title": "桂枝汤方解思维导图", "title_en": "Guizhi Decoction Formula Breakdown Mind Map", "subtitle": "一张图看懂桂枝汤的君臣佐使配伍与作用机理", "subtitle_en": "Understand the sovereign-minister-assistant-courier compatibility and mechanism of Guizhi Decoction in one diagram", "description": "学方剂最头疼的就是面对一长串药名和君臣佐使的复杂关系，教材翻来翻去脑子还是一团浆糊。\n\n这个案例以《伤寒论》经典方「桂枝汤」为例，输入方剂组成和配伍信息，一键生成层次清晰的方解思维导图——君药桂枝解肌发表、臣药芍药益阴敛营、佐药姜枣调和脾胃、使药炙甘草调和诸药，功效、主治、配伍特点、用法一图尽览。\n\n适合：中医学学生复习备考、中医爱好者自学方剂、方剂研究者梳理配伍逻辑、中医教学工作者制作讲义。", "description_en": "The most frustrating part of learning herbal formulas is facing a long list of ingredients with complex sovereign-minister-assistant-courier relationships, flipping through textbooks without a clear mental picture.\n\nThis case takes the classic Shang Han Lun formula \"Guizhi Decoction\" as an example—input the composition and compatibility data, and generate a well-structured formula breakdown mind map in one click: sovereign herb Guizhi releases the exterior, minister herb Shaoyao nourishes yin and secures the defensive, assistants Jiangzao harmonize spleen and stomach, and courier Zhigancao mediates all herbs—covering functions, indications, compatibility features, and usage all in one view.\n\nIdeal for: TCM students preparing for exams, TCM enthusiasts self-studying formulas, formula researchers analyzing compatibility logic, and TCM educators creating teaching materials.", "categories": [ "education", "productivity" ], "tags": [ "方解", "思维导图", "桂枝汤", "伤寒论", "中药" ], "tags_en": [ "formula breakdown", "mind map", "Guizhi Decoction", "Shang Han Lun", "Chinese herbal medicine" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为「桂枝汤」做一张方解思维导图（SVG）。组成：桂枝 9g（君）、芍药 9g（臣）、生姜 9g（佐）、大枣 3 枚（佐）、炙甘草 6g（使）。功效：解肌发表、调和营卫。主治：外感风寒表虚证（恶风、发热、汗出、脉浮缓）。用法：水煎温服，服后啜热稀粥，温覆取微汗。来源：《伤寒论》。", "prompt_en": "Please create a formula breakdown mind map (SVG) for \"Guizhi Decoction\". Composition: Guizhi (Cinnamon Twig) 9g (Sovereign), Shaoyao (White Peony Root) 9g (Minister), Shengjiang (Fresh Ginger) 9g (Assistant), Dazao (Jujube) 3 pieces (Assistant), Zhigancao (Honey-fried Licorice Root) 6g (Courier). Functions: Release the exterior and resolve the muscle layer, harmonize the defensive and nutritive qi. Indications: External wind-cold with exterior deficiency pattern (aversion to wind, fever, sweating, floating and moderate pulse). Usage: Decoct in water and take warm; after taking, sip hot thin congee, cover with blankets to induce slight sweating. Source: Shang Han Lun (Treatise on Cold Damage).", "skills": [ { "id": "tcm-formula-breakdown-visual", "name": "方剂方解结构化可视化", "description": "将中药方剂的结构化信息（组成、配伍角色、功效、主治、用法等）渲染为层次清晰的SVG思维导图", "skill_id": "skill_2085355033552953344" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "usage_count": 0, "created_at": "2026-08-04T06:44:57.000Z", "updated_at": "2026-08-04T06:44:57.000Z", "artifacts": [ { "file": "output.html", "label": "桂枝汤方解思维导图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "teacher-semester-start-workbench", "artifact_type": "html", "title": "高校教师开学任务清零工作台", "title_en": "Semester-Start Task Clearing Workbench", "subtitle": "五大板块 · 逾期预警 · 清零进度一屏统管，开学季冲刺在线同步", "subtitle_en": "Five hubs, overdue alerts and clearing progress on one screen — sprint through semester start with live sync", "description": "高校老师一到开学就被备课、组会、课题、行政事务追着跑：事情散在群通知、邮箱和脑子里，谁快到期、谁已经逾期全靠记性。这个工作台把教学、科研、指导学生、事务、成长五大板块的任务放进同一张表，逾期标红、今天到期高亮、总体清零进度一屏看清。演示数据是一位高校旅游管理教授开学季的 20 项任务，从第一周课件、组会安排到课题申报、学者档案全覆盖。适合高校教师、研究生导师，以及任何想用一张表管住开学冲刺季的人。", "description_en": "When a new semester starts, college professors get chased by course prep, group meetings, grant deadlines and admin paperwork — everything scattered across group chats and inboxes, with due dates living only in memory. This workbench puts all tasks from five hubs (teaching, research, student mentoring, affairs, growth) into one table: overdue items turn red, due-today items get highlighted, and the overall clearing progress is visible at a glance. The demo data covers 20 real-to-life tasks of a tourism-management professor during semester start, from Week 1 courseware to grant applications and scholar profiles. Great for college teachers, graduate advisors, and anyone wrangling a semester-start sprint with one table.", "tags": [ "高校教师", "开学季", "任务管理", "教学科研", "进度跟踪" ], "tags_en": [ "college teaching", "semester start", "task management", "teaching & research", "progress tracking" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/SzymywKLxiqKbzNOk2Nmu1，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个高校教师的开学任务清零工作台，浅色学术风、蓝紫渐变主色，面向高校教师与研究生导师，用一张「高校教师开学任务清零工作台」数据表管住开学季全部待办。表字段：事项名称、所属中枢（教学中枢/科研中枢/指导学生中枢/事务中枢/成长中枢）、所属模块（备课进度/作业批改/考核安排/文献管理/论文写作/课题申报/选题管理/组会管理/日程管理/材料整理/成果展示/学术影响力）、状态（未开始/进行中/已完成/已搁置）、进度、优先级（P0/P1/P2）、截止日期、关联对象、关键指标、下一步行动。页面包含：顶栏同步状态；欢迎卡（教授问候与场景日期）；清零总览（总体进度环形图、全部/进行中/已完成/已逾期四张可点击筛选的统计卡、「今天要处理」区按逾期天数排序、可折叠的添加任务表单，选中选题模块时显示脱敏提示）；五大板块完成情况环形卡片，点击跳转对应清单；事项清单（板块页签 + 模块筛选 + 按分组展示的任务卡，支持改状态、改进度、删除，进度满 100 自动完成）；侧栏板块导航与清空数据入口。内置 20 条开学季演示任务，所有增删改实时写回数据表。PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/SzymywKLxiqKbzNOk2Nmu1 with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows:/n/nBuild a semester-start task clearing workbench for college professors, in a light academic style with a blue-purple gradient, for college teachers and graduate advisors, managing every semester-start to-do with one \"Semester-Start Task Clearing\" table. Fields: task name, hub (Teaching / Research / Student Mentoring / Affairs / Growth), module (Course prep / Grading / Assessments / Literature / Paper writing / Grant proposals / Topic mgmt / Group meetings / Scheduling / Paperwork / Achievements / Academic presence), status (Not started / In progress / Done / On hold), progress, priority (P0/P1/P2), due date, related target, key metric, next action. The page includes: a header with sync status; a welcome card (professor greeting with scenario date); a clearing overview (overall progress ring, four clickable stat cards for all / in progress / done / overdue, a \"to handle today\" list sorted by overdue days, a collapsible add-task form with a privacy note when the topic module is selected); five hub ring cards that jump to the hub's list; a task list (hub tabs + module filter + grouped task cards supporting status change, progress edit and delete, auto-complete at 100% progress); a sidebar with hub navigation and clear-all entry. Ships with 20 semester-start demo tasks; every create, update and delete is written back to the table in real time. responsive on PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "高校教师开学任务清零工作台（自包含演示）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "张河清教授", "name_en": "Prof. Zhang Heqing", "avatar": "", "badge": "community" }, "source": "community", "skill_id": null, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 98, "created_at": "2026-09-04T15:30:00+08:00", "updated_at": "2026-09-04T16:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "teacher-workbench", "title": "教师工作台", "title_en": "Teacher Workspace", "subtitle": "把课堂记录、成绩分析与班务安排集中到一个在线工作台", "subtitle_en": "One online workspace for classroom records, grade analysis, and homeroom routines", "description": "老师在课堂上要快速记下学生表现、打卡和成绩，课后又要整理待办、补测、家校沟通和班务，信息常常分散在纸张、表格和聊天记录里。\n\n这个案例演示了一个教师可直接使用的在线工作台：今天课程与待办、快速批量记录、完成率和成绩概览、学生名册，以及考勤、家校联系、沟通和班务都围绕同一套数据表展开。\n\n适合：小学及初中教师管理多个班级的日常课堂记录；班主任协调考勤、家校沟通和班务；教研团队展示可持续维护的教学数据工作流。", "description_en": "Teachers need to capture participation, check-ins, and grades during class, then organize tasks, retests, family communication, and homeroom routines after class. Those details are often scattered across paper, spreadsheets, and messages.\n\nThis case demonstrates an online workspace teachers can use directly: today’s classes and tasks, rapid batch recording, completion and grade insights, a student roster, plus attendance, family contacts, communication, and class duties built around one connected set of data tables.\n\nSuitable for: primary and secondary teachers managing daily records across multiple classes; homeroom teachers coordinating attendance, family communication, and duties; and teaching teams demonstrating a maintainable classroom-data workflow.", "categories": [ "workspace" ], "tags": [ "教师工作台", "课堂记录", "成绩分析", "班务管理" ], "tags_en": [ "Teacher Workspace", "Classroom Records", "Grade Analysis", "Homeroom Management" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "完整交互工作台" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/FCtGCs9SZKxZEWrQfS29BE?source=2，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n搭建一个给中国中小学教师长期使用的「教师工作台」，支持一位教师管理多个班级。目标是在课堂上 30 秒内完成一轮学生记录，课后直接查看汇总、成绩和待办。\n\n建立并关联 10 张在线表：学生表、课表、项目表、学业记录表、表现记录表、待办表、考勤表、家长联系表、沟通记录表、班务表。所有汇总都由原始记录即时计算，新增、修改、删除需要写回资料库并在页面中同步。\n\n页面包含：\n1. 今日：按当天星期展示课程、未完成待办和待补测学生；\n2. 记一笔：选择班级、日期和项目后，批量选择学生记录完成、成绩、过关或表现分；\n3. 汇总：展示完成率、最近考试均分、及格率、表现趋势和待补测；\n4. 成绩：按班级和考试项目展示学生最新有效成绩；\n5. 学生：按班级浏览、搜索学生及标签；\n6. 班主任：展示考勤异常、家长联系、沟通记录和班务安排。\n\n界面使用浅色、清爽、高可读性的教师办公风格；桌面端为宽屏多栏布局，手机端自动单列；中文为默认语言，并提供完整英文界面。示例数据必须匿名化，家长电话、备注和沟通内容只出现在班主任相关页面。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to make a one-click copy of https://www.workbuddy.cn/space/d/FCtGCs9SZKxZEWrQfS29BE?source=2 with online data storage and bidirectional sync. Prioritize the one-click replication flow; if that is unavailable, build it as follows: Build a long-term Teacher Workspace for primary and secondary school teachers in China. One teacher should be able to manage multiple classes. The goal is to complete a round of student records in 30 seconds during class, then review summaries, grades, and tasks after class. Create and connect ten online tables: Students, Schedule, Projects, Academic Records, Behavior Records, Tasks, Attendance, Family Contacts, Communication Records, and Class Duties. Calculate every summary from raw records in real time. Creates, updates, and deletes must write back to Library and synchronize in the page. Include:\n1. Today: daily schedule, unfinished tasks, and students pending retest;\n2. Quick Record: choose a class, date, and project, then select students to record completion, scores, passing status, or behavior points in batches;\n3. Insights: completion rate, latest exam mean, pass rate, behavior trends, and pending retests;\n4. Grades: latest valid score for each student by class and assessment;\n5. Roster: browse and search students by class and tags;\n6. Homeroom: attendance exceptions, family contacts, communication records, and class duties. Use a light, clear, highly readable teacher-workflow style. Use a wide multi-column desktop layout and a responsive single-column mobile layout. Chinese is the default language and the English interface must be complete. Demo data must be anonymized; family phone numbers, notes, and communication content may appear only on homeroom-related pages. Deliver a single HTML file Online storage and bidirectional sync must remain the only data product model.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929" } ], "mcps": [], "creator": { "name": "AI康康老师", "name_en": "AI Kangkang Teacher", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "community", "usage_count": 0, "created_at": "2026-08-26T14:22:15+08:00", "updated_at": "2026-08-26T14:22:15+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "team-weekly-report", "title": "带团队必备：一键出销售团队管理周报", "title_en": "Team Weekly Report", "subtitle": "业绩概览、成员排名、管理建议一键成报", "subtitle_en": "Performance overview, member ranking and management advice — one click to a weekly report", "description": "接入或上传团队业绩数据，自动生成业绩概览、成员排名和管理建议周报，团队管理一目了然。", "description_en": "Connect or upload team performance data and it generates a weekly report with a performance overview, member rankings and management suggestions, giving team leads full clarity.", "categories": [ "data" ], "tags": [ "团队管理", "业绩周报", "销售管理", "B2B贸易" ], "tags_en": [ "Team Management", "Weekly Report", "Sales Management", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "销售团队管理周报（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易运营专家。请帮我生成本团队的经营管理周报。\n\n【数据来源】\n- 我已绑定业务系统：读取{周期，如 2026-08-18~08-24}团队与成员业绩数据；\n- 我上传了业绩表：请读取并映射到成交量、销售额、毛利、回款、新客、目标达成等口径；\n- 我暂时没有数据：说明需要哪些字段并给示例周报。\n\n【要求】\n1. 团队业绩概览：本周成交量、销售额、毛利、回款率、周目标达成，附环比；\n2. 成员业绩排名表：排名、业务员、成交量、销售额、新客数、目标达成率；\n3. 管理建议：查询【行业信息资产库】里这个行业的团队管理口径，给出标杆复制、重点帮扶、下周目标、风险跟进4类建议。\n\n若数据源未就绪或返回不足，请以通用团队管理方法论为主产出并说明。请生成完整团队管理周报。", "skills": [ { "id": "team-management", "name": "团队管理", "description": "TradeBuddy把“老板想看团队每个人干得怎么样、谁该提醒”的诉求，做成客观、可追溯、不伤人的人效视图。数据带样本量与口径说明，仅老板/管理员可见。", "skill_id": "skill_2091156624101240832" } ], "experts": [ { "id": "yunying-expert", "name": "B2B贸易运营专家", "description": "懂B2B贸易的运营专家：面向老板与管理员，做生意体检、看团队人效、批量扫风险、沉淀企业知识。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_D7992fbSsXNx" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tech-conference-landing", "title": "科技大会活动落地页", "title_en": "Tech Conference Landing Page", "subtitle": "倒计时 + 演讲者阵容 + 视差滚动，打造品牌氛围感", "subtitle_en": "Countdown + speaker lineup + parallax scroll for brand atmosphere", "description": "活动页设计往往面临信息堆砌、缺乏品牌调性的困境，导致用户停留时间短、报名转化低。本案例以 Apple + Mistral AI 极简美学为灵感，打造一个深色沉浸式科技大会落地页：实时倒计时营造紧迫感、演讲者卡片悬浮发光彰显科技质感、视差滚动引导用户自然浏览至报名表单。适合活动运营、品牌市场团队、独立开发者快速生成高转化率的品牌级活动页。", "description_en": "Event pages often suffer from information overload and lack of brand identity, leading to low engagement and poor registration conversion. This case draws inspiration from Apple + Mistral AI minimalist aesthetics to create a dark, immersive tech conference landing page: real-time countdown creates urgency, speaker cards with hover glow effects convey tech sophistication, and parallax scrolling naturally guides users to the registration form. Ideal for event operations, brand marketing teams, and indie developers who need a high-conversion branded event page quickly.", "categories": [ "business", "creation" ], "tags": [ "活动页", "科技大会", "视差滚动", "品牌设计", "报名转化" ], "tags_en": [ "Event Page", "Tech Conference", "Parallax", "Brand Design", "Registration" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "科技大会落地页" } ], "prompt": "请生成一个 AI 科技大会活动落地页（单文件 HTML，内联无外部依赖），Apple + Mistral 极简风格，大量留白、高端科技氛围。\n\n模块：\n1. Hero：大会名「AI Frontier 2026」+ 副标题「探索 AI 的下一个前沿」+ 日期地点 + 实时倒计时 + 报名 CTA\n2. 演讲者阵容：6 位嘉宾（姓名/职位/公司）\n3. 日程：时间线布局，分上午/下午\n4. 亮点数字：参会者 2000+ / 演讲 30+ / Workshop 12 / 合作伙伴 50+（滚动数字递增）\n5. 报名表单：姓名/邮箱/公司/职位 + 基础校验\n6. Footer：社交图标 + 版权", "prompt_en": "Please generate a landing page for an AI tech conference (single-file HTML, inlined with no external dependencies), in an Apple + Mistral minimalist style, with lots of whitespace and a high-end tech atmosphere.\n\nSections:\n1. Hero: conference name \"AI Frontier 2026\" + subtitle \"Exploring AI's Next Frontier\" + date and location + a live countdown + a registration CTA\n2. Speaker lineup: 6 guests (name / title / company)\n3. Agenda: a timeline layout, split into morning / afternoon\n4. Highlight numbers: attendees 2000+ / talks 30+ / Workshops 12 / partners 50+ (numbers counting up on scroll)\n5. Registration form: name / email / company / title + basic validation\n6. Footer: social icons + copyright", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 8, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-05-03T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": true, "recommended_order": 17 }, { "id": "tech-service-review", "title": "技术服务合同风险审查", "title_en": "Tech Service Contract Risk Review", "subtitle": "锁定知识产权归属、验收标准与延期赔付的关键漏洞", "subtitle_en": "Identify key weaknesses in IP ownership, acceptance criteria, and delay compensation", "description": "技术服务合同的核心在于交付成果的定义与验收标准的量化。本案例演示专家如何从企业数字化转型咨询+系统开发服务合同中，精准定位知识产权归属不明确、服务需求明细待填写、分期付款未绑定里程碑节点、保密义务赔偿未量化等高风险点，确保甲方投入的安全性与乙方权益的平衡。\n\n适合：CTO/CIO、技术采购负责人、SaaS产品经理、创业公司创始人。", "description_en": "The core of tech-service contracts lies in defining deliverables and quantifying acceptance criteria. This case demonstrates how the expert pinpoints high-risk issues in an enterprise digital-transformation + system-development agreement: undefined IP ownership, pending service details, milestone-unbound staged payments, unquantified confidentiality damages.\n\nSuitable for: CTOs/CIOs, tech-procurement leads, SaaS PMs, startup founders.", "categories": [ "business" ], "tags": [ "知识产权", "验收标准", "付款里程碑" ], "tags_en": [ "IP Ownership", "Acceptance Criteria", "Payment Milestones" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请对我提供的《技术服务合同》进行深度风险审查。重点关注知识产权归属条款是否明确（源代码、文档、数据）、各阶段交付物与验收标准是否量化具体、四期付款是否绑定明确的里程碑节点（原型确认/上线验收/质保期满）、延期赔偿比例是否合理且设有上限、保密义务范围与违约赔偿金额是否量化。输出高/中/低分级风险清单。", "prompt_en": "Please perform a deep risk review on my Technical Service Agreement. Focus on IP-ownership clarity (source code, docs, data), quantified deliverables & acceptance criteria per phase, whether 4-stage payments bind clear milestones, reasonableness of delay damages with caps, and whether confidentiality scope & breach damages are quantified. Output a tiered risk list.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T16:56:11+08:00", "updated_at": "2026-08-13T18:08:23+08:00", "artifacts": [ { "file": "output.html", "label": "基于 V1.3 真实审查任务的互动审查页面" }, { "file": "技术服务合同_完整审查报告_V1.6.html", "label": "可视化商务版完整审查报告（HTML）" }, { "file": "技术服务合同_完整审查报告_V1.3.md", "label": "完整审查报告数据版（Markdown）" }, { "file": "技术服务合同_审查批注_V1.3.docx", "label": "腾讯电子签官方导出的 V1.3 审查批注合同" }, { "file": "技术服务合同_审查摘要_V1.3.xlsx", "label": "腾讯电子签官方导出的 V1.3 审查摘要 Excel" }, { "file": "真实审查结果_V1.3.json", "label": "腾讯电子签 V1.3 审查任务原始返回数据" }, { "file": "技术服务合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-campus-interview-prep", "title": "腾讯校招面试准备", "title_en": "Tencent Campus Interview Preparation", "subtitle": "围绕目标岗位生成面试准备计划和模拟追问清单", "subtitle_en": "Create interview preparation plans and follow-up question lists for target roles.", "description": "痛点：同学常把面试准备做成题库背诵，忽略岗位 JD、项目表达和真实追问逻辑。\n演示内容：本案例以算法-多模态方向官网 JD 为输入，结合腾讯校招面试辅导材料，输出 1v1 面试准备路径、项目深挖问题和 10 天计划。\n适用人群：已进入笔试或面试前准备阶段，需要针对腾讯技术/算法方向进行系统复盘的候选人。", "description_en": "Pain point: Students often prepare by memorizing question banks while ignoring the JD, project storytelling, and realistic follow-up logic.\nDemo: This case uses an official multimodal algorithm JD and Tencent campus interview coaching materials to generate a 1:1 interview plan, project deep-dive questions, and a 10-day schedule.\nAudience: Candidates preparing for Tencent engineering or algorithm interviews after application or written tests.", "tags": [ "腾讯校招", "面试准备", "算法", "模拟追问" ], "tags_en": [ "Tencent Campus Recruitment", "Interview Preparation", "Algorithm", "Mock Follow-up" ], "artifact_type": "md", "categories": [ "education" ], "prompt": "我准备腾讯校招“算法-多模态方向”的 1v1 面试，请你作为腾讯校招求职教练，基于官网 JD 和面试辅导材料，帮我制定准备计划。\n\n我的背景：硕士，做过图文检索/多模态理解项目，熟悉 Python、PyTorch、Transformer，论文和竞赛一般，但有完整项目落地经历；面试预计 10 天后。\n\n请输出：\n1. 先判断是否缺少关键信息，必要时用 2-4 个选项追问。\n2. 基于官网 JD 拆解面试官通常会关注的能力。\n3. 给出项目深挖准备框架。\n4. 生成 12 个可能的追问问题，并说明每类问题考察什么。\n5. 制定 10 天准备计划。\n\n注意：不要承诺题目命中率，不要说“一定会问”；用“通常会关注”“建议准备”等柔性表达。不要编造通过率、录取率、HC、薪酬。", "prompt_en": "I'm preparing for a 1:1 interview for Tencent's campus recruitment \"Algorithm - Multimodal direction.\" Please act as a Tencent campus recruitment job coach and, based on the official website JD and interview coaching materials, help me create a preparation plan.\n\nMy background: Master's degree, have worked on image-text retrieval / multimodal understanding projects, familiar with Python, PyTorch, and Transformer, average on papers and competitions, but with complete project-implementation experience; the interview is expected in about 10 days.\n\nPlease output:\n1. First determine whether key information is missing, and if necessary follow up with 2-4 options.\n2. Based on the official website JD, break down the competencies interviewers typically focus on.\n3. Provide a framework for preparing to deep-dive into projects.\n4. Generate 12 possible follow-up questions, and explain what each category of question is assessing.\n5. Develop a 10-day preparation plan.\n\nNote: Do not promise question hit rates, do not say \"will definitely be asked\"; use soft phrasing like \"typically focuses on\" and \"recommend preparing.\" Do not fabricate pass rates, offer rates, headcount, or compensation.", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "面试准备计划", "label_en": "Interview Preparation Plan" } ], "cover_image": "cover.png", "quality_score": 90, "skills": [ { "id": "tencent-campus-recruit", "name": "腾讯校园招聘", "description": "面向 WorkBuddy 上架使用的腾讯校园招聘助手，覆盖流程制度、招聘动态、岗位方向、简历优化、面试陪练和防诈骗场景。", "skill_id": "skill_2054903442024890368" } ], "experts": [], "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 100, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-campus-job-match", "title": "腾讯校招岗位匹配", "title_en": "Tencent Campus Job Matching", "subtitle": "基于官网岗位与 JD，生成岗位推荐和准备路径", "subtitle_en": "Generate role recommendations and preparation steps from official Tencent campus JDs.", "description": "痛点：很多同学只知道想投腾讯，但不知道产品、技术、算法等方向哪个更匹配，也容易被非官方岗位信息干扰。\n演示内容：本案例基于 join.qq.com 官网岗位数据，先澄清候选人背景，再推荐 3 个真实岗位，并拆解匹配理由、风险点和 7 天准备计划。\n适用人群：准备投递腾讯校招或实习、需要快速选方向并制定行动计划的应届生。", "description_en": "Pain point: Many students want to apply to Tencent but are unsure which product, engineering, or algorithm roles fit them, and may be misled by unofficial job information.\nDemo: This case uses official join.qq.com job data, clarifies the candidate profile, recommends three real roles, and breaks down fit, gaps, and a 7-day preparation plan.\nAudience: Students preparing for Tencent campus recruitment or internships who need role selection and an actionable plan.", "tags": [ "腾讯校招", "岗位匹配", "JD解读", "求职准备" ], "tags_en": [ "Tencent Campus Recruitment", "Job Matching", "JD Analysis", "Career Preparation" ], "artifact_type": "md", "categories": [ "education", "search" ], "prompt": "我正在准备腾讯校园招聘，请你作为腾讯校招求职教练，基于 join.qq.com 官网岗位/JD 数据帮我完成岗位匹配。\n\n我的背景：计算机相关专业硕士，做过大模型应用、数据分析和后端服务项目；目标城市深圳/北京/上海；希望在产品经理（技术背景）、后台开发、算法/多模态之间选择；当前处于投递前准备阶段。\n\n请按腾讯校园招聘 Skill 的规则执行：\n1. 先判断信息是否足够；若缺少关键条件，用 2-4 个选项追问，不要猜。\n2. 用官网岗位脚本查询真实岗位和 JD，不要编造岗位、HC、通过率、薪酬或 offer 档位。\n3. 推荐 3-5 个真实岗位，输出岗位名称、工作地、招聘类型、投递链接、匹配理由和风险点。\n4. 对最推荐岗位拆解硬性条件、软性素质、加分项、简历侧重和面试准备。\n5. 最后给 7 天求职准备计划，并提醒个人流程信息以 join.qq.com 校招官网 offer 鹅为准。", "prompt_en": "I'm preparing for Tencent's campus recruitment. Please act as a Tencent campus-recruitment job coach and help me with job matching based on the position/JD data from the join.qq.com official site.\n\nMy background: a master's degree in a computer-related major, with project experience in LLM applications, data analysis, and backend services; target cities Shenzhen/Beijing/Shanghai; I want to choose among Product Manager (technical background), backend development, and algorithm/multimodal roles; I'm currently in the pre-application preparation stage.\n\nPlease follow the rules of the Tencent campus recruitment Skill:\n1. First judge whether the information is sufficient; if key conditions are missing, ask follow-up questions with 2-4 options—don't guess.\n2. Use the official-site job script to query real positions and JDs; don't fabricate positions, headcount, pass rates, compensation, or offer tiers.\n3. Recommend 3-5 real positions, outputting the job title, work location, recruitment type, application link, match rationale, and risk points.\n4. For the most recommended position, break down the hard requirements, soft qualities, bonus points, resume emphasis, and interview prep.\n5. Finally provide a 7-day job-prep plan, and remind me that personal process information is authoritative on the join.qq.com campus recruitment official site's \"Offer E\" (offer goose).", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "岗位匹配报告", "label_en": "Job Matching Report" } ], "cover_image": "cover.png", "quality_score": 92, "skills": [ { "id": "tencent-campus-recruit", "name": "腾讯校园招聘", "description": "面向 WorkBuddy 上架使用的腾讯校园招聘助手，覆盖流程制度、招聘动态、岗位方向、简历优化、面试陪练和防诈骗场景。", "skill_id": "skill_2054903442024890368" } ], "experts": [], "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 100, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-lit-search-case-1", "title": "PCSK9靶点评估与管线调研", "title_en": "Target Assessment and Pipeline Research for PCSK9", "subtitle": "一键获取心血管靶点全景：遗传证据、作用机制、管线资产一览", "subtitle_en": "Full-spectrum cardiovascular target view: genetics, mechanism, and pipeline assets in one report", "description": "靶点评估需要同时覆盖人类遗传学、成药性、临床试验进展三条维度，但手工逐一检索 PubMed、ClinVar、ClinicalTrials 极为耗时，且难以保证不遗漏。腾讯生命科学文献检索专家可联查多权威来源，自动去重并按维度分层呈现，附完整 PMID / DOI / NCT 溯源。本案例以 PCSK9 在动脉粥样硬化性心血管病中的靶点评估为示例，产出完整的靶点情报 HTML 报告，涵盖人类遗传、作用机制、成药性、在研及终止资产管线。适合药物研发团队、医学事务及靶点选择阶段的研究人员使用。", "description_en": "Target assessment requires simultaneous coverage of human genetics, tractability, and clinical trial progress — yet manually searching PubMed, ClinVar, and ClinicalTrials is time-consuming and error-prone. The Tencent Life Sciences Literature Search Expert federates authoritative sources, deduplicates results, and presents findings in stratified layers with full PMID/DOI/NCT traceability. This case uses PCSK9 in atherosclerotic cardiovascular disease as a demonstration, delivering a complete HTML target intelligence report covering human genetics, mechanism, tractability, and active/discontinued pipeline assets. Ideal for drug discovery teams, medical affairs, and researchers at the target selection stage.", "categories": [ "search", "education" ], "tags": [ "靶点评估", "药物管线", "文献检索", "PCSK9", "心血管" ], "tags_en": [ "Target Assessment", "Drug Pipeline", "Literature Search", "PCSK9", "Cardiovascular" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "tencent-lit-search-output-1.html", "artifacts": [ { "file": "tencent-lit-search-output-1.html", "label": "PCSK9 靶点情报报告" } ], "prompt": "请深度评估 PCSK9 在动脉粥样硬化性心血管病中的靶点价值，并梳理截至检索日公开来源中的在研与终止资产；分开呈现人类遗传、作用机制、成药性、试验阶段和申办方，附 PMID、DOI、NCT。", "prompt_en": "Please deeply evaluate PCSK9 as a therapeutic target in atherosclerotic cardiovascular disease, and map all active and discontinued assets from public sources as of the retrieval date. Present human genetics, mechanism, tractability, trial phase, and sponsor separately, with PMID, DOI, and NCT numbers.", "experts": [ { "id": "tencent-lit-search", "name": "生命科学文献检索", "description": "生命科学证据检索专家。覆盖药研、临床、生信及遗传与罕见病，按场景联查专业来源，结果分层去重可溯源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "843d1587f82b", "expert_id": "ex_DXuzV9trsD3U" } ], "skills": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-24T16:32:00+08:00", "updated_at": "2026-08-25T12:18:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-meeting-workbench", "title": "腾讯会议工作台", "title_en": "Tencent Meeting Workbench", "subtitle": "日历视图 · 冲突提醒 · 录制与纪要一站管理", "subtitle_en": "Calendar view, conflict alerts, recordings & minutes in one place", "description": "每周被「这周开了几个会、跟谁撞了、昨天那场复录录在哪」问倒时，最缺的是一份能随手翻的会议地图和一份不用点十几个按钮就能拿到的纪要。\n\n这个案例展示了一个专属于自己的「腾讯会议工作台」：顶部四宫格总览（倒计时下一场、本周会议数、本月会议数、本月总时长），主体是带热力图和冲突标记的月历（时间冲突标红、前后间隔 ≤15 分钟标黄、密度从少到多），右侧今日会议表清晰区分已结束/进行中/未开始，附带录制回看与权限申请入口。中英双语（?lang=en 切换英文）。开箱即用，无需连接真实账号即可在内置演示数据下查看完整布局与交互。\n\n适合：周会密集的团队负责人、需要快速回顾历史会议/纪要的人、想拥有一个可视化会议日历而不只是看通知的人、希望把会议冲突和间隔紧提前看见的人。", "description_en": "When \"how many meetings this week, who overlapped, where is yesterday's recap\" throws you off, the missing piece is a glanceable meeting map plus minutes that don't take ten clicks to fetch.\n\nThis case shows a personal Tencent Meeting Workbench: a top four-tile overview (next-meeting countdown, this-week count, this-month count, this-month total time), a main month calendar with a heatmap and conflict markers (red for time overlap, yellow for ≤15-minute tight gaps, density from light to heavy), and a right-side today list that clearly separates ended / live / upcoming with recording playback and access-request entry. Bilingual in Chinese and English (?lang=en switches to English). It works out of the box with built-in demo data, no live account required, so you can preview the full layout and interactions immediately.\n\nBest for: team leads with packed weekly syncs, anyone needing quick access to past meetings and minutes, people who want a visual meeting calendar instead of just notifications, and those who want to spot meeting conflicts and tight gaps in advance.", "categories": [ "workspace" ], "tags": [ "会议管理", "日历视图", "冲突提醒", "会议纪要", "工作台" ], "tags_en": [ "Meeting Management", "Calendar View", "Conflict Alerts", "Meeting Minutes", "Workbench" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "腾讯会议工作台（演示版 HTML）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "请优先使用资料库能力读取「腾讯会议工作台」公开空间中的文档（资料库中找不到时，直接打开链接读取）：\n\n1.《工作台复刻指南（从这里开始）》 https://www.workbuddy.cn/space/d/1KBZphvL7Za0z9omlSoxpC\n2.《serve.js 源码（逐字落地）》 https://www.workbuddy.cn/space/d/rgTO4R5LgEO5wWf7LH1evO\n3.《tmeet CLI 集成约定（前端数据协议）》 https://www.workbuddy.cn/space/d/hvCUksf6ynJFMOsscrdMgi\n\n然后在我的工作台目录（index.html 所在目录）把《serve.js 源码（逐字落地）》中的代码逐字保存为 serve.js（已存在则跳过），在后台启动它（node serve.js，默认 http://localhost:8787），并打开 http://localhost:8787/index.html —— 页面应显示「腾讯会议已连接」并同步我的真实会议数据。\n\n复刻完成后，请引导我在 WorkBuddy 左侧「专家 / 连接器」面板中主动完成「腾讯会议」连接器授权，以便同步真实会议数据。", "prompt_en": "Use the library capability first to read the documents in the public Tencent Meeting Workbench library (open the links directly if not found in the library):\n\n1. Workbench Replica Guide (Start Here) https://www.workbuddy.cn/space/d/1KBZphvL7Za0z9omlSoxpC\n2. serve.js Source (Copy Verbatim) https://www.workbuddy.cn/space/d/rgTO4R5LgEO5wWf7LH1evO\n3. tmeet CLI Integration Spec (Frontend Data Protocol) https://www.workbuddy.cn/space/d/hvCUksf6ynJFMOsscrdMgi\n\nThen save the serve.js source into my workbench directory (the directory containing index.html), start it in the background (node serve.js, default http://localhost:8787), and open http://localhost:8787/index.html - the page should show Tencent Meeting Connected and sync my real meeting data.\n\nOnce the replica is complete, guide me to proactively authorize the Tencent Meeting connector in the Experts / Connectors panel on the left, so my real meeting data can be synced.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T15:50:00+08:00", "updated_at": "2026-08-27T16:45:00+08:00", "source": "official", "is_recommended": false, "recommended_order": 0 }, { "id": "the-very-hungry-caterpillar-grade5-english", "title": "《The Very Hungry Cat", "title_en": "《The Very Hungry Cat", "subtitle": "小学五年级英语 · 《The Very Hungry Caterpillar》绘本阅读课", "subtitle_en": "小学五年级英语 · 《The Very Hungry Caterpillar》绘本阅读课", "description": "五年级英语绘本《The Very Hungry Caterpillar》互动课件，用动画和拖拽练习学星期与食物词汇。", "description_en": "五年级英语绘本《The Very Hungry Caterpillar》互动课件，用动画和拖拽练习学星期与食物词汇。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "the-very-hungry-caterpillar-grade5-english.html", "artifacts": [ { "file": "the-very-hungry-caterpillar-grade5-english.html", "label": "《The Very Hungry Cat｜互动教学网页" } ], "prompt": "帮我做一个五年级英语绘本《The Very Hungry Caterpillar》的互动教学网页，要有绘本动画和拖拽排序练习。", "prompt_en": "帮我做一个五年级英语绘本《The Very Hungry Caterpillar》的互动教学网页，要有绘本动画和拖拽排序练习。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tiered-coupon-plan", "title": "满减优惠券方案设计", "title_en": "Tiered Spend-and-Save Coupon Plan", "subtitle": "客单价分布驱动的满减券组合设计", "subtitle_en": "A practical Weimob playbook for tiered spend-and-save coupon plan", "description": "做满减优惠券方案设计时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“满减优惠券方案设计”展开：结合近 30 天客单价分布设计满减门槛、面额、有效期和使用规则，并评估对毛利及订单量的影响。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、促销活动运营人员、经营分析团队。", "description_en": "The hardest part of tiered spend-and-save coupon plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates tiered spend-and-save coupon plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Promotion Operations specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "满减优惠券", "客单价", "门槛设计", "毛利测算", "订单提升" ], "tags_en": [ "Tiered Spend-and-Save Coupon Plan", "Mall Operations", "Promotion Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "结合近30天客单价分布，帮我设计满减优惠券的门槛和面额组合，设置有效期及使用规则，并测算对毛利和订单量的影响。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a tiered spend-and-save coupon plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tiered-membership-benefits-design", "title": "分级会员权益设计", "title_en": "Tiered Membership Benefits Design", "subtitle": "三档会员差异化权益与触达方案", "subtitle_en": "A practical Weimob playbook for tiered membership benefits design", "description": "做分级会员权益设计时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“分级会员权益设计”展开：为普通、银卡和金卡会员设计差异化折扣、积分倍率和专属券，并形成配置及触达方案。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、积分与会员权益运营人员、经营分析团队。", "description_en": "The hardest part of tiered membership benefits design is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates tiered membership benefits design through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Points and Membership Benefits specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "会员等级", "差异化权益", "会员折扣", "积分倍率", "专属券" ], "tags_en": [ "Tiered Membership Benefits Design", "Customer Operations", "Points and Membership Benefits", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我为普通会员、银卡会员和金卡会员设计差异化权益，包括折扣、积分倍率和专属券，并给出配置路径和触达话术。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a tiered membership benefits design covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" }, { "id": "member-ops", "name": "会员权益运营", "description": "查询积分与储值方案、流水和客户权益变动，支持按客户信息、业务编号及明确时间范围筛选。", "skill_id": "skill_2087080649681588224" }, { "id": "customer-profile", "name": "客户资料查询", "description": "查询客户资料和客户列表，支持按身份信息检索，并按消费、积分、余额及企微关系等条件筛选。", "skill_id": "skill_2087078752245129216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tiffany-study-workbench", "title": "Tiffany 的学习工作台", "title_en": "Tiffany's Study Workbench", "subtitle": "把预习、错题、资料与亲子激励收进一个学习空间", "subtitle_en": "Bring preview notes, mistakes, materials, and family motivation into one study space", "description": "学习资料、错题复盘和每日计划散落在不同地方时，孩子很难看见自己的进步，家长也不容易给到及时、具体的鼓励。\n\n这个案例展示了一个面向中学生的学习工作台：首页汇总今日任务、课表、考试倒计时和心情；预习区、错题本、专项突破和学科资料库分别沉淀学习过程；积分、奖励兑换与家长留言板把正反馈自然融入日常。数据落在资料库数据表，支持在线存储与双向同步。页面内置中英双语，追加 ?lang=en 即可切换英文。\n\n适合：需要建立自主学习节奏的学生、想把学习陪伴做得更具体的家长，以及希望快速搭建一套可持续使用的家庭学习工作台的人。", "description_en": "When study materials, mistake reviews, and daily plans live in different places, students struggle to see their progress and parents struggle to give timely, specific encouragement.\n\nThis case demonstrates a study workbench for middle-school students: the dashboard brings together today's tasks, schedule, exam countdown, and mood; Preview Zone, Mistake Notebook, Breakthrough, and Subject Library capture the learning process; points, rewards, and a family message board make positive feedback part of everyday learning. Data lives in WorkBuddy Library tables with online storage and two-way sync. The page includes built-in Chinese and English UI; append ?lang=en to switch to English.\n\nBest for: students building a self-directed study rhythm, parents who want more concrete ways to support learning, and anyone who wants to quickly create a sustainable family study workbench.", "categories": [ "workspace" ], "tags": [ "学习工作台", "错题管理", "预习计划", "亲子激励", "双语" ], "tags_en": [ "Study Workbench", "Mistake Management", "Preview Planning", "Family Motivation", "Bilingual" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "Tiffany 的学习工作台（中英双语单文件 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/KSR5qI0MECB44ZjU6YhTf9，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「学习工作台」，面向中学生与家长，用结构化数据表管理学习过程与亲子激励。数据表：错题本（题目、学科、年级、单元、错因、正确解法、掌握状态、下次复习）、学科资料（标题、学科、类型、内容）、爸妈留言（作者、内容、时间、状态、类型）、积分流水（原因、变动额、时间）、预习区（学科、单元、笔记、疑问、状态）、专项突破（学科、薄弱点、目标、状态）、任务清单（文本、类型、日期、已完成、完成时间、学科）。\n\n页面与交互：① 首页仪表盘——今日任务勾选完成即自动加积分、输入回车快速新增、可删除，显示完成进度；考试倒计时按日期排序、临近 7 天高亮、支持添加考试（名称+日期+科目）；每周课表按周一到周五逐节展示，点击格子可改科目；每日一句可换一句或自定义；今日心情打卡可附一句话。② 预习区——按学科新增预习（单元、笔记、疑问），可标记已复习、编辑删除。③ 错题本——录入错题（题目、错因、正确解法、掌握状态、下次复习日期），支持按学科分组和掌握状态看板两种视图、今日待复习筛选，标记已掌握记积分。④ 专项突破——为薄弱学科建立目标，进度在未开始/进行中/已达成间流转。⑤ 学科资料库——收集公式、笔记、试卷等内容，支持图片与外链。⑥ 家长留言板——积分明细全部留痕（任务/错题/预习/打卡自动加减分）；孩子用积分按价格申请兑换奖励，兑换自动扣分；爸妈可给孩子留言；设置里可改名字、各行为积分规则与奖励清单（名称+积分，一行一个）。PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 6 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/KSR5qI0MECB44ZjU6YhTf9 with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Study Workbench\" for middle-school students and their parents, managing the learning process and family motivation in structured tables. Tables: Mistake Notebook (question, subject, grade, unit, mistake cause, correct solution, mastery status, next review), Subject Library (title, subject, type, content), Parent Messages (author, content, time, status, type), Points Log (reason, delta, time), Preview Zone (subject, unit, notes, questions, status), Breakthrough (subject, weak point, goal, status), Task List (text, type, date, done, completion time, subject). Pages and interactions: ① Dashboard — checking off a today-task auto-awards points, press Enter to add tasks quickly, deletable, with completion progress; exam countdown sorted by date with a 7-day approaching highlight, exams addable by name + date + subject; weekly timetable shown period by period, click a cell to change the subject; the daily quote can be shuffled or customized; today's mood check-in with an optional note. ② Preview Zone — add previews by subject (unit, notes, questions), mark reviewed, edit and delete. ③ Mistake Notebook — record mistakes (question, cause, correct solution, mastery status, next review date), browse by subject grouping or a mastery board, filter due-today, marking mastered awards points. ④ Breakthrough — set goals for weak subjects, progress flows through Not Started / In Progress / Achieved. ⑤ Subject Library — collect formulas, notes, and papers, with images and links. ⑥ Family Board — a full points log (tasks / mistakes / previews / check-ins auto add or subtract); the child redeems rewards at listed prices with points deducted automatically; parents leave messages; Settings edits the name, per-action point rules, and the reward list (name + price, one per line). responsive on PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus 6 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿和生活管理，支持响应式 HTML、本地数据持久化与一键部署。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd9b7744a331", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "Tiffany的一家", "name_en": "Tiffany's Family", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "source": "community", "usage_count": 0, "created_at": "2026-09-04T22:28:48+08:00", "updated_at": "2026-09-04T22:50:55+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "time-management", "title": "时间黑洞诊断与方法动手实操", "title_en": "时间黑洞诊断与方法动手实操", "subtitle": "自我管理 · 时间黑洞诊断与方法动手实操", "subtitle_en": "自我管理 · 时间黑洞诊断与方法动手实操", "description": "时间管理互动课件，用时间黑洞自查、四象限拖拽分级和番茄钟三个动手模块，把方法落到练习。", "description_en": "时间管理互动课件，用时间黑洞自查、四象限拖拽分级和番茄钟三个动手模块，把方法落到练习。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "time-management.html", "artifacts": [ { "file": "time-management.html", "label": "时间黑洞诊断与方法动手实操｜互动教学网页" } ], "prompt": "帮我做一个时间管理主题的互动教学网页，要有时间黑洞自查、四象限拖拽分级和番茄钟。", "prompt_en": "帮我做一个时间管理主题的互动教学网页，要有时间黑洞自查、四象限拖拽分级和番茄钟。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "top-product-contribution-30d", "title": "近30天成交商品贡献分析", "title_en": "30-Day Product Contribution Analysis", "subtitle": "成交商品销量与支付订单贡献度分析", "subtitle_en": "A practical Weimob playbook for 30-day product contribution analysis", "description": "做近30天成交商品贡献分析时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“近30天成交商品贡献分析”展开：汇总近 30 天实际有成交的商品，按销量和支付订单数分析商品贡献集中度，识别常态热销与偶发大单驱动商品。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、商品销售诊断运营人员、经营分析团队。", "description_en": "The hardest part of 30-day product contribution analysis is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates 30-day product contribution analysis through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Product Sales Diagnosis specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "成交商品", "销量贡献", "订单贡献", "大单识别", "商品运营" ], "tags_en": [ "30-Day Product Contribution Analysis", "Business Diagnosis", "Product Sales Diagnosis", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "汇总近30天实际有成交的商品，按销量和支付订单数分析贡献度，识别哪些商品属于常态热销、哪些主要由偶发大单拉动，并给出运营建议。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a 30-day product contribution analysis covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" }, { "id": "goods-ops", "name": "商品运营", "description": "查询和筛选商品、查看详情、排序及类目分组；支持将创建商品、调价、改库存和上下架等操作委派处理。", "skill_id": "skill_2087079138104320000" }, { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "topic-ai-overview", "title": "人工智能产业扫描报告", "title_en": "AI Industry Scan Report", "subtitle": "算力-模型-应用全链条，看懂 AI 产业全景与易方达产品", "subtitle_en": "Compute-model-application chain, AI panorama and E Fund products at a glance", "description": "AI 天天上头条，但算力、大模型、应用到底什么关系，有哪些指数和基金可以跟踪？这份案例把人工智能产业拆成一张全景图：指数表现打底、产业链分层拆解、近期观点原样引述、易方达 AI 相关产品展示。\n\n案例演示了 AI 按「行业全景扫描」分支真实取数：从指数直通车过滤人工智能相关指数并拉取收益、回撤、估值指标，检索易方达公众号产业链投教文章梳理算力-模型-应用上下游，引述近期观点，汇总易方达 AI 主题 ETF 与场外基金，产品业绩基准由官网脚本批量抓取。\n\n适合：关注 AI 赛道的投资者。", "description_en": "AI makes headlines every day, but how do compute, large models, and applications relate to each other, and which indexes and funds can track the theme? This case turns the AI industry into a panorama: index performance as the base, a layered breakdown of the industry chain, verbatim quotes of recent views, and E Fund AI-related products on display.\n\nThe case demonstrates the AI \"Industry Panorama Scan\" branch with real data: filtering AI-related indexes from the index expressway and pulling return, drawdown, and valuation metrics; retrieving educational articles on the industry chain from E Fund's official WeChat accounts to map the compute-model-application value chain; quoting recent views; and compiling E Fund AI-themed ETFs and open-end funds, with fund benchmarks batch-fetched from the official website.\n\nBest for: investors tracking the AI theme.", "categories": [ "finance", "search" ], "tags": [ "人工智能", "行业全景", "算力", "大模型" ], "tags_en": [ "AI", "Industry Panorama", "Compute", "Large Models" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "按「行业全景扫描」分支生成人工智能产业全景报告：行业表现（主指数 KPI 五方块与相关指数）、产业链结构、相关观点、易方达相关基金。严格遵循 skill 的 HTML 报告模板与排版规范，输出单文件 HTML（无外部资源、直链链接），所有数据真实查询并注明来源，缺失标「数据暂缺」，不预测后市、不给投资建议。", "prompt_en": "Following the \"Industry Panorama Scan\" branch, produce a panorama report on the AI industry: industry performance (main index KPI boxes plus related indexes), industry chain, related views, and E Fund related funds. Strictly follow the skill's HTML report template and layout spec, output a single-file HTML (no external resources, direct links), all data from real queries with sources, mark missing data as \"data unavailable\", no market predictions and no investment advice.", "skills": [ { "id": "efunds-topic-overview", "name": "易方达主题概念速览", "description": "当用户提到某个行业、投资策略、资产类别、热门主题等，想获取相关信息、辅助理解概念时使用。基于易方达基金各官方微信公众号已发布文章，自动围绕关键词梳理整合相关资讯，输出结构化内容（包含概念阐释、市场动态、相关表现等），帮助用户快速了解某一主题的相关内容。", "skill_id": "skill_2092171456670191616" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-25T22:03:16+08:00", "updated_at": "2026-08-26T09:35:00+08:00", "artifacts": [ { "file": "output.html", "label": "人工智能产业全景报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "topic-dividend-overview", "title": "红利策略怎么看", "title_en": "How to Read the Dividend Strategy", "subtitle": "高股息逻辑+指数估值+基金产品，看懂红利策略值不值得关注", "subtitle_en": "High-dividend logic, index valuation and fund products in one report", "description": "「红利」「高股息」这几年反复被提起，但红利策略到底是什么、现在估值贵不贵、有哪些产品可以关注？这份案例把红利策略讲透：概念逻辑、近期动态、相关指数表现和易方达红利产品一次列全。\n\n案例演示了 AI 按「通用模板报告」分支真实取数：检索易方达公众号投教文章摘录红利策略定义，引述近一年红利相关动态，从指数直通车拉取中证红利等指数收益、回撤、估值指标，列出易方达红利相关 ETF 与场外基金，业绩基准由官网脚本批量抓取。\n\n适合：想了解红利策略概念的投资者。", "description_en": "\"Dividend\" and \"high-yield\" are recurring buzzwords these years, but what exactly is the dividend strategy, is it expensive now, and which products are worth watching? This case makes the dividend strategy crystal clear: the concept logic, recent dynamics, related index performance, and E Fund's dividend products — all listed in one go.\n\nThe case demonstrates AI pulling real data following the \"Template Report\" branch: retrieving educational articles from E Fund's WeChat official account to extract the dividend strategy definition, quoting the past year's dividend-related dynamics, pulling yield / drawdown / valuation metrics for the CSI Dividend and other indices from \"Index Express\" (指数直通车), listing E Fund's dividend-related ETFs and off-exchange funds, with performance benchmarks batch-collected by an official-website script.\n\nBest for: investors who want a clear picture of the dividend strategy concept.", "categories": [ "finance", "search" ], "tags": [ "红利", "高股息", "投资策略", "估值" ], "tags_en": [ "Dividend", "High Yield", "Investment Strategy", "Valuation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "按「通用模板报告」分支生成红利策略报告：概念综述、相关动态、相关表现（指数六列表）、易方达相关基金。严格遵循 skill 的 HTML 报告模板与排版规范，输出单文件 HTML（无外部资源、直链链接），所有数据真实查询并注明来源，缺失标「数据暂缺」，不预测后市、不给投资建议。", "prompt_en": "Following the \"Template Report\" branch, produce a report on the dividend strategy: concept overview, related dynamics, related performance (six-column index table), and E Fund related funds. Strictly follow the skill's HTML report template and layout spec, output a single-file HTML (no external resources, direct links), all data from real queries with sources, mark missing data as \"data unavailable\", no market predictions and no investment advice.", "skills": [ { "id": "efunds-topic-overview", "name": "易方达主题概念速览", "description": "当用户提到某个行业、投资策略、资产类别、热门主题等，想获取相关信息、辅助理解概念时使用。基于易方达基金各官方微信公众号已发布文章，自动围绕关键词梳理整合相关资讯，输出结构化内容（包含概念阐释、市场动态、相关表现等），帮助用户快速了解某一主题的相关内容。", "skill_id": "skill_2092171456670191616" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-25T22:03:16+08:00", "updated_at": "2026-08-31T17:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "红利策略报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "trade-marketing-content", "title": "一套营销内容，公众号朋友圈海报全搞定", "title_en": "Trade Marketing Content", "subtitle": "公众号推文＋朋友圈文案＋海报文案一整套", "subtitle_en": "A full set — official-account article, Moments copy and poster text", "description": "输入公司和主题，自动生成公众号推文、朋友圈文案、海报文案一整套贸易推广物料，拿来就能用。", "description_en": "Enter your company and theme: it generates a complete trade promotion kit — official-account article, Moments copy and poster text — ready to use as-is.", "categories": [ "creation" ], "tags": [ "营销内容", "文案生成", "获客", "B2B贸易" ], "tags_en": [ "Marketing Content", "Copywriting", "Customer Acquisition", "B2B Trade" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "贸易营销内容套件（HTML 版）" } ], "prompt": "你是TradeBuddy出品的B2B贸易销售专家。请帮我生成一套营销推广内容。\n\n【背景】\n- 公司：{如 华东建材贸易}\n- 主推：{如 螺纹/高线/板卷现货，当天送达}\n- 主题/卖点：{如 现货充足、含税到货、材质质保齐全}\n\n【要求】\n1. 结合当前行情氛围（可调用【找钢行情MCP】），查询【行业信息资产库】里这个行业的营销话术口径；\n2. 生成一整套：①公众号推文（标题+导语）②朋友圈短文案3条 ③海报文案（主标题+副标题+CTA，可交设计出图）；\n3. 文案要贴合钢铁/建材贸易场景、突出现货和信任感、口语接地气、有紧迫感。\n\n请生成完整营销内容套件。", "skills": [ { "id": "marketing-content-generation", "name": "营销内容生成", "description": "TradeBuddy把“我要推这批货 / 这个活动”变成一句话就能拿到的一组可直接发出去的营销内容——朋友圈、社群、公众号、B2B 平台、短视频，一次给齐多版。", "skill_id": "skill_2090731933854892032" } ], "experts": [ { "id": "xiaoshou-expert", "name": "B2B贸易销售专家", "description": "懂B2B贸易的销售专家：发货盘、写文案、出报价、盯商机、经营客户、看业绩、谈价成单，助你多成单。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_42GwnCqPUv7a" } ], "mcps": [], "creator": { "name": "找钢TradeBuddy官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-25T12:35:00+08:00", "updated_at": "2026-08-25T12:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "trademark-law-2019-vs-2026-comparison", "title": "商标法 2026 修订前后对比与要点提示", "title_en": "Trademark Law 2026 Amendment: Before-and-After Comparison & Key Takeaways", "subtitle": "用法律法规检索 Skill 快速生成现行法与待施行法的结构化对比", "subtitle_en": "Use the legal research Skill to quickly produce a structured comparison of current and upcoming trademark law", "description": "企业法务、知识产权律师和商标代理人经常接到「修法了，快速给我一份新旧法对比」的需求，但手动翻阅条文、核对条号费时费力，还容易遗漏关键变化。本案例演示如何通过法律法规检索 Skill，分别调取现行有效的《商标法（2019 修正）》与尚未生效的《商标法（2026 修订）》原文，自动生成一份包含监管体制、诚实信用、恶意注册、商标代理、驰名商标、侵权赔偿、误导公众使用等重点主题的前后对比表和实务要点提示 HTML 报告。适合需要快速完成修法解读、客户汇报、合规培训或内部风控提示的法律从业者、企业知识产权负责人和商标代理机构。", "description_en": "Corporate counsel, IP lawyers, and trademark agents often get asked to produce a quick before-and-after comparison when trademark law is amended, but manually cross-referencing articles and article numbers is time-consuming and error-prone. This case demonstrates how to use the legal research Skill to retrieve the current Trademark Law (2019 Amendment, currently effective) and the not-yet-effective 2026 Amendment, then automatically generate a structured HTML report covering institutional supervision, good-faith principles, bad-faith registration, trademark agency practice, well-known marks, infringement damages, misleading public use, and practical tips. It is designed for legal professionals, corporate IP managers, and trademark agencies who need to deliver amendment briefs, client updates, compliance training, or internal risk alerts quickly.", "prompt": "请使用法律法规检索 Skill，分别检索《中华人民共和国商标法（2019 修正）》（现行有效）和《中华人民共和国商标法（2026 修订）》（尚未生效，计划 2027-01-01 施行）中以下主题的条文：立法目的、监管体制、诚实信用原则、恶意注册、商标代理、驰名商标保护、侵权赔偿、侵权行为类型、误导公众使用、虚假诉讼处罚。将两部法律的对应条号、条文内容整理成前后对比表，提炼 8-10 条实务要点提示，并输出为单文件 HTML（CSS 内联、无外部资源、无 hash 锚点），顶部显著标注 2026 修订版的时效状态为尚未生效。", "prompt_en": "Use the legal research Skill to retrieve the following topics from the Trademark Law of the People's Republic of China (2019 Amendment, currently effective) and the same law (2026 Amendment, not yet effective, scheduled to take effect on 2027-01-01): legislative purpose, supervision system, good-faith principle, bad-faith registration, trademark agency practice, well-known trademark protection, infringement damages, types of infringement acts, misleading public use, and penalties for bad-faith litigation. Compile the article numbers and text into a side-by-side comparison table, extract 8-10 practical tips, and output a single-file HTML with inline CSS, no external resources, and no hash anchors. Clearly label the 2026 Amendment as not yet effective at the top.", "artifact_type": "html", "categories": [ "search", "education", "business" ], "tags": [ "商标法", "知识产权", "法律法规检索", "修订对比", "合规提示", "HTML报告" ], "tags_en": [ "trademark law", "intellectual property", "legal research", "amendment comparison", "compliance tips", "HTML report" ], "preview": "comparison.html", "artifacts": [ { "file": "comparison.html", "label": "商标法 2026 修订前后对比与要点提示", "label_en": "Trademark Law 2026 Amendment Comparison & Key Takeaways" } ], "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "source": "community", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 86, "usage_count": 0, "created_at": "2026-09-04T16:55:00+08:00", "updated_at": "2026-09-04T16:55:00+08:00", "skills": [], "experts": [ { "id": "ip-team-lead", "name": "知识产权专家团", "description": "覆盖知识产权条款专项审查、权利归属与授权许可分析、侵权担保与责任评估、权利链风险识别与修改建议的知识产权法律专家团队", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "1bcf88d993f1", "expert_id": "ex_4IBuGIfQnBYT" } ], "mcps": [], "is_recommended": false, "recommended_order": 0 }, { "id": "travel-itinerary-planner", "title": "AI 旅游行程规划助手", "title_en": "AI Travel Itinerary Planner", "subtitle": "一句话生成可交互的多日行程页，含时间轴 + 地图打点连线", "subtitle_en": "One-click interactive multi-day itinerary with timeline and map markers", "description": "输入城市与天数，AI 自动规划兼顾历史文化、美食、自然、夜生活的多日行程，并输出一个开箱即用的单 html 页面：顶部柔光渐变 banner + 玻璃拟态数据徽章（天数 / POI 数 / 总预算），左侧时间轴卡片按景点 / 美食 / 文化 / 购物四类分色并以虚线连接，右侧腾讯地图按天打点连线，顶部 Tab 切换 DAY1/2/3，点击卡片地图自动飞到对应位置。无需任何构建，双击浏览器即可使用。", "description_en": "Input a city and number of days, AI auto-generates a multi-day itinerary covering history, food, nature, and nightlife. Outputs a single ready-to-use HTML page with a soft gradient banner, glassmorphism badges (days / POIs / budget), color-coded timeline cards, Tencent Map markers with route lines, and Day tabs. Click any card to fly the map to that location. No build required — just double-click and open in browser.", "categories": [ "travel" ], "tags": [ "旅游规划", "行程", "腾讯地图", "时间轴", "数据可视化", "单html" ], "tags_en": [ "travel planning", "itinerary", "Tencent Map", "timeline", "data visualization", "single html" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我做一个【成都】3 日旅游行程规划单文件 HTML，景点和行程由你自行规划，兼顾历史文化、美食、自然、夜生活。视觉参考 Stripe 风格（清爽、柔光渐变）。\n\n页面：顶部 banner 居中放大标题，右侧用徽章展示天数 / POI 数 / 总预算；下方左侧为每天的时间轴卡片列表（POI 分景点/美食/文化/购物四类用不同配色，按时间顺序），右侧为腾讯地图按顺序打点连线；顶部 Tab 切换 DAY1/2/3，点击左侧卡片地图飞到对应位置。所有 CSS/JS 内联，输出单一 HTML 文件可直接打开。", "prompt_en": "Help me make a single-file HTML for a 3-day travel itinerary in [Chengdu]; you plan the attractions and itinerary yourself, balancing history & culture, food, nature, and nightlife. Visual reference: Stripe style (clean, soft gradients).\n\nPage: a centered, enlarged title in the top banner, with badges on the right showing number of days / POI count / total budget; below, the left side is a list of timeline cards for each day (POIs color-coded into four categories — attractions/food/culture/shopping — in chronological order), and the right side is a Tencent Map with markers connected in sequence; tabs at the top switch between DAY1/2/3, and clicking a card on the left flies the map to the corresponding location. All CSS/JS inlined, output a single HTML file that can be opened directly.", "skills": [ { "id": "tencentmap-map-assistant", "name": "腾讯地图助手", "description": "腾讯地图MCP集，提供地点搜索、路线规划、地图可视化等能力", "skill_id": "skill_2062731548497326080" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": true, "featured_order": 4, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-06-05T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "行程规划HTML页面", "label_en": "Itinerary HTML Page" } ], "is_recommended": true, "recommended_order": 11 }, { "id": "tree-planting-problem", "title": "植树问题（间隔与棵数）", "title_en": "植树问题（间隔与棵数）", "subtitle": "小学五年级数学广角 · 植树问题（间隔与棵数）", "subtitle_en": "小学五年级数学广角 · 植树问题（间隔与棵数）", "description": "五年级数学《植树问题》互动课件，用种树模拟器演示间隔数与棵数的一一对应关系。", "description_en": "五年级数学《植树问题》互动课件，用种树模拟器演示间隔数与棵数的一一对应关系。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "tree-planting-problem.html", "artifacts": [ { "file": "tree-planting-problem.html", "label": "植树问题（间隔与棵数）｜互动教学网页" } ], "prompt": "帮我做一个小学五年级数学《植树问题》的互动教学网页，要能一键种树并实时看间隔数和棵数的关系。", "prompt_en": "帮我做一个小学五年级数学《植树问题》的互动教学网页，要能一键种树并实时看间隔数和棵数的关系。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "trend-issues-deep-column", "title": "当智能变得廉价", "title_en": "When Intelligence Becomes Cheap", "subtitle": "用一篇高级深度专栏，写清 AI 时代判断力的价值迁移", "subtitle_en": "A premium long-form column on why judgment gains value in the age of AI", "description": "热点很多、观点更快，但真正难的是把趋势写出证据、立场与余韵，而不是追着热词跑。\n\n这个案例围绕“当智能变得廉价”完成了一套可直接使用的趋势专栏内容包：12 个标题、3 版导语、完整长文、8 条重点引文、3 期后续选题，以及一份带判断罗盘与数据关系图的杂志式阅读页。全文基于 Microsoft、世界经济论坛、Stanford HAI 与 Anthropic 的公开资料，明确区分事实、分析与样本边界。\n\n适合：品牌内容负责人、知识型创作者、商业媒体编辑、企业文化与思想领导力团队。", "description_en": "Trend cycles move quickly, but the hard part is turning a topic into an evidence-based point of view with lasting resonance instead of chasing buzzwords.\n\nThis case delivers a ready-to-use editorial package around “When Intelligence Becomes Cheap”: 12 headline options, three leads, a complete long-form essay, eight pull quotes, three follow-up issues, and a magazine-style reading page with a judgment compass and a data relationship graphic. The work draws on public sources from Microsoft, the World Economic Forum, Stanford HAI, and Anthropic, while clearly separating facts, analysis, and sample limitations.\n\nBest for: brand content leads, knowledge creators, business-media editors, and corporate thought-leadership teams.", "categories": [ "creation", "writing" ], "tags": [ "趋势专栏", "深度写作", "品牌叙事", "杂志排版" ], "tags_en": [ "Trend Column", "Long-form Writing", "Brand Storytelling", "Editorial Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "杂志式趋势专栏阅读页" }, { "file": "column-package.md", "label": "标题组、导语、专栏成稿与引文内容包" } ], "prompt": "围绕趋势议题《当智能变得廉价：判断力为何重新成为稀缺能力》，为关注 AI、组织变化与个人成长的知识工作者写一套深度专栏。请检索并使用 Microsoft《2025 Work Trend Index》、World Economic Forum《Future of Jobs Report 2025》、Stanford HAI《2025 AI Index Report》与 Anthropic 2025 年 9 月《Economic Index》的真实公开资料，注明来源与访问日期，区分相关性、因果性和平台样本边界。输出 12 个分方向标题、短中长三版导语、约 3000 字完整长文、8 条重点引文和 3 期后续选题；核心论证聚焦定义问题、选择证据、设定边界、承担后果。另生成单文件、资源全内联、可离线打开的杂志式 HTML 阅读页，视觉采用克制的高端纸媒风格，并用内联图形表达核心框架。", "prompt_en": "Create an in-depth editorial package around the trend topic “When Intelligence Becomes Cheap: Why Judgment Is Becoming Scarce Again” for knowledge workers interested in AI, organizational change, and personal growth. Search and use real public materials from Microsoft’s 2025 Work Trend Index, the World Economic Forum’s Future of Jobs Report 2025, Stanford HAI’s 2025 AI Index Report, and Anthropic’s September 2025 Economic Index. Cite each source with an access date while distinguishing correlation, causality, and platform-specific sample limits. Deliver 12 headlines grouped by direction, three leads in short, medium, and long forms, an approximately 3,000-Chinese-character feature, eight pull quotes, and three follow-up issues. Center the argument on defining problems, selecting evidence, setting boundaries, and owning consequences. Also produce a single-file, fully inlined, offline-ready magazine-style HTML reading page with a restrained premium editorial aesthetic and inline graphics that explain the core framework.", "skills": [], "experts": [ { "id": "ContentCreator", "name": "文爆爆", "description": "擅长创作引人入胜的多平台内容，让品牌故事触达目标受众", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "10c87ab42060", "expert_id": "ex_nchiDpT84L9o" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "official", "usage_count": 0, "created_at": "2026-09-08T15:12:01Z", "updated_at": "2026-09-08T15:35:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "turn-after-sales-into-repeat-purchases", "title": "把售后变成复购机会", "title_en": "Turn After-Sales Service into Repeat Purchases", "subtitle": "售后场景安抚、挽回与二次购买沟通策略", "subtitle_en": "A practical Weimob playbook for turn after-sales service into repeat purchases", "description": "做把售后变成复购机会时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“把售后变成复购机会”展开：围绕退款、退货、售后咨询等场景，设计既能安抚客户、又有机会促成再次购买的沟通策略和可直接使用的话术。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、门店导购负责人、导购任务与话术运营人员、经营分析团队。", "description_en": "The hardest part of turn after-sales service into repeat purchases is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates turn after-sales service into repeat purchases through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Store and Sales Associate Operations teams, Sales Associate Tasks and Scripts specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "售后沟通", "退款安抚", "客户挽回", "复购机会", "导购话术" ], "tags_en": [ "Turn After-Sales Service into Repeat Purchases", "Store and Sales Associate Operations", "Sales Associate Tasks and Scripts", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "围绕退款、退货、售后咨询等场景，设计一套既能安抚客户、又有机会促成再次购买的沟通策略和可直接使用的话术。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a turn after-sales service into repeat purchases covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "order-management", "name": "订单与售后管理", "description": "查询订单和售后单的列表与详情，支持按时间、状态、支付、配送方式及关键词进行筛选。", "skill_id": "skill_2087081152876433408" }, { "id": "customer-follow-up", "name": "客户跟进话术", "description": "为导购生成可直接发送的客户跟进话术，支持商品、活动和优惠券推荐及多轮改写。", "skill_id": "skill_2087078594478350336" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ui-designer-inspiration-cards", "title": "UI 设计灵感卡片", "title_en": "UI Design Inspiration Cards", "subtitle": "一句话主题，六套看得见的视觉方向", "subtitle_en": "One-line brief, six visual directions you can actually see", "description": "接手一个新项目要做视觉，最怕的就是没方向——脑子里只有「高级一点」「有质感一点」这种模糊词，落到配色、字体和圆角上还是无从下手。这套案例演示了把「晨间习惯打卡 App」这样一句话主题交给 UI 设计师，得到一本可翻阅的灵感手册：6 个差异明显的视觉方向（极简留白、编辑排版、暗色专注、温暖有机、新粗野主义、玻璃拟态），每个方向都给出配色板、字体搭配、形态参数，还有一块用该风格真实渲染的打卡界面示意，风格看得见摸得着；结尾附六方向气质定位图和选型建议，色板点击可复制色值。适合：正在给新产品定视觉方向的产品经理和设计师、需要快速产出多套风格提案的独立开发者、想学习设计系统参数化表达的前端工程师，以及接了设计需求但缺少灵感的创作者。", "description_en": "Starting visual design on a new project with zero direction is painful—all you have are vague words like \"make it premium\" or \"give it texture,\" and they collapse the moment you face concrete choices of color, type, and radius. This case shows what happens when you hand a one-line brief—\"a morning habit check-in app\"—to a UI designer: you get a browsable inspiration handbook with 6 distinctly different visual directions (quiet minimal, editorial, dark focus, warm organic, neo-brutalism, glassmorphism). Each direction includes a color palette, type pairing, form parameters, and a live check-in screen rendered in that style so the look is visible, not imagined. A positioning map and selection guide wrap it up, and every swatch is click-to-copy. Ideal for: PMs and designers setting the visual direction for a new product, indie developers who need multiple style proposals fast, front-end engineers learning parameterized design-system thinking, and creators who took on a design task but need inspiration.", "categories": [ "creation" ], "tags": [ "UI 设计", "灵感卡片", "视觉方向", "设计系统", "配色方案" ], "tags_en": [ "UI Design", "Inspiration Cards", "Visual Direction", "Design System", "Color Palette" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我想为「晨间习惯打卡 App」做视觉设计，但现在完全没有方向。请你作为专业 UI 设计师，帮我生成一套设计灵感卡片（单文件 HTML，可直接离线打开）：\n\n1. 给出 6 个差异明显的视觉方向（例如：极简留白、编辑排版、暗色专注、温暖有机、新粗野主义、玻璃拟态）\n2. 每个方向做成一张完整的灵感卡片，包含：设计理念、配色板（含色值与用途标注）、字体搭配与字号比例、圆角/描边/阴影/留白等形态参数、情绪关键词、适用场景\n3. 关键要求：每个方向配一块「组件微缩示意」——用该方向的视觉语言真实渲染一个打卡界面（习惯列表 + 主按钮），让风格看得见\n4. 结尾给一张六方向气质定位图（冷静—温暖 × 克制—张扬两个维度）和简短的选择建议\n5. 页面本身要有设计刊物级的排版质感，像一本可翻阅的灵感手册\n\n把主题换成你自己的产品（如「手冲咖啡电商」「夜跑 App」）即可复用同样的方法。", "prompt_en": "I need to do the visual design for a \"morning habit check-in app\" but have no direction at all yet. Acting as a professional UI designer, generate a set of design inspiration cards for me (single-file HTML, openable offline):\n\n1. Provide 6 clearly different visual directions (e.g., quiet minimal, editorial, dark focus, warm organic, neo-brutalism, glassmorphism)\n2. Make each direction a complete inspiration card containing: design philosophy, color palette (with hex values and usage labels), type pairing and type scale, form parameters such as radius/stroke/shadow/whitespace, mood keywords, and suitable scenarios\n3. Key requirement: each direction includes a \"mini component specimen\"—actually render a check-in screen (habit list + primary button) in that direction's visual language, so the style is visible\n4. End with a positioning map of the six directions (calm–warm × restrained–bold axes) plus brief selection advice\n5. The page itself should have design-publication-grade typographic quality, like a browsable inspiration handbook\n\nSwap in your own product topic (e.g., \"pour-over coffee e-commerce\" or \"night running app\") to reuse the same method.", "skills": [], "experts": [ { "id": "UiDesigner", "name": "像素君", "name_en": "Pixel Sensei", "description": "精通设计系统与组件库、追求像素级完美的 UI 设计专家", "description_en": "UI design expert with deep mastery of design systems and component libraries, pursuing pixel-perfect quality", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_eWDRfCKP9AY2" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-09-07T23:31:42+08:00", "updated_at": "2026-09-08T15:27:00+08:00", "artifacts": [ { "file": "output.html", "label": "设计灵感卡片（六个视觉方向 · 单文件 HTML）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ultra-wide-fundus-ai-diagnosis", "title": "超广角眼底照 AI 多病种筛查报告", "title_en": "Ultra-Wide Fundus AI Multi-Disease Screening Report", "subtitle": "一张超广角眼底照，一次调用跑完 47 维体征 + 23 项疾病筛查，输出带病灶坐标的完整报告", "subtitle_en": "One ultra-wide fundus photo, 47 sign dimensions and 23 diseases screened in a single call, with a full report including lesion coordinates", "description": "拿到医院拍的眼底照片，最难受的是看不懂——一张灰扑扑的圆形图，医生一句「有点问题，再观察」，回家只能自己瞎搜。想找个人认真解读，挂号又得等上两周。\n\n这个案例把一张真实的左眼超广角眼底彩照交给腾讯觅影眼科 AI，实测约 54 秒返回结构化结果：23 项疾病推测诊断检出 3 项疑似（视网膜裂孔、玻璃体内硅油/气体填充物、增殖期糖网 PDR），47 维体征逐项判别得到 6 项阳性、30 项阴性、11 项不适用，还给出 4 个带坐标和置信度的病灶检测框（视网膜纤维膜 3 处，置信度 0.50/0.48/0.34）。报告没有停在罗列结果，而是把这些散点串成一条判读链——新生血管加纤维膜指向 PDR，光凝斑和硅油说明既往已治疗，纤维膜牵拉造成裂孔，而「视网膜脱离」为阴性意味着还在能用激光封孔预防脱离的时间窗内。同时如实标注了两个子模型结论冲突的低置信项（黄斑前膜 0.37 vs 分类阴性），建议以 OCT 复核。\n\n适合：拿到眼底照片想读懂结果的患者及家属、需要快速初筛的基层与体检机构医生、糖尿病患者的眼底年度随访、眼科 AI 产品的能力验证、医学影像 AI 的落地评估。", "description_en": "The hardest part about getting a fundus photo from the hospital is that you cannot read it. You get a dim circular image, the doctor says \"something looks off, let's watch it,\" and you go home to search blindly. Booking a specialist for a proper reading means waiting two more weeks.\n\nThis case hands a real left-eye ultra-wide fundus photo to the Tencent Miying ophthalmic AI and gets structured results back in about 54 seconds of measured wall-clock time: 3 of 23 diseases flagged as suspected (retinal hole, intravitreal silicone oil/gas tamponade, proliferative diabetic retinopathy), 47 sign dimensions resolved into 6 positive, 30 negative and 11 not-applicable, plus 4 lesion detection boxes with coordinates and confidence scores (three retinal fibrous membranes at 0.50, 0.48 and 0.34). The report does not stop at listing findings — it connects them into a reading chain: neovascularization plus fibrous membrane points to PDR, laser scars and silicone oil indicate prior treatment, membrane traction produced the retinal hole, and a negative retinal detachment means the case is still inside the window where laser can seal the hole and prevent detachment. It also honestly flags a low-confidence conflict between two sub-models (epiretinal membrane at 0.37 versus a negative classification) and recommends OCT confirmation.\n\nGood for: patients and families trying to understand a fundus photo, primary-care and health-checkup physicians who need fast triage, annual retinal follow-up for diabetic patients, capability validation of ophthalmic AI products, and deployment evaluation of medical imaging AI.", "categories": [ "data", "productivity" ], "tags": [ "眼底AI", "医学影像", "多病种筛查", "糖网筛查", "报告解读" ], "tags_en": [ "Fundus AI", "Medical Imaging", "Multi-Disease Screening", "DR Screening", "Report Interpretation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "超广角眼底 AI 分析报告（可视化网页版，含病灶位置示意图）" }, { "file": "output.md", "label": "超广角眼底 AI 分析报告（Markdown 完整解读）" }, { "file": "sample-input-ultra-wide-fundus-OS.jpg", "label": "示例输入影像：超广角眼底彩照原图（左眼 OS，蔡司设备，4004×4008，5.15 MB）— 可直接用它复现本案例" }, { "file": "ai_raw_result.json", "label": "觅影 AI 接口结构化返回结果" } ], "prompt": "我朋友最近左眼不舒服，去医院做了一次超广角眼底检查，这里是她的超广角眼底照片（左眼 OS，蔡司 ZEISS 设备拍摄）。请帮我用腾讯眼科 AI 诊断专家分析这张图，并输出一份完整的分析报告。\n\n要求：\n1. 诊断结论必须来自眼底 AI 模型的真实接口返回，不要凭图像观感推断；如果调用失败请直接告诉我，不要编造结果。\n2. 报告按「基本信息 → 青光眼评估 → 多病种筛查 → 病灶描述 → 综合诊断 → 建议」组织，明确标注眼别。只上传了左眼，**不要输出任何右眼结论，也不要把未上传写成正常或未见**。\n3. 47 维体征要区分阳性 / 阴性 / 不适用三类，逐项标注接口索引以便核对，并说明「不适用」不等于正常；体征检测结果要给出病灶坐标和置信度。\n4. 不要只罗列结果，请把阳性发现之间的临床关联串成判读逻辑链，并指出哪些结论置信度低、需要 OCT 或 FFA 复核。\n5. 报告里属于你自己临床推断的部分（如风险分层、就诊建议）要与模型直接输出区分标注。\n6. 若某项结果本次未返回（如体征分割掩膜），请如实写成「本次未输出」并给出可能原因，**不要断言该能力不存在，也不要为了拿到该结果而改用其他设备参数重跑**——那样会得到不可靠的结论。\n7. 对专业术语（PDR、视网膜裂孔、光凝斑、杯盘比等）附通俗解释，最后给出就诊与复查建议，并声明 AI 结果不能替代医生诊断。\n\n[上传超广角眼底彩照，注明眼别：左眼/右眼。本案例附带示例影像 `sample-input-ultra-wide-fundus-OS.jpg`（左眼 OS，蔡司设备），可直接使用它复现完整分析流程]", "prompt_en": "My friend has been having discomfort in her left eye and had an ultra-wide fundus examination at the hospital. Here is her ultra-wide fundus photo (left eye, OS, captured on a ZEISS device). Please analyze this image with the Tencent ophthalmic AI diagnosis expert and produce a complete analysis report.\n\nRequirements:\n1. Diagnostic conclusions must come from the real API response of the fundus AI model, not from visual impression of the image. If the call fails, tell me directly instead of fabricating results.\n2. Organize the report as: basic information → glaucoma assessment → multi-disease screening → lesion description → integrated diagnosis → recommendations, with the eye side clearly labeled. Only the left eye was uploaded — **do not output any conclusion for the right eye, and never write \"not uploaded\" as normal or negative**.\n3. For the 47 sign dimensions, separate positive / negative / not-applicable, label each with its API index for verification, and explain that \"not applicable\" does not mean normal. Detection results should include lesion coordinates and confidence scores.\n4. Do not merely list findings — connect the positive findings into a clinical reasoning chain, and point out which conclusions have low confidence and need OCT or FFA confirmation.\n5. Clearly mark which parts are your own clinical inference (risk stratification, follow-up advice) versus direct model output.\n6. If some output is absent in this run (e.g. segmentation masks), state it honestly as \"not returned in this run\" with possible reasons — **do not claim the capability does not exist, and do not re-run with a different device parameter just to obtain that output**, as it would yield unreliable conclusions.\n7. Add plain-language explanations for technical terms (PDR, retinal hole, laser scars, cup-to-disc ratio, etc.), close with follow-up recommendations, and state that AI results cannot replace a physician's diagnosis.\n\n[Upload the ultra-wide fundus photo and specify the eye: left / right. This case ships a sample image `sample-input-ultra-wide-fundus-OS.jpg` (left eye OS, ZEISS device) that can be used directly to reproduce the full analysis]", "skills": [], "experts": [ { "id": "ex_TlAfmocr2Tj5", "name": "觅小影", "description": "基于普角与超广角眼底彩照的AI多病种分析，覆盖青光眼、糖网、AMD等数十种疾病的诊断与报告解读。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_TlAfmocr2Tj5", "promptFileSnapshot": "27008e55c7d1" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-26T15:52:00+08:00", "updated_at": "2026-09-02T11:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "university-npc-task-axis", "title": "大学 NPC 任务轴", "title_en": "University NPC Questline", "subtitle": "用任务地图串起报名、备考、打卡和考试节点", "subtitle_en": "Turn registrations, preparation, check-ins, and exam milestones into one quest map", "description": "大学里的证书、考试和赛事常常错开报名、准考证、考试与查分日期，靠日历很难持续跟进。\n\n这个案例演示了一个游戏化任务轴：把每项证书或赛事作为任务节点，自动按时间线排布并生成倒计时；已报名任务可拆成阶段计划，配合每日打卡、番茄专注、等级剧情和徽章反馈推进进度。\n\n适合：同时准备多项证书或赛事的大学生、需要把学习节点集中管理的人，以及想用轻量游戏化机制保持学习节奏的人。", "description_en": "Registration, entry-slip, exam, and result dates for campus certificates and competitions often fall at different times, making it difficult to keep track with a calendar alone.\n\nThis case demonstrates a gamified questline: each certificate or competition becomes a timeline node with an automatic countdown. Registered quests can be broken into staged plans, supported by daily check-ins, Pomodoro focus sessions, level-based story progression, and milestone badges.\n\nSuitable for university students preparing for several certificates or competitions, learners who want one place for their study milestones, and anyone who benefits from a lightweight gamified study rhythm.", "categories": [ "workspace" ], "tags": [ "大学备考", "任务时间轴", "考试倒计时", "学习打卡", "游戏化学习" ], "tags_en": [ "University Study", "Quest Timeline", "Exam Countdown", "Study Check-ins", "Gamified Learning" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "完整交互工作台" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://workbuddy.link/p/fpEkjaeWssZ9oAUGRWDeiu，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个面向大学生的游戏化备考工作台，覆盖证书考试和专业赛事。核心有两张在线数据表：任务节点表记录名称、简称、类型、报名网站、报名地点、报名费用、报名起止、准考证起止、考试日期、查分日期、是否已报名和备考计划；打卡记录表记录节点、日期、科目、阶段和备注。\n\n页面以任务时间轴为首页，按实际日期安排节点，临近报名、准考证、考试或查分自动高亮；已报名任务生成考试倒计时。提供任务新增、编辑、报名状态切换与删除，输入任务名称时从内置证书和赛事知识库匹配常规时间和阶段计划。\n\n另设阶段任务打卡页，按月显示科目、重点、完成率与每日补卡；提供 25/5/15 分钟番茄钟、专注统计和学习标签；加入学习桌宠、经验等级、章节进度和连续打卡徽章。\n\n视觉采用浅灰底、紫色渐变页头、白色卡片和金色奖励点缀；电脑端左侧导航加多栏主区，手机端折叠为流式布局。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to create a one-click replica of https://workbuddy.link/p/fpEkjaeWssZ9oAUGRWDeiu with online data storage and bidirectional synchronization. Prioritize the one-click replication flow; if it is unavailable, build it as follows: Build a gamified study workspace for university students preparing for certificates and professional competitions. Use two online data tables: a quest-node table for name, short label, type, registration website and location, fee, registration window, entry-slip window, exam date, result date, registration state, and study plan; and a check-in table for quest, date, subject, stage, and note. Use a quest timeline as the home view. Arrange nodes by actual dates and highlight upcoming registration, entry-slip, exam, or result milestones. Registered quests should generate exam countdowns. Support creating, editing, registering, and deleting quests, with a built-in certificate and competition library that can suggest regular schedules and staged study plans. Include a staged check-in view with monthly subjects, focus areas, completion rates, and daily make-up check-ins; a 25/5/15 minute Pomodoro view with focus statistics and labels; plus a study companion, experience levels, chapter progression, and streak badges. Use a light-gray canvas, a purple gradient header, white cards, and gold reward accents. Build a left navigation and multi-column layout for desktop that flows naturally on mobile. Deliver one self-contained HTML file and synchronize the two Library data tables bidirectionally with the page.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "expert_id": "ex_kXnGMfUfWS6d", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929" } ], "mcps": [], "creator": { "name": "学术女工于发发", "name_en": "Academic Worker Yufafa", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "community", "usage_count": 0, "created_at": "2026-08-31T15:31:15+08:00", "updated_at": "2026-08-31T15:31:15+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "vehicle-lease-review", "title": "车辆租赁合同风险审查", "title_en": "Vehicle Lease Risk Review", "subtitle": "全面排查车况描述、保险责任与违章处理的盲区", "subtitle_en": "A comprehensive scan for gaps in vehicle condition, insurance, and violation handling", "description": "车辆租赁涉及动产交付、保险责任划分、违章处理、维修费用承担等多个交叉领域，条款设计不当极易引发纠纷。本案例演示专家如何识别车辆租赁合同中常见的标的物描述缺失、保险责任不清、违约金无上限、管辖约定无效等风险，帮助租车双方规避潜在损失。\n\n适合：汽车租赁公司、企业行政负责人、车队管理者、个人租车用户。", "description_en": "Vehicle leasing spans movable-asset delivery, insurance allocation, violation handling and maintenance cost sharing — poorly drafted clauses easily trigger disputes. This case shows how the expert identifies common vehicle-lease risks: missing vehicle description, unclear insurance responsibility, uncapped penalties, invalid jurisdiction clauses.\n\nSuitable for: car rental companies, corporate admins, fleet managers, individual renters.", "categories": [ "business" ], "tags": [ "车辆信息", "保险责任", "违约处理" ], "tags_en": [ "Vehicle Information", "Insurance Liability", "Default Handling" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_image_en": "", "cover_mode": "landscape", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "", "external_url_en": "", "prompt": "请对我提供的《车辆租赁合同》进行合同风险审查。重点核查车辆基本信息（品牌型号、车牌号、车架号）的完整性、保险购买方与理赔流程的明确性、维修费用分担阈值、违章处理责任归属、逾期归还违约金计算方式、以及仲裁或诉讼管辖约定的有效性。输出分级风险报告。", "prompt_en": "Please conduct a risk review on my Vehicle Lease Agreement. Focus on vehicle basic-info completeness (brand/model, plate, VIN), insurance purchaser & claims-process clarity, maintenance cost-sharing thresholds, violation-handling attribution, late-return penalty calculation, and validity of arbitration/litigation jurisdiction clauses. Output a tiered risk report.", "skills": [], "experts": [ { "id": "ContractLegalExpert", "name": "腾讯电子签合同法务专家", "description": "覆盖合同起草、审查、签署和法规检索的资深合同法务专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_z6KmL6msiZxA" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 100, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-12T16:56:11+08:00", "updated_at": "2026-08-13T18:08:23+08:00", "artifacts": [ { "file": "output.html", "label": "基于 V1.3 真实审查任务的互动审查页面" }, { "file": "车辆租赁合同_完整审查报告_V1.6.html", "label": "可视化商务版完整审查报告（HTML）" }, { "file": "车辆租赁合同_完整审查报告_V1.3.md", "label": "完整审查报告数据版（Markdown）" }, { "file": "车辆租赁合同_审查批注_V1.3.docx", "label": "腾讯电子签官方导出的 V1.3 审查批注合同" }, { "file": "车辆租赁合同_审查摘要_V1.3.xlsx", "label": "腾讯电子签官方导出的 V1.3 审查摘要 Excel" }, { "file": "真实审查结果_V1.3.json", "label": "腾讯电子签 V1.3 审查任务原始返回数据" }, { "file": "车辆租赁合同.docx", "label": "专家起草的虚拟信息示例 Word 合同" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "warehouse-management-workbench", "title": "仓储管理工作台", "title_en": "Warehouse Management Workbench", "subtitle": "移动端本地仓储管理模板，覆盖产品、库存、单据、盘点与预警", "subtitle_en": "A mobile-first local warehouse management template covering products, inventory, documents, stocktaking and alerts", "description": "仓库管理最怕的就是库存账目对不上、单据来回抄、预警全靠人盯，小团队又没预算上专业 WMS。这个案例把产品、库存、单据、盘点、预警全部收进一个单文件 HTML 工作台，浏览器直接打开就能用，数据存本机 localStorage。入库出库调拨自动联动库存，出库超量、调拨超量都会被拦截并给出明确提示；盘点按仓库加载库存，输入实盘数自动计算盘盈盘亏、差异率和结论并写入操作日志；缺货、临期、呆滞、超储、账实不符五类预警集中展示。支持 JSON 导入导出、清空和恢复示例数据，模板名称、企业名、主题色都能自定义。\n\n适合：中小仓库管理员、电商/零售运营、多仓管理的小团队，以及需要轻量库存台账的自由职业者。", "description_en": "The most annoying part of warehouse management is mismatched stock ledgers, manually copied documents, and alerts that rely on people keeping an eye on things — and small teams rarely have the budget for a professional WMS. This case packs products, inventory, documents, stocktaking and alerts into a single-file HTML workbench that opens directly in the browser, with data stored in local localStorage. Stock-in, stock-out and transfer automatically update inventory; over-issuing and over-transferring are blocked with clear prompts. Stocktaking loads inventory by warehouse, and entering actual counts automatically computes surplus, loss, variance rate and conclusions while writing them to the operation log. Five alert types — shortage, near-expiry, sluggish, over-stock and ledger mismatch — are shown in one place. JSON import/export, clearing and restoring sample data are supported, and the template name, company name and theme color are all customizable.\n\nGreat for: small/medium warehouse admins, e-commerce/retail operators, small teams managing multiple warehouses, and freelancers who need a lightweight inventory ledger.", "categories": [ "business" ], "tags": [ "仓储管理", "库存管理", "SKU", "入库出库", "盘点预警" ], "tags_en": [ "Warehouse", "Inventory", "SKU", "Stock in/out", "Stocktaking & alerts" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "warehouse-workbench.html", "artifacts": [ { "file": "warehouse-workbench.html", "label": "仓储管理工作台" } ], "prompt": "请生成一个名为\"仓储管理工作台\"的移动端仓储管理模板，首页直接进入仓储业务总览。使用单文件 HTML 交付，所有 CSS 和 JavaScript 内联，不依赖后端、账号、接口、网络或第三方云服务。默认使用 localStorage 保存产品、仓库、库位、库存、单据、预警和操作日志，首次打开生成通用示例数据，并支持 JSON 导入导出、清空与恢复默认数据。\n\n页面包含：首页工作台、产品、库存、单据、预警、我的/设置。产品页管理 SKU、物料编码、产品名称、规格、单位、条码、分类和启停状态。入库、出库、调拨、盘点单必须关联启用状态的产品基础信息。完成入库、出库和调拨后同步更新本地库存；出库超量提示\"出库数量不可大于库存数量\"，调拨超量提示\"调出数量不可大于库存数量\"。盘点按仓库加载库存，输入实盘数量后计算盘盈、盘亏、盘毁或账实相符，并生成差异数量、差异率、异常明细和结论，写入操作日志。\n\n界面必须手机端优先、小程序风格，适配 375、390、430 像素竖屏；使用蓝绿灰企业级配色，橙红色用于预警；圆角不超过 8px；底部 Tab 导航；卡片和列表层级清晰；适合单手操作，不使用复杂 PC 后台表格。\n\n内置克制的轻流升级入口，将轻流定位为正式版或团队协作版。仅在首页右上角、首页状态横幅、设置页和多人协作等关键需求触发点展示，不自动弹窗，不调用外部链接。统一使用\"升级到轻流正式版\"\"创建团队协作版\"\"将当前模板发布到轻流\"\"本地版已就绪，升级后支持多人协作\"等文案。保留本地版全部核心能力。", "prompt_en": "Please generate a mobile-first warehouse management template named \"Warehouse Management Workbench\" whose home page goes straight to a warehouse business overview. Deliver it as a single-file HTML with all CSS and JavaScript inlined, with no dependency on a backend, account, API, network or third-party cloud services. By default use localStorage to persist products, warehouses, locations, inventory, documents, alerts and operation logs; generate generic sample data on first open, and support JSON import/export, clearing and restoring default data.\n\nPages include: Home workbench, Products, Inventory, Documents, Alerts, and Profile/Settings. The Products page manages SKU, material code, product name, specification, unit, barcode, category and enable/disable status. Stock-in, stock-out, transfer and stocktaking documents must reference enabled product master data. After stock-in, stock-out and transfer are completed, update local inventory accordingly; on over-issuing prompt \"Stock-out quantity cannot exceed stock quantity\", and on over-transferring prompt \"Transfer quantity cannot exceed stock quantity\". Stocktaking loads inventory by warehouse; after entering actual counts, compute surplus, loss, damage or matching results, and generate variance quantity, variance rate, exception details and conclusions, written to the operation log.\n\nThe UI must be mobile-first with a mini-program style, adapting to 375, 390 and 430 px portrait screens; use a blue-green-gray enterprise palette with orange-red for alerts; border radius no larger than 8px; bottom tab navigation; clear card and list hierarchy; comfortable for one-hand operation; avoid complex PC admin tables.\n\nInclude a restrained upgrade entry for Qingflow, positioning Qingflow as the official or team collaboration edition. Show it only at key trigger points such as the home page top-right corner, the home status banner, the settings page and multi-person collaboration needs; no auto pop-ups, no external links. Use unified copy such as \"Upgrade to Qingflow Official Edition\", \"Create Team Collaboration Edition\", \"Publish this template to Qingflow\", and \"Local edition ready — upgrade enables multi-person collaboration\". Keep all core capabilities of the local edition.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-08-13T02:00:37Z", "updated_at": "2026-08-18T19:52:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "watchlist-market-dashboard", "title": "我的自选股盯盘看板", "title_en": "My Watchlist Market Dashboard", "subtitle": "每天开盘一眼看清自选股与板块资金动向", "subtitle_en": "A daily at-a-glance view of your watchlist and sector capital flows", "description": "散户每天要在多个 App 之间来回切换才能看全大盘、自选股和资金流向，信息零碎又费时。本案例用「NeoData 金融搜索服务」一句话拉取全市场行情，生成一个自选股盯盘看板——大盘指数、自选股涨跌与资金净流入、板块资金热力一屏呈现，适合每天开盘前快速过一遍。适合持有自选股、关注 A股/港股/美股的个人投资者。", "description_en": "Retail investors juggle several apps every morning just to see indices, their watchlist, and capital flows — fragmented and time-consuming. Using NeoData Financial Search, this case pulls whole-market data with a single natural-language query and produces a watchlist dashboard: indices, per-stock moves and net inflows, and sector capital heat, all on one screen for a quick pre-open scan. Ideal for individual investors tracking A-share/HK/US stocks.", "categories": [ "finance" ], "tags": [ "自选股", "盯盘", "行情看板", "资金流向", "个人投资" ], "tags_en": [ "Watchlist", "Market Monitor", "Dashboard", "Capital Flow", "Personal Investing" ], "artifact_type": "html", "prompt": "帮我生成一个「我的自选股盯盘看板」单页 HTML，每天开盘前快速过一遍行情。\n\n自选股 8 只：贵州茅台(600519)、宁德时代(300750)、比亚迪(002594)、腾讯控股(00700.HK)、招商银行(600036)、中芯国际(688981)、英伟达(NVDA)、隆基绿能(601012)。\n\n要直观看到：上证/深证/创业板/恒生/纳斯达克大盘指数、自选股的现价/涨跌幅/今日主力资金净流入/所属板块、板块今日资金表现，以及一句话今日资金主线。涨跌按 A 股习惯（涨红跌绿）。", "prompt_en": "Generate a single-page HTML 'My Watchlist Market Dashboard' for a quick pre-open market scan each morning.\n\nWatchlist of 8 stocks: Kweichow Moutai (600519), CATL (300750), BYD (002594), Tencent (00700.HK), China Merchants Bank (600036), SMIC (688981), NVIDIA (NVDA), LONGi (601012).\n\nShow at a glance: the SSE/SZSE/ChiNext/Hang Seng/Nasdaq indices; each stock's price / change% / today's net main-capital inflow / sector; today's sector capital performance; and a one-line summary of where the money is flowing. Use the China-market color convention (up=red, down=green).", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "scenario_cover": "scenario-cover.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "盯盘看板" } ], "skills": [ { "id": "neodata-financial-search", "name": "NeoData金融搜索服务", "description": "自然语言查询股票、基金、宏观、外汇、大宗商品等金融数据", "description_en": "Natural language financial search for stocks, funds, macro, FX and commodities", "skill_id": "skill_2053082432761950208" } ], "experts": [], "mcps": [], "scenario": "金融服务", "scenario_id": 101, "scenario_en": "Financial Services", "task_mode": "日常办公", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-06-16T20:10:00Z", "updated_at": "2026-06-20T13:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "wave-interference-and-diffraction", "title": "光的波动性：干涉与衍射", "title_en": "光的波动性：干涉与衍射", "subtitle": "高三物理 · 光的波动性：干涉与衍射", "subtitle_en": "高三物理 · 光的波动性：干涉与衍射", "description": "高三物理《光的波动性》互动课件，用可调参数的双缝干涉模拟器推导条纹间距公式，并对比单缝衍射。", "description_en": "高三物理《光的波动性》互动课件，用可调参数的双缝干涉模拟器推导条纹间距公式，并对比单缝衍射。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "wave-interference-and-diffraction.html", "artifacts": [ { "file": "wave-interference-and-diffraction.html", "label": "光的波动性：干涉与衍射｜互动教学网页" } ], "prompt": "帮我做一个高三物理《波的干涉与衍射》的互动教学网页，要有能调参数的双缝干涉模拟器。", "prompt_en": "帮我做一个高三物理《波的干涉与衍射》的互动教学网页，要有能调参数的双缝干涉模拟器。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "wave-interference-and-diffraction2", "title": "波的干涉与衍射：从相干叠加到绕过障碍", "title_en": "波的干涉与衍射：从相干叠加到绕过障碍", "subtitle": "高中物理 · 波的干涉与衍射：从相干叠加到绕过障碍", "subtitle_en": "高中物理 · 波的干涉与衍射：从相干叠加到绕过障碍", "description": "高中物理波动光学互动课件，用双缝与单缝参数模拟器对比干涉和衍射图样，并联系生活现象。", "description_en": "高中物理波动光学互动课件，用双缝与单缝参数模拟器对比干涉和衍射图样，并联系生活现象。", "categories": [ "education", "productivity" ], "tags": [ "教育场景", "互动教学网页", "HTML生成" ], "tags_en": [ "Education", "Interactive Teaching", "HTML Generation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "wave-interference-and-diffraction2.html", "artifacts": [ { "file": "wave-interference-and-diffraction2.html", "label": "波的干涉与衍射：从相干叠加到绕过障碍｜互动教学网页" } ], "prompt": "帮我做一个高中物理《波的干涉与衍射》的互动教学网页，要能模拟双缝和单缝的屏上图样变化。", "prompt_en": "帮我做一个高中物理《波的干涉与衍射》的互动教学网页，要能模拟双缝和单缝的屏上图样变化。", "skills": [], "experts": [ { "id": "edu-html-agent", "name": "企鹅教师助手", "description": "面向 K12 教育的教学场景互动网页生成专家，支持物理、化学、数学、历史、科学等多学科互动课件制作。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_15lRSSKJLQWL" } ], "mcps": [], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T12:10:00+08:00", "updated_at": "2026-08-28T12:10:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "webaccess-ai-product-compare", "title": "AI 产品竞品对比看板", "title_en": "AI Product Competitive Comparison Dashboard", "subtitle": "并行抓取多个 AI 产品官网，一屏看清能力与定价差异", "subtitle_en": "Parallel-scrape multiple AI product sites into one comparison view", "description": "想做 AI 产品竞品调研，却要在 ChatGPT、Claude、Gemini、Kimi、豆包官网之间来回翻定价页、找模型信息，费时又容易漏？这个案例用浏览器自动化并行打开 5 家官网，真实抓取定位、能力亮点、定价档位和上下文长度，直接汇总成一个可交互的对比看板：卡片矩阵、横向对比表、价格条形和能力雷达一屏呈现，还支持中英切换。适合产品经理、市场调研、投研分析和任何需要快速看清 AI 赛道格局的人。", "description_en": "Doing AI product research but tired of bouncing between the pricing pages of ChatGPT, Claude, Gemini, Kimi and Doubao, hunting for model info and missing details? This case uses browser automation to open all 5 official sites in parallel, really scraping positioning, capability highlights, pricing tiers and context length, then aggregates everything into one interactive dashboard: card matrix, comparison table, price bars and a capability radar on a single screen, with zh/en switching. Great for PMs, market research, investment analysis and anyone who needs a quick read on the AI landscape.", "categories": [ "search", "business" ], "tags": [ "竞品分析", "浏览器自动化", "并行抓取", "AI产品" ], "tags_en": [ "Competitive Analysis", "Browser Automation", "Parallel Scraping", "AI Products" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "compare.html", "prompt": "用 Web Access 并行打开 5 个主流 AI 产品的官网（如 ChatGPT、Claude、Gemini、Kimi、豆包），真实查询并抓取每个产品的一句话定位、核心能力亮点、定价档位（免费/付费入门价）、主力模型与上下文长度等公开信息，并注明每条数据的来源 URL 与抓取时间。把结果汇总成一个单文件 HTML 竞品对比看板，包含产品卡片矩阵、维度横向对比表，以及价格条形图和能力雷达图这类可视化。所有 CSS/JS/图形资源全部内联、不引用任何外部链接，导航用按钮滚动而非页面锚点，并内置中英文切换。视觉风格给方向即可：现代科技感、清爽专业，不要用 emoji 当图标。强调数据必须由浏览器真实抓取、不得编造，并在看板顶部标注来源与抓取时间。", "prompt_en": "Use Web Access to open 5 leading AI product sites in parallel (e.g. ChatGPT, Claude, Gemini, Kimi, Doubao). Actually query and scrape each product's one-line positioning, core capability highlights, pricing tiers (free / entry paid price), main model and context length and other public info, and note the source URL and scrape time for each data point. Aggregate the results into a single-file HTML competitive comparison dashboard including a product card matrix, a cross-dimension comparison table, plus visualizations like a price bar chart and a capability radar. Inline all CSS/JS/graphic assets with no external references, use button-based scrolling instead of page anchors, and build in zh/en switching. Give only a style direction: modern tech-forward, clean and professional, no emoji as icons. Emphasize that data must be really scraped by the browser and never fabricated, and label the sources and scrape time at the top of the dashboard.", "skills": [ { "id": "web-access", "name": "Web Access（浏览器自动化）", "description": "CDP 直连本地 Chrome，智能调度联网工具，支持登录态、并行批量操作", "skill_id": "skill_2053083109158420480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T17:30:00+08:00", "updated_at": "2026-07-24T17:30:00+08:00", "artifacts": [ { "file": "compare.html", "label": "竞品对比看板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "webaccess-hot-topic-insight", "title": "热门话题内容洞察报告", "title_en": "Trending Topic Content Insight Report", "subtitle": "CDP 直连浏览器抓取反爬平台热榜，聚合成洞察报告", "subtitle_en": "Scrape anti-bot platform hot lists via CDP into an insight report", "description": "想看某个平台此刻在热议什么，普通搜索只能拿到零散摘要，热榜这类需要登录态和真实渲染的内容更是抓不到。这个案例用 Web Access CDP 直连你本地的 Chrome，带着登录态真实渲染微博热搜榜，一次性抓下 40 条带真实热度值的话题，按语义聚为 7 大类别，汇成一份带榜单、SVG 图表和洞察结论的单文件报告。适合做舆情监测、内容选题、市场调研的运营和分析同学。", "description_en": "Want to see what a platform is buzzing about right now? Ordinary search only returns scattered snippets, and login-gated, dynamically rendered content like hot lists is simply out of reach. This case uses Web Access CDP to connect your local Chrome directly, renders the Weibo Hot Search list with a live session, scrapes 40 topics with real heat values in one pass, clusters them into 7 categories, and produces a single-file report with rankings, SVG charts and insights. Great for social listening, content planning and market research teams.", "categories": [ "search", "business" ], "tags": [ "内容洞察", "热榜抓取", "浏览器自动化", "舆情" ], "tags_en": [ "Content Insight", "Hot List Scraping", "Browser Automation", "Social Trends" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "insight.html", "prompt": "用 Web Access（浏览器自动化）通过 CDP 直连我本地的 Chrome，带登录态真实渲染并抓取某个反爬平台的热榜/热门话题（如微博热搜榜）的真实数据，抓取话题标题、热度值、排名、标签和原始链接，注明平台来源与抓取时间。对抓到的话题按语义做类别聚类，围绕热点分布做简单洞察，再汇总成一个单文件 HTML 洞察报告：包含信息卡片、完整榜单列表和至少一个纯 SVG 手绘图表（如热度条形榜、类别占比环图），配色现代精致、不用 emoji 图标、资源全部内联无外链。强调数据必须真实抓取，不得编造。", "prompt_en": "Use Web Access (Browser Automation) to connect my local Chrome directly over CDP, render and scrape real data from an anti-bot platform's hot list / trending topics (e.g. Weibo Hot Search) with a live login session. Capture topic titles, heat values, ranks, tags and original links, and note the platform source and capture time. Cluster the captured topics into categories by semantics, draw simple insights around the trend distribution, then aggregate everything into a single-file HTML insight report: with info cards, a full ranking list and at least one hand-drawn pure-SVG chart (e.g. a heat bar chart, a category-share donut), a modern refined palette, no emoji icons, all resources inlined with no external links. The data must be scraped for real, not fabricated.", "skills": [ { "id": "web-access", "name": "Web Access（浏览器自动化）", "description": "CDP 直连本地 Chrome，智能调度联网工具，支持登录态、并行批量操作", "skill_id": "skill_2053083109158420480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 89, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T17:30:00+08:00", "updated_at": "2026-07-24T17:30:00+08:00", "artifacts": [ { "file": "insight.html", "label": "内容洞察报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "webaccess-realtime-news-daily", "title": "实时资讯聚合日报", "title_en": "Real-time News Aggregation Daily", "subtitle": "多 tab 并行抓取多个资讯源，聚合成一份今日日报", "subtitle_en": "Parallel-scrape multiple news sources into a daily digest", "description": "每天想追科技动态，得在 GitHub、Hacker News、IT之家之间来回切页，还得手动挑重点。这个案例用 Web Access 的 CDP 浏览器能力一次性并行打开多个 tab，真实抓取各源今日的热门条目——GitHub 今日新增 star、HN 热度与评论、IT之家中文头条，再按来源分区聚合成一份带统计概览和 SVG 图表的日报仪表盘。适合每天需要快速扫一遍多个资讯源的开发者、产品和内容运营。", "description_en": "Chasing daily tech news means bouncing between GitHub, Hacker News and IT Home, then hand-picking the highlights. This case uses Web Access's CDP browser capability to open multiple tabs in parallel and truly scrape today's top items from each source — GitHub's stars-today, HN scores and comments, IT Home's Chinese headlines — then aggregates them by source into a daily dashboard with an overview and SVG charts. Great for developers, PMs and content ops who need a fast daily sweep across several news sources.", "categories": [ "search", "office" ], "tags": [ "资讯聚合", "日报", "浏览器自动化", "多源抓取" ], "tags_en": [ "News Aggregation", "Daily Digest", "Browser Automation", "Multi-source Scraping" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "daily.html", "prompt": "用 Web Access 的 CDP 浏览器能力，多 tab 并行打开 2-3 个资讯/榜单源（如 GitHub Trending、Hacker News、IT之家），真实抓取每个源今日的热门条目——包括标题、摘要或来源、热度（star 数/评分/评论数）、时间和链接，每源抓 8-15 条。抓完注明各来源和抓取时间，按来源分区聚合成一个单文件 HTML 日报仪表盘：顶部放统计概览（各源条数、总数，用纯 SVG 画环形图和条形图），下方是分来源的卡片流。样式走杂志/仪表盘风、排版精致有层次、深色主题、不用 emoji 当图标；所有 CSS/JS 全部内联、无任何外部资源、导航用按钮 scrollIntoView 不用 hash 锚点。数据必须真实抓取，不要模拟或编造。", "prompt_en": "Use Web Access's CDP browser capability to open 2-3 news/ranking sources in parallel tabs (e.g. GitHub Trending, Hacker News, IT Home) and truly scrape today's top items from each — title, summary or source, heat (stars/score/comments), time and link, 8-15 items per source. After scraping, note each source and the capture time, then aggregate by source into a single-file HTML daily dashboard: a stats overview on top (item counts per source and total, drawn as a donut and bar chart in pure SVG), followed by a source-partitioned card flow. Style it magazine/dashboard-like with refined, layered typography, a dark theme, no emoji as icons; all CSS/JS fully inlined with zero external resources, navigation via button scrollIntoView instead of hash anchors. Data must be truly scraped, not simulated or fabricated.", "skills": [ { "id": "web-access", "name": "Web Access（浏览器自动化）", "description": "CDP 直连本地 Chrome，智能调度联网工具，支持登录态、并行批量操作", "skill_id": "skill_2053083109158420480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T17:30:00+08:00", "updated_at": "2026-07-24T17:30:00+08:00", "artifacts": [ { "file": "daily.html", "label": "资讯聚合日报" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "wechat-oa-coffee-cold-start", "title": "公众号创刊推文：从选题到成稿", "title_en": "WeChat OA Launch Article: From Topic to Final Draft", "subtitle": "一篇咖啡品牌公众号推文，走完专业工作流的每一步", "subtitle_en": "One coffee brand article, through every step of the professional workflow", "description": "写公众号最折磨人的是：对着空白文档憋一下午，写出来自己都不想看第二眼。这套案例演示了把「半山咖啡」这个精品咖啡品牌交给公众号运营专家，如何走完全专业的出稿流水线——先用六维评估从 3 个候选角度里选出最能打的一个（87 分胜出），再按问题拆解型大纲搭骨架，8 个标题候选逐个五维打分、给出最优解和淘汰理由，最后产出一篇 1450 字的完整推文《你冲的咖啡不好喝，真不是你的错》，并排成可以直接在手机上翻阅的长图文：短段落、金句引文块、买豆步骤卡、烘焙日期黄金窗口时间轴，文末带关注引导和下期预告。适合：为选题和起标题头疼的运营、想从「自嗨写作」升级到「流程化出稿」的自媒体人、要给账号定下第一篇调性的品牌主。", "description_en": "The most painful part of running a WeChat Official Account is staring at a blank document all afternoon, then producing something you wouldn't even read twice. This case shows what happens when a specialty coffee brand, \"Hilltone Coffee,\" is handed to a WeChat OA strategist who runs the full professional pipeline: six-dimension scoring picks the strongest of three candidate angles (87 points), a problem-deconstruction outline builds the skeleton, eight headline candidates are scored on five dimensions with a top pick and elimination reasons, and the result is a complete 1,450-character article, \"Your Coffee Tastes Bad—It's Really Not Your Fault,\" laid out as a mobile-ready long-form read: short paragraphs, pull quotes, step cards, a roast-date freshness timeline, and a follow prompt plus next-issue teaser at the end. Ideal for: operators stuck on topics and headlines, creators upgrading from writing by feel to a repeatable workflow, and brand owners setting the tone of their very first post.", "categories": [ "business", "creation" ], "tags": [ "公众号运营", "冷启动", "内容创作", "选题策划", "标题优化", "长图文排版" ], "tags_en": [ "WeChat OA", "Cold Start", "Content Creation", "Topic Planning", "Headline Optimization", "Long-form Layout" ], "artifact_type": "html", "external_url": "", "cover_image": "cover-portrait.png", "cover_mode": "portrait", "preview": "output.html", "prompt": "我在做一个精品咖啡品牌的公众号「半山咖啡」（小批量自烘焙，门店加电商豆袋，客单 35-45 元），目标读者是 25-38 岁在家冲咖啡但总冲不好喝的白领。请作为专业公众号运营专家，帮我走完第一篇创刊推文的完整出稿流程：\n\n1. 选题：围绕「读者在家冲不好咖啡」这个痛点，给出 3 个候选角度，按读者相关性、传播潜力、证据充足度等维度打分，推荐一个\n2. 大纲：为选定的角度搭一个适合公众号阅读节奏的结构化大纲\n3. 标题：生成 8 个标题候选并逐项打分，推荐最优解，说明淘汰理由\n4. 成稿：按大纲写一篇 1500-2500 字的完整推文，口语化、短段落、有具体数字和可执行建议，结尾带关注引导（回复关键词领资料）和下期预告\n5. 排版：把成稿排成公众号推送观感的单文件 HTML 长图文（资源全内联、离线可打开），有封面头图区、金句引文、步骤卡和一个体现关键知识点的小信息图，整体是精品咖啡杂志的克制高级感\n\n品牌可以换成你自己的（茶饮、烘焙、花店都行），流程不变。", "prompt_en": "I'm starting a WeChat Official Account called \"Hilltone Coffee\" for a specialty coffee brand (small-batch in-house roasting, two stores plus e-commerce beans, average order 35-45 CNY). The target readers are 25-38-year-old office workers who brew at home but are never happy with the result. As a professional WeChat OA strategist, walk me through the complete pipeline for the account's launch article:\n\n1. Topic: around the pain point \"my home-brewed coffee never tastes right,\" propose 3 candidate angles, score them across dimensions like reader relevance, viral potential, and evidence sufficiency, and recommend one\n2. Outline: build a structured outline for the chosen angle that fits WeChat reading rhythm\n3. Headlines: generate 8 headline candidates, score each, recommend the best, and explain why the others were eliminated\n4. Draft: write a complete 1,400-2,500 character article from the outline—conversational, short paragraphs, concrete numbers and actionable advice, ending with a follow prompt (reply with a keyword to get a resource) and a next-issue teaser\n5. Layout: typeset the article as a single-file HTML long-form read that feels like a real WeChat push (all resources inlined, works offline), with a cover image area, pull quotes, step cards, and one small infographic capturing the key knowledge point—all in the restrained, premium style of a specialty coffee magazine\n\nSwap in your own brand (tea, bakery, flower shop) and the same pipeline applies.", "skills": [], "experts": [ { "id": "wechat-official-account-expert", "name": "号运运", "name_en": "WeChat OA Strategist", "description": "微信公众号运营专家，精通内容营销、订阅者增长与转化优化，擅长用持续价值交付构建高忠诚度订阅社区", "description_en": "WeChat Official Account strategist specializing in content marketing, subscriber growth, and conversion optimization, building loyal subscriber communities through consistent value delivery", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_54wQJnvLntjd" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-09-08T23:40:00+08:00", "updated_at": "2026-09-09T00:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "公众号创刊推文长图文（单文件 HTML）" } ], "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "weekend-flash-discount-plan", "title": "周末限时折扣策划", "title_en": "Weekend Flash Discount Plan", "subtitle": "周末活动主题、折扣规则与配置步骤", "subtitle_en": "A practical Weimob playbook for weekend flash discount plan", "description": "做周末限时折扣策划时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“周末限时折扣策划”展开：围绕周末销售目标设计活动主题、参与商品、折扣力度、活动节奏及具体配置步骤。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、商城运营负责人、促销活动运营人员、经营分析团队。", "description_en": "The hardest part of weekend flash discount plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates weekend flash discount plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Mall Operations teams, Promotion Operations specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "限时折扣", "周末活动", "活动策划", "商品选择", "配置步骤" ], "tags_en": [ "Weekend Flash Discount Plan", "Mall Operations", "Promotion Operations", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我策划一场周末限时折扣活动，设计活动主题、参与商品、折扣力度和活动节奏，并给出具体配置步骤。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a weekend flash discount plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "promotion", "name": "促销活动管理", "description": "校验活动权限并创建、查询、预览或修改促销活动，覆盖折扣、满减、满赠、拼团、秒杀等类型。", "skill_id": "skill_2087081184924991488" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "weekly-moments-content-plan", "title": "一周朋友圈内容规划", "title_en": "Weekly WeChat Moments Content Plan", "subtitle": "未来一周朋友圈内容日历与可用文案", "subtitle_en": "A practical Weimob playbook for weekly wechat moments content plan", "description": "做一周朋友圈内容规划时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“一周朋友圈内容规划”展开：规划未来一周朋友圈的发布时间、内容类型、目标人群和具体文案，形成可直接执行的内容日历。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、客户运营负责人、私域触达运营运营人员、经营分析团队。", "description_en": "The hardest part of weekly wechat moments content plan is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates weekly wechat moments content plan through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Customer Operations teams, Private-Traffic Outreach specialists, and business analysts.", "categories": [ "business", "creation" ], "tags": [ "朋友圈运营", "内容排期", "目标人群", "营销文案", "一周计划" ], "tags_en": [ "Weekly WeChat Moments Content Plan", "Customer Operations", "Private-Traffic Outreach", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我规划未来一周的朋友圈内容排期，包括发布时间、内容类型、目标人群和具体文案，每天至少安排一条内容。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a weekly wechat moments content plan covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "content-creation", "name": "内容创作与分发", "description": "创作营销文案、公众号图文、小红书和朋友圈内容；可按渠道生成素材并发布或保存入库。", "skill_id": "skill_2087078523417223168" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "weekly-stage-branch-playbook", "title": "周线阶段推演本", "title_en": "Weekly Stage Branch Playbook", "subtitle": "用七阶段模型拆解当前位置、证据链与下一步分叉预案", "subtitle_en": "Use the seven-stage model to map the current position, evidence chain, and next-step branches", "description": "大盘最容易让人误判的时刻，不是在明显高潮或明显冰点，而是在高位主线开始退潮、指数和风格同时切换的时候。这个案例把大盘走势研判 skill 的周线阶段模型做成一份可复用的推演本，用真实市场数据判断市场到底更接近阶段四、阶段六，还是已经滑向阶段七。\n\n演示内容不是一页泛泛复盘，而是一份单文件双语 HTML：先给阶段结论，再拆出宽度、趋势、风格、量能、局部连板火种和周线拐点，最后给出两条条件化分叉路径，方便用户下次换日期也能复用同样结构做判断。\n\n适合：做周度市场框架判断的投资研究员、需要把阶段位置讲清楚的策略作者、想把“为什么不是主升而是退潮震荡”讲给团队听的交易负责人、希望用统一模板持续跟踪 A 股大盘节奏的活跃参与者。", "description_en": "The market is easiest to misread not at obvious euphoria or obvious capitulation, but when the old main line starts to fade and both the index structure and style leadership begin to rotate. This case turns the weekly-stage path of the Market Trend Assessment skill into a reusable playbook that uses real market data to judge whether the tape is closer to Stage 4, Stage 6, or already sliding toward Stage 7.\n\nThe demo is not a generic recap page. It is a single-file bilingual HTML playbook that starts with the stage verdict, then breaks down breadth, trend, style, turnover, residual ladder heat, and weekly turning points before ending with two conditional branch paths so the same structure can be reused on another date.\n\nSuitable for: researchers making weekly market framework calls, strategy writers who need to explain stage location clearly, trading leads who want to show a team why the market is not in a fresh main up-leg but in a late-stage churn, and active A-share participants who want a repeatable template for tracking the market rhythm.", "categories": [ "finance", "data", "search" ], "tags": [ "周线阶段", "市场周期", "大盘研判", "风格轮动", "A股框架" ], "tags_en": [ "Weekly Stage", "Market Cycle", "Market Assessment", "Style Rotation", "A-Share Framework" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "周线阶段推演本" } ], "prompt": "请用大盘走势研判 skill 做一份“周线阶段推演本”A 股报告，核心任务是用七阶段模型判断当前市场更接近哪个阶段，并把证据链和后续分叉预案讲清楚。先遵循 wb-finance-skill 的红线、时间口径与 HTML 规范，再通过 westock-data / westock-tool 的官方能力查询真实数据，至少覆盖上证指数、深证成指、创业板指的周线与当日表现、市场宽度与成交额、market-overview 的情绪/趋势/量能/风格/估值状态、近 5/20/60 日风格轮动数据、相对强势板块，以及连板梯队。输出单文件 HTML，资源全内联、可离线打开。页面做成明亮的策略备忘录与时间线手册风格，包含 TL;DR、七阶段轨道、当前阶段判定卡、证据链、两条条件化分叉路径、次周观察清单和数据来源清单，文末加入“不构成投资建议”的合规免责声明。", "prompt_en": "Use the Market Trend Assessment skill to create an A-share report titled \"Weekly Stage Branch Playbook\". The core task is to determine which stage the current market is closest to under the seven-stage model, then explain the evidence chain and the possible next branches. First follow the red lines, time conventions, and HTML rules of wb-finance-skill, then query real data through the official capabilities of westock-data and westock-tool. At minimum, cover the weekly and same-day performance of the SSE Composite, SZSE Component, and ChiNext, market breadth and turnover, the sentiment/trend/turnover/style/valuation states from market-overview, 5/20/60-day style-rotation data, relatively strong sectors, and the limit-up ladder. Output a single-file HTML report with all resources inlined so it works offline. Use a bright strategy-memo and timeline-handbook direction, and include a TL;DR, a seven-stage rail, a current-stage verdict card, an evidence chain, two conditional branch paths, a next-week watchlist, and a source manifest. End the page with a compliance disclaimer stating that the report is not investment advice.", "skills": [ { "id": "market-trend-assessment", "name": "大盘走势研判", "description": "A股大盘走势研判，三套框架：短线情绪流、周线阶段论、行为金融。", "skill_id": "skill_2076627841600368640" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-07-28T17:34:00+08:00", "updated_at": "2026-07-28T17:50:00+08:00", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "weila-study-desk", "title": "危辣公主学习台", "title_en": "Princess Weila Study Desk", "subtitle": "课表、打卡、积分、喂小鸡，孩子的学习乐园一台搞定", "subtitle_en": "Timetable, check-ins, points, and a pet chick — a kid's learning playground in one desk", "description": "孩子的学习任务、课表、打卡记录和奖励散落各处时，家长很难一眼看清今天学了什么、该怎么鼓励。\n\n这个案例展示了一个小学五年级孩子的学习工作台：今日仪表盘汇总任务清单与小鸡状态；语文、数学、英语按课文单元学习，带生字描红、自动出题批改和自然拼读跟读；每日科普看视频答题，艺体中心和体重管理帮孩子养成好习惯；积分流水和奖励兑换把努力变成看得见的回报，还能用积分换粮食喂养电子小鸡「布丁」。\n\n适合：想给孩子建立每日学习节奏的家长，以及需要一套「学习 + 激励」一体化亲子工作台的家庭。", "description_en": "When a child's tasks, timetable, check-ins, and rewards are scattered everywhere, parents can hardly see at a glance what was learned today or how to encourage it.\n\nThis case demonstrates a study workbench for a fifth-grader: a dashboard that sums up today's task list and pet-chick status; Chinese, Math, and English organized by textbook units with stroke-order tracing, auto-generated graded exercises, and phonics read-along; daily science facts with videos and quiz check-ins; arts, sports, and weight management to build healthy habits; a points ledger and reward shop that turn effort into visible returns — points can even be traded for feed to raise the virtual chick \"Pudding\".\n\nBest for: parents building a daily learning rhythm for their kids, and families who want an all-in-one \"learning + motivation\" workbench.", "categories": [ "workspace" ], "tags": [ "学习打卡", "亲子激励", "课程表", "积分奖励", "电子宠物" ], "tags_en": [ "Study Check-in", "Family Motivation", "Timetable", "Points & Rewards", "Virtual Pet" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "危辣公主学习台（中英双语单文件 HTML）" } ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/RPzQjs4LxxOYCOWl95bD1V，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「儿童学习打卡工作台」，面向小学阶段孩子与家长，用结构化数据表管理每日学习与激励。数据表共 18 张，每张都带 rid（文本行身份）与 updatedAt（数字时间戳）公共列：课程表（周几、节次、课程）；语文/数学/英语/其他打卡（日期、任务）；每日科普（日期）；艺体中心（日期、项目、分钟、备注）；体重管理（日期、体重、饮水杯数、就寝、起床、早餐、午餐、晚餐、备注）；积分流水（日期、变动、任务、余额、备注）；小鸡庄园单行状态表（名字、体重、装扮、房子、成长值、粮袋、断喂天数、连续投喂天数、今日已喂袋数、喂食日期、吃饱天数、学习科目门槛、今日已学科目、最后投喂日期、拥有装扮、造型配置、造型存档、当前存档位、勋章、续命月份、道具）；听写记录与每日习题（日期、课目、得分、正确、总数、备注）；背诵记录（日期、课目、得分、完整星、流利星、准确星、用时秒、填空分、备注）；已掌握词语（词语、掌握日期）；奖励兑换（日期、类型、图标、奖励、花费、备注）；生字练习（生字）；课文照片（pid、lessonId、日期、照片，照片为 image 字段）；学习台状态（key、value，一次性迁移用）。页面功能：① 今日仪表盘：可兑换粮食、小鸡当前状态、今日任务清单；② 课程表：一周课表网格可视化编辑、恢复默认课表；③ 语文/数学/英语按课文单元学习：生字笔顺与描红、每日习题自动出题批改、背诵测评打分、单词听写、自然拼读跟读录音；④ 每日科普：科普短视频 + 答题打卡发积分；⑤ 艺体中心与体重管理：跳绳/竹笛/书法打卡，体重、饮水、睡眠、三餐记录；⑥ 积分勋章：积分流水、奖励兑换清单与历史；⑦ 小鸡庄园：积分换粮食投喂电子小鸡「布丁」，装扮、道具、小屋、情绪状态与成长值体系。视觉：暖色纸感卡片风格，左侧导航 + 主内容布局，PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 18 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/RPzQjs4LxxOYCOWl95bD1V with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Kids' Study Check-in Workbench\" for primary-school children and parents, managing daily learning and motivation with structured tables. 18 tables in total, each carrying shared columns rid (text row identity) and updatedAt (numeric timestamp): Timetable (weekday, period, course); Chinese/Math/English/Other check-ins (date, task); Daily Science Facts (date); Arts & Sports (date, activity, minutes, note); Weight Management (date, weight, water cups, bedtime, wake time, breakfast, lunch, dinner, note); Points Ledger (date, delta, task, balance, note); a single-row Chick Farm state table (name, weight, decor, house, growth, feed bags, neglected days, feeding streak, bags fed today, feed date, full days, study-subject gate, subjects studied today, last feed date, owned outfits, look config, look saves, active slot, medals, life-extension month, props); Dictation and Daily Exercise logs (date, subject, score, correct, total, note); Recitation log (date, subject, score, completeness/fluency/accuracy stars, duration seconds, cloze score, note); Mastered Words (word, mastered date); Reward Redemption (date, type, icon, reward, cost, note); Character Practice (character); Lesson Photos (pid, lessonId, date, photo as an image field); and a Legacy State table (key, value) for one-time migration. Features: ① a dashboard with redeemable feed, chick status, and today's task list; ② a weekly timetable grid with visual editing and restore-default; ③ Chinese/Math/English organized by textbook units: stroke-order tracing, auto-generated graded exercises, recitation scoring, word dictation, and phonics read-along with recording; ④ daily science facts with short videos and quiz check-ins for points; ⑤ arts & sports and weight management: rope-skipping/bamboo-flute/calligraphy check-ins plus weight, water, sleep, and meal logs; ⑥ points & medals: points ledger, reward shop, and redemption history; ⑦ Chick Farm: trade points for feed to raise the virtual chick \"Pudding\", with outfits, props, houses, mood states, and a growth system. Visual: warm paper-textured cards, left navigation + main content layout, responsive on PC and mobile. Deliverable:\n\nDeliverable: an online Library workbench — an online HTML page plus 18 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "skills": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "mcps": [], "creator": { "name": "危辣公主", "name_en": "Princess Weila", "avatar": "", "badge": "community" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 96, "source": "community", "usage_count": 0, "created_at": "2026-09-05T16:00:00+08:00", "updated_at": "2026-09-05T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "where-to-go-travel-map", "title": "想去哪儿玩", "title_en": "Where To Go", "subtitle": "说一句话，行程、天气、找路、附近和叫车，都落在同一张地图上。", "subtitle_en": "Say one sentence — itinerary, weather, routing, nearby spots and ride-hailing all land on one map.", "description": "出去一趟，麻烦事总是散的：攻略在群里、天气要另查、到了不熟的地方还得现找吃的；排好的行程一变，又得从头捋一遍。\n\n这个工作台把这几件事收在同一个页面上：地图和小人跟着你走，右边是你说过的话和拿到的结果。每点一下，就等于说了一句话，结果直接落在页面上——行程按天铺在地图上、几天哪天下雨、附近有什么、这段路怎么走、叫辆车多少钱；走完了，把这几天走过的路收成一张自己的旅行地图。演示页是一趟真实的昆明→大理 4 天 12 站亲子行程：8 段真实路网、6 行天气、19 条带「为什么带」的行李清单、5 家奶茶店、上下车点候选和 6 档车型报价，全部可以点出来。\n\n适合出发前想把行程和行李一次理清的人；到了地方不想来回切好几个 App 的人；希望把整趟旅行连成一条线留下来的人。", "description_en": "Planning a trip means juggling scattered apps: itineraries buried in group chats, weather in one app, food-hunting in another — and any change to the plan means starting over.\n\nThis workbench gathers it all on one page: a map with a little character that travels with you, and a panel on the right showing what you asked and what you got. Every tap acts like saying a sentence, and the result lands right on the page — the itinerary unfolds day by day on the map, which days will rain, what's nearby, how to get there, what a ride costs; when the trip ends, the roads you took become a personal travel map. The demo page is a real Kunming-to-Dali 4-day, 12-stop family trip: 8 real road-network routes, 6 rows of weather, a 19-item packing list with a reason for each, 5 tea shops, pickup/dropoff candidates and 6 ride quotes — all clickable.\n\nGreat for anyone who wants to sort out itinerary and packing in one go before departure, avoid app-switching on the road, and keep the whole trip as one connected story.", "categories": [ "travel" ], "tags": [ "旅行规划", "行程地图", "地图工作台", "亲子游" ], "tags_en": [ "Trip Planning", "Itinerary Map", "Map Workbench", "Family Trip" ], "artifact_type": "link", "cover_image": "cover.png", "cover_mode": "landscape", "external_url": "https://workbuddy.link/p/D2oBPDI1BSQb1HCiRWRCNl", "preview": "", "artifacts": [], "prompt": "帮我做一个叫「想去哪儿玩」的网页工作台，一整页：左侧地图 + 角色，右侧面板，底部一排按钮。要能把这几件事都用起来：\n\n1. 行前：说一句想去哪玩几天 → 行程按天铺在地图上；问一句这几天的天气 → 逐日给到；再说一句「帮我准备这趟旅行的行李」→ 给出该带什么的建议，每条带一句「为什么带」。\n2. 行程地图：点某一天就只看那天，每段距离和耗时标出来；扫码能在腾讯地图小程序里打开、导航。\n3. 找路：说「从 A 到 B 怎么走」，给出驾车 / 公交 / 步行 / 骑行的走法和距离耗时。\n4. 附近：说「我附近有什么奶茶店」，按当前位置给结果，点落在图上、列表带地址和距离。\n5. 叫车：说一句去哪，先给上车点候选（不替用户选），选完给下车点候选，再询价、给几档车型和价格。\n6. 收尾：说「把我这次行程生成一张个人旅行地图」，出一张独立网页：上面是这几天连起来的路线，下面是按天的行程记录。\n\n技术上有一件事必须做对：页面本体不直连模型，中间要有一个本地小服务——它绑定当前会话，页面点按钮 → 本地服务把这句话发给当前会话 → 干完之后把结果回传给本地服务 → 本地服务用 SSE 把状态推回页面。发消息、收结果、状态持久化、断线重连这几件事的具体做法，参考 skill workbuddy-web-workbench；行程、天气、搜周边、路线规划与个人旅行地图用腾讯地图·地图助手，叫车用腾讯出行服务打车。\n\n先给一版能跑通的最小页面（行前 + 行程地图两块），再补后面四块。", "prompt_en": "Build a web workbench called \"Where To Go\" — one full page: map + character on the left, panel on the right, a row of buttons at the bottom. Make all of the following work:\n\n1. Pre-trip: say where you want to go and for how many days → the itinerary unfolds on the map day by day; ask about the weather → day-by-day forecast; say \"help me pack for this trip\" → a packing list with one \"why bring it\" line per item.\n2. Itinerary map: tap a day to see only that day, with distance and duration labeled on each leg; scan a QR code to open the trip in the Tencent Maps mini program and navigate.\n3. Routing: say \"how do I get from A to B\" → driving / transit / walking / cycling options with distance and duration.\n4. Nearby: say \"any bubble tea shops near me\" → results based on the current location, pins on the map, list with address and distance.\n5. Ride-hailing: say where you want to go → pickup-point candidates first (never choose for the user), then dropoff candidates, then fare quotes with several car classes and prices.\n6. Wrap-up: say \"turn this trip into a personal travel map\" → a standalone page with the connected route on top and the day-by-day itinerary below.\n\nOne technical thing must be done right: the page itself never calls the model directly. There is a small local service in between — it binds to the current session; clicking a button on the page sends that sentence to the session through the service, results come back to the service, and the service pushes status to the page via SSE. For the specifics of sending messages, receiving results, state persistence and reconnection, refer to the skill workbuddy-web-workbench; use Tencent Maps Assistant for itinerary, weather, nearby search, routing and the personal travel map, and Tencent Ride Service for ride-hailing.\n\nDeliver a minimal working page first (pre-trip + itinerary map), then add the remaining four blocks.", "scenario": "", "scenario_id": 0, "scenario_use": "", "task_mode": "", "experts": [], "skills": [ { "id": "tencentmap-map-assistant", "name": "腾讯地图·地图助手", "description": "一句自然语言调用腾讯地图全套能力：AI 旅游攻略、地点搜索、路线规划、天气查询，并可生成自定义地图（网页 / 小程序指南）", "slug": "tencentmap-map-assistant" }, { "id": "tencent-ride-skill", "name": "腾讯出行服务打车", "description": "腾讯出行服务叫车助手：处理「帮我叫车」「从 A 地去 B 地」这类请求，覆盖叫车、下单与订单查询", "slug": "tencent-ride-skill" }, { "id": "workbuddy-web-workbench", "name": "WorkBuddy Web Workbench", "description": "教怎么搭出「网页 ↔ 会话」框架：页面不直连模型，经本地小服务绑定会话、发消息、收结果、SSE 推状态", "slug": "workbuddy-web-workbench" } ], "mcps": [], "creator": { "name": "腾讯地图", "name_en": "Tencent Maps", "avatar": "", "badge": "community" }, "source": "community", "quality_score": 98, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-24T08:10:00Z", "updated_at": "2026-09-24T08:20:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "wx-channel-grain-storage-script", "title": "杂粮防潮收纳爆款脚本", "title_en": "Viral Grain Storage Video Script", "subtitle": "90 秒讲透痛点、实测与自然转化", "subtitle_en": "A 90-second script connecting pain points, proof, and a natural conversion", "description": "杂粮开封后最烦的不是不好找，而是袋口夹不严、开封日期记不住，回潮后口感也跟着变。\n\n这个案例围绕“谷序可视密封杂粮罐”生成一条 90 秒视频号口播脚本：前 8 秒用“夹住了不等于密封”抓住注意力，中段依次展示旋盖、日期转盘、透明余量和叠放效果，最后用“先量柜子再购买”自然收口。产物包含连续时间轴、分段画面、逐句台词、节奏证据和完整口播稿。\n\n适合：家居收纳品牌、厨房用品商家、视频号带货达人、家庭生活类内容运营。", "description_en": "Opened grain bags are frustrating not only because they are hard to organize, but because clips do not fully seal them, opening dates are forgotten, and humidity can affect texture.\n\nThis case creates a 90-second WeChat Channels spoken-word script for the Guxu transparent grain-storage jars. The first eight seconds use the contrast that clipping is not the same as sealing, followed by visual demonstrations of the screw lid, date dial, visible fill level, and stackable design. It closes naturally by asking viewers to measure their cabinet before buying. The deliverable includes a continuous timeline, shot directions, line-by-line dialogue, pacing evidence, and a complete voice-over.\n\nIdeal for: home-organization brands, kitchenware merchants, WeChat Channels commerce creators, and household-lifestyle content teams.", "categories": [ "creation" ], "tags": [ "视频号脚本", "厨房收纳", "口播带货", "爆款结构" ], "tags_en": [ "WeChat Channels script", "kitchen organization", "spoken commerce", "viral structure" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "视频号爆款口播脚本" } ], "prompt": "请生成一条可直接拍摄的视频号爆款口播脚本。创作目的为带货与种草，行业是厨房收纳，目标受众为 30—45 岁家庭主理人，选题是“开封杂粮如何防潮并保持橱柜整齐”。产品为“谷序可视密封杂粮罐”，规格 3 个装、售价 59 元，卖点包括食品接触级 PP 材质、旋拧密封盖、日期转盘、透明可视、宽口便于倾倒和可叠放。请选用痛点解决型结构，总时长控制在 90 秒左右；Hook 要在前 3 秒指出“夹子封袋不等于密封”的反常识，中段用真实可拍动作依次展示受潮与凌乱痛点、旋盖与日期转盘、叠放前后对比，CTA 提醒用户先测量橱柜尺寸再决定购买。每个段落分别写画面与台词，时间轴必须连续；口语要像朋友分享，短句为主，用“...”标停顿，不使用绝对化词、医疗功效或虚假促销，并在价格后写清“59 元/3 个”。最终交付结构化分段表、节奏说明、标签与口播全文 Markdown。", "prompt_en": "Create a ready-to-film viral spoken-word script for WeChat Channels. The goal is commerce and product seeding in the kitchen-organization category, targeting household decision-makers aged 30–45. The topic is how to keep opened grains dry and neatly organized. The product is the Guxu transparent sealed grain-storage jar, sold as a set of three for RMB 59, with food-contact-grade PP, a screw-on sealing lid, a date dial, a transparent body, a wide mouth for easy pouring, and a stackable design. Use a pain-point-solution structure and keep the total duration around 90 seconds. In the first three seconds, use the counterintuitive hook that clipping a bag is not the same as sealing it. In the body, use filmable actions to show dampness and clutter, the screw lid and date dial, and a before-and-after stacking comparison. The CTA should remind viewers to measure their cabinet before deciding to buy. Write visuals and dialogue separately for every segment, with a continuous timeline. Keep the voice conversational, favor short sentences, use “...” for pauses, avoid absolute claims, medical claims, and fabricated promotions, and state the price as “RMB 59 / 3 jars.” Deliver a Markdown document with a structured segment table, pacing notes, tags, and the full voice-over.", "skills": [ { "id": "lingyi-wx-viral-script-generator", "name": "视频号爆款脚本生成", "description": "【零一数科·出品】视频号爆款脚本生成（免费版）。免费免登录出脚本，按爆款结构生成口播脚本。", "slug": "lingyi-wx-viral-script-generator" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-08-28T08:37:17Z", "updated_at": "2026-08-31T16:31:11+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "wx-channel-magnetic-stand-oral-script", "title": "磁吸支架口播种草脚本", "title_en": "Magnetic Stand Seeding Script", "subtitle": "会议、追剧、桌拍三场景的 85 秒克制种草", "subtitle_en": "An 85-second restrained pitch across meetings, viewing, and desk filming", "description": "桌面支架类内容很容易拍成参数堆砌：厚度、转轴和横竖切换全说了，观众却不知道这些参数分别解决什么问题。\n\n这个案例以职场人的线上会议、追剧和桌面拍摄为三条真实场景线索，在付费版视频号口播 Skill 官方结果基础上制作一条约 85 秒的优化脚本。产物包含选题策略、内容边界、Hook—中段—CTA 分段表、完整口播、逐段机位动作字幕、拍摄清单与六维质检；核心只使用约 12 毫米折叠厚度、一秒横竖切换和 360 度阻尼转轴等已知信息。\n\n适合：数码配件商家、效率好物达人、视频号带货运营、职场生活类账号。", "description_en": "Desk-stand content often becomes a list of specifications: thickness, hinge, and orientation switching are all mentioned, yet viewers still cannot tell which problem each feature solves.\n\nThis case follows three real workplace scenarios—online meetings, watching videos, and desk filming—and turns the paid WeChat Channels Skill's official result into an optimized script of about 85 seconds. The deliverable includes topic strategy, claim boundaries, a Hook–body–CTA segment table, the full voice-over, shot/action/caption directions for every segment, a filming checklist, and six-dimension QC. It uses only known inputs: the approximately 12 mm folded profile, one-second orientation switching, and a 360-degree resistance hinge.\n\nIdeal for: mobile-accessory merchants, productivity-product creators, WeChat Channels commerce teams, and workplace-lifestyle accounts.", "categories": [ "creation" ], "tags": [ "视频号口播", "磁吸支架", "种草脚本", "带货转化" ], "tags_en": [ "WeChat Channels voice-over", "magnetic phone stand", "product seeding script", "commerce conversion" ], "artifact_type": "html", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "artifacts": [ { "file": "output.html", "label": "优化版视频号口播种草脚本报告" } ], "prompt": "请为视频号生成 1 条可直接拍摄的口播种草脚本。创作模式为原创，创作方向为达人带货，行业是家居百货，营销目的为短视频带货，发布平台为视频号，使用快速版，目标时长约 85 秒。产品是“轻稳磁吸手机支架”，售价 79元/个，卖点包括一秒磁吸横竖切换、折叠后约 12 毫米厚、360 度阻尼转轴；目标用户为 25—40 岁需要线上会议、追剧和桌面拍摄的职场人。账号人设是 30 岁效率好物信息型博主，语气克制，不虚构个人经历。必须同时覆盖线上会议、追剧和桌面拍摄三个场景；禁止扩写抗颠簸、不掉落、耐久、承重、磁力强弱或其他未提供数据的性能；价格在台词和字幕中统一写“79元/个”；CTA 只提醒先确认手机或保护壳的磁吸兼容性并按需考虑。分段表的每一段都必须分别写明机位、动作和字幕。最终输出完整 Markdown 报告，包含生成概览、选题策略、内容边界、Hook—中段—CTA 分段表、口播全文、拍摄清单、质检评分和发布前核对。", "prompt_en": "Generate one ready-to-film spoken product-seeding script for WeChat Channels. Use original-creation mode, creator-commerce direction, the home-goods industry, short-video sales as the campaign goal, WeChat Channels as the platform, the fast depth setting, and a target duration of about 85 seconds. The product is the Qingwen magnetic phone stand at RMB 79 per unit. Its known inputs are one-second magnetic portrait/landscape switching, an approximately 12 mm folded profile, and a 360-degree resistance hinge. The audience is working professionals aged 25–40 who need online meetings, video viewing, and desk filming. Use a restrained, information-led creator voice and do not invent personal experience. Cover all three scenarios. Do not add claims about vibration resistance, drop prevention, durability, load, magnetic strength, or any unsupported performance. Always state the price as RMB 79 per unit. The CTA may only ask viewers to confirm magnetic compatibility and consider the product according to actual need. Every segment-table row must separately specify camera position, action, and on-screen caption. Deliver a complete Markdown report with an overview, topic strategy, claim boundaries, Hook–body–CTA segment table, full voice-over, filming checklist, quality score, and pre-publication checks.", "skills": [ { "id": "lingyi-wx-video-oral-seeding-script-plus", "name": "视频号口播脚本生成", "description": "【零一数科·出品】视频号口播脚本生成（付费版），支持原创与基于视频号分享链接的二创。", "slug": "lingyi-wx-video-oral-seeding-script-plus" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-28T10:08:20Z", "updated_at": "2026-08-31T16:31:11+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "xhs-drug-note-compliance-review", "title": "小红书药品笔记合规审查", "title_en": "Xiaohongshu Drug Note Compliance Review", "subtitle": "逐条揪出违规点、定风险等级，附可直接过审的改写稿", "subtitle_en": "Pinpoint every violation, grade the risk, and get a ready-to-publish compliant rewrite", "description": "做医药营销最头疼的就是：种草笔记写得越有冲击力，越容易踩《广告法》和平台规则的红线，轻则被小红书限流下架，重则面临监管处罚。\n\n这个案例演示了把一篇典型的糖尿病「神药」推广笔记交给医疗营销合规专家：一次性识别出 13 条内容违规 + 8 条平台专项违规（绝对化用语、治愈率承诺、患者证言、虚构临床数据、处方药大众传播等），整体定为最高档的严重违规，按小红书和抖音的专项规则逐条标注法规依据，并直接给出一版保留传播力、能在小红书过审的合规改写稿，最终产出一份排版精美的单文件 HTML 审查报告。\n\n适合：药企市场部与合规团队、医疗健康 MCN 内容运营、医药电商文案、投放代理商的审核同学。", "description_en": "The biggest headache in pharma marketing: the punchier your seeding note, the easier it crosses the red lines of China's Advertising Law and platform rules — risking takedowns on Xiaohongshu at best, and regulatory penalties at worst.\n\nThis case shows a typical diabetes \"miracle drug\" promo note reviewed by the healthcare marketing compliance expert: it flags 13 content violations plus 8 platform-specific ones in a single pass — absolute claims, cure-rate promises, patient testimonials, fabricated clinical data, prescription-drug mass-media promotion — grades the note at the most severe risk level, annotates every item with its legal basis and the specific rules of Xiaohongshu and Douyin, then delivers a compliant rewrite that keeps the appeal and can pass Xiaohongshu review — all wrapped in a beautifully typeset single-file HTML report.\n\nBest for: pharma marketing and compliance teams, healthcare MCN content operators, pharma e-commerce copywriters, and reviewers at media agencies.", "categories": [ "business" ], "tags": [ "医疗合规", "广告法", "小红书", "违规改写", "医药营销" ], "tags_en": [ "Healthcare Compliance", "Advertising Law", "Xiaohongshu", "Compliant Rewrite", "Pharma Marketing" ], "artifact_type": "html", "cover_mode": "landscape", "preview": "compliance-review-report.html", "artifacts": [ { "file": "compliance-review-report.html", "label": "合规审查报告（HTML 版，含违规清单、风险定级与改写稿）" } ], "prompt": "请按小红书和抖音的平台规则，审核以下推广笔记。请指出所有违规点（含平台专项规则），评估风险等级，并给出能在小红书过审的改写版本。\n\n【输出形式】将完整审查报告输出为排版美观的单文件 HTML 页面（所有样式内联、不引用任何外部资源、可离线直接打开），结构包含：检测概要、违规详情（按风险等级从高到低排序，每条含命中原文/违规原因/修改建议）、平台专项审核（小红书/抖音）、合规改写稿（逐段给出，可直接发布）、综合建议。排版风格：浅色专业风格，信息层级清晰；风险等级用色彩体系区分（如严重=红、高=橙、中=琥珀）；违规条目用卡片式布局，命中原文用引用样式、法规依据用标签样式；合规改写稿的原文与改写后对照呈现；适配桌面与窄屏。\n\n【投放平台】小红书种草笔记\n【笔记内容】\n“【主标题】 彻底告别XX痛！最新特效神药“糖克欣”震撼上市，3天稳糖，永不复发！\n\n【正文】\n您还在为长期的糖尿病并发症折磨而痛苦吗？还在担心终身服药对肝肾的毒副作用吗？\n由国际顶尖医学研究院最新研发的专利新药——“糖克欣胶囊”，现已正式登陆国内市场！它是目前唯一能够彻底根治糖尿病的特效药物。\n* 3天见效： 服用当天，血糖断崖式下跌；只需3天，空腹血糖永久稳定在 4.5 mmol/L 的黄金状态。\n* 安全无毒： 100%纯天然植物提取，绝无任何化学添加，没有任何副作用，比吃大米饭还要安全！\n* 停药不反弹： 彻底修复胰岛细胞。针对1000名晚期糖尿病患者的临床试验证明，治愈率高达99.8%，停药后绝不反弹！\n\n【真实案例分享】\n北京的张大爷（72岁）：“我有20年糖尿病史，打胰岛素都没用。吃了‘糖克欣’两个疗程，血糖完全正常了，现在大鱼大肉随便吃，医生说我的胰岛功能比年轻人还好！”\n\n【限时福利】\n前100名致电的患者，可享受“买3疗程送2疗程”的内部优惠，并由“国家级控糖专家”为您提供一对一免费保证治疗指导。名额有限，立刻拨打热线：400-XXX-XXXX！”", "prompt_en": "Please review the following promotional note against the platform rules of Xiaohongshu and Douyin. Identify all violations (including platform-specific rules), assess the risk level, and provide a rewritten version that can pass Xiaohongshu's review.\n\n[Output format] Deliver the full review report as a beautifully typeset single-file HTML page (all styles inlined, no external resources, opens offline). The structure should include: detection summary, violation details (sorted by risk level from high to low, each with the quoted original text, violation reason, and fix suggestion), platform-specific review (Xiaohongshu / Douyin), a compliant rewrite (section by section, ready to publish), and overall recommendations. Styling direction: a light, professional look with clear information hierarchy; a color system distinguishing risk levels (e.g., critical=red, high=orange, medium=amber); card-style layout for each violation, with quoted original text in blockquote style and legal bases as tags; the compliant rewrite presented side-by-side against the original; responsive on desktop and narrow screens.\n\n[Platform] Xiaohongshu seeding note\n[Note content]\n\"[Headline] Say goodbye to XX pain forever! The latest miracle drug 'Tangkexin' launches with a bang — stable blood sugar in 3 days, never relapses!\n\n[Body]\nAre you still suffering from long-term diabetes complications? Still worried about the liver and kidney side effects of lifelong medication?\nDeveloped by a top international medical research institute, the patented new drug 'Tangkexin Capsules' has officially entered the domestic market! It is the only specific drug that can completely cure diabetes.\n* Effective in 3 days: blood sugar plummets on the first day; in just 3 days, fasting glucose stays permanently at the golden level of 4.5 mmol/L.\n* Safe and non-toxic: 100% pure natural plant extract, zero chemical additives, no side effects — safer than eating rice!\n* No rebound after stopping: completely repairs islet cells. A clinical trial on 1,000 late-stage diabetes patients proves a cure rate of 99.8%, with no rebound after discontinuation!\n\n[Real case sharing]\nUncle Zhang from Beijing (72): 'I had diabetes for 20 years and even insulin didn't work. After two courses of Tangkexin, my blood sugar is completely normal. I can eat whatever I want now, and my doctor says my islet function is better than young people's!'\n\n[Limited-time offer]\nThe first 100 callers get an internal 'buy 3 courses, get 2 free' deal, plus one-on-one free treatment guidance from a 'national-level glucose control expert'. Limited slots — call the hotline now: 400-XXX-XXXX!\"", "skills": [], "experts": [ { "id": "nges-healthcare-compliance-lite", "name": "腾讯健康NGES合规专家", "description": "医疗营销合规专家，审核药品广告、医美种草、保健品声称，识别合规风险并支持改写为合规版本。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ytn7WGJKoya8", "promptFileSnapshot": "f4a36dc03d8d" } ], "mcps": [], "creator": { "name": "腾讯健康NGES官方", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 98, "source": "official", "usage_count": 0, "created_at": "2026-08-12T18:18:20+08:00", "updated_at": "2026-08-18T18:42:00+08:00", "cover_image": "cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "xiaoe-revenue-30-day-dashboard", "title": "最近30天店铺营收看板", "title_en": "30-Day Shop Revenue Dashboard", "subtitle": "把访问、下单与支付信号放进一页经营看板", "subtitle_en": "Bring visits, orders, and payments into one operating dashboard", "description": "每天看店铺数据，最烦的是指标分散、口径又容易混在一起。这个案例调用小鹅通营收分析能力，查询最近完整 30 天的真实店铺交易数据，整理出核心指标、每日趋势、业务解读与数据限制，并生成一份零售橙红色的离线 HTML 看板。\n\n适合：小鹅通店铺经营者、零售业务负责人、营收分析人员、需要快速复盘店铺表现的产品和运营团队。", "description_en": "The hardest part of reviewing shop performance is having scattered metrics and mixed definitions. This case uses Xiaoe revenue analysis capabilities to query real transaction data for the most recent complete 30 days, then turns the core metrics, daily trend, business interpretation, and data limitations into an offline HTML dashboard with a retail orange-red visual style.\n\nSuitable for: Xiaoe shop operators, retail business owners, revenue analysts, and product or operations teams that need a fast shop-performance review.", "categories": [ "data" ], "tags": [ "营收分析", "店铺经营", "数据看板", "零售运营" ], "tags_en": [ "Revenue analysis", "Shop operations", "Data dashboard", "Retail operations" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "分析最近完整 30 天的店铺营收表现，并生成一个 HTML 页面。请使用「营收经营分析」技能查询真实店铺数据，先明确绝对起止日期，再展示营收总览中的访问、下单与成功支付指标、每日趋势、数据事实、业务解读和可执行建议；页面注明数据来源、统计口径、数据完整性与金额单位限制，不把支付金额表述为结算或到账金额。页面采用零售橙红色的专业看板风格，要求单文件 HTML、CSS/JS 全内联、可离线打开，并支持中文和英文界面切换。", "prompt_en": "Analyze the shop's revenue performance for the most recent complete 30 days and generate an HTML page. Use the 营收经营分析 (Revenue Operations Analysis) skill to query real shop data, state the absolute start and end dates first, and present the revenue summary metrics for visits, orders, and successful payments, daily trends, data facts, business interpretation, and actionable recommendations. Label the data source, metric definitions, data completeness, and amount-unit limitations, and do not describe payment amount as settlement or funds received. Use a professional dashboard style in retail orange-red, with a single-file HTML, fully inline CSS/JS, offline support, and Chinese/English interface switching.", "skills": [ { "id": "revenue-operations", "skill_id": "skill_2092433508315860992", "name": "营收经营分析", "description": "对店铺进行营收经营情况的分析", "source": "" } ], "experts": [ { "id": "xiaoe-data-analyst", "name": "小鹅数据分析师", "description": "查询小鹅通平台直播、课程、交易等多维经营数据，组合分析并导出报表，用数据支撑商家决策", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_1M2LsZX1RbK9" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-26T17:14:29+08:00", "updated_at": "2026-08-31T09:00:00+08:00", "artifacts": [ { "file": "output.html", "label": "最近完整30天店铺营收 HTML 看板" } ], "is_recommended": false, "recommended_order": 0 }, { "title": "5 分钟，把今天的疲惫放下", "subtitle": "不用任何道具，在办公室就能跟着做的呼吸引导", "description": "面向教师、社工等一线工作者的自助放松练习脚本。输入身份与压力场景，一次生成 5 分钟分步呼吸引导：每步标注时长（准备 30 秒 / 觉察呼吸 1 分钟 / 4-6 呼吸 2 分钟 / 身体扫描 1 分钟 / 收尾 30 秒），语言安静舒缓，开头交代准备姿势、结尾给出回到工作状态的方式。", "artifact_type": "html", "prompt": "我是{身份，如：初三班主任}，最近{压力场景，如：带毕业班压力很大}。请带我做一个 5 分钟的放松练习：以呼吸练习为主，分步骤给出引导语（每步标注时长），语言安静舒缓，不需要任何道具，在办公室就能完成。开头告诉我准备姿势，结尾告诉我如何收尾回到工作状态。结尾附免责说明。", "skills": [], "experts": [], "categories": [ "productivity", "education" ], "tags": [ "心理健康", "放松练习", "教师", "自我照顾" ], "id": "xinqing-5min-relaxation", "title_en": "5 Minutes to Set Today Fatigue Down", "subtitle_en": "A no-equipment breathing guide you can follow right at your desk", "description_en": "A self-guided relaxation script for frontline workers like teachers and social workers. Enter your role and stress scenario to generate a 5-minute step-by-step breathing guide: each step labeled with duration (prep 30s / breath awareness 1min / 4-6 breathing 2min / body scan 1min / wrap-up 30s), calm and soothing language, with starting posture at the top and a way to return to a working state at the end.", "tags_en": [ "Mental Health", "Relaxation Exercise", "Teachers", "Self-Care" ], "prompt_en": "I am a {role, e.g.: homeroom teacher for 9th grade} and lately {stress scenario, e.g.: the pressure of leading a graduating class is intense}. Guide me through a 5-minute relaxation exercise: mainly breathing, with step-by-step guidance (each step labeled with duration), calm and soothing language, no props needed, doable right in the office. Tell me the starting posture at the beginning and how to wrap up and return to a working state at the end. End with a disclaimer.", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "quality_score": 0, "artifacts": [ { "file": "output.html", "label": "5分钟放松引导页" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-17T16:17:00+08:00", "updated_at": "2026-08-17T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "title": "一篇「考前紧张」科普推文，带循证依据区", "subtitle": "从选题到成稿，产出可直接排版发布的心理科普长文", "description": "面向心理科普内容生产者的推文写作。输入受众与主题，一次生成 1500 字以内的完整科普长文：真实场景开头、3 个有循证依据的知识点、2 个可立即操作的小练习、「什么时候需要找大人帮忙」提示，并在文末附标注来源类型的简明依据区。语气平等温暖，不说教不制造焦虑。", "artifact_type": "html", "prompt": "请为{受众，如：初中生}写一篇关于{主题，如：考前紧张}的心理科普推文，1500 字以内。要求：1）开头用一个真实生活场景引入；2）包含 3 个有循证依据的知识点和 2 个可立即操作的小练习；3）语气平等温暖，不说教、不制造焦虑；4）结尾附「什么时候需要找大人帮忙」提示；5）文末附「简明依据区」，标注每个知识点的来源类型；6）不做诊断、不承诺疗效。输出为可直接排版的正文。", "skills": [], "experts": [], "categories": [ "creation", "education" ], "tags": [ "心理健康", "科普内容", "公众号", "内容生产" ], "id": "xinqing-exam-anxiety-article", "title_en": "An Exam-Anxiety Science Article, With an Evidence Section", "subtitle_en": "Go from topic to a ready-to-publish long-form mental-health science article", "description_en": "Long-form article writing for mental-health science content producers. Enter your audience and topic to generate, in one pass, a complete science article under 1500 words: a real-life opening, 3 evidence-based knowledge points, 2 immediately actionable mini-exercises, a when to ask an adult for help note, and a brief evidence section at the end labeling source types for each point. Equal, warm tone, not preachy, not anxiety-inducing.", "tags_en": [ "Mental Health", "Science Content", "Official Account", "Content Production" ], "prompt_en": "Write a mental-health science article about {topic, e.g.: exam anxiety} for {audience, e.g.: middle schoolers}, under 1500 words. Requirements: 1) Open with a real-life scenario; 2) Include 3 evidence-based knowledge points and 2 immediately actionable mini-exercises; 3) Equal, warm tone, not preachy, not anxiety-inducing; 4) End with a when to ask an adult for help note; 5) Append a brief evidence section labeling the source type for each knowledge point; 6) No diagnosis, no promises of efficacy. Output as ready-to-publish body text.", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "quality_score": 0, "artifacts": [ { "file": "output.html", "label": "考前紧张科普推文" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-17T16:17:00+08:00", "updated_at": "2026-08-17T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "title": "孩子不肯上学，先找老师还是医院？", "subtitle": "把「先找谁、什么情况升级到哪」理成一张清晰路径表", "description": "面向家长的求助路径导航。输入孩子年级与具体问题，一次生成由近及远的三级求助路径表（学校心理老师 / 儿少心理门诊 / 青少年服务热线），每一级标明适用情况、开口话术和需要准备的材料，并单独标注需要跳过逐级路径立即求助的情形。", "artifact_type": "html", "prompt": "孩子{年级}，出现{问题描述，如：成绩掉得厉害、不肯上学}。请帮我梳理求助路径：第一步找谁、什么情况升级到哪一级（学校老师／心理老师／专业机构／医院／热线），每一步该怎么开口、需要准备什么信息。输出为分步骤说明加一张简明路径对照表，并单独说明哪些情况需要立即求助、不必逐级走。结尾附免责说明。", "skills": [], "experts": [], "categories": [ "productivity", "education" ], "tags": [ "心理健康", "求助路径", "家长", "拒学" ], "id": "xinqing-help-seeking-path", "title_en": "My Child Refuses School — Teacher First, or Doctor?", "subtitle_en": "Turn who to call first and when to escalate into one clear path table", "description_en": "A help-seeking path navigator for parents. Enter your child grade and the specific problem to generate, in one pass, a three-tier escalation path (school counselor / child and adolescent mental health clinic / youth service hotline), with each tier noting applicable situations, how to broach the conversation, and materials to prepare, plus a separate callout for situations that require bypassing the tiered path and seeking help immediately.", "tags_en": [ "Mental Health", "Help-Seeking Path", "Parents", "School Refusal" ], "prompt_en": "My child is in {grade} and is showing {problem description, e.g.: grades dropping sharply, refusing to go to school}. Help me map out a help-seeking path: who to approach first, when to escalate to the next level (school teacher / school counselor / professional service / hospital / hotline), how to broach it at each step, and what information to prepare. Output as step-by-step instructions plus a concise path comparison table, and separately note which situations require immediate help without going through the levels sequentially. End with a disclaimer.", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "quality_score": 0, "artifacts": [ { "file": "output.html", "label": "求助路径导航页" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-17T16:17:00+08:00", "updated_at": "2026-08-17T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "title": "开口三句就呛起来，陪我练一遍怎么聊", "subtitle": "AI 扮演孩子陪你对练，练完还给你逐句复盘", "description": "面向家长的亲子沟通实战演练。输入冲突场景后，AI 先帮你理清这段对话真正的目标，再扮演你孩子的年龄与语气对练 3-4 个回合（反应真实、不刻意配合），最后逐句复盘：哪句说得好、哪句会让孩子关上门，并给出可直接替换的说法。", "artifact_type": "html", "prompt": "我是一位{年级}孩子的家长。我们最近总是在{冲突场景，如：写作业、玩手机、成绩}这件事上说不到三句就吵起来。请你扮演我的孩子，用他那个年纪真实的语气和反应跟我对话，让我练习怎么开口。要求：1）先帮我理清这段对话我真正想达到什么目标；2）扮演孩子与我对练 3-4 个回合，反应要真实，不要过分配合；3）结束后告诉我哪几句说得好、哪几句会让孩子关上门，给出可替换的说法；4）结尾附免责说明。", "skills": [], "experts": [], "categories": [ "creation", "education" ], "tags": [ "心理健康", "亲子沟通", "家长", "对话演练" ], "id": "xinqing-parent-child-dialogue", "title_en": "Three Sentences In and Already Fighting — Practice the Talk With Me", "subtitle_en": "AI plays your kid for practice, then gives you a line-by-line debrief", "description_en": "A hands-on parent-child communication drill for parents. After you describe a conflict scenario, the AI first helps you clarify the real goal of the conversation, then role-plays your child at the right age and tone for 3-4 rounds (reactions feel real, not overly cooperative), and finally does a line-by-line debrief: which lines landed well, which would make the child shut down, and ready-to-use replacement phrasing.", "tags_en": [ "Mental Health", "Parent-Child Communication", "Parents", "Dialogue Practice" ], "prompt_en": "I am a parent of a {grade} child. We keep getting into fights within three sentences over {conflict scenario, e.g.: homework, phone use, grades}. Please role-play my child, using the real tone and reactions of someone that age, so I can practice how to open the conversation. Requirements: 1) First help me clarify what I truly want to achieve in this talk; 2) Role-play the child for 3-4 rounds, reactions must feel real, not overly cooperative; 3) Afterward, tell me which lines worked well and which would make the child shut down, provide replacement phrasing; 4) End with a disclaimer.", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "quality_score": 0, "artifacts": [ { "file": "output.html", "label": "亲子对话演练页" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-17T16:17:00+08:00", "updated_at": "2026-08-17T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "title": "孩子一回家就关门，正常吗？", "subtitle": "用发展心理学读懂青春期行为，拿到今天就能试的三个做法", "description": "面向初中生家长的心理科普问答。输入孩子的年级和具体表现，一次生成：行为背后的心理发展规律解释、是否属于青春期正常范围的判断依据、3 个当天可操作的具体做法，以及需要寻求专业帮助的信号清单。全文语气温和不说教，不含诊断内容。", "artifact_type": "html", "prompt": "我是一位初中生家长。孩子今年{年级}，最近{具体表现，如：一回家就关房门，问什么都不耐烦}。请用家长能听懂的话，解释这种行为背后的心理发展规律，帮我判断这在青春期算不算正常，并给我 3 个今天就能试的具体做法。最后请告诉我：出现哪些信号时，这就不再是「正常青春期表现」，需要寻求专业帮助。要求：语气温和不说教，不超过 800 字，不做诊断，结尾附免责说明。", "skills": [], "experts": [], "categories": [ "creation", "education" ], "tags": [ "心理健康", "家长", "青春期", "科普问答" ], "id": "xinqing-parent-psychoeducation", "title_en": "My Kid Shuts the Door the Moment They Get Home — Is That Normal?", "subtitle_en": "Decode teenage behavior with developmental psychology and walk away with 3 things to try today", "description_en": "A mental-health Q&A for parents of middle schoolers. Enter your child grade and specific behaviors to generate, in one pass: an explanation of the developmental psychology behind the behavior, criteria to judge whether it falls within normal adolescence, 3 concrete actions you can try today, and a checklist of signals that call for professional help. Warm, non-preachy tone throughout, with no diagnostic content.", "tags_en": [ "Mental Health", "Parents", "Adolescence", "Q&A" ], "prompt_en": "I am a parent of a middle schooler. My child is in {grade} and recently {specific behavior, e.g.: shuts the bedroom door the moment they get home, gets irritated at every question}. In plain language a parent can understand, explain the developmental psychology behind this behavior, help me judge whether this falls within normal adolescence, and give me 3 concrete actions I can try today. Finally, tell me: what signals would indicate this is no longer normal teenage behavior and requires professional help. Requirements: warm tone, not preachy, under 800 words, no diagnosis, end with a disclaimer.", "preview": "output.html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "quality_score": 0, "artifacts": [ { "file": "output.html", "label": "心理科普问答页" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-08-17T16:17:00+08:00", "updated_at": "2026-08-17T16:17:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "yesterday-four-metric-business-snapshot", "title": "昨日经营四指标速览", "title_en": "Yesterday's Four-Metric Business Snapshot", "subtitle": "昨日四项经营指标对比与异常波动诊断", "subtitle_en": "A practical Weimob playbook for yesterday's four-metric business snapshot", "description": "做昨日经营四指标速览时，最难的是从分散的经营数据中快速找到真正值得行动的问题。\n\n本案例围绕“昨日经营四指标速览”展开：汇总当前店铺昨日的 GMV、订单量、转化率和客单价，并与前一日及近 7 个完整自然日的日均水平比较，帮助经营者快速发现销售与成交效率的异常变化。产物以单文件 HTML 报告呈现关键指标、诊断结论和下一步建议。\n\n适合：微盟商户、经营诊断负责人、经营数据概览运营人员、经营分析团队。", "description_en": "The hardest part of yesterday's four-metric business snapshot is turning fragmented commerce data into a clear business decision.\n\nThis case demonstrates yesterday's four-metric business snapshot through a self-contained HTML report that presents key metrics, findings, causes, and actionable next steps.\n\nBest for: Weimob merchants, Business Diagnosis teams, Business Overview specialists, and business analysts.", "categories": [ "business", "data" ], "tags": [ "GMV", "订单量", "转化率", "客单价", "异常波动" ], "tags_en": [ "Yesterday's Four-Metric Business Snapshot", "Business Diagnosis", "Business Overview", "Operational Insight", "Action Plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML经营分析报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我汇总昨天的GMV、订单量、转化率和客单价，与前一日及近7个完整自然日的日均值对比，指出需要重点关注的异常波动。生成html报告", "prompt_en": "Using the relevant Weimob store data, complete a yesterday's four-metric business snapshot covering the same business scope and decision objective as the Chinese prompt. Identify key metrics and anomalies, explain the likely business causes, and provide actionable next steps. Deliver the result as a self-contained HTML report with all resources embedded and no external dependencies.", "skills": [ { "id": "data-analytics", "name": "经营数据分析", "description": "分析经营指标并生成图表或报告，覆盖日报、趋势、对比、排行、促销活动与导购任务分析。", "skill_id": "skill_2087078945183113216" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 100, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-18T14:39:16+08:00", "updated_at": "2026-08-18T14:39:16+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "youxiao-link-up-workbench", "artifact_type": "html", "title": "幼小衔接·我不怕_学习成长工作台", "title_en": "School-Ready Hero · Learning Workbench", "subtitle": "语文·数学·英语·体能四科闯关，怪兽战斗+勋章激励，家长后台全掌握", "subtitle_en": "Chinese, Math, English and Fitness quests with monster battles, medal rewards and a full parent dashboard", "description": "幼小衔接阶段最难的不是学会多少字，而是每天愿意坐下来练 20 分钟。这个工作台把语文（古诗跟读、写字描红、拼音闯关）、数学（凑十破十能量碎片、找规律、舒尔特方格）、英语（单词寻宝、动画剧场、AI 跟读）和体能（跳绳、仰卧起坐）装进一个闯关页面：每完成一个任务就获得能量碎片，达成档位就能变身假面骑士击败怪兽，同科目每 3 次练习自动授予迪迦/赛罗/泽塔勋章，能量还能合成收藏卡牌。家长辅助后台可以调节难度档位、管理屏幕时间、写鼓励便签、手动补录练习记录，所有练习数据按日期归档，孩子练了什么、达成什么档位一眼可查。演示数据为 3 天 24 条练习记录，覆盖四科练习、勋章授予与体能打卡。适合幼升小家长、低年级老师以及做幼儿童习习惯养成的自学者。", "description_en": "At school-readiness age the hardest part is not how much a child learns, but sitting down to practice 20 minutes every day. This workbench packs Chinese (poem reciting, character tracing, pinyin quests), Math (make-10 energy shards, pattern training, Schulte grid), English (word treasure, cartoon theater, AI recite) and Fitness (rope skipping, sit-ups) into one quest page: every finished task earns energy shards, reaching a tier transforms into a Rider to defeat a monster, every 3 practices in a subject auto-grants a Tiga/Sero/Zeta medal, and shards can synthesize collectible cards. The parent dashboard adjusts difficulty tiers, manages screen time, leaves encouragement notes and manually back-fills practice records — everything is archived by date, so what was practiced and which tier was reached is always one glance away. Demo data covers 24 records over 3 days, including subject practice, medal grants and fitness check-ins. Great for parents of rising first-graders, lower-grade teachers, and anyone building practice habits for young kids.", "tags": [ "幼小衔接", "游戏化学习", "习惯养成", "亲子共育", "家长后台" ], "tags_en": [ "school readiness", "gamified learning", "habit building", "parenting", "parent dashboard" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/GqJJLKryDgSz3is4N3u1Xk，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「幼小衔接·我不怕」学习成长工作台，卡通童趣风格，面向幼升小孩子与家长，用一张「练习记录」表管住全部练习数据。表字段：日期（日期）、记录类型（选择：学习记录/勋章成就/体能打卡/家长设置）、科目（选择：语文/数学/英语/体能）、板块、任务名称、完成数量、正确数、总题数、用时分钟、达成档位（选择：入门/合格/良好/优秀/双优）、勋章类型（选择：迪迦勋章/赛罗勋章/泽塔勋章/假面骑士勋章）、记录方式（选择：系统自动/手动录入/拍照打卡）、备注。页面包含：今日任务首页（四科任务建议、今日完成统计、四科入口卡对应奥特曼/骑士角色）；语文乐园（古诗跟读与原文展示、写字描红笔画引导、拼音闯关选择题）；数学能量站（凑十破十收集能量碎片、找规律训练、舒尔特方格 16 格计时）；英语星球（单词寻宝、动画剧场对话、AI 跟读句子练习）；体能闯关（跳绳与仰卧起坐按个数分档，热身提醒，达成档位变身假面骑士与怪兽战斗，击败棉花糖小怪等怪兽升级，支持高阶 BOSS 战）；成就勋章墙（同科目每 3 次练习自动授予对应勋章、能量碎片合成收藏卡牌、勋章进度展示）；家长辅助后台（难度档位调节、屏幕时间管理、鼓励便签、手动录入练习记录、按类型筛选与删除记录）。所有练习完成后自动写入数据表并触发勋章判定，内置 3 天 24 条演示数据。PC 与移动端自适应。\n\n交付：资料库在线工作台——在线 HTML 页面 + 数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Please use the Library capability to build a same-style copy of https://www.workbuddy.cn/space/d/GqJJLKryDgSz3is4N3u1Xk with data stored online and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"School-Ready Hero\" learning workbench in a playful cartoon style for rising first-graders and their parents, managing all practice data in one \"Practice Records\" table. Fields: date (date), record type (select: learning record / medal milestone / fitness check-in / parent settings), subject (select: Chinese / Math / English / Fitness), section, task name, completed count, correct count, total questions, minutes spent, tier reached (select: Rookie / Pass / Good / Great / Star), medal type (select: Tiga / Sero / Zeta / Rider medal), entry mode (select: auto / manual entry / photo check-in), note. The page includes: a Today's Tasks home (four-subject task suggestions, daily completion stats, four subject entry cards mapped to Ultraman/Rider characters); Chinese Fun Land (poem reciting with original text, character tracing with stroke guidance, pinyin multiple-choice quests); Math Energy Station (make-10 energy shard collection, pattern training, timed 4x4 Schulte grid); English Planet (word treasure hunt, cartoon theater dialogues, AI sentence reciting); Fitness Challenge (rope skipping and sit-ups tiered by reps per minute, warm-up reminder, transform into a Rider to battle monsters when a tier is reached, defeat the Marshmallow Blob and friends to level up, with an Elite BOSS mode); Trophy Wall (a medal auto-granted for every 3 practices in a subject, energy shards synthesized into collectible cards, medal progress); and a Parent Corner (difficulty tier control, screen time management, encouragement notes, manual record entry, filter and delete records by type). Every finished practice writes back to the table and triggers medal logic; ships with 24 demo records across 3 days. responsive on PC and mobile.\n\nDeliverable: an online Library workbench — an online HTML page plus the data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "幼小衔接·我不怕_学习成长工作台（自包含演示）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "工作台搭建师", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "布瑞的不累老爸", "name_en": "Burui De Bulei Laoba", "avatar": "", "badge": "community" }, "source": "community", "skill_id": null, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 90, "created_at": "2026-09-04T15:30:00+08:00", "updated_at": "2026-09-04T15:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "zhuge-liang-footprints-map", "title": "诸葛亮行迹指南地图", "title_en": "Zhuge Liang Footprints Map", "subtitle": "七个人生阶段串起十六处三国足迹的互动地图", "subtitle_en": "An interactive map linking 16 Three Kingdoms sites across 7 life stages", "description": "想系统地走一遍诸葛亮的人生足迹，却要在大堆三国攻略里逐个搜坐标、查开放信息？「躬耕南阳」到底在哪、街亭古战场怎么去、五丈原和武侯墓怎么串成一条线，查起来费时费力。\n\n这个案例把诸葛亮从琅琊出生到星落五丈原的人生轨迹搬进一张可交互地图：7 个人生阶段、16 处真实历史遗迹、13 座城市，按阶段筛选、沿迹巡游、迁徙飞线一键切换，每处足迹都带历史背景注解，扫码还能存成腾讯地图小程序随时导航。\n\n适合三国历史爱好者、人文旅行规划者、研学与课堂展示，也适合想复用同款做「人物足迹地图」「主题路线地图」的任何人。", "description_en": "Want to trace Zhuge Liang's life journey but stuck searching coordinates site by site across scattered guides? Where exactly is Longzhong, how do you reach Jieting, and how do Wuzhangyuan and the Wuhou Tomb fit into one route?\n\nThis case turns Zhuge Liang's life path — from his birth in Langya to his death at Wuzhangyuan — into one interactive map: 7 life stages, 16 real historical sites across 13 cities, with stage filtering, a guided tour mode, and migration arcs one tap away. Every footprint carries historical notes, and a QR code saves it as a Tencent Maps mini-program for on-the-go navigation.\n\nGreat for Three Kingdoms history fans, cultural trip planners, classrooms and study tours — and anyone who wants to reuse the same format for figure-themed or route-themed map guides.", "categories": [ "travel", "education" ], "tags": [ "腾讯地图", "诸葛亮", "三国历史", "人文旅行", "地图指南" ], "tags_en": [ "Tencent Maps", "Zhuge Liang", "Three Kingdoms history", "cultural travel", "map guide" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "用腾讯地图的自定义地图能力，制作一份「诸葛亮行迹指南」人物足迹地图页，输出一份可直接打开的单页 HTML。数据用真实历史地点，按人生阶段分 7 组共 16 处足迹、跨 13 座城市：琅琊少年·阳都故里（诸葛亮故里纪念馆、沂南汉墓博物馆）、躬耕隆中·十年隐居（古隆中风景区、襄阳古城）、初出茅庐·博望与新野（汉桑城公园、博望坡遗址）、联吴抗曹·柴桑与赤壁（三国赤壁古战场）、入蜀定蜀·白帝与剑门（奉节白帝城瞿塘峡、剑门关风景区）、北伐中原·汉中祁山（石门栈道、定军山、祁山堡、街亭古战场遗址）、星落五丈原（五丈原诸葛亮庙、武侯墓风景区、成都武侯祠博物馆），每处足迹含名称、坐标、历史背景注解与所属阶段。页面布局：左侧标题卡与阶段分组列表（序号徽标、阶段简介、足迹卡片，支持按阶段筛选、单阶段独占与查看全部），右侧全幅腾讯地图渲染点位与轨迹连线，点击列表或标记联动聚焦并弹出信息窗；工具条提供「全部阶段、沿迹巡游、迁徙飞线」三种模式，巡游按顺序自动飞行聚焦各足迹，飞线以弧线展示跨阶段迁徙示意，底部附图例与小程序指南二维码。交付：一份可直接双击打开的单页 HTML 地图指南。", "prompt_en": "Use Tencent Maps' custom map capability to create a \"Zhuge Liang Footprints Guide\" interactive map page, delivered as a single HTML file that opens directly in a browser. Use real historical sites, organized into 7 life-stage groups totaling 16 footprints across 13 cities: Langya Youth – Yangdu Hometown (Zhuge Liang Hometown Memorial, Yinan Han Tomb Museum); Farming in Longzhong – Ten Years in Seclusion (Gulongzhong Scenic Area, Xiangyang Ancient City); First Campaign – Bowang and Xinye (Hansang City Park, Bowang Slope Historic Site); Alliance at Chibi – Chaisang and Red Cliffs (Three Kingdoms Chibi Ancient Battlefield); Entering Shu – Baidi City and Jianmen (Fengjie Baidi City Qutang Gorge, Jianmenguan Scenic Area); Northern Expeditions – Hanzhong and Qishan (Shimen Plank Road, Dingjun Mountain, Qishan Fortress, Jieting Ancient Battlefield); Fall at Wuzhangyuan (Wuzhangyuan Zhuge Liang Temple, Wuhou Tomb Scenic Area, Chengdu Wuhou Shrine Museum). Each footprint includes a name, coordinates, historical notes, and its life stage. Layout: a left panel with the title card and stage-grouped list (numbered badges, stage intro, footprint cards, with per-stage filtering, solo-stage mode, and show-all), and a full-bleed Tencent map on the right rendering all markers and route lines, with list-marker linked focus and info windows. The toolbar offers three modes — All Stages, Guided Tour (auto-flying through footprints in order), and Migration Arcs (arcs showing cross-stage movements), plus a legend and a mini-program guide QR code. Deliverable: a single-page HTML map guide that opens by double-click.", "skills": [ { "id": "tencentmap-map-assistant", "name": "腾讯地图·地图助手", "description": "腾讯地图·地图助手 Skill，一句自然语言调用腾讯地图全套能力，开箱即用。提供 AI 旅游攻略、地点搜索（含评分/人均/营业时间）、关键词提示、路线规划（驾车/步行/公交/骑行）、地址解析与逆解析、行政区划、IP 定位、距离计算、天气查询，并可生成自定义地图（网页 / 小程序指南）。涉及找地点、规划路线、旅游行程、查天气、坐标转换等出行场景时使用。", "slug": "tencentmap-map-assistant" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "腾讯地图", "name_en": "Tencent Maps", "avatar": "", "badge": "community" }, "source": "community", "quality_score": 97, "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-09-24T08:10:00Z", "updated_at": "2026-09-24T08:30:00Z", "artifacts": [ { "file": "output.html", "label": "诸葛亮行迹指南地图页" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "zhuyao-tdee-workbench", "artifact_type": "html", "title": "打倒那个猪妖 · TDEE 减肥搭子", "title_en": "Slay the Pig Demon · TDEE Weight-Loss Companion", "subtitle": "口述三餐打卡、晨起称重，引擎拿真实体重反向校准你的专属每日消耗，每瘦一公斤猪妖血条掉一截", "subtitle_en": "Log meals by voice and weigh in each morning — the engine reverse-calibrates your true daily burn from real weigh-ins, and the pig demon loses HP with every kilo you shed", "description": "减肥有两道坎：一是不知道自己每天到底消耗多少——网上的公式套谁都是同一个数；二是坚持不下去，每年上亿国人尝试减肥，九成中途放弃。这个工作台的创建人曾是国家二级运动员，用 8 个月从 210 斤减到 140 斤，把亲测有效的打法做成了产品。它不是一张减脂记录表，而是一台会自我校准的秤：每天一句话口述三餐和运动、早上称个体重，引擎以 Mifflin-St Jeor 公式起步，每周拿真实掉秤数据对账反向校准（1kg 脂肪约等于 7700 千卡），约两周收敛出专属于你的每日消耗（TDEE），误差从 6.4% 压到 4% 以内，打卡越久越准。四个战场各司其职：打倒猪妖把目标体重变成猪妖血量，每瘦一公斤血条掉一截，11 态像素猪妖陪你作战；口述打卡一句话解析热量与营养素，近 7 日饮食健康雷达 6 维评分一眼看穿；我的数据提供身体档案、体重趋势与校准引擎收敛曲线；体质进化用数据回答你是猎豹还是树懒，坚持会升级，还能生成分享海报。10 套主题配色随机切换，数据落在资料库数据表，支持在线存储与双向同步。适合：想搞清自己真实代谢、告别「喝水都胖」困惑的减肥人，需要游戏化正反馈才能坚持的打卡党，以及喜欢用数据管理身材的健身爱好者。", "description_en": "Losing weight has two hurdles: you never know how many calories you actually burn each day — generic formulas give everyone the same number — and sticking with it is hard, with over a hundred million people in China trying to lose weight each year and ninety percent giving up. This workbench was built by a former national second-tier athlete who shed 35 kg (210 to 140 jin) in 8 months, turning a personally proven method into a product. It is not a fat-loss log sheet but a scale that calibrates itself: each day, describe your meals and exercise in one sentence and log your morning weight; the engine starts from the Mifflin-St Jeor estimate, then reconciles recorded intake against actual weight change (1 kg of fat ≈ 7,700 kcal) every week to reverse-calibrate your personal TDEE, converging in about two weeks with error squeezed from 6.4% to under 4% — the longer you check in, the more accurate it gets. Four battlefields: Slay the Pig Demon turns your target weight into the demon's HP bar, dropping with every kilo lost, joined by an 11-state pixel demon; Voice Check-in parses calories and nutrients from one sentence, with a 7-day diet health radar scoring 6 dimensions; My Data offers body profile, weight trend and the calibration convergence curve; Body Evolution answers with data whether you are a cheetah or a sloth, leveling up as you persist, with a shareable poster. Ten switchable color themes, with all data stored in Library database tables featuring online storage and two-way sync. Great for anyone who wants to understand their real metabolism, check-in users who need gamified feedback to stay on track, and fitness lovers who manage their body with data.", "tags": [ "TDEE", "减肥打卡", "口述记录", "代谢校准", "游戏化减肥" ], "tags_en": [ "TDEE", "weight-loss check-in", "voice logging", "metabolic calibration", "gamified fitness" ], "prompt": "请使用资料库的能力帮我一键做同款 https://www.workbuddy.cn/space/d/HBTff2KNrtAUpfLTLQv8XN，需要实现数据在线存储和双向同步。优先走一键做同款的链路；若没有能力，则按以下方式搭建：\n\n做一个「打倒那个猪妖」TDEE 减肥搭子工作台，数据落在资料库数据表，共三张表：① 身体档案（昵称、性别、年龄、身高(cm)、当前体重(kg)、目标体重(kg)、初始BMR(kcal)、基础日TDEE(kcal)、主要运动、每周次数、单次时长(小时)）——首次使用弹出建档向导，一分钟填完生成专属档案；② 每日记录（日期、晨重(kg)、摄入热量(kcal)、运动消耗(kcal)、是否训练日、当日盈余(kcal)、数据来源、备注）——同日再次打卡走增量更新补全，当日已有晨重或口述时自动回填表单；③ 常吃食物（食物名、分类、单位、每单位热量(kcal)、常用份量参考）——作为口述解析的热量参照库。\n\n核心机制：引擎以 Mifflin-St Jeor 公式估算每日总消耗（TDEE）起步，之后每天拿「记录的摄入热量 vs 实际体重变化（1kg 脂肪约 7700 千卡）」对账，每周反向校准专属于用户的个人 TDEE，误差持续收敛，打卡越久越准。四个模块：① 打倒猪妖——目标体重即猪妖血量（血量＝总目标进度），每瘦一公斤血条掉一截，实时展示已减重量、今日热量缺口与预计击败天数，击败后进入庆祝态；② 口述打卡——一句话描述三餐和运动，自动解析热量与营养素并写入当日记录，展示今日战况与近 7 日饮食健康雷达（6 维度评分）；③ 我的数据——身体档案、体重趋势、TDEE 校准收敛曲线（预估值逐步贴向真实值）；④ 体质进化——基于记录数据给出猎豹、候鸟、棕熊、树懒、仓鼠等代谢体质画像与升级路径，体质变化自动留痕，支持生成体质进化分享海报。视觉走像素游戏风（像素猪妖随进度切换状态、军事风像素指挥官头像），内置多套可随机切换的配色主题；未设定目标时用演示数据展示完整效果，建档后切换为真实记录。\n\n交付：资料库在线工作台——在线 HTML 页面 + 7 张数据表部署到资料库空间，页面绑定数据表读写，数据在线存储、支持双向同步。", "prompt_en": "Use the Library capability to build me a same-style copy of https://www.workbuddy.cn/space/d/HBTff2KNrtAUpfLTLQv8XN with online data storage and two-way sync. Prefer the one-click same-style flow; if unavailable, build it as follows: Build a \"Slay the Pig Demon\" TDEE weight-loss companion workbench backed by Library database tables, with three tables: (1) Body Profile (nickname, gender, age, height(cm), current weight(kg), target weight(kg), baseline BMR(kcal), baseline daily TDEE(kcal), main sport, sessions per week, hours per session) — an onboarding wizard pops up on first use and creates a personal profile in one minute; (2) Daily Record (date, morning weight(kg), calorie intake(kcal), exercise burn(kcal), training day, daily surplus(kcal), data source, notes) — checking in again on the same day merges into the existing record incrementally, and the form auto-fills when today's weight or narration already exists; (3) Frequent Foods (food name, category, unit, calories per unit(kcal), portion reference) — the calorie reference library for narrated-meal parsing. Core mechanism: the engine starts from a Mifflin-St Jeor estimate of daily total energy expenditure (TDEE), then reconciles recorded calorie intake against actual weight change (1 kg of fat ≈ 7,700 kcal) each day and reverse-calibrates the user's personal TDEE weekly, with error converging over time — the longer the check-in streak, the more accurate it gets. Four modules: (1) Slay the Pig Demon — target weight is the demon's HP (HP = total goal progress), dropping with every kilo lost, live-tracking weight lost, today's calorie gap and projected days to victory, with a celebratory state upon defeat; (2) Voice Check-in — describe meals and exercise in one sentence, auto-parse calories and nutrients into the daily record, and show today's battle report plus a 7-day diet health radar (6-dimension scoring); (3) My Data — body profile, weight trend and the TDEE calibration convergence curve (the estimate gradually hugging the true value); (4) Body Evolution — data-driven metabolic body-type portraits (cheetah, migratory bird, brown bear, sloth, hamster, etc.) with leveling paths, automatic history of body-type changes, and a shareable body-evolution poster. Visuals follow a pixel-game style (the pixel pig demon changes state with progress, military-style pixel commander avatars) with multiple randomly switchable color themes; when no goal is set, demo data showcases the full experience and switches to real records after onboarding.\n\nDeliverable: an online Library workbench — an online HTML page plus 7 data tables deployed to the Library space, with the page bound to the tables for online storage and two-way sync.", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "打倒那个猪妖 · TDEE 减肥搭子（自包含单文件，内置中英双语）" } ], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "categories": [ "workspace" ], "scenario": "", "scenario_use": "", "task_mode": "", "mcps": [], "experts": [ { "id": "WorkspaceBuilder", "name": "WorkspaceBuilder", "description": "为不同人群定制专属数字工作台，覆盖学习备考、职场效率、自媒体创作、宝妈育儿、生活管理五大场景，PC/移动端双适配，一键部署即用", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c505c9bc8929", "expert_id": "ex_kXnGMfUfWS6d" } ], "skills": [], "creator": { "name": "浅蓝", "name_en": "Qianlan", "avatar": "", "badge": "community" }, "source": "community", "sort_order": 100, "is_featured": false, "featured_order": 0, "usage_count": 0, "quality_score": 92, "created_at": "2026-09-10T11:15:05+08:00", "updated_at": "2026-09-10T19:05:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "dapagliflozin-usage", "title": "达格列净片用法用量查询", "title_en": "Dapagliflozin Dosage Lookup", "subtitle": "一句话问出口，按肾功能分层的剂量一次说清", "subtitle_en": "Ask once, get eGFR-stratified dosing in a single pass", "description": "降糖药里最不敢自己乱吃的，是剂量要跟着肾功能走的那一类。达格列净片能不能吃、吃多少，取决于肾功能检查结果落在哪一档，患者自己根本记不住，翻说明书又被一整张表格劝退。\n\n这个案例把用法用量查询交给用药专家：用户只发一句「帮我查一下达格列净片的用法用量」，专家调官方药品库拉到安达唐®（10mg*30 片、阿斯利康）的电子说明书原文，一次就产出一张网页——先交代查到的是哪一盒药，再把用法用量拆成四条人能读的话：开始治疗前要查什么、推荐剂量按什么定、eGFR 低于 45 为什么不建议用来降糖、肝功能受损要不要调量。页面注明数据来源，附扫码进小程序看完整说明书的入口。不追问、不分多轮。\n\n适合：2 型糖尿病长期服药人群、心衰或慢性肾病患者、替父母管药的家庭照护者、做患者教育的社区医生与药师。", "description_en": "The hardest glucose-lowering drugs to self-manage are the ones whose dose follows kidney function. Whether dapagliflozin is appropriate, and at what dose, depends on which band your kidney-function result falls into — impossible for a patient to remember, and the package insert presents it as a dense table most people give up on.\n\nThis case hands the lookup to the medication expert. The user sends one line — \"help me look up the dosage for dapagliflozin\" — and the expert queries the official drug database for the electronic package insert of Forxiga (10 mg x 30 tablets, AstraZeneca), then produces a single web page in one pass: which pack was found, followed by the dosage section broken into four readable points — what to check before starting, what the recommended dose depends on, why it is not recommended for glucose control below an eGFR of 45, and whether liver impairment changes the dose. The page states its data source and includes a scan-code entry to the full insert. No follow-up questions, no multi-turn back and forth.\n\nBest for: people on long-term type 2 diabetes medication, patients with heart failure or chronic kidney disease, family caregivers managing medicines for parents, and community doctors and pharmacists doing patient education.", "categories": [ "education", "search" ], "tags": [ "用法用量查询", "慢病用药", "降糖药", "达格列净", "肾功能分层" ], "tags_en": [ "Dosage Lookup", "Chronic Care", "Glucose-lowering", "Dapagliflozin", "Renal Stratification" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "达格列净片用法用量说明页（按肾功能分层给剂量 + 小程序码入口）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是腾讯健康药箱用药专家。用户用一句大白话问某个药怎么吃，你要在**这一轮内直接产出成品页面**，不追问、不拆成多轮对话。\n\n用户输入：帮我查一下达格列净片的用法用量\n\n请一次做完以下四步：\n1. 用官方药品查询能力检索该药品，取匹配度最高的那一条，读取其电子说明书的「用法用量」章节原文。\n2. 页面开头交代查到的是哪一盒药：商品名（通用名）、规格、生产企业。该药若存在多个厂家或规格版本，在页面内用一句话说明本页展示的是哪个版本即可，**不要反问用户**。\n3. 把「用法用量」原文拆成几条人能读懂的话，每条一个小标题加一句说明：开始治疗前要评估什么、推荐剂量由什么决定、哪些情况不建议用于某个治疗目标、肝肾功能受损要不要调整剂量。其中「不建议使用」这类内容单独做成警示条目，视觉上和其他条目区分开。\n4. 页面注明数据来源，结尾提示具体剂量请遵医嘱，并给出扫码进入腾讯健康药箱小程序查看完整说明书的入口。\n\n数据要求：剂量、适应症、禁忌等一律以查到的说明书原文为准，不要凭印象补数字；查不到的内容如实说明并提示以药品实际标签为准。\n\n输出要求：单文件 HTML，CSS 与图片资源全部内联、可离线打开；医疗蓝的清爽风格，适老化阅读（正文字号偏大、行距宽松）。只讲用法用量这一件事，不要把适应症、不良反应、成份等章节一股脑铺开。", "prompt_en": "You are the Tencent Health Medicine Chest medication expert. The user asks in plain everyday language how to take a drug, and you must **produce the finished page within this single turn** — no follow-up questions, no splitting into a multi-turn dialogue.\n\nUser input: help me look up the dosage for dapagliflozin\n\nComplete all four steps in one pass:\n1. Use your official drug lookup capability to search for the drug, take the best-matching result, and read the verbatim \"dosage and administration\" section of its electronic package insert.\n2. Open the page by stating which pack was found: brand name (generic name), strength, manufacturer. If the drug has multiple manufacturers or strengths, simply note on the page which version is shown — **do not ask the user**.\n3. Break the dosage text into a few readable points, each with a short heading and a one-line explanation: what to assess before starting treatment, what determines the recommended dose, situations where it is not recommended for a given treatment goal, and whether impaired kidney or liver function changes the dose. Render \"not recommended\" content as a separate warning item, visually distinct from the rest.\n4. State the data source on the page, close with a note that the exact dose should follow the prescriber, and include a scan-code entry to the Tencent Health Medicine Chest mini-program for the full package insert.\n\nData requirements: dosage, indications, and contraindications must follow the retrieved insert text verbatim; never fill in numbers from memory. If something cannot be found, say so and defer to the physical drug label.\n\nOutput requirements: a single-file HTML page with all CSS and images inlined so it opens offline; a clean medical-blue style, readable for older users (larger body text, generous line spacing). Cover only the dosage question — do not dump the indications, adverse reactions, or ingredients sections.", "skills": [], "experts": [ { "id": "eyao-medicine-chest-expert", "name": "腾讯健康药箱用药专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_zNvIxWe9Yekr", "description": "基于中国药品监管数据中心与药品上市许可持有人官方电子说明书，输入药名即可查药品注册信息与说明书章节，并支持用药提醒与打卡管理", "promptFileSnapshot": "cc05ad881502" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-28T13:00:00Z", "updated_at": "2026-08-31T11:40:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-drop-now", "title": "基金下跌了，我该怎么办？", "title_en": "My Fund Dropped — What Should I Do?", "subtitle": "小达养基笔记：下跌时的行为纠偏与应对框架", "subtitle_en": "Xiaoda's Fund Notes: behavioral correction and a response framework for downturns", "description": "【痛点】基金账户一片绿、净值持续回撤时，很多投资者会被恐慌和\"损失厌恶\"劫持，要么冲动割肉、要么越跌越补，事后往往追悔。他们缺的不是代码，而是一套能让自己冷静下来的判断框架。【演示】本案例用易方达投教基地「小达养基笔记」技能，基于财富学堂投教知识库，以三段式结构输出一份自包含图文回答：先用真实行情图还原\"近一月 A 股走出了怎样的行情\"，再拆解\"恐慌时大脑在干什么\"（追涨杀跌、锚定、损失厌恶等认知偏误），最后给出\"不同情况、不同选择\"的情景化应对（按持基质量、仓位、资金来源分情形），全程不替用户做买卖决策。【适用人群】正经历基金回撤、情绪化操作倾向明显的基金/理财用户，以及需要投教级合规引导来稳定心态的投资者。", "description_en": "Pain point: when a fund account turns red and net value keeps falling, many investors are hijacked by panic and loss aversion — either impulsively cutting losses or averaging down at every dip, and often regretting it afterward. What they lack is not a ticker code but a calm, structured decision framework. Demo: this case uses the E Fund investor-education assistant \"Xiaoda's Fund Notes\" to produce a self-contained illustrated answer grounded in the Fortune Academy knowledge base, in a three-part structure: first a real chart showing \"how A-shares have moved over the past month\", then \"what your brain does when panicking\" (cognitive biases like herding, anchoring, and loss aversion), and finally \"different situations, different choices\" — scenario-based responses by holding quality, position size, and funding source — never making the buy/sell decision for the user. Target audience: fund/wealth-management users currently suffering drawdowns or prone to emotional trading, and anyone needing a compliant, educational nudge to steady their mindset.", "categories": [ "finance", "education" ], "tags": [ "基金", "基金下跌", "投教问答", "行为金融", "投资心态" ], "tags_en": [ "Funds", "Fund Drawdown", "Investor Education Q&A", "Behavioral Finance", "Investing Mindset" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "基金下跌了，我该怎么办.html", "preview_en": "", "external_url": "", "artifacts": [ { "file": "基金下跌了，我该怎么办.html", "label": "完整图文回答（含A股走势图与行为纠偏）" } ], "prompt": "用 efunds-edu-query 回答：基金下跌了，我该怎么办？", "prompt_en": "Use the efunds-edu-query skill to answer: My fund dropped, what should I do?", "skills": [ { "id": "efunds-edu-query", "name": "易方达投教问答", "description": "易方达投教基地「小达养基笔记」智能问答助手——基于财富学堂投教知识库，以三段式结构精准回答基金/ETF/黄金等投资品种的概念、选择与时机类问题，全文零第三人称转述、开场即观点，不处理纯行情查询、个股推荐、K线分析等非投教类问题。", "skill_id": "skill_2089898953429123072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-27T11:19:00+08:00", "updated_at": "2026-08-27T11:19:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "infographic-startup-funding", "title": "AI 创业融资数据信息图", "title_en": "AI Startup Funding Infographic", "subtitle": "2025 年全球 AI 创业公司融资全景可视化", "subtitle_en": "2025 Global AI Startup Funding Landscape Visualization", "description": "快速了解 AI 行业的资金流向？本案例将复杂的融资数据转化为一张精美的交互式信息图，包含融资总额、头部公司排名、赛道分布饼图和季度趋势折线图，配合动态数字计数器和现代卡片布局，让投资数据一目了然。适合投资人、创业者和行业分析师。", "description_en": "Want to quickly understand AI industry funding flows? This case transforms complex funding data into a beautiful interactive infographic with total amounts, top company rankings, category distribution, and quarterly trends. Perfect for investors, entrepreneurs, and industry analysts.", "categories": [ "finance" ], "tags": [ "信息图", "AI融资", "数据可视化", "创业投资" ], "tags_en": [ "Infographic", "AI Funding", "Data Visualization", "Startup Investment" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI 融资信息图页面" } ], "prompt": "请创建一个「2025 年 AI 创业融资全景」信息图页面（单文件 HTML，内联无外部依赖），深蓝+电紫配色、卡片式布局、数字动画计数、响应式。\n\n图表全部用 SVG 绘制（饼图/折线图/柱状图）。包含关键统计：融资总额、Top 5 公司、按赛道分类的融资分布、季度趋势。数据可模拟但需合理可信。", "prompt_en": "Please create a \"2025 AI Startup Funding Panorama\" infographic page (single-file HTML, fully inline with no external dependencies), with a deep blue + electric purple color scheme, a card-based layout, count-up number animations, and responsive design.\n\nAll charts should be drawn with SVG (pie charts / line charts / bar charts). Include key statistics: total funding raised, the Top 5 companies, the distribution of funding by sector, and quarterly trends. The data may be simulated but must be reasonable and credible.", "skills": [ { "id": "infographic-maker", "name": "Infographic Maker", "description": "把文章、概念和数据提炼成信息图", "skill_id": "skill_2053082171331530752" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-recommend-master-jiangsu", "title": "画像推荐：硕士应届的两条赛道对比", "title_en": "Profile-Based Recommendation: Two Tracks for Master's Graduates", "subtitle": "同一份画像分别投进公务员和事业单位岗位库，看总量与结构差异", "subtitle_en": "Run one profile against both civil-service and public-institution databases", "description": "很多硕士应届生同时在准备公务员和事业单位，但没人告诉他们两边到底有多少岗位能报、推荐出来的岗位长什么样。\n\n本案例把「硕士研究生 + 2027 届 + 江苏」这组画像分别投进两个岗位库，拿到 6,988 与 8,817 两个总量，逐条列出推荐值最高的岗位，并统计推荐理由标签（薪资区间 / 加班强度 / 平台优势）的分布，最后如实说明推荐值不是录取概率、接口不返回专业与报名时间。\n\n适合：同时备考多条赛道的硕士应届生、帮学生做选岗规划的老师。", "description_en": "Many master's students prepare for both civil-service and public-institution exams, yet nobody tells them how many posts each track actually offers.\n\nThis case runs one profile — master's degree, class of 2027, Jiangsu — against both databases, returning 6,988 and 8,817 eligible posts. It lists the highest-scoring recommendations, breaks down the reason tags (salary band, overtime intensity, platform advantage), and states plainly that the recommendation score is not an admission probability.\n\nFor: master's students weighing multiple tracks, and counsellors planning applications with them.", "categories": [ "search" ], "tags": [ "职位推荐", "硕士应届", "江苏", "赛道对比" ], "tags_en": [ "position recommendation", "master's graduate", "Jiangsu", "track comparison" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "硕士应届画像的双赛道推荐对照", "label_en": "Dual-track recommendation for a master's graduate profile" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我是 2027 届硕士研究生，想在江苏考编。帮我做一次对照分析：1) 分别按「公务员」和「事业单位」两个考试类型，用同一组画像（硕士研究生 + 江苏 + 2027 年毕业）真实查询可推荐岗位，各给出总数和推荐值最高的 5 个岗位（岗位、单位、地区、招录人数、推荐值、推荐理由）；2) 统计两批结果的推荐理由标签分布，说明薪资区间、加班强度、平台优势各自代表什么；3) 明确说明推荐值不等于录取概率；4) 接口没有的信息（学历/专业/政治面貌要求、报录比、报名时间）如实说明需查公告原文，不要推测。", "prompt_en": "I'm a 2027 master's graduate planning to take civil-service exams in Jiangsu. Run a comparison: 1) query eligible posts with the same profile (master's degree, Jiangsu, graduating 2027) under both \"civil servant\" and \"public institution\" exam types, giving the total and the top 5 recommendations for each (post, unit, region, headcount, score, reasons); 2) tabulate the distribution of reason tags across both result sets and explain what salary band, overtime intensity and platform advantage each mean; 3) state clearly that the recommendation score is not an admission probability; 4) for data the interface does not return (education/major/political-status requirements, competition ratio, application dates), say so honestly and do not guess.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "patient-edu-reference-audit", "title": "患教参考文献专项审核", "title_en": "Patient Education Reference Quality Audit", "subtitle": "逐条评估文献来源级别、时效性与引用准确性", "subtitle_en": "Grade each reference by source level, timeliness, and citation accuracy", "description": "患教文章里引用了 5 篇文献，编辑最怕两种坑：一是把不含数据的官方公告当成药代数据的出处，二是文献清单和正文上标对不上、具体数值无法溯源。\n\n这个案例把一篇新型安眠药（DORA 类）患教文章交给 AI 做参考文献专项审核，逐条给出来源级别、时效性与引用准确性结论。产物顶部是专项审核结论（文献 5 篇、🔴 需修改 1、🟡 建议优化 3），下方是 5 张文献审核卡片：[1] 国家药监局批准公告被错误用于支撑半衰期（公告不含任何药代数据）；[2] 莱博雷生临床应用中国专家共识（P0，2026）；[3] 法赞雷生片说明书（P4，半衰期 2~4h 与原文一致）；[4] 达利雷生 FDA 说明书（P4，半衰期约 8h 与原文一致）；[5] 失眠症诊断和治疗指南 2025 版（P0）。每条都标出文中引用点、与原文核对结果和改写建议，文末按严重程度分组展开全部明细。\n\n适合：医学内容编辑、患教内容审核员、药企医学事务与 Market Access 人员、健康科普作者、医学编辑团队负责人。", "description_en": "A patient education article cites 5 references — the two traps editors fear most are using an official approval notice (which carries no PK data) as the source for pharmacokinetic numbers, and a reference list that does not line up with in-text citations so figures cannot be traced.\n\nThis case hands a DORA-class hypnotic patient education article to AI for a dedicated reference audit, giving a source-level, timeliness, and citation-accuracy verdict for each reference. The deliverable opens with the audit conclusion (5 references, 1 must-fix, 3 suggested), followed by 5 reference cards: [1] the NMPA approval notice wrongly used to support half-life (the notice carries no PK data); [2] Chinese expert consensus on lemborexant (P0, 2026); [3] fazeren (fazanorexant) package insert (P4, half-life 2~4h matches); [4] daridorexant FDA label (P4, half-life ~8h matches); [5] Insomnia Diagnosis and Treatment Guideline 2025 (P0). Each card shows the in-text citation points, the check against the source, and a rewrite suggestion, with all findings grouped at the end.\n\nBest for: medical content editors, patient-education reviewers, pharma medical-affairs and market-access teams, health science writers, and editorial team leads.", "categories": [ "education", "business" ], "tags": [ "参考文献审核", "来源级别", "时效性", "引用准确性", "患教审核" ], "tags_en": [ "Reference Audit", "Source Level", "Timeliness", "Citation Accuracy", "Patient Ed Review" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "参考文献专项审核报告（逐条文献卡片 + 核对结论）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是医学科普内容审核助手，请对我提供的患教文章做一次参考文献专项审核，逐条评估每篇文献的来源级别、时效性与引用准确性，并输出为一份可直接打开的审核报告网页（不要分多轮追问，一次输出到底）。\n\n待审核内容：{粘贴患教文章正文，含参考文献列表}\n参考文献原文：{上传该文章引用的全部参考文献 PDF（指南/共识/说明书/期刊论文原文），供逐条溯源核对}\n\n请按以下方式逐条审核参考文献：\n1. 来源级别：依据指南/共识(P0) > 系统评价/Meta(P1) > 大样本RCT(P2) > 权威期刊论著/队列(P3) > 政府或权威机构官方发布、可查证中文期刊(P4) > 专家意见/可查证报刊科普(P5) 的优先级，给出每篇的级别与判定理由\n2. 时效性：标注发表/更新年份，判断是否在近5年内、是否存在更新版本可用\n3. 引用准确性：列出文中每个标注该文献的句子，逐项核对文中所引数据/表述是否能在文献原文中溯源（写明原文位置与原文表述），指出引用错配、冗余标注或无法溯源之处\n4. 著录格式：检查文献列表是否缺卷期页、中英文标点是否统一、是否给出通用名对照\n\n网页需包含以下部分：\n- 头部专项审核结论：文献总数、需修改与建议优化项数\n- 逐条文献审核卡片：每篇含题录、来源级别徽章、时效性、文中引用点、与原文核对结果、修改建议\n- 文末明细：按「需修改」「建议优化」分组展开\n\n呈现要求：单文件 HTML，CSS/图片资源全部内联，可离线打开。\n审核意见撰写要求：涉及数据与文献不一致时，必须写明文献中的原始位置（如 p9 半衰期）与原始数值，并给出按原文改写后的建议表述；每条修改建议须具体可执行。\n只发现问题、给修改建议，不做「可以发布」的发布决策。", "prompt_en": "You are a medical science communication content reviewer. Run a dedicated reference audit on the patient education article I provide — grade each reference by source level, timeliness, and citation accuracy — and output a ready-to-open review report web page in one pass.\n\nContent to review: {paste the patient education article, including the reference list}\nReference sources: {upload the full-text PDFs of every reference cited (guidelines/consensus/package inserts/journal articles) so each citation can be traced}\n\nAudit each reference as follows:\n1. Source level: assign a level per the priority guideline/consensus(P0) > systematic review/Meta(P1) > large RCT(P2) > authoritative journal article/cohort(P3) > government or authoritative official release, verifiable Chinese journal(P4) > expert opinion, verifiable newspaper/popular-science(P5), with the rationale.\n2. Timeliness: state the publication/update year and whether it is within 5 years or a newer version exists.\n3. Citation accuracy: list every in-text sentence citing that reference, verify item by item whether the cited data/statement can be traced to the source (give the original location and wording), and flag mismatches, redundant citations, or untraceable claims.\n4. Bibliography format: check for missing volume/issue/page, punctuation consistency, and generic-name cross-reference.\n\nThe page must contain: a header audit verdict (total references, must-fix and suggested counts); per-reference cards (bibliographic entry, source-level badge, timeliness, in-text citation points, check against source, suggested fix); and an end detail section grouped by must-fix / suggested.\nSingle-file HTML, fully inlined, offline-openable. Find issues and give fixes only — no publish decision.", "skills": [], "experts": [ { "id": "patient-education-content-review-assistant", "name": "腾讯健康药箱-私域患教内容审核专家", "description": "对患教文章、用药指导、疾病科普做六维度质量审核，核对数据与参考文献一致性，输出结构化审核报告与可执行修改建议", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "37c580b50c18", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-24T08:00:00Z", "updated_at": "2026-08-25T19:48:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ped-private-domain-copy-review", "title": "私域社群短文话术合规快审", "title_en": "Quick Compliance Review for Private-Domain Social Copy", "subtitle": "10 条社群推送文案逐条过审，秒判通过/修改/驳回", "subtitle_en": "10 social media copy items reviewed one by one — instant Pass/Modify/Reject verdicts", "description": "药企私域社群每天要发很多条推送文案——打卡提醒、用药指导、活动推广混杂在一起。运营同学最头疼的是：这条能不能发？那句有没有违规？一条条人工看太慢，漏了一条被投诉就麻烦了。\n\n本案例把 10 条微信社群推送文案（含商品名推荐、用户见证、直播送药、团购优惠等高风险场景）交给「腾讯健康药箱-私域患教内容审核助手」，它逐条快速审核，重点抓：处方药隐性广告（违反《广告法》第15条）、可识别隐私信息、个案当普遍化、疗效断言、诱导性促销、停药换药建议等。每条文案输出判定 + 问题列表 + 修改后版本。\n\n适合: 药企私域运营、社群内容审核员、医药合规专员、互联网医院内容风控。", "description_en": "Pharma private-domain groups push dozens of copy items daily — medication reminders, usage tips, promotional offers all mixed together. Operations teams' biggest headache: Can this line be published? Is that sentence compliant? Manual review is too slow; missing one problematic line can lead to complaints.\n\nThis case hands 10 WeChat social copy items (including high-risk scenarios like branded drug recommendations, user testimonials, drug giveaways, group-buy discounts) to the Tencent Health Pill-Box Private-Domain Patient-Education Content Review Assistant. It reviews each item rapidly, focusing on: hidden prescription drug ads (violating Ad Law Art.15), identifiable privacy info, case-to-generalization, efficacy claims, induced promotions, discontinuation/switching advice. Each item gets a verdict + issue list + revised version.\n\nSuitable for: pharma private-domain operations, community content reviewers, pharma compliance officers, internet-hospital content moderators.", "categories": [ "business", "productivity" ], "tags": [ "患教审核", "社群文案", "合规快审", "广告法" ], "tags_en": [ "Patient Education Review", "Social Copy", "Quick Compliance", "Ad Law" ], "artifact_type": "html", "preview": "output.html", "prompt": "我是药企私域运营，每周要往糖尿病社群发十几条推送。上次被合规抽查退回两条，说踩了红线，我到现在也没搞清楚边界在哪。这是本周排期的 5 条文案，帮我逐条判：能发、要改、还是必须撤？要改的直接给我写好替换版本。\n\n**01 打卡提醒**\n糖友们早上好！今天是您服用格泰平® 二甲双胍缓释片的第 23 天啦～连续打卡药效更稳定！还没打卡的赶紧的！群里张阿姨已连续打卡 60 天，血糖从 9.3 降到 6.1，效果太好了！\n\n**02 直播活动**\n本周五晚 8 点，北京协和医院内分泌科陈主任直播讲《糖尿病并发症早期信号》！福利：前 100 名送免费血糖仪；抽奖送格泰平® 一个月用量（价值 298 元）。仅限本群，名额有限，先到先得！\n\n**03 患者见证**\n刘姐去年 7 月确诊 2 型糖尿病，空腹血糖 11.8。服用格泰平® 配合饮食运动，3 个月后复查空腹 6.1、糖化 6.3%，医生说控制得非常好！附刘姐体检报告对比图（已获授权）。\n\n**04 新品促销**\n格泰平® 新包装上线！现在下单首发优惠：原价 298 元/盒，限时 198 元/盒！团购 3 盒起再打 8 折！点击购买（本内容仅用于用药管理辅助，不作为诊疗依据）\n\n**05 用药知识**\n长期吃一种药会耐药吗？二甲双胍不会产生传统耐药性。若原方案\"不管用了\"，可考虑：医生指导下加用第二种药；换用新型制剂（如格泰平® 缓释片血药浓度更平稳）；评估是否启动胰岛素。请咨询主治医生。\n\n我特别想知道：提药品商品名到底能不能提、患者故事配数据算不算违规、直播送药风险多大。最后给我一张汇总表，我拿去跟合规同事对齐。", "prompt_en": "I handle private-domain operations at a pharma company, pushing a dozen-plus messages weekly to our diabetes community. Compliance spot-checked us last time and bounced two for crossing red lines — I still don't understand where the boundary sits. Here are 5 pieces scheduled this week. Judge each: ship, revise, or pull? For revisions, write me the replacement copy directly.\n\n**01 Check-in reminder**\nMorning! Today is day 23 on your Getaiping® Metformin ER Tablets — consistent check-ins keep the effect steadier! Haven't checked in? Hurry! Aunt Zhang has a 60-day streak and her glucose dropped from 9.3 to 6.1 — amazing results!\n\n**02 Livestream event**\nThis Friday 8pm, Director Chen from Peking Union Medical College Hospital's endocrinology department presents \"Early Warning Signs of Diabetic Complications\"! Perks: free glucose meter for the first 100; lucky draw for a one-month supply of Getaiping® (worth ¥298). Group members only, limited slots, first come first served!\n\n**03 Patient testimonial**\nSister Liu was diagnosed with type 2 diabetes last July, fasting glucose 11.8. On Getaiping® plus diet and exercise, three months later fasting was 6.1 and HbA1c 6.3% — her doctor said control was excellent! Attached: before/after lab report comparison (consent obtained).\n\n**04 Product promotion**\nGetaiping® new packaging is here! Launch offer: was ¥298/box, now ¥198/box for a limited time! Buy 3+ for another 20% off! Tap to buy. (This content supports medication management only and is not a basis for diagnosis or treatment.)\n\n**05 Medication knowledge**\nDoes long-term single-drug use cause resistance? Metformin doesn't cause resistance in the traditional sense. If your regimen \"stops working,\" consider: adding a second agent under physician guidance; switching to a newer formulation (e.g., Getaiping® ER offers steadier plasma levels); evaluating whether to start insulin. Consult your treating physician.\n\nI especially want to know: can we name branded drug products at all, does a patient story with numbers count as a violation, and how risky is giving away medication in a livestream? Finish with a summary table I can take to align with the compliance team.", "skills": [], "experts": [ { "id": "patient-education-content-review-word-assistant", "name": "腾讯健康药箱-私域患教内容审核助手", "description": "基于国家药监局权威数据源，专攻私域社群短文话术合规快审，识别处方药推广/隐私/个案普遍化/疗效断言/诱导性表述五大红线。能力来源于官方产品截图。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "", "promptSourceNote": "官方 prompt 文件 HTTP 404 缺失；本次基于 9 张官方产品截图提取能力基线执行", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "official", "usage_count": 0, "artifacts": [ { "file": "output.html", "label": "社群推送文案快审报告（10条：3通过/1修改/6驳回）" }, { "file": "output.md", "label": "社群推送文案快审报告（10条：3通过/1修改/6驳回）（Markdown 版）" } ], "cover_image": "cover.jpg", "cover_mode": "landscape", "cover_source": "官方参考图片 image0（六维度患教内容审核封面）", "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-31T12:04:40+08:00", "theme_label": "腾讯蓝 · 私域社群话术快审", "theme_color": "#0052D9", "is_recommended": false, "recommended_order": 0 }, { "id": "qingming-scroll-research", "title": "清明上河图深度考证", "title_en": "Deep Verification of Along the River During the Qingming Festival", "subtitle": "北宋张择端风俗画的内容、技法、流变与馆藏全维度考证", "subtitle_en": "Full-dimensional verification of content, technique, transmission and collection of Zhang Zeduan's Northern Song genre painting", "description": "一幅传世名画背后的版本流变、技法考据、馆藏流转常常众说纷纭。这个案例演示了如何借助腾讯探元文物库与知识库，并行检索历代《清明上河图》摹本与张择端相关馆藏，再由主代理亲自做 2-3 次多源 WebSearch 交叉核验，覆盖画作内容与历史背景、艺术技法与材料、版本流变与后世摹本、当前馆藏与保护四个维度，最终交付包含配图与图表的 Markdown/HTML/DOCX/PDF 四格式深度考证报告。\n\n适合：书画研究者、博物馆策展与社教人员、宋代社会史与城市史爱好者、艺术史与绘画专业学生、文化遗产数字化保护从业者。", "description_en": "The provenance, technique, version transmission and museum custody of a传世 masterpiece are often debated. This case demonstrates how to leverage the Tencent Tanyuan relic and knowledge bases for parallel retrieval of historical copies of Along the River During the Qingming Festival and Zhang Zeduan-related collections, then run 2-3 multi-source WebSearch passes for cross-verification across four dimensions—painting content and historical context, artistic technique and materials, version transmission and later copies, current collection and conservation—delivering a cross-checked four-format (MD/HTML/DOCX/PDF) deep verification report with embedded images and charts.\n\nBest for: painting and calligraphy researchers, curators and docents, Song dynasty social and urban history enthusiasts, art-history students, and cultural heritage digital conservation practitioners.", "categories": [ "business" ], "tags": [ "清明上河图", "宋代风俗画", "版本流变", "深度考证", "书画馆藏" ], "tags_en": [ "Along the River During Qingming Festival", "Song Genre Painting", "Version Transmission", "Deep Verification", "Painting Collection" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告" } ], "prompt": "请按「腾讯探元学术研究-深度考证」专家的工作流，深度研究「清明上河图」。\n\n研究范围（覆盖四维）：\n- 内容与历史背景：北宋汴京城市生活、市井百态、虹桥漕运场景\n- 艺术技法与材料：绢本设色、界画与白描结合、人物建筑比例、色彩与晕染技法\n- 版本流变与摹本：宋本/明本/清院本对比、各代摹本传承谱系、仇英/沈源等后世摹本\n- 馆藏与保护现状：故宫博物院藏北宋张择端本状态、数字化保护与展陈实践\n\n交付要求（强制）：\n1. 探元两库 MCP（search_movable_relics + search_relics）并行检索 + 主代理 2-3 次 WebSearch 多源核验\n2. 配图 ≥ 2 张（建议 3-6 张），优先探元两库 cover；图表 ≥ 1 张（建议 2-3 张，如版本流变时间线/技法对比表/汴京城市结构图）\n3. 四格式报告：Markdown 母版 → HTML → DOCX → PDF，10 章节固定结构\n4. 仅结论不含建议；参考文献只列网络公开来源；主要结论同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Deep Verification' expert workflow to deeply research 'Along the River During the Qingming Festival'.\n\nResearch scope (four dimensions):\n- Content and historical context: Northern Song Bianjing urban life, market scenes, Hongqiao canal transport\n- Artistic technique and materials: silk-based painting with color, jiehua (ruled-line) + baimiao combination, figure/architecture proportion, coloring techniques\n- Version transmission and copies: Song/Ming/Qing court-edition comparison, copy lineage, Qiu Ying/Shen Yuan later copies\n- Collection and conservation status: Palace Museum Song-dynasty Zhang Zeduan original, digital conservation and display\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 2-3 multi-source WebSearch passes by main agent\n2. ≥2 images (3-6 recommended), Tanyuan covers preferred; ≥1 chart (2-3 recommended: version timeline/technique comparison/Bianjing urban structure)\n3. Four-format reports: Markdown master → HTML → DOCX → PDF, fixed 10-section structure\n4. Conclusions only, no recommendations; references only list public web sources; main conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research", "name": "文博深度考证专家", "description": "基于探元文物库与知识库并行采集，由主代理直接联网多源交叉核验，交付四格式深度考证报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e5dcea690c95", "expert_id": "ex_RNmBcUuH9fOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-14T16:47:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "taxue-aroma-feasibility-analysis", "title": "香薰新品投放可行性评估", "title_en": "Ad Feasibility Assessment for a Scented Candle Launch", "subtitle": "5-20 万首投腾讯广告前，先花 30 秒看个底", "subtitle_en": "A 30-second read before your first Tencent Ads campaign with a 50K–200K budget", "description": "第一次投腾讯广告最悬的不是「钱够不够」，是「这钱到底能不能花」：品类适不适合投、预算够不够、可能要承担的成本、哪里会踩坑——问一圈人说法都不一样，自己又拿不准。\n\n这个案例把「评估踏雪香薰投腾讯广告可不可行」这句话直接变成一份五维打分的可行性评估报告：总分与结论（能不能投）、建议怎么投、预算怎么花、主要风险和避坑建议，合成一份离线可看、可打印的单文件 HTML 报告。\n\n适合：首次投放腾讯广告的广告主、预算有限想先看底的中小品牌主理人、需要快速判断投放可行性的市场负责人。", "description_en": "The hardest part of a first Tencent Ads campaign isn't whether the budget is enough — it's whether the money is even worth spending: is the category even viable here, is the budget realistic, what's the cost you may need to absorb, and what pitfalls will you hit. Ask five people, get five different answers, and you're still not sure.\n\nThis case turns a single sentence — \"Assess whether promoting our scented candle Taxue on Tencent Ads is viable\" — into a five-dimension scored feasibility report. It covers the verdict (can we run it), a recommended approach, how the budget should be split, the main risks, and how to avoid them, delivered as a single-file offline HTML report.\n\nGood for: advertisers running Tencent Ads for the first time, indie brand founders with limited budgets who want a baseline read, and marketing leads who need a quick go/no-go assessment.", "categories": [ "business" ], "tags": [ "广告", "营销", "腾讯广告", "可行性评估" ], "tags_en": [ "Advertising", "Marketing", "Tencent Ads", "Feasibility Assessment" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "踏雪香薰投放可行性评估报告" } ], "prompt": "帮我评估「踏雪香薰」投腾讯广告可不可行，出一份可行性分析报告。\n\n- 产品/服务：踏雪香薰——东方美学与情绪疗愈香氛品牌。以\"雪\"为意象，主打静谧、清冽、祛除烦躁的嗅觉体验，为快节奏生活中的都市人打造一片\"精神留白\"与独处空间。\n\n- 核心卖点：\n - 自然植萃与静音燃烧：天然手工大豆蜡与蜂蜡混合配方，搭配木质烛芯，燃烧时伴有轻微松木焚裂声，带来视觉、听觉与嗅觉的多重疗愈\n - 极简冷感东方美学 packaging：瓶身采用哑光磨砂冰川质感玻璃或微粗糙感素烧陶瓷，兼具空间摆件属性与收藏价值\n\n- 产品线与售价：\n - 香薰蜡烛：168-288 元\n - 香薰配件（藤条/扩香石）：58-228 元\n - 随身香氛（车载/香囊）：58-288 元\n - 礼盒装：188-588 元\n\n- 测试目标：验证情绪疗愈香氛在腾讯广告的兴趣电商链路，沉淀\"夜晚独处 + 东方美学\"心智占位\n- 关注指标：ROI 达成（兴趣电商链路回本优先）+ 品牌声量（与小红书等种草心智协同）\n- 成本预期：单条成交客单 CPL 50-150 元（兴趣电商低客单区间）\n- 总预算：5-20 万\n- 投放周期：1 个月\n- 上线时间：2 周内\n\n- 线上门面：微信小店（视频号小店同步）+ 腾讯官方落地页（用于积累企微私域）\n- 开发资源：零开发基础（小团队主要做内容与品牌，无专职研发，需用平台自动归因能力）\n- 文案风险点：「治愈」「祛除烦躁」类强效果承诺词需自检；东方美学意境表达避免过度玄化\n- 投放历史：从没投过腾讯广告（首次投放，团队精力集中在线下买手店与电商平台）\n\n请直接生成《投放可行性评估》网页报告（HTML），包含：总分与结论、能不能投、建议怎么投、预算怎么花、主要风险和避坑建议。", "prompt_en": "Assess whether promoting our scented candle brand \"Taxue\" on Tencent Ads is viable, and produce a feasibility analysis report.\n\n- Product/Service: Taxue is an Eastern-aesthetic, emotional-healing fragrance brand. Centered on the imagery of \"snow,\" it focuses on a quiet, crisp, and stress-relieving scent experience, creating a small space of mental \"negative space\" and solitude for urban dwellers living at a fast pace.\n\n- Core selling points:\n - Natural plant extracts and silent burning: a handcrafted blend of soy wax and beeswax with a wooden wick; the gentle pine-crackling sound of the wick adds a multi-sensory healing touch (visual, auditory, olfactory)\n - Minimalist cool-toned Eastern aesthetic packaging: matte frosted glacier-textured glass or slightly coarse unglazed ceramic, doubling as a decorative object and a collectible\n\n- Product line and pricing:\n - Scented candles: RMB 168–288\n - Accessories (rattan diffusers/aromatherapy stones): RMB 58–228\n - Carry-on fragrance (car clips/sachets): RMB 58–288\n - Gift sets: RMB 188–588\n\n- Test goal: validate emotional-healing fragrance performance on Tencent Ads' interest-based e-commerce path, and anchor the \"evening solitude + Eastern aesthetic\" mind-share\n- KPIs: ROI achievement (interest e-commerce path prioritized for break-even) + brand buzz (synergy with Xiaohongshu and other content-driven mental-share channels)\n- Cost expectation: per-order CPL between RMB 50–150 (interest e-commerce low-AOV range)\n- Total budget: RMB 50,000–200,000\n- Campaign duration: 1 month\n- Launch timeline: within 2 weeks\n\n- Online storefronts: WeChat Mini Shop (synced with Video Account Mini Shop) + Tencent official landing page (to accumulate enterprise-WeChat private-domain contacts)\n- Engineering resources: zero in-house development (small team focused on content and brand, no dedicated engineers; must rely on platform-native attribution)\n- Copywriting risk notes: strong-effect claims such as \"healing\" and \"stress relief\" must be self-checked; Eastern-aesthetic expressions should avoid being overly abstract\n- Tencent Ads history: never run Tencent Ads before (first-time advertiser; the team has been focused on offline concept stores and e-commerce platforms)\n\nPlease produce an \"Ad Feasibility Assessment\" web report (HTML) directly, covering: total score and verdict, whether to launch, recommended approach, how the budget should be split, main risks, and how to avoid them.", "skills": [], "experts": [ { "id": "media-placement-expert", "name": "齐亮·金手指媒介经理", "description": "把「我想投广告」变成可行性评估和可执行的投放方案。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cfc94777867b", "expert_id": "ex_c5aizcWZaV8m" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 101, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T07:26:56Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-methods", "title": "基金进阶知识", "title_en": "Advanced Fund Knowledge", "subtitle": "小达养基笔记：核心-卫星、决策框架与再平衡", "subtitle_en": "Xiaoda's Fund Notes: core-satellite, decision framework and rebalancing", "description": "【痛点】很多投资者不是缺\"买什么\"的点子，而是缺一套能长期复用、能管住情绪的方法论——结果行情一来就乱、一跌就慌。【演示】本案例用易方达投教基地「小达养基笔记」技能，基于财富学堂投教知识库，输出一份系统化的基金投资方法图文：先讲\"核心-卫星\"如何搭好仓位结构，再点出投资前要管住大脑的四个陷阱（认知偏误），给出可落地的\"五步投资决策框架\"，并拆解资产配置的再平衡怎么做、核心仓如何用股债搭配搭建，最后总结配置的典型误区，全程不替用户做买卖决策。【适用人群】有一定基金认知、想从\"零散操作\"升级到\"方法体系\"的投资者，以及需要投教级合规方法论梳理的用户。", "description_en": "Pain point: many investors don't lack ideas of \"what to buy\" — they lack a reusable, emotion-proof methodology, so they scramble when markets move and panic when prices fall. Demo: this case uses the E Fund investor-education assistant \"Xiaoda's Fund Notes\" to produce a systematic, illustrated guide to fund-investing methods grounded in the Fortune Academy knowledge base: first how to structure positions with a \"core-satellite\" approach, then the four mental traps to guard against before investing (cognitive biases), a practical five-step investment decision framework, how to rebalance asset allocation, how to build the core sleeve with stock-bond pairing, and finally the typical allocation pitfalls — never making the buy/sell decision for the user. Target audience: investors with some fund knowledge who want to upgrade from ad-hoc trades to a method system, and anyone needing a compliant, educational methodology overview.", "categories": [ "finance", "education" ], "tags": [ "基金", "投资方法", "投教问答", "资产配置", "核心-卫星" ], "tags_en": [ "Funds", "Investing Method", "Investor Education Q&A", "Asset Allocation", "Core-Satellite" ], "artifact_type": "html", "cover_image": "cover.jpg", "cover_mode": "landscape", "preview": "基金进阶知识.html", "preview_en": "", "external_url": "", "artifacts": [ { "file": "基金进阶知识.html", "label": "完整图文回答（方法论 + 决策框架）" } ], "prompt": "用 efunds-edu-query 回答：有没有我需要知道的基金投资方法？", "prompt_en": "Use the efunds-edu-query skill to answer: Are there fund investing methods I should know?", "skills": [ { "id": "efunds-edu-query", "name": "易方达投教问答", "description": "易方达投教基地「小达养基笔记」智能问答助手——基于财富学堂投教知识库，以三段式结构精准回答基金/ETF/黄金等投资品种的概念、选择与时机类问题，全文零第三人称转述、开场即观点，不处理纯行情查询、个股推荐、K线分析等非投教类问题。", "skill_id": "skill_2089898953429123072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T15:15:00+08:00", "updated_at": "2026-08-28T15:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "mazdutide-info", "title": "玛仕度肽注射液药品查询", "title_en": "Mazdutide Injection Info Lookup", "subtitle": "全球首个 GCG/GLP-1 双靶新药，一次搞懂是什么药", "subtitle_en": "The first GCG/GLP-1 dual agonist, understood in one pass", "description": "新机制、新名字的代谢病用药越来越多，用户拿到一个陌生药名时最朴素的诉求只有两个：这到底是个什么药、它能治什么。可说明书一上来就是十几个章节，剂量还要按周数一档档加，普通人根本读不完，更记不住哪一周该打多少。\n\n这个案例把「是什么药」型查询交给用药专家：用户问一句「帮我查一下玛仕度肽注射液，这是什么药」，一次就拿到完整速读页——先锁定是哪一盒（信尔美®、信达生物、三个规格），再讲清双靶机制（GLP-1 受体管降糖减重、GCG 受体管能量消耗与肝脂代谢），然后把两项已获批适应症分开列（长期体重控制、2 型糖尿病血糖控制），并把最容易记错的剂量滴定表（2mg 起始→4mg→维持 4 或 6mg）和漏打处理规则做成对照表。禁忌、警示语、特殊人群、贮藏条件在侧栏一并给全，文末注明数据来源并留扫码入口。\n\n适合：正在用或准备用 GLP-1 类药物的患者、关注减重与控糖新药的人群、替家人查药的家庭照护者，以及需要对新药做快速核对的社区医生与药师。", "description_en": "New-mechanism metabolic drugs keep arriving, and when users meet an unfamiliar name they really only ask two things: what kind of drug is this, and what does it treat. The package insert answers neither quickly — it opens with a dozen dense sections, and the dose has to be escalated week by week, which is exactly what people forget.\n\nThis case hands the \"what is this drug\" question to the medication expert. A user asks \"help me look up mazdutide injection, what is this drug\" and gets one complete primer page: it pins down the product (Xinermei, Innovent Bio, three strengths), explains the dual-target mechanism (the GLP-1 receptor arm for glucose and weight, the GCG receptor arm for energy expenditure and hepatic fat metabolism), then lists both approved indications separately (long-term weight management, and glycaemic control in adult type 2 diabetes). The easily-misremembered titration schedule (start at 2mg, step to 4mg, maintain at 4 or 6mg) and the missed-dose rules are laid out as comparison tables. Contraindications, the boxed warning, special populations, and storage conditions sit in the sidebar, and the page closes with the data source and a scan-code entry.\n\nBest for: patients on or starting GLP-1 class drugs, anyone tracking new weight-loss and glucose-lowering agents, family caregivers looking up medicines for relatives, and community doctors and pharmacists who need a fast, accurate check on a new drug.", "categories": [ "education", "search" ], "tags": [ "药品查询", "GLP-1", "玛仕度肽", "减重降糖", "剂量滴定", "新药速查" ], "tags_en": [ "Drug Lookup", "GLP-1", "Mazdutide", "Weight & Glucose", "Dose Titration", "New Drug Reference" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "玛仕度肽注射液速读页（是什么 / 能治什么 / 剂量滴定 / 漏打处理 + 小程序码入口）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是腾讯健康药箱用药专家，帮用户认识一个陌生药品。用户会拿一个没听过的药名来问「这是什么药」，你要**一次性**输出一个读完就懂的完整速读页面，不追问、不分多轮。\n\n用户输入：{用户想了解的药品，例如「帮我查一下玛仕度肽注射液，这是什么药」}\n\n请一次完成以下全部内容：\n\n1. **确认是哪一盒药**：商品名（通用名）、上市许可持有人、剂型、是否处方药，并列出全部规格；规格有多档时说明每档的用途差异。\n2. **这是什么药**：讲清它属于哪一类、作用在什么靶点或机制上。若是多靶点药物，把每个靶点各自负责什么疗效分开讲。再补一段制剂与给药特点（给药途径、用药频次、可否自行注射或服用、禁止的给药方式、给药时间是否受进餐影响）。\n3. **能治什么**：把**全部已获批适应症**逐项分开列出，不要只写其中一项。每项适应症下的细分场景（如单药治疗与联合治疗、或不同 BMI 与合并症门槛）必须分条写清，不合并成一句话。\n4. **剂量怎么用**：如该药需要剂量滴定或分档给药，做成按阶段（周数或分层指标）对照的表格，写清每档剂量、该档的目的、进入下一档的前置条件与最大维持剂量。\n5. **漏用怎么办**：如说明书有漏用处理规则，按距上次给药的时间区间做成对照表。\n6. **特殊人群**：肝肾功能不全、老年、儿童青少年、孕妇与哺乳期分别说明是否需要调整剂量或禁用，并区分「无需调整」「数据有限」「禁用」三类结论。\n7. **安全信息**：禁忌人群、说明书警示语、常见不良反应、贮藏条件（含温度与特殊保存期限）。\n8. **数据来源与入口**：注明数据来源，并在结尾给出扫码进入腾讯健康药箱小程序查看完整说明书的入口。\n\n数据要求：必须真实查询该药品的官方电子说明书或权威药品数据库，机制、适应症、规格、剂量、禁忌一律以查到的说明书原文为准。查不到的项目如实说明并提示以药品实际标签为准，**不要凭印象补内容、不要写「示意」或「模拟」数据**。涉及具体剂量处应提示遵医嘱。\n\n输出形式：单文件 HTML 页面，所有 CSS 与图片资源内联、可离线打开，不引用任何外部链接或 CDN。整体走克制的蓝色医疗风，采用主栏 + 侧栏的卡片式布局，把核心结论（用法、剂量）前置到首屏，安全信息收在侧栏；信息密度要高但层级清晰，不使用 emoji 作图标。页面顶部以用药专家形象作为标识。文末附免责声明，说明内容仅供用药参考、不构成诊疗建议。", "prompt_en": "You are the Tencent Health Medicine Box medication expert, helping users understand an unfamiliar drug. Users bring a name they have never heard and ask \"what is this drug\"; you must produce a complete primer page in **one shot** — no follow-up questions, no multi-turn dialogue.\n\nUser input: {the drug the user asks about, e.g. \"help me look up mazdutide injection, what is this drug\"}\n\nComplete all of the following in a single pass:\n\n1. **Confirm which product it is**: brand name (generic name), marketing authorisation holder, dosage form, prescription status, and every available strength; when several strengths exist, explain how their roles differ.\n2. **What this drug is**: state its class and the target or mechanism it acts on. For multi-target agents, describe what each target contributes separately. Then add the formulation and administration facts (route, dosing frequency, whether self-administration is allowed, prohibited routes, and whether timing depends on meals).\n3. **What it treats**: list **every approved indication** as a separate item — do not cover only one of them. Sub-scenarios under each indication (monotherapy vs combination, or differing BMI and comorbidity thresholds) must be itemised rather than collapsed into one sentence.\n4. **How it is dosed**: if the drug requires titration or tiered dosing, build a stage-by-stage table (by week or by a stratifying metric) stating each dose level, its purpose, the precondition for stepping up, and the maximum maintenance dose.\n5. **Missed doses**: if the label defines missed-dose handling, present it as a table keyed by the interval since the last dose.\n6. **Special populations**: address hepatic and renal impairment, elderly, children and adolescents, pregnancy and lactation, stating whether dose adjustment or contraindication applies, and distinguishing \"no adjustment needed\", \"limited data\", and \"contraindicated\".\n7. **Safety information**: contraindicated populations, the labelled warning, common adverse reactions, and storage conditions (including temperature and any special holding period).\n8. **Source and entry point**: state the data source, and close with a scan-code entry to the Tencent Health Medicine Box mini-program for the full package insert.\n\nData requirements: you must perform a real lookup of the drug's official electronic package insert or an authoritative drug database. Mechanism, indications, strengths, doses, and contraindications must follow the retrieved label text. Where something cannot be found, say so and defer to the physical drug label — **do not fill in content from memory and do not present \"illustrative\" or \"simulated\" figures**. Wherever specific doses appear, remind the reader to follow medical advice.\n\nOutput form: a single-file HTML page with all CSS and images inlined so it opens offline, referencing no external links or CDNs. Use a restrained blue medical style with a main-column plus sidebar card layout, promoting the core conclusions (administration and dose) to the first screen and keeping safety information in the sidebar. Information density should be high but clearly tiered; do not use emoji as icons. Place the medication-expert character at the top of the page as its identity mark. Close with a disclaimer stating the content is for medication reference only and does not constitute medical advice.", "skills": [], "experts": [ { "id": "eyao-medicine-chest-expert", "name": "腾讯健康药箱用药专家", "description": "管药这件事的助手：帮你查药品说明书、设置用药提醒、记录每天吃药，让你管好每顿药", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "", "name_en": "Medication Expert", "description_en": "Your medication companion: look up package inserts, set dose reminders, and log every dose you take.", "expert_id": "ex_zNvIxWe9Yekr" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 86, "usage_count": 0, "created_at": "2026-08-29T22:10:00Z", "updated_at": "2026-08-31T12:45:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "patient-edu-science-logic-audit", "title": "患教科学性与医学逻辑审核", "title_en": "Patient Education Scientific & Medical-Logic Audit", "subtitle": "逐项核查数据来源、机制描述与因果关系表述", "subtitle_en": "Verify data sourcing, mechanism description, and causal claims item by item", "description": "中医科普最容易踩的两个雷：一是把文献里的方子改写后和原方对不上，二是食疗推荐漏了关键人群禁忌。\n\n这个案例把一篇「秋包伏」中药饮片调理患教文章交给 AI 做科学性与医学逻辑专项审核，逐项核查数据来源、机制描述与因果关系。产物顶部是专项审核结论（🔴 需修改 2、🟡 建议优化 4，覆盖科学性与医学逻辑两个维度），下方是 6 张问题卡片：🔴 文中「罗汉果粥」配方与文献[3]原方不符——原方六味含罗汉果 1/3 个、小黄粟米 50g，文中的「基础方」把这两味丢了且全无剂量，方名叫「罗汉果粥」却不含罗汉果；🔴 祛湿组推荐薏苡仁却未提示文献[1]明确「孕期不要服用薏米」这一关键禁忌；🟡 文献[3]为小儿秋季咳嗽经验却直接套用成人、多篇文献合并标注扩大化、培土生金术语与原文献不一致、糖尿病患者等特殊人群未提示。每条都指到具体段落并给出可执行修改建议，文末按严重程度分组展开明细。\n\n适合：中医科普编辑、患教内容审核员、中医院公众号运营、健康科普自媒体、医学编辑团队负责人。", "description_en": "TCM popular-science articles hit two classic traps: rewriting a sourced formula so it no longer matches the original, and recommending food therapy while omitting a key population contraindication.\n\nThis case hands an article on autumn 'dryness' (qiubao fu) herbal-regimen patient education to AI for a dedicated scientific-accuracy and medical-logic audit, verifying data sourcing, mechanism description, and causal claims item by item. The deliverable opens with the verdict (2 must-fix, 4 suggested, across the scientific and medical-logic dimensions), followed by 6 issue cards: 🔴 the 'luohanguo congee' formula does not match reference [3] — the six-herb original uses 1/3 luohanguo and 50 g millet — both dropped in the text, with no dosages given, so a congee named after luohanguo contains none; 🔴 the dampness-removing group recommends coix seed without flagging reference [1]'s explicit 'do not take coix during pregnancy'; plus 4 suggested items (pediatric experience applied to adults, over-broad multi-reference citation, terminology mismatch, missing special-population notes). Each points to a specific paragraph with an actionable fix; all findings grouped at the end.\n\nBest for: TCM science-communication editors, patient-education reviewers, hospital WeChat-account operators, health content creators, and editorial team leads.", "categories": [ "education", "business" ], "tags": [ "科学性审核", "医学逻辑", "中医科普", "数据溯源", "因果关系" ], "tags_en": [ "Scientific Accuracy", "Medical Logic", "TCM Popular Science", "Data Sourcing", "Causal Claims" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "科学性与医学逻辑专项审核报告（逐条问题卡片 + 核对结论）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是医学科普内容审核助手，请对我提供的患教文章做一次科学性与医学逻辑专项审核，逐项核查数据来源、机制描述与因果关系表述，并输出为一份可直接打开的审核报告网页（不要分多轮追问，一次输出到底）。\n\n待审核内容：{粘贴患教文章正文，含参考文献列表}\n参考文献原文：{上传该文章引用的全部参考文献 PDF（指南/共识/说明书/期刊论文原文），供逐条溯源核对}\n\n请重点审核以下两类问题：\n一、科学性\n1. 数据 / 配方溯源：文中引用的方药、剂量、食疗方是否与所引文献原文一致，指出与原方不符、剂量缺失或张冠李戴之处\n2. 机制描述：中医病机 / 治则表述是否准确，术语与所引文献是否一致\n3. 引用标注：每个标注的文献是否真正支撑该表述，是否存在标注扩大化\n二、医学逻辑\n1. 因果关系：相关性能否被表述为因果，机制推论是否成立\n2. 证据外推：文献人群（如小儿）与文章受众（如成人）是否匹配，外推是否说明\n3. 关键安全信息：食疗 / 用药推荐是否遗漏特殊人群禁忌（如孕妇、糖尿病、脾胃虚寒），缺失是否致害\n\n网页需包含以下部分：\n- 头部专项审核结论：需修改与建议优化项数，以及科学性与医学逻辑两个维度的各自结论\n- 逐条问题卡片：每条含严重程度（🔴/🟡）、所属维度、位置（段落）、问题描述、修改建议\n- 文末明细：按「需修改」「建议优化」分组展开\n\n呈现要求：单文件 HTML，CSS/图片资源全部内联，可离线打开。\n审核意见撰写要求：涉及与文献不符时，必须回到文献原文，逐味/逐句写明原始配方与剂量、或原句表述，并与文中写法做「文中：X ／ 原文：Y」的对照，再给出按原文改写后的建议表述；每条修改建议须具体可执行。\n只发现问题、给修改建议，不做「可以发布」的发布决策。", "prompt_en": "You are a medical science communication content reviewer. Run a dedicated scientific-accuracy and medical-logic audit on the patient education article I provide — verify data sourcing, mechanism description, and causal claims item by item — and output a ready-to-open review report web page in one pass.\n\nContent to review: {paste the patient education article, including the reference list}\nReference sources: {upload the full-text PDFs of every reference cited (guidelines/consensus/package inserts/journal articles) so each citation can be traced}\n\nFocus on two classes of issues:\nA. Scientific accuracy: (1) formula/data sourcing — do cited formulas, dosages, food therapies match the referenced source, flag mismatches, missing dosages, or misattribution; (2) mechanism description — are TCM pathogenesis/principle statements accurate and terminology consistent with sources; (3) citation — does each cited reference actually support the claim, or is citation over-broad.\nB. Medical logic: (1) causality — is correlation presented as causation, do mechanism inferences hold; (2) extrapolation — does the source population (e.g., children) match the article audience (e.g., adults), is extrapolation noted; (3) critical safety — do therapy/drug recommendations omit special-population contraindications (pregnancy, diabetes, cold-spleen), and would omission cause harm.\n\nThe page must contain: a header verdict (must-fix and suggested counts, plus a conclusion for each of the scientific and medical-logic dimensions); per-issue cards (severity 🔴/🟡, dimension, location/paragraph, problem, suggested fix); and an end detail section grouped by must-fix / suggested.\nSingle-file HTML, fully inlined, offline-openable. When flagging a mismatch, go back to the source text, state the original formula with dosages (or the original sentence) herb by herb, present it as \"text: X / source: Y\", then give a rewrite; fixes must be actionable. Find issues and give fixes only — no publish decision.", "skills": [], "experts": [ { "id": "patient-education-content-review-assistant", "name": "腾讯健康药箱-私域患教内容审核专家", "description": "对患教文章、用药指导、疾病科普做六维度质量审核，核对数据与参考文献一致性，输出结构化审核报告与可执行修改建议", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "37c580b50c18", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-24T08:05:00Z", "updated_at": "2026-08-25T19:48:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ped-reference-consistency-audit", "title": "患教参考文献 P0-P5 分级专项核验", "title_en": "P0-P5 Graded Reference Audit for Patient-Education Articles", "subtitle": "引用的文献逐条验真伪、查数据、评级别，输出核验清单", "subtitle_en": "Verify every cited reference for authenticity, data accuracy, and authority — output a graded audit checklist", "description": "患教文章里引用的文献是科学性的基石，但运营同学不是医学研究员，很难判断一篇引用到底靠不靠谱：这篇 SCI 论文是真的吗？正文里的数据和原文一致吗？这篇指南是最新的吗？引用一个过时的结论会不会误导患者？\n\n本案例把一篇引用了 7 篇文献的高血压患教文章交给「腾讯健康药箱-私域患教内容审核助手」，它对每篇参考文献做 P0-P5 分级逐条核验：文献真实性（DOI/期刊/标题可查）、权威性评估（SCI 分区/影响因子/是否国家级指南）、时效性（近 5 年）、正文数据一致性（引用的数据与原文是否吻合）、引用适当性。输出《参考文献核验清单》表格 + 每篇详细核验报告。\n\n适合: 药企医学部、患教内容质控员、医学编辑、循证医学专员。这是唯一以「审计形态」呈现的患教审核案例，与六维审核和社群快审形成差异化互补。", "description_en": "Cited references are the scientific foundation of patient-education articles — but operations staff aren't medical researchers. How can they tell if a citation is reliable: Is this SCI paper real? Does the data in the article match the original? Is this guideline up-to-date? Could citing outdated conclusions mislead patients?\n\nThis case hands a hypertension patient-education article citing 7 references to the Tencent Health Pill-Box Private-Domain Patient-Education Content Review Assistant. It performs a P0-P5 graded audit on each reference: authenticity (DOI/journal/title verifiable), authority assessment (SCI IF / national guideline or not), timeliness (within 5 years?), data consistency (does quoted data match original?), citation appropriateness. Outputs a Reference Audit Checklist table + detailed per-reference report.\n\nSuitable for: pharma medical affairs, patient-education QC, medical editors, EBM specialists. This is the only \"audit-style\" patient-education review case, complementing the six-dimension review and social-copy quick-review cases.", "categories": [ "business", "productivity" ], "tags": [ "患教审核", "文献核验", "P0-P5分级", "循证医学" ], "tags_en": [ "Patient Education Review", "Reference Audit", "P0-P5 Grading", "EBM" ], "artifact_type": "html", "preview": "output.html", "prompt": "我是医学编辑，这篇高血压患教文章挂了 5 篇参考文献，看着挺唬人。但我怀疑有几篇是作者随手编的——期刊名对不上、年份可疑，正文数字也跟我印象中的原文对不上。帮我一篇一篇查：真的存在吗？够权威吗？太老了吗？正文引用跟原文对得上吗？哪几篇必须换掉？\n\n**正文引用（节选）**\n- 我国成人高血压患病率 27.5%，患者约 2.45 亿，控制率仅 16.8%【1】\n- 覆盖 31 省、45 万人的调查显示 18 岁以上居民患病率 23.2%【2】\n- ALLHAT 研究纳入超 4 万人，噻嗪类利尿剂预防心血管事件不劣于钙拮抗剂与 ACEI【3】\n- SPRINT 研究收缩压目标 120 mmHg，主要终点风险降低 25%，全因死亡降 27%【4】\n- DASH 饮食使收缩压平均下降 11.4 mmHg，效果接近单药治疗【5】\n\n**参考文献**\n1. 《中国高血压防治指南（2024 年版）》，中国高血压联盟\n2. Wang Z, et al. China Hypertension Survey 2012–2015. *Circulation*. 2018\n3. ALLHAT Officers. Major outcomes in high-risk hypertensive patients. *JAMA*. 2002\n4. SPRINT Research Group. Intensive versus Standard Blood-Pressure Control. *NEJM*. 2015\n5. Appel LJ, et al. Effects of Dietary Patterns on Blood Pressure. *NEJM*. 2001\n\n麻烦给每篇打个可用性等级（能直接用／可用但要留意／必须换掉），单列一节说清\"正文数字 vs 原文数字\"哪里不一致。最后给我替换检索建议——不能用的那些，我该去找什么文献补上。", "prompt_en": "I'm a medical editor. This hypertension patient-education article cites 5 references, and they look impressive. But I suspect some were fabricated — journal names don't match, years look off, and the body figures don't line up with what I recall. Verify each: does it exist? Authoritative enough? Too dated? Do the cited figures match the source? Which must be replaced?\n\n**Body citations (excerpt)**\n- Adult hypertension prevalence in China is 27.5%, about 245 million patients, with only 16.8% controlled [1]\n- A survey across 31 provinces and 450,000 people found prevalence among adults 18+ at 23.2% [2]\n- ALLHAT enrolled over 40,000 patients; thiazide diuretics were non-inferior to calcium channel blockers and ACE inhibitors [3]\n- SPRINT targeted 120 mmHg systolic, reporting 25% primary-endpoint risk reduction and 27% lower all-cause mortality [4]\n- The DASH diet lowers systolic pressure by 11.4 mmHg on average, approaching monotherapy [5]\n\n**References**\n1. Chinese Guidelines for the Prevention and Treatment of Hypertension (2024 edition), Chinese Hypertension League\n2. Wang Z, et al. China Hypertension Survey 2012–2015. *Circulation*. 2018\n3. ALLHAT Officers. Major outcomes in high-risk hypertensive patients. *JAMA*. 2002\n4. SPRINT Research Group. Intensive versus Standard Blood-Pressure Control. *NEJM*. 2015\n5. Appel LJ, et al. Effects of Dietary Patterns on Blood Pressure. *NEJM*. 2001\n\nAssign each reference a usability grade (use as-is / usable with caveats / must replace), and add a dedicated section on where \"body figure vs. source figure\" diverges. Finish with replacement search guidance — for the ones I can't use, what literature should I find instead?", "skills": [], "experts": [ { "id": "patient-education-content-review-word-assistant", "name": "腾讯健康药箱-私域患教内容审核助手", "description": "基于国家药监局权威数据源，对患教文章参考文献做 P0-P5 四级分类专项核验（真实性/权威性/时效性/数据一致性/引用适当性五维）。能力来源于官方产品截图（image4: 参考文献 P0-P5 分级审核封面）。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "", "promptSourceNote": "官方 prompt 文件 HTTP 404 缺失；本次基于 9 张官方产品截图提取能力基线执行", "expert_id": "ex_SkxanTjk7Wvv" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 84, "source": "official", "usage_count": 0, "artifacts": [ { "file": "output.html", "label": "高血压患教参考文献核验清单（7篇：P0×3/P1×2/P3×2/P5×0）" }, { "file": "output.md", "label": "高血压患教参考文献核验清单（7篇：P0×3/P1×2/P3×2/P5×0）（Markdown 版）" } ], "cover_image": "cover.jpg", "cover_mode": "landscape", "cover_source": "官方参考图片 image4（患教参考文献P0-P5分级专项审核封面，完美匹配案例主题）", "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-31T12:05:46+08:00", "theme_label": "学术靛 · 参考文献 P0-P5 核验", "theme_color": "#5b4bd6", "is_recommended": false, "recommended_order": 0 }, { "id": "pharma-full-insight-report", "title": "药品全景数据洞察报告", "title_en": "Pharmaceutical Full-Dimension Data Insight Report", "subtitle": "一键生成覆盖九大维度的完整 HTML 数据洞察报告，含总览页、企业侧与药箱侧策略建议", "subtitle_en": "One-click generate a complete HTML data insight report covering nine dimensions, with overview page and strategy recommendations for both enterprise and medicine-chest sides", "description": "药企做药品数据分析时，数据通常散落在 BI 平台、公开市场报告与内部运营系统中。一次完整的分析需要跨系统取数、对齐统计口径，再将结果整理为汇报材料，往往需要数个工作日才能交付。\n\n这个案例将全流程交由 AI 完成：用户只需指定药品名与时间范围（如「帮我生成 xx 药品 2026 上半年的数据洞察报告」），AI 即从药箱 BI 平台获取产品数据、从公开渠道检索市场数据，按两大层面九个维度（市场规模与药箱用户规模、月度销量与用户走势、SKU 销量与扫码分布、地域分布、竞品关联分析；药箱人群画像、功能使用情况、科普访问、用药打卡）自动生成完整的 HTML 报告。每个维度包含可视化图表、中性客观的数据洞察，以及按企业侧/药箱侧分组的策略建议，最终汇总为总览页的全景洞察。\n\n适合：药企市场部与医学部数据分析师、腾讯健康药箱运营团队、需要定期产出药品数据报告的产品经理、从事药企数字化解决方案的销售与售前顾问。", "description_en": "For pharma companies, drug data is typically scattered across BI platforms, public market reports, and internal operations systems. A complete analysis requires pulling data across systems, aligning metric definitions, and compiling results into presentable materials—often taking several business days to deliver.\n\nThis case hands the entire workflow to AI: the user only needs to specify the drug name and time range (e.g., \"generate a data insight report for xx drug for H1 2026\"), and AI retrieves product data from the medicine chest BI platform, collects market data from public sources, and automatically generates a complete HTML report covering two layers and nine dimensions (market scale & user scale, monthly sales & user trends, SKU sales & scan distribution, geographic distribution, competitor correlation analysis; user profiling, feature usage, science-popularization visits, medication check-ins). Each dimension includes visualization charts, neutral objective data insights, and strategy recommendations grouped by enterprise / medicine-chest side, culminating in an overview page with panoramic insights.\n\nBest for: pharma market and medical affairs data analysts, Tencent Health Medicine Chest operations teams, product managers who need to regularly produce drug data reports, and sales/pre-sales consultants working on pharma digital solutions.", "categories": [ "data", "business" ], "tags": [ "数据洞察报告", "药企分析", "九大维度", "HTML报告", "策略建议" ], "tags_en": [ "Data Insight Report", "Pharma Analysis", "Nine Dimensions", "HTML Report", "Strategy Recommendations" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "药品全景数据洞察报告（九大维度完整 HTML 报告，含图表+洞察+策略建议+总览页）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "请基于药箱产品数据 + 公开市场数据，为「xx 药品」生成一份数据洞察报告，数据范围为 2026.01 — 2026.06。要求：\n1. 报告完整覆盖：两大层面（整体市场规模和药箱用户分析、药箱功能使用分析）、九个维度（1.1 市场规模和药箱用户规模、1.2 月度销量和药箱用户月度走势、1.3 Sku 销量和药箱用户扫码分布、1.4 地域销量和药箱用户地域分布、1.5 竞品规模和药箱扫码关联用药分析；2.1 药箱人群画像、2.2 药箱功能使用整体情况、2.3 药箱科普访问情况、2.4 药箱用药打卡情况）\n2. 数据来源：产品数据全部来源于药箱BI 平台数据，不可编造；网络数据每个数据标注真实可追溯信源（若某一小节没有查询到相关数据，禁止编造数据，直接备注「暂未查询到xxx数据，可直接补充真实数据报告进行分析）\n3. 洞察要求：基于当前查询到的数据进行洞察总结。统一采用中性、客观的陈述式描述，只陈述数据事实、量级、结构、趋势与对比关系，不做褒贬评价。\n4. 策略建议：针对各维度的数据洞察，分别针对药企侧和药箱侧给出 2–3 条可落地的市场/产品运营动作。\n5. 报告产出形式：将所有数据和分析内容汇总成一个完整的 html 格式的报告，各维度按统一结构产出：图表数据 + 数据洞察文字 + 策略建议（按执行主体企业 / 药箱分组）；并从所有结论中总结关键信息汇总到总览页，生成数据的全景洞察。", "prompt_en": "Based on medicine chest product data + public market data, generate a data insight report for \"xx Drug\" with data range 2026.01 — 2026.06. Requirements:\n1. Full coverage of report: two layers (overall market scale and medicine-chest user analysis, medicine-chest feature usage analysis), nine dimensions (1.1 market scale and user scale, 1.2 monthly sales and user trends, 1.3 SKU sales and scan distribution, 1.4 geographic sales and user distribution, 1.5 competitor scale and correlated medication analysis; 2.1 user profiling, 2.2 overall feature usage, 2.3 science-popularization visits, 2.4 medication check-in status)\n2. Data sources: All product data comes from the Medicine Chest BI platform—do not fabricate. For web data, each data point must cite a traceable source (if a section has no relevant data found, do not fabricate; note \"Data for xxx not yet queried from public channels; supplement with real data report for analysis\")\n3. Insight requirements: Summarize insights based on retrieved data. Use neutral, objective descriptive language—state only data facts, magnitudes, structures, trends, and comparative relationships; no evaluative judgments.\n4. Strategy recommendations: For each dimension's insight, provide 2–3 actionable market/product operation actions for both the pharma enterprise side and the medicine-chest side.\n5. Output format: Compile all data and analysis into a single complete HTML report. Each dimension follows a unified structure: chart data + insight text + strategy recommendations (grouped by executor: enterprise / medicine-chest); summarize key findings from all conclusions into an overview page generating a panoramic data insight.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T06:44:00Z", "updated_at": "2026-09-04T08:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-campus-resume-optimization", "title": "腾讯校招简历优化", "title_en": "Tencent Campus Resume Optimization", "subtitle": "围绕真实 JD，把项目经历改成更有匹配度的表达", "subtitle_en": "Rewrite project experience around real JDs for stronger role fit.", "description": "痛点：很多同学简历写成技术或经历罗列，和目标岗位 JD 的匹配关系不清，面试官很难快速看到亮点。\n演示内容：本案例以腾讯官网后台开发 JD 为依据，结合 STAR 方法和简历正直原则，输出诊断、改写示例和投递前检查清单。\n适用人群：已有简历初稿、准备投递腾讯技术方向或想针对具体 JD 做定制优化的同学。", "description_en": "Pain point: Many resumes list technologies and experiences without showing clear JD fit, making it hard for interviewers to see the candidate's strengths.\nDemo: This case uses an official Tencent backend development JD, applies STAR writing and integrity rules, and outputs diagnosis, rewrite examples, and a pre-submission checklist.\nAudience: Students with a draft resume who are applying for Tencent engineering roles and need JD-specific optimization.", "tags": [ "腾讯校招", "简历优化", "STAR", "后台开发" ], "tags_en": [ "Tencent Campus Recruitment", "Resume Optimization", "STAR", "Backend Development" ], "artifact_type": "md", "categories": [ "education", "creation" ], "prompt": "我准备投递腾讯校园招聘的“软件开发-后台开发方向”，请你作为腾讯校招求职教练，基于官网 JD 和简历优化指南帮我优化简历表达。\n\n目标岗位：软件开发-后台开发方向，应届毕业生。\n候选人经历：做过分布式任务调度项目，负责调度器、任务状态流转、失败重试和监控告警；使用 Java/Go、MySQL、Redis、消息队列；另有高并发接口压测和性能优化经历。\n\n请输出：\n1. 先判断是否缺少必要信息，必要时用 2-4 个选项追问。\n2. 基于官网 JD 拆出岗位关键词。\n3. 诊断当前简历表达可能存在的问题。\n4. 用 STAR 方法改写 2 段项目经历。\n5. 给出技术栈、项目成果、面试追问准备和投递前检查清单。\n\n注意：不要虚构项目、数据、奖项或实习；只能在真实经历基础上优化表达。不要提供薪酬、HC、通过率等敏感信息。", "prompt_en": "I'm preparing to apply for Tencent's campus recruitment \"Software Development - Backend Development\" track. Please act as a Tencent campus recruitment job coach and help me optimize my resume's wording based on the official JD and resume optimization guidelines.\n\nTarget position: Software Development - Backend Development, new graduate.\nCandidate experience: built a distributed task scheduling project, responsible for the scheduler, task state transitions, failure retries, and monitoring & alerting; used Java/Go, MySQL, Redis, message queues; also has experience with high-concurrency interface stress testing and performance optimization.\n\nPlease output:\n1. First determine whether necessary information is missing, and if needed, ask follow-up questions with 2-4 options.\n2. Extract the position keywords based on the official JD.\n3. Diagnose potential problems with the current resume's wording.\n4. Rewrite 2 project experiences using the STAR method.\n5. Provide a tech stack, project outcomes, interview follow-up question prep, and a pre-submission checklist.\n\nNote: Do not fabricate projects, data, awards, or internships; only optimize the wording based on real experience. Do not provide sensitive information such as salary, headcount (HC), or pass rates.", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "简历优化报告", "label_en": "Resume Optimization Report" } ], "cover_image": "cover.png", "quality_score": 91, "skills": [ { "id": "tencent-campus-recruit", "name": "腾讯校园招聘", "description": "面向 WorkBuddy 上架使用的腾讯校园招聘助手，覆盖流程制度、招聘动态、岗位方向、简历优化、面试陪练和防诈骗场景。", "skill_id": "skill_2054903442024890368" } ], "experts": [], "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 102, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "terracotta-army-research", "title": "秦始皇兵马俑深度考证", "title_en": "Deep Verification of the Terracotta Army of Qin Shi Huang", "subtitle": "秦代陶俑军阵的发掘、工艺、保护与文化价值全维度考证", "subtitle_en": "Full-dimensional verification of excavation, craftsmanship, conservation and cultural value of the Qin dynasty terracotta warrior pits", "description": "一座世界文化遗产的发掘史、工艺密码、保护难题与文化解读往往是庞杂的研究体系。这个案例演示了如何借助腾讯探元文物库与知识库，并行检索秦始皇兵马俑相关馆藏与学术资源，再由主代理亲自做 2-3 次多源 WebSearch 交叉核验，覆盖发现与发掘历程、军阵编制与陶俑工艺、青铜兵器与彩绘保护、文化价值与申遗四个维度，最终交付包含配图与图表的 Markdown/HTML/DOCX/PDF 四格式深度考证报告。\n\n适合：秦汉考古研究员、博物馆策展与社教人员、世界文化遗产保护从业者、古代军事史爱好者、艺术史与雕塑专业学生、博物馆数字化保护工程师。", "description_en": "The excavation history, craftsmanship codes, conservation challenges and cultural interpretation of a World Cultural Heritage site constitute a vast research system. This case demonstrates how to leverage the Tencent Tanyuan relic and knowledge bases for parallel retrieval of Terracotta Army collections and academic resources, then run 2-3 multi-source WebSearch passes for cross-verification across four dimensions—discovery and excavation process, military formation and pottery craftsmanship, bronze weapons and painted-figure conservation, cultural value and UNESCO application—delivering a cross-checked four-format (MD/HTML/DOCX/PDF) deep verification report with embedded images and charts.\n\nBest for: Qin-Han archaeology researchers, curators and docents, World Heritage conservation practitioners, ancient military history enthusiasts, art-history and sculpture students, and museum digital conservation engineers.", "categories": [ "business" ], "tags": [ "秦始皇兵马俑", "秦代陶俑", "世界遗产", "深度考证", "文物保护" ], "tags_en": [ "Terracotta Army", "Qin Dynasty Pottery Figures", "World Heritage", "Deep Verification", "Cultural Relic Conservation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化报告" } ], "prompt": "请按「腾讯探元学术研究-深度考证」专家的工作流，深度研究「秦始皇兵马俑」。\n\n研究范围（覆盖四维）：\n- 发现与发掘历程：1974 年临潼西杨村发现、一号/二号/三号坑与铜车马发掘序列\n- 军阵编制与陶俑工艺：步兵/骑兵/车兵方阵、陶俑烧制与组装工艺、人俑个性化造型\n- 青铜兵器与彩绘保护：青铜剑铬盐氧化处理、彩绘脱落机理与 PEG 保护技术\n- 文化价值与申遗：1987 年 UNESCO 世界遗产、秦代军事与政治文化、当代展示与数字化保护\n\n交付要求（强制）：\n1. 探元两库 MCP（search_movable_relics + search_relics）并行检索 + 主代理 2-3 次 WebSearch 多源核验\n2. 配图 ≥ 2 张（建议 3-6 张），优先探元两库 cover；图表 ≥ 1 张（建议 2-3 张，如发掘时间线/军阵编制图/保护技术对比）\n3. 四格式报告：Markdown 母版 → HTML → DOCX → PDF，10 章节固定结构\n4. 仅结论不含建议；参考文献只列网络公开来源；主要结论同步输出到对话窗口", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Deep Verification' expert workflow to deeply research the 'Terracotta Army of Qin Shi Huang'.\n\nResearch scope (four dimensions):\n- Discovery and excavation process: 1974 Lintong Xiyangcun discovery, Pit 1/2/3 and bronze chariots excavation sequence\n- Military formation and pottery craftsmanship: infantry/cavalry/chariot formations, pottery firing and assembly, individualized modeling\n- Bronze weapons and painted-figure conservation: bronze swords chromium salt oxidation, polychrome peeling mechanism and PEG protection technology\n- Cultural value and UNESCO: 1987 UNESCO World Heritage, Qin military and political culture, contemporary display and digital conservation\n\nDelivery requirements (mandatory):\n1. Parallel retrieval via Tanyuan dual bases MCP (search_movable_relics + search_relics) + 2-3 multi-source WebSearch passes by main agent\n2. ≥2 images (3-6 recommended), Tanyuan covers preferred; ≥1 chart (2-3 recommended: excavation timeline/military formation diagram/conservation technology comparison)\n3. Four-format reports: Markdown master → HTML → DOCX → PDF, fixed 10-section structure\n4. Conclusions only, no recommendations; references only list public web sources; main conclusions sync to chat window", "skills": [], "experts": [ { "id": "tanyuan-deep-research", "name": "文博深度考证专家", "description": "基于探元文物库与知识库并行采集，由主代理直接联网多源交叉核验，交付四格式深度考证报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "e5dcea690c95", "expert_id": "ex_RNmBcUuH9fOC" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 102, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-08-14T16:47:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "fund-newbie-buy", "title": "基金新手入门指南", "title_en": "Fund Beginner's Guide", "subtitle": "小达养基笔记：从分类、风险匹配到定投入门", "subtitle_en": "Xiaoda's Fund Notes: from fund types and risk fit to dollar-cost-averaging", "description": "【痛点】基金新手常被\"买哪只、去哪买、怎么投\"三件事卡住，又容易被营销话术带偏，还没开始就交了学费。【演示】本案例用易方达投教基地「小达养基笔记」技能，基于财富学堂投教知识库，以新手友好的方式讲清四件事：先说\"基金是什么、有哪些分类\"，帮新手建立基础认知；再强调\"基金有风险，先匹配自己的风险承受能力\"，把风险匹配放在动手之前；接着讲\"怎么买\"——渠道、认购/申购、价格与费率一次说清；最后落到\"怎么投\"——建议从定投开始，降低择时压力。全程不替用户做买卖决策。【适用人群】零基础或入门级基金投资者，想先建立正确认知再动手的新手，以及需要投教级合规入门引导的用户。", "description_en": "Pain point: fund beginners get stuck on three things — which one to buy, where to buy, how to invest — and are easily swayed by marketing, paying tuition before they even start. Demo: this case uses the E Fund investor-education assistant \"Xiaoda's Fund Notes\" to explain four things in a beginner-friendly way grounded in the Fortune Academy knowledge base: what a fund is and its main categories (basic literacy first); that funds carry risk and you should match your risk tolerance before acting; how to buy — channels, subscription/redemption, pricing and fees, all in one place; and how to invest — starting with dollar-cost averaging to reduce timing pressure. Never makes the buy/sell decision for the user. Target audience: zero-base or entry-level fund investors who want correct cognition before acting, and anyone needing a compliant, educational onboarding guide.", "categories": [ "finance", "education" ], "tags": [ "基金", "新手入门", "投教问答", "基金定投", "怎么买基金" ], "tags_en": [ "Funds", "Beginner", "Investor Education Q&A", "Dollar-Cost Averaging", "How to Buy Funds" ], "artifact_type": "html", "cover_image": "cover.jpg", "cover_mode": "landscape", "preview": "基金新手入门指南.html", "preview_en": "", "external_url": "", "artifacts": [ { "file": "基金新手入门指南.html", "label": "完整图文回答（入门四步：认知/风险/购买/定投）" } ], "prompt": "用 efunds-edu-query 回答：我是新手小白，基金怎么买、怎么投？", "prompt_en": "Use the efunds-edu-query skill to answer: I'm a beginner, how do I buy and invest in funds?", "skills": [ { "id": "efunds-edu-query", "name": "易方达投教问答", "description": "易方达投教基地「小达养基笔记」智能问答助手——基于财富学堂投教知识库，以三段式结构精准回答基金/ETF/黄金等投资品种的概念、选择与时机类问题，全文零第三人称转述、开场即观点，不处理纯行情查询、个股推荐、K线分析等非投教类问题。", "skill_id": "skill_2089898953429123072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 103, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T15:15:00+08:00", "updated_at": "2026-08-28T15:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "kpi-sales-dashboard", "title": "销售 KPI 数据仪表盘", "title_en": "Sales KPI Data Dashboard", "subtitle": "输入销售数据，生成纯 CSS/SVG 实现的 KPI 仪表盘页面", "subtitle_en": "Input sales data, generate a KPI dashboard page using pure CSS/SVG", "description": "销售团队每周要开复盘会，每月要交经营数据，每季度要给管理层汇报——每次都要从 CRM 导出数据再手动做图表，费时费力还不好看。\n\n这个案例演示了输入销售团队的核心 KPI 数据（总营收、新签客户、续约率、客单价、团队排名），AI 自动生成一个纯 CSS/SVG 实现的销售 KPI 仪表盘 HTML 页面：居中布局（max-width 720px），顶部 4 个 KPI 大数字指标卡、月度营收趋势柱状图、团队成员销售排行榜、客户转化漏斗图。全部纯 CSS/SVG 绘制，零外部依赖。\n\n适合：销售主管、销售运营、BD 团队、需要定期汇报销售数据的团队。", "description_en": "Sales teams need weekly reviews, monthly reports, and quarterly executive presentations — every time requiring CRM data exports and manual chart creation, which is time-consuming and rarely looks professional.\n\nThis case demonstrates inputting sales team core KPI data (total revenue, new customers, renewal rate, average deal size, team rankings), with AI generating a pure CSS/SVG sales KPI dashboard HTML page: centered layout (max-width 720px), top 4 KPI metric cards, monthly revenue trend bar chart, team member sales leaderboard, customer conversion funnel. All pure CSS/SVG, zero external dependencies.\n\nBest for: Sales managers, sales operations, BD teams, any team needing regular sales data reporting.", "categories": [ "data", "business" ], "tags": [ "销售仪表盘", "KPI", "数据可视化", "销售分析", "CSS图表" ], "tags_en": [ "Sales Dashboard", "KPI", "Data Visualization", "Sales Analysis", "CSS Charts" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个销售 KPI 仪表盘页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，现代精美。\n\n直观展示本月 KPI 达成、营收趋势、团队排行与转化漏斗。填入你的数据：\n- 核心 KPI：总营收、新签客户、续约率、平均客单价（含目标、达成率与环比）\n- 月度营收趋势\n- 团队成员销售排行\n- 客户转化漏斗\n\n用纯 CSS/SVG 绘制图表，零外部依赖。", "prompt_en": "Generate a sales KPI dashboard page (single-file HTML, inlined with no external dependencies). The interface text should be in Simplified Chinese, modern and polished.\n\nIntuitively display this month's KPI achievement, revenue trend, team ranking, and conversion funnel. Fill in your own data:\n- Core KPIs: total revenue, new customers signed, renewal rate, average deal size (with targets, attainment, and MoM change)\n- Monthly revenue trend\n- Team member sales ranking\n- Customer conversion funnel\n\nDraw all charts in pure CSS/SVG with zero external dependencies.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 103, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "KPI 销售仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "markitdown-word-to-markdown", "title": "Word 计划书一键转 Markdown", "title_en": "Turn a Word Plan into Clean Markdown in One Shot", "subtitle": "多级标题、表格、列表原样保留，直接贴进 wiki", "subtitle_en": "Headings, tables and lists preserved—ready for your wiki", "description": "写好的 Word 计划书，复制到 Markdown 编辑器里格式全乱了？标题掉级、表格散成一坨、项目符号变纯文本，光重排就得折腾半小时。\n\n这个案例演示了用 MarkItDown 把一份《新产品上市推广计划书》.docx 一条命令转成干净的 Markdown：一级到三级标题层级完整保留，渠道预算表、上市时间节点表都转成标准 Markdown 表格，项目符号列表和有序列表原样还原，全程零乱码、零控制字符。\n\n适合经常要把 Word 文档搬进团队 wiki、Notion、语雀，或者想把文档喂给 AI 做二次编辑的产品、运营和文档同学。", "description_en": "Finished a plan in Word, but pasting it into a Markdown editor wrecks the formatting? Headings collapse, tables fall apart, bullet points turn into plain text—reformatting alone eats half an hour.\n\nThis case shows MarkItDown converting a \"New Product Launch Plan\" .docx into clean Markdown with a single command: heading levels one through three stay intact, the channel-budget table and launch-timeline table become proper Markdown tables, and both bulleted and numbered lists are faithfully restored—no garbled characters, no control characters.\n\nPerfect for product, ops and documentation folks who regularly move Word docs into a team wiki, Notion or Yuque, or want to feed a document into AI for further editing.", "categories": [ "productivity" ], "tags": [ "Word转Markdown", "文档转换", "办公", "Markdown" ], "tags_en": [ "Word to Markdown", "Document Conversion", "Office", "Markdown" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "把这份《新产品上市推广计划书》Word 文档（.docx，含多级标题、渠道预算表、上市时间节点表、项目符号列表）转成 Markdown，完整保留标题层级、表格和列表结构，方便贴进团队 wiki 或喂给 AI 做二次编辑。", "prompt_en": "Convert this \"New Product Launch Plan\" Word document (.docx, with multi-level headings, a channel-budget table, a launch-timeline table, and bulleted lists) into Markdown, fully preserving the heading levels, tables, and list structure, so it can be pasted into a team wiki or fed to an AI for further editing.", "skills": [ { "id": "markitdown-skill", "name": "MarkItDown", "description": "文档转 Markdown(PDF/Word/PPT/图片OCR/音频转写/网页)", "skill_id": "skill_2053083553754644480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 99, "sort_order": 103, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T00:00:00Z", "updated_at": "2026-07-24T11:04:33Z", "artifacts": [ { "file": "output.html", "label": "上市推广计划书（HTML 渲染预览）" }, { "file": "output.md", "label": "MarkItDown 转换原文（output.md）" }, { "file": "product-launch-plan.docx", "label": "源文件（Word 上市推广计划书）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "pharma-single-dimension-insight", "title": "单维度洞察报告", "title_en": "Single-Dimension Insight Report", "subtitle": "针对某一指定维度生成独立的图表+洞察+策略子报告", "subtitle_en": "Generate a standalone chart + insight + strategy sub-report for any specified dimension", "description": "并非每次分析都需要一份数十页的全景报告。实际工作中，往往只需弄清某一个维度的数据表现，例如本月扫码人数的环比变化、各省份用户分布情况，或竞品在药箱内的关联用药情况。但在 BI 平台中定位指标、导出数据并制作图表，仍需耗费可观的时间成本。\n\n这个案例让 AI 针对单个维度快速产出结果：用户指定药品名与维度名（如「帮我生成 xx 药品 2026 上半年整体市场规模和药箱用户分析的数据报告」），AI 只聚焦该维度，产出包含可视化图表、数据洞察结论，以及下一步策略建议（按企业侧/药箱侧分组）的独立 HTML 子报告。生成完成后，系统会主动询问是否需要补齐其他维度，进而拼装为完整版报告。\n\n适合：需要快速查看单一指标趋势的药企运营人员、在多维度分析前先做单维探查的数据分析师、希望先评估子报告效果再决定是否生成完整版的药企决策者、以及开展功能模块专项分析的药箱产品经理。", "description_en": "Not every analysis calls for a dozens-of-pages panoramic report. In practice, teams often only need to understand one specific dimension—the month-over-month change in scan counts, the geographic distribution of users, or how competitor drugs correlate within the medicine chest. Yet locating the metric in a BI platform, exporting the data, and building charts still carries a considerable time cost.\n\nThis case enables AI to deliver results rapidly for a single dimension: the user specifies the drug name and dimension (e.g., \"generate a data report for xx drug H1 2026 overall market scale and medicine-chest user analysis\"), and AI focuses solely on that dimension, producing a standalone HTML sub-report with visualization charts, data insight conclusions, and next-step strategy recommendations (grouped by enterprise / medicine-chest). Once generated, the system proactively inquires whether other dimensions should be added to assemble the full report.\n\nBest for: pharma operations staff needing quick single-metric trend checks, data analysts conducting single-dimension exploration before multi-dimensional analysis, pharma decision-makers who want to evaluate a sub-report before producing the full version, and medicine-chest product managers running feature-specific analyses.", "categories": [ "data", "business" ], "tags": [ "单维度分析", "数据洞察", "图表报告", "策略建议", "药箱数据" ], "tags_en": [ "Single-Dimension Analysis", "Data Insight", "Chart Report", "Strategy Recommendations", "Medicine-Chest Data" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "单维度数据洞察报告（指定维度的图表+洞察+策略子报告 HTML）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我生成xx 药品2026 上半年「整体市场规模和药箱用户分析」的数据报告，生成内容包括可视化图表、数据洞察结论、以及下一步策略建议（按执行主体企业 / 药箱分组）。", "prompt_en": "Generate a data report for xx drug H1 2026 \"Overall Market Scale and Medicine-Chest User Analysis\". The output should include visualization charts, data insight conclusions, and next-step strategy recommendations (grouped by executor: enterprise / medicine-chest).", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 103, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T06:44:00Z", "updated_at": "2026-09-04T08:55:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "shexiang-baoxin-checkin", "title": "一句话打卡麝香保心丸", "title_en": "One-line Check-in for Shexiang Baoxin Wan", "subtitle": "说一句「吃过了」，方案自动匹配、今日进度一眼看清", "subtitle_en": "Say you took it — the regimen is matched automatically and today's progress is clear at a glance", "description": "长期吃药最难的不是记得怎么吃，而是吃了之后记得记。一天三次、常年不断的方案下，事后回想、漏记、断档几乎是常态；等到复诊被问「最近按时吃了吗」，谁也说不清楚。\n\n这个案例把打卡这件事压缩成一句话：用户只说「帮我记录今天已经吃过麝香保心丸了」，不用报剂量。AI 从这句话里认出药名，在用药档案里匹配到「每天 3 次 · 每次 2 丸」，按当前时段落库并返回确认页——顶部复述本次打卡（今天 18:00，今日第 3 次），下方把药品、用法、本次时间三项摆开供核对，配今日 3 个时间点的完成进度；再往下是日常服用与发作时含服两套方案的区分、漏服不补量的提醒、以及孕妇禁用等服药期间要留意的几条，末尾给出扫码进小程序查看完整打卡日历的入口。\n\n适合：冠心病等需长期规律服药的慢病人群、记不住有没有吃药的中老年用户、替父母代打卡的家庭照护者、做慢病随访管理的药企与医疗机构运营。", "description_en": "The hard part of long-term medication is not knowing how to take it — it is remembering to log that you did. On a three-times-daily, year-round regimen, recalling later, missing entries and gaps in the record are the norm; by the time a follow-up asks \"have you been taking it on schedule?\", nobody can answer with confidence.\n\nThis case compresses the check-in into a single sentence. The user only says \"log that I took Shexiang Baoxin Wan today\" — no dose needed. The assistant picks the drug name out of that sentence, matches \"3 times daily, 2 pills each\" from the user's medication record, files the entry under the current time slot and returns a confirmation page: the header restates this check-in (today 18:00, the third dose of the day); below it the drug, the regimen and this dose's time are laid out for verification, alongside completion progress across today's three time slots; further down it separates the daily regimen from sublingual use during an episode, warns against doubling up after a missed dose, and lists precautions such as the contraindication in pregnancy, closing with a scan-code entry to the mini-program's full check-in calendar.\n\nBest for: patients on long-term regimens such as coronary heart disease, older users who cannot recall whether they took a dose, family caregivers logging on a parent's behalf, and chronic-care follow-up teams at pharma companies and healthcare providers.", "categories": [ "education", "productivity" ], "tags": [ "用药打卡", "慢病管理", "中老年用药", "麝香保心丸", "服药记录" ], "tags_en": [ "Medication Check-in", "Chronic Care", "Older Adult Care", "Shexiang Baoxin Wan", "Dose Log" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "麝香保心丸打卡确认页（本次记录核对 + 今日进度 + 用药方案 + 注意事项 + 小程序码入口）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "你是用药专家，帮用户把「吃过了」这件事记下来。用户会用一句大白话告诉你他刚吃了什么药，你要立刻完成打卡，并给出一个看得见、能核对的确认页。\n\n用户输入：{用户想记录的服药事实，例如「帮我记录今天已经吃过麝香保心丸了」}\n\n请完成：\n1. 从用户这句话里认出药品名，在用户已有的用药档案中匹配对应方案（每天几次、每次多少），并按当前时段判断这是今天的第几次。匹配不到方案时，只按用户明确说出的信息记录，不要自行推测剂量。\n2. 顶部用一句话复述本次打卡：药品名、匹配到的用法、本次时间、今日第几次。\n3. 把药品、用法、本次时间三项并排列出来供用户核对；再给出今日各时间点的完成进度，已打卡、本次、未打卡要一眼能区分。\n4. 说明该药的用药方案，区分「日常按时服用」与「症状发作时含服」两种情形，并提示漏服时不要补量。\n5. 列出服药期间需要留意的关键安全信息（禁忌人群、症状加重需就医、合并用药先问医师、贮藏），用法用量与禁忌须取自官方电子说明书原文，不做医学解读。\n6. 结尾给出扫码进入小程序查看和管理完整打卡记录的入口。\n\n输出为单文件 HTML 页面，CSS 与图片资源全部内联、可离线打开；蓝色系医疗风格，左右两栏布局（左侧记录与方案、右侧机制说明与注意事项），打卡成功状态一眼可见。只处理这一次打卡确认，不追问其他信息、不扩展成多轮对话。", "prompt_en": "You are a medication expert who logs the fact that a user has taken a dose. The user tells you in plain language what they just took; you complete the check-in immediately and return a confirmation page they can see and verify.\n\nUser input: {the medication fact to log, e.g. \"log that I took Shexiang Baoxin Wan today\"}\n\nDo the following:\n1. Identify the drug name from the sentence, match the corresponding regimen (how many times daily, how much each time) in the user's medication record, and determine which dose of the day this is from the current time slot. If no regimen matches, record only what the user explicitly stated and do not infer a dose.\n2. At the top, restate this check-in in one sentence: drug name, the matched regimen, this dose's time, and which dose of the day it is.\n3. Lay out the drug, the regimen and this dose's time side by side for verification; then show completion progress across today's time slots, making logged / current / not-yet-logged visually distinct.\n4. Explain the regimen, separating routine scheduled dosing from sublingual use during an episode, and warn not to double up after a missed dose.\n5. List the key safety information for the treatment period (contraindicated populations, when worsening symptoms require care, asking a physician before combining drugs, storage). Dosage and contraindications must come from the official electronic package insert verbatim, with no medical interpretation added.\n6. Close with a scan-code entry to the mini-program for reviewing and managing the full check-in history.\n\nOutput a single-file HTML page with all CSS and images inlined so it opens offline; a blue clinical style with a two-column layout (records and regimen on the left, how-it-works and precautions on the right), where the successful check-in state is visible at a glance. Handle only this one check-in confirmation; do not ask follow-up questions or expand into a multi-turn dialogue.", "skills": [], "experts": [ { "id": "eyao-medicine-chest-expert", "name": "腾讯健康药箱用药专家", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_zNvIxWe9Yekr", "description": "管药这件事的助手：帮你查药品说明书、设置用药提醒、记录每天吃药，让你管好每顿药", "promptFileSnapshot": "" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 103, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-28T13:00:00Z", "updated_at": "2026-08-31T03:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "markitdown-ppt-to-markdown", "title": "PPT 一键转 Markdown，会议纪要不用手敲", "title_en": "Turn PPT into Markdown in One Click, No More Retyping Meeting Notes", "subtitle": "把季度复盘汇报 PPT 逐页转成干净的 Markdown，表格照样保留", "subtitle_en": "Convert a quarterly review deck into clean Markdown page by page, tables included", "description": "开完复盘会，一份几十页的 PPT 摆在面前，想整理成文字纪要或者喂给 AI 做总结，却发现只能一页页复制粘贴，表格更是排版全乱——费时又容易漏。\n\n这个案例把一份真实的《Q3 季度业务复盘汇报》PPTX 交给 MarkItDown，逐页转成结构清晰的 Markdown：每页带幻灯片编号注释、标题和要点原样保留，分品类销售表格也完整转成 Markdown 表格，全程真实执行、零乱码。\n\n适合要写会议纪要的运营和产品、需要把演示稿内容二次加工的职场人，以及想把 PPT 内容丢给大模型做提炼总结的所有人。", "description_en": "After a review meeting, you're staring at a deck of dozens of slides. You want to turn it into text notes or feed it to an AI for a summary, but you end up copy-pasting slide by slide—and the tables come out a mess. Slow and error-prone.\n\nThis case hands a real \"Q3 Business Review\" PPTX to MarkItDown and converts it page by page into clean, structured Markdown: each slide keeps its number comment, title and bullet points, and the category sales table is fully preserved as a Markdown table. Fully executed, zero garbled characters.\n\nGreat for ops and product folks writing meeting notes, professionals repurposing presentation content, and anyone who wants to feed PPT content to an LLM for summarization.", "categories": [ "productivity" ], "tags": [ "PPT转Markdown", "演示文稿", "办公", "会议纪要" ], "tags_en": [ "PPT to Markdown", "Presentation", "Office", "Meeting Notes" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "把这份《Q3 季度业务复盘汇报》PowerPoint（.pptx，约 8-10 页，含封面、议程、核心指标、分品类销售表格和下季度计划）转成 Markdown，逐页保留标题和要点文本、表格结构也保留，方便整理成会议纪要或喂给 AI 做总结。", "prompt_en": "Convert this \"Q3 Business Review\" PowerPoint (.pptx, about 8-10 slides, including cover, agenda, core metrics, a category sales table, and next-quarter plans) into Markdown. Keep the title and bullet text for each slide, and preserve the table structure, so it can be turned into meeting notes or fed to an AI for summarization.", "skills": [ { "id": "markitdown-skill", "name": "MarkItDown", "description": "文档转 Markdown(PDF/Word/PPT/图片OCR/音频转写/网页)", "skill_id": "skill_2053083553754644480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 100, "sort_order": 104, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T00:00:00Z", "updated_at": "2026-07-24T11:04:33Z", "artifacts": [ { "file": "output.html", "label": "业务复盘（HTML 渲染预览）" }, { "file": "output.md", "label": "MarkItDown 转换原文（output.md）" }, { "file": "q3-business-review.pptx", "label": "源文件（Q3 业务复盘 PPT）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "pharma-data-quick-query", "title": "药品数据快速查询", "title_en": "Pharmaceutical Data Quick Query", "subtitle": "一句话查数，秒级返回单一指标", "subtitle_en": "Ask in one sentence, get a single metric back in seconds", "description": "最轻量的数据需求往往只需要一个数值，例如某药品上月的扫码人数、日均访问量或打卡率。若为此专门登录 BI 平台、设置筛选条件并等待查询加载，所需成本远高于信息本身的价值。\n\n这个案例将取数简化为一句自然语言：用户以日常用语提问（如「帮我查询 xx 药品 2026 年 1-6 月的药箱月均扫码人数」），AI 直接从药箱数据平台返回精确数值，并附带指标说明与时间范围。查询完成后，系统会进一步询问是否需要生成可视化图表或完整的数据洞察分析报告，实现从「获取单个数值」到「产出完整报告」的平滑延展。\n\n适合：需要快速核实单一 KPI 指标的药企运营与销售、会议中需要即时查询数据的药企客户经理、撰写日报周报时需要填充具体数值的产品与运营人员、以及希望在深入分析前先评估数据质量的各类用户。", "description_en": "The lightest data requirement is often a single figure—last month's scan count for a drug, daily active user volume, or the check-in rate. Logging into a BI platform, configuring filters, and waiting for queries to load simply to obtain one number carries a cost well beyond the value of the information itself.\n\nThis case reduces data retrieval to a single natural-language sentence: the user asks in everyday language (e.g., \"look up the monthly average scan count for xx drug from Jan–Jun 2026\"), and AI returns the exact value directly from the medicine chest data platform, along with the metric description and time range. Once the query is complete, the system further inquires whether to generate visualization charts or a full data insight analysis report—enabling a smooth extension from \"retrieve a single figure\" to \"produce a complete report.\"\n\nBest for: pharma operations and sales needing quick single-KPI verification, account managers requiring on-the-spot data lookups during meetings, product and operations staff filling in specific figures for daily or weekly reports, and any user who wants to assess data quality before committing to deeper analysis.", "categories": [ "data", "search" ], "tags": [ "数据查询", "快速查数", "药箱指标", "KPI查询", "自然语言" ], "tags_en": [ "Data Quick Query", "Fast Lookup", "Medicine-Chest Metrics", "KPI Query", "Natural Language" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "药品数据快速查询结果页（问题+答案+引导生成图表/报告的 HTML 页面）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查询xx 药品2026 年1-6 月的药箱月均扫码人数。", "prompt_en": "Look up the monthly average medicine-chest scan count for xx drug from January to June 2026.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 104, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-31T06:44:00Z", "updated_at": "2026-08-31T06:44:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "trtc-expert-showcase", "title": "腾讯云实时音视频专家能力全景展示", "subtitle": "一站式完成 TRTC 问题的「预检 → 诊断 → 治理」全链路", "description": "TRTC 应用一旦上线，开发、运维每天都会遇到各种问题：用户反馈卡顿、看不到画面、加入房间失败、不知道资产有多少、配置是否合规——这些问题以前需要在控制台里点 5~10 个菜单、翻客户端日志、甚至跨多个产品页面才能搞清楚。本案例完整展示腾讯云实时音视频专家的三大核心能力：基础指标查询（应用清单 / 推流参数 / 在房状态 / SDK 版本）、通话诊断（五段法定位黑屏/卡顿/断流根因）、云端巡检解读（风险下钻 + 与现网问题交叉验证），全部用自然语言一句话完成。适合 TRTC 应用的开发、客服一线、运维 SRE，作为日常处理 TRTC 问题的标准入口与方法论参考。", "tags": [ "实时音视频", "TRTC", "故障诊断", "云端巡检", "通话查询" ], "categories": [ "dev" ], "artifact_type": "md", "cover_image": "cover.png", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "专家能力全景展示报告", "label_en": "Expert Capability Showcase Report" } ], "prompt": "请扮演腾讯云实时音视频专家（tencent-rtc-expert，依托 cloudq Skill 后端），完成一份 TRTC「能力全景展示」报告，用样例 SdkAppId 演示（参数可用占位符），串联以下四部分：\n\n一、基础指标查询：完成若干典型查询，每个用「自然语言提问 → 数据表 → 典型用途」三段式展示。\n二、通话诊断：以「观众看不到主播画面」为例，按采集→推流→网络→拉流→渲染五段法分段诊断，每段给关键指标与健康判定，锁定异常环节并给出根因与分级处置建议。\n三、云端巡检解读：拉取最新巡检结果，解读警告项并下钻到房间/用户粒度，与第二部分诊断交叉验证，按紧急/重要/长期分级处置。\n四、能力价值小结：说明三段能力的协作关系，对比传统方式与专家方式，给出可复用模板。\n\n输出 Markdown 报告，每部分独立成章。", "prompt_en": "Acting as the Tencent Cloud RTC expert (tencent-rtc-expert, backed by the cloudq skill), produce a TRTC \"capability showcase\" report using a sample SdkAppId (parameters may be placeholders), covering four parts:\n\n1. Basic metric queries: run several typical queries, each shown as \"natural-language question → data table → typical use\".\n2. Call diagnosis: using \"viewers can't see the host's video\" as the case, diagnose in five stages (capture → publish → network → pull → render), with key metrics and a health verdict per stage, pinpoint the faulty stage, and give root cause plus tiered fixes.\n3. Cloud inspection review: pull the latest inspection results, interpret warnings and drill down to room/user level, cross-validate against part 2, and triage as urgent/important/long-term.\n4. Capability value summary: explain how the three capabilities work together, compare the traditional vs expert approach, and give a reusable template.\n\nOutput a Markdown report with each part as its own chapter.", "skills": [], "experts": [ { "id": "TencentRtcExpert", "name": "腾讯云实时音视频专家", "description": "专注于 TRTC 领域的实时音视频专家，提供通话查询、单次通话诊断、云端巡检解读与故障排查（黑屏/卡顿/断流/回声）。依托 cloudq Skill 后端能力，按 SdkAppId / RoomId / UserId 定位单次通话问题。", "expert_id": "ex_2Qd4D2kPRBuZ" } ], "quality_score": 93, "source": "official", "created_at": "2026-05-11T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_featured": false, "usage_count": 0, "sort_order": 104, "mcps": [], "scenario": "", "task_mode": "", "cover_mode": "landscape", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "tags_en": [ "Real-Time Audio/Video", "TRTC", "Troubleshooting", "Cloud Inspection", "Call Query" ], "title_en": "Tencent Cloud TRTC Expert Capability Showcase", "subtitle_en": "One-stop TRTC issue resolution: Pre-Check → Diagnosis → Governance", "description_en": "Once a TRTC application goes live, developers and operators face daily issues: user-reported lag, no video, room join failures, missing asset info, configuration compliance — problems that previously required clicking through 5-10 console menus, digging through client logs, and navigating across multiple product pages. This case demonstrates the three core capabilities of the Tencent Cloud TRTC Expert: basic metric queries (app list, push parameters, room presence, SDK versions), call diagnostics (five-phase root cause analysis for black screen, lag, and disconnection), and cloud inspection interpretation (risk drill-down with cross-validation against production issues) — all accomplished with a single natural-language prompt. Ideal for TRTC developers, frontline support, and SRE engineers as a standard entry point and methodology reference for daily TRTC issue resolution.", "is_recommended": false, "recommended_order": 0 }, { "id": "markitdown-batch-office-to-markdown", "title": "散落的办公文档，一键汇成 Markdown 知识库", "title_en": "Turn Scattered Office Docs into One Markdown Knowledge Base", "subtitle": "Word/Excel/PPT 批量转换 + 合并成带索引的统一知识库", "subtitle_en": "Batch-convert Word/Excel/PPT and merge into one indexed knowledge base", "description": "项目做久了，Word 立项书、Excel 数据表、PPT 复盘散落一地，格式五花八门——想全文检索搜不了，想丢给 AI 做问答又喂不进去，翻来翻去还容易漏。\n\n这个案例用 MarkItDown 把一个「云图 SaaS 客户成功项目」里的三份不同格式文档一次性批量转成 Markdown，再合并成一份带「文档索引」的统一知识库 output.md：表格、要点、分节标题都原样保留，纯文本、零乱码，直接归档或喂 AI 都行。\n\n适合要整理项目资料、搭建团队知识库、给 RAG/AI 问答准备语料的产品、运营、客户成功和知识管理同学。", "description_en": "After a project runs for a while, Word proposals, Excel sheets and PowerPoint reviews pile up in every format imaginable — you can't full-text search them, you can't feed them to an AI for Q&A, and flipping through them one by one means things slip through the cracks.\n\nThis case uses MarkItDown to batch-convert three differently formatted documents from a fictional \"Skymap SaaS Customer Success project\" into Markdown in one go, then merges them into a single indexed knowledge base, output.md: tables, bullet points and section headings are preserved as-is, plain text with zero garbled characters, ready to archive or feed straight to an AI.\n\nPerfect for product, ops, customer success and knowledge-management folks who need to organize project files, build a team knowledge base, or prepare corpora for RAG/AI Q&A.", "categories": [ "productivity" ], "tags": [ "批量转换", "知识库", "办公文档", "Markdown" ], "tags_en": [ "Batch Conversion", "Knowledge Base", "Office Docs", "Markdown" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我这个项目里散落着 Word 立项书、Excel 客户数据表、PowerPoint 复盘汇报。请用 MarkItDown 把它们批量转成 Markdown，再合并成一份带文档索引的统一知识库 Markdown 文件，方便归档、全文检索和喂给 AI 做问答。", "prompt_en": "My project has a Word proposal, an Excel customer data sheet and a PowerPoint review scattered around. Please use MarkItDown to batch-convert them into Markdown, then merge them into a single unified knowledge-base Markdown file with a document index, so it's easy to archive, full-text search and feed to an AI for Q&A.", "skills": [ { "id": "markitdown-skill", "name": "MarkItDown", "description": "文档转 Markdown(PDF/Word/PPT/图片OCR/音频转写/网页)", "skill_id": "skill_2053083553754644480" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 98, "sort_order": 105, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-07-24T00:00:00Z", "updated_at": "2026-07-24T11:04:33Z", "artifacts": [ { "file": "output.html", "label": "统一知识库（HTML 渲染预览）" }, { "file": "output.md", "label": "MarkItDown 合并转换原文（output.md）" }, { "file": "项目立项书.docx", "label": "源文件·Word 立项书" }, { "file": "客户健康度数据.xlsx", "label": "源文件·Excel 客户数据" }, { "file": "季度复盘汇报.pptx", "label": "源文件·PPT 复盘汇报" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "revealjs-startup-pitch", "title": "创业融资路演 Slides", "title_en": "Startup Pitch Deck Slides", "subtitle": "一键生成 Seed 轮路演演示文稿", "subtitle_en": "One-click Seed round pitch deck generation", "description": "融资路演做 PPT 要反复改十几版，设计师排不上档期，最终效果还不够专业？用 deck-generator 技能直接生成一套 12-15 页的完整路演演示文稿，涵盖痛点、方案、市场、商业模式、财务预测、团队等核心板块，支持 AI 配图和专业动画效果。\n\n本案例以一家 AI 客服 SaaS 初创公司为例，演示了 deck-generator 如何自动生成包含封面、问题定义、解决方案、产品演示、TAM/SAM/SOM 市场分析、竞品对比、团队介绍、财务预测和融资需求的完整路演 PPTX。\n\n适合：创业者、CEO、融资顾问。", "description_en": "Pitch deck iterations are endless and designers are always bottlenecked. Use deck-generator to generate a complete 12-15 page pitch deck PPTX covering pain points, solution, market size, business model, financials, and team with AI-generated images.\n\nThis case uses an AI customer service SaaS startup scenario to demonstrate how deck-generator auto-generates a professional pitch deck with all essential sections.\n\nPerfect for founders, CEOs, and fundraising advisors.", "categories": [ "creation", "business" ], "tags": [ "路演", "融资", "创业", "Pitch Deck", "演示文稿" ], "tags_en": [ "Pitch Deck", "Fundraising", "Startup", "Presentation", "Slides" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "创业融资路演Reveal.js 演示" } ], "prompt": "帮我生成一份创业融资路演演示文稿。我们是一家做 AI 客服 SaaS 的初创公司，目标融 Seed 轮 500 万美元。包含：痛点分析、解决方案、产品 Demo 截图描述、市场规模（TAM/SAM/SOM）、商业模式、竞品对比、团队介绍、财务预测、融资需求与资金用途。共 12-15 页。请使用 deck-generator 生成专业的 PPTX 演示文稿。", "prompt_en": "Help me create a startup fundraising pitch deck. We are a startup building AI customer service SaaS, aiming to raise a $5 million Seed round. Include: pain point analysis, solution, product demo screenshot descriptions, market size (TAM/SAM/SOM), business model, competitive comparison, team introduction, financial projections, and the funding ask with use of funds. 12-15 slides total. Please use deck-generator to produce a professional PPTX presentation.", "skills": [ { "id": "deck-generator", "name": "deck-generator", "description": "AI 驱动的演示文稿生成，统一视觉风格的幻灯片", "description_en": "AI-powered slide deck generation with consistent visual styling", "skill_id": "skill_2053082023531843584" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 96, "sort_order": 105, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "landing-page-ai-product", "title": "AI 产品高转化落地页", "title_en": "High-Converting AI Product Landing Page", "subtitle": "SaaS 产品首页的完整设计与转化优化", "subtitle_en": "Complete design and conversion optimization for SaaS product homepage", "description": "SaaS 产品上线了但官网转化率不到 1%，Hero 区没吸引力、CTA 不够突出、社会证明缺失？用前端开发 Skill 生成一个高转化率的 AI 写作助手产品首页——Hero + 功能亮点 + 使用流程 + 客户评价 + 定价 + FAQ，响应式设计开箱即用。\n\n适合产品经理、增长负责人、独立开发者。", "description_en": "SaaS product launched but website conversion rate is under 1%? Hero section lacks punch, CTAs are buried, social proof is missing. Use the Frontend Dev Skill to generate a high-converting AI writing assistant homepage — Hero + Features + How It Works + Testimonials + Pricing + FAQ with responsive design. Perfect for PMs, growth leads, and indie devs.", "categories": [ "creation", "business" ], "tags": [ "落地页", "SaaS", "转化优化", "产品首页", "响应式设计" ], "tags_en": [ "Landing Page", "SaaS", "Conversion Optimization", "Product Homepage", "Responsive Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为 AI 写作助手 SaaS「笔流 AI」（WriteFlow AI）设计一个高转化率落地页（单文件 HTML，内联无外部依赖），界面文字以简体中文为主，现代简约、响应式。\n\n区块：\n1. Hero：大标题「AI 驱动的写作革命」+ 副标题 + 「免费试用 / 观看演示」双 CTA + 产品截图占位\n2. 信任背书：50,000+ 用户 / 4.9 星 / 500+ 企业 / 99.9% 可用\n3. 功能亮点：智能续写、多语言翻译、风格调整、SEO 优化 四个卡片\n4. 使用流程：输入→AI 处理→获得结果 三步\n5. 客户评价：3 条（含头像/姓名/职位/星级）\n6. 定价：免费版 / 专业版 ¥99 月 / 企业版 ¥299 月，专业版高亮\n7. FAQ：5 条手风琴\n8. 底部 CTA + 邮箱输入\n9. Footer", "prompt_en": "Please design a high-conversion landing page for the AI writing assistant SaaS \"笔流 AI\" (WriteFlow AI) (single-file HTML, fully inline with no external dependencies), with UI text primarily in Simplified Chinese, in a modern, minimalist, responsive style.\n\nSections:\n1. Hero: a large headline \"AI-Driven Writing Revolution\" + a subheading + dual CTAs (\"Free Trial / Watch Demo\") + a product screenshot placeholder\n2. Trust badges: 50,000+ users / 4.9 stars / 500+ enterprises / 99.9% availability\n3. Feature highlights: four cards — smart continuation, multilingual translation, style adjustment, SEO optimization\n4. How it works: three steps — input → AI processing → get results\n5. Customer testimonials: 3 of them (with avatar / name / job title / star rating)\n6. Pricing: Free / Pro ¥99 per month / Enterprise ¥299 per month, with the Pro plan highlighted\n7. FAQ: 5 items in an accordion\n8. Bottom CTA + email input\n9. Footer", "skills": [ { "id": "landing-page-design", "name": "落地页设计", "description": "高转化率落地页设计与优化", "slug": "landing-page-design" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 106, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "AI 产品落地页" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "humanizer-academic-polish", "title": "学术论文「去 AI 味」润色", "subtitle": "AI 检测率从 98% 降到 23%，保留学术严谨性", "description": "用 AI 写论文最大的问题就是一眼 AI。这个案例展示了如何用「去 AI 味」Skill 对 AI 生成的学术论文摘要进行深度润色，把 GPTZero 检测的 AI 概率从 98% 降到 23%。\n\n润色过程不是简单的同义替换，而是系统性地调整：拆分均匀长句为长短参差的节奏、去掉 AI 高频空洞词（comprehensive / significant / remarkable）替换为具体数据、增加作者的主观判断和不确定性表达、添加自我限定（limitations）让论文更可信。\n\n案例完整展示了原文 vs 润色后的对比、4 个维度的检测数据变化、以及逐条改动分析，可以直接作为润色参考模板使用。\n\n适合：研究生、博士生、科研工作者、论文写作辅导、任何需要降低 AI 痕迹的写作场景。", "categories": [ "education", "productivity" ], "tags": [ "去AI味", "论文润色", "学术写作", "AI检测", "降重", "Humanizer", "GPTZero" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "学术论文去 AI 味润色对比报告" } ], "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 107, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-04-10T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "prompt": "请对以下 AI 生成的学术论文摘要进行「去 AI 味」润色。\n\n要求：\n1. 保留所有学术要点和结论\n2. 句式长短参差，避免每句都是复合长句\n3. 去掉 AI 常用的空洞修饰词\n4. 适当加入作者的主观判断和不确定性表达\n5. 用具体数字替代模糊描述\n6. 输出润色前后对比和 AI 检测指标变化\n\n原文（粘贴 AI 生成的论文段落）：\n...", "prompt_en": "Please \"de-AI\" and polish the following AI-generated academic paper abstract.\n\nRequirements:\n1. Preserve all academic points and conclusions\n2. Vary sentence length so they aren't all long, complex sentences\n3. Remove the empty filler words AI tends to use\n4. Appropriately add the author's subjective judgments and hedging\n5. Replace vague descriptions with specific numbers\n6. Output a before-and-after comparison and the change in AI-detection metrics\n\nOriginal text (paste the AI-generated paragraph):\n...", "title_en": "Academic Paper De-AI Polish", "subtitle_en": "Reduce AI detection rate from 98% to 23% while preserving academic rigor", "description_en": "The biggest problem with AI-written papers is they scream \"AI.\" This case shows how the Humanizer Skill deeply polishes AI-generated academic abstracts, bringing GPTZero detection probability from 98% down to 23%.\n\nThe process goes beyond simple synonym swaps — it systematically adjusts: breaks uniform-length sentences into varied rhythms, replaces hollow AI buzzwords (comprehensive / significant / remarkable) with specific data, adds the author's subjective judgment and hedging expressions, and inserts self-limitations to make the paper more credible.\n\nThe case includes a full before-vs-after comparison, detection score changes across 4 dimensions, and a line-by-line edit analysis — usable as a polishing reference template.\n\nBest for: Graduate students, PhD candidates, researchers, academic writing tutors, anyone who needs to reduce AI traces in writing.", "tags_en": [ "De-AI", "Paper Polishing", "Academic Writing", "AI Detection", "Paraphrasing", "Humanizer", "GPTZero" ], "is_recommended": false, "recommended_order": 0 }, { "id": "shader-particle-galaxy", "title": "GLSL 粒子星系动画", "title_en": "GLSL Particle Galaxy Animation", "subtitle": "基于 WebGL 的实时交互式星系粒子效果", "subtitle_en": "Real-time interactive galaxy particle effects powered by WebGL", "description": "想要在网页中创建令人惊叹的太空视觉效果？本案例展示如何使用 GLSL Shader 编写实时粒子星系动画，通过噪声函数生成自然的粒子运动轨迹，配合色彩渐变和深度模糊，创造出逼真的银河旋臂效果。适合前端开发者、创意编程爱好者和游戏开发学习者。", "description_en": "Want to create stunning space visuals on the web? This case demonstrates how to use GLSL Shaders to create real-time particle galaxy animations with noise-driven trajectories, color gradients, and depth blur for realistic spiral arm effects. Perfect for frontend developers, creative coders, and game dev learners.", "categories": [ "creation", "dev" ], "tags": [ "WebGL", "Shader", "粒子动画", "创意编程" ], "tags_en": [ "WebGL", "Shader", "Particle Animation", "Creative Coding" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "GLSL 粒子星系动画页面" } ], "prompt": "请创建一个交互式粒子星系动画页面（单文件 HTML，内联无外部依赖，含内联 vertex/fragment shader 和 JS，打开即运行）。\n\n要求：纯 WebGL + GLSL 实现（不依赖 Three.js）；5000+ 粒子形成旋臂结构；用 simplex/perlin noise 驱动运动；支持鼠标拖拽旋转视角；粒子颜色从蓝紫核心渐变到金色外缘；星点闪烁；深色太空背景。页面底部附简短技术说明。", "prompt_en": "Please create an interactive particle-galaxy animation page (single-file HTML, inlined with no external dependencies, including inline vertex/fragment shaders and JS, running as soon as it's opened).\n\nRequirements: implemented in pure WebGL + GLSL (without relying on Three.js); 5000+ particles forming a spiral-arm structure; motion driven by simplex/perlin noise; support for rotating the viewpoint via mouse drag; particle colors gradating from a blue-purple core to a gold outer edge; twinkling star points; a dark space background. Include a brief technical note at the bottom of the page.", "skills": [ { "id": "shader-dev", "name": "GLSL Shader开发", "description": "GLSL Shader 视觉效果开发", "skill_id": "skill_2053082684416987136" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 108, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T20:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ai-agent-product-handbook", "title": "AI Agent 产品设计实战手册", "subtitle": "从零到一，用 AI 写一本 5 章 3 万字的产品设计手册", "description": "想体系化沉淀产品方法论，但一想到要写 3 万字就犯怵。这个案例展示了如何用 FBS-BookWriter（福帮手）从一句话主题出发，经过联网调研、大纲生成、逐章写作、分层审校，最终产出一本 5 章结构化的「AI Agent 产品设计实战手册」。\n\n手册覆盖了 Agent 产品的核心设计要素：从用户意图理解到多轮对话设计、从 Skill/插件生态构建到安全合规、从 MVP 验证到规模化运营。每章不是泛泛而谈，而是给出了具体的设计框架、真实案例拆解和可复用的 Checklist。整本手册自动完成了事实核查、引用标注和中文排版。\n\n适合：产品经理、AI 创业者、技术团队负责人、产品方法论沉淀者、想出书但没时间写的人。", "categories": [ "productivity" ], "tags": [ "写书", "手册", "AI Agent", "产品设计", "长文写作" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI Agent 产品设计实战手册（全文）" } ], "prompt": "请用福帮手帮我写一本《AI Agent 产品设计实战手册》。\n\n主题：面向产品经理和 AI 创业者的 Agent 产品设计方法论，从 0 到 1 覆盖设计、构建、验证、运营全链路。\n\n要求：\n1. 目标读者：有 1–3 年经验的产品经理、AI 方向创业者\n2. 全书约 5 章、3 万字左右\n3. 章节结构参考：\n - 第一章：AI Agent 产品的本质与设计范式\n - 第二章：用户意图理解与多轮对话设计\n - 第三章：Skill 生态与插件化架构设计\n - 第四章：安全合规与边界设计\n - 第五章：从 MVP 到规模化运营\n4. 每章包含核心概念框架、真实产品案例拆解、可复用的 Checklist 或决策矩阵、本章小结与行动建议\n5. 写作风格：专业但不学术，像资深 PM 做团队分享\n6. 联网查证关键数据\n7. 输出 Markdown 格式，中文排版规范", "prompt_en": "Please use FBS-BookWriter to write a book titled \"AI Agent Product Design in Practice\".\n\nTheme: an Agent product design methodology for product managers and AI entrepreneurs, covering the full 0-to-1 lifecycle — design, building, validation, operations.\n\nRequirements:\n1. Target readers: product managers with 1-3 years of experience, and AI-focused entrepreneurs\n2. About 5 chapters, roughly 30,000 words\n3. Reference chapter structure:\n - Chapter 1: The Essence and Design Paradigms of AI Agent Products\n - Chapter 2: Understanding User Intent and Designing Multi-turn Dialogue\n - Chapter 3: Skill Ecosystem and Pluggable Architecture Design\n - Chapter 4: Security, Compliance, and Boundary Design\n - Chapter 5: From MVP to Scaled Operations\n4. Each chapter includes a core conceptual framework, real product-case breakdowns, a reusable checklist or decision matrix, and a chapter summary with action recommendations\n5. Writing style: professional but not academic, like a senior PM giving an internal team talk\n6. Verify key data online\n7. Output in Markdown, following Chinese typesetting conventions", "skills": [ { "id": "fbs-bookwriter", "name": "fbs-bookwriter", "description": "福帮手出品 | 高质量长文档手稿工具链：书籍、手册、白皮书、行业指南、长篇报道、深度专题；支持联网查证（宿主允许时启用，离线自动降级）、S/P/C/B 分层审校、中文排版与 MD/HTML 交付。触发词：福帮手、福帮手写书skill、福帮手写书、写书、出书、写长篇、写手册、写白皮书、写行业指南、协作写书、定大纲、写章节、封面、插图、排版构建、导出、去AI味、质量自检、图文书、写报道、写深度稿、写特稿、写专题、写调查报道、写长文、拆书改写、海外本地化改写、爆款结构改写、激活原料、原料盘点、整理素材", "description_en": "Dual-channel long-form writing workflow for books, manuals, whitepapers, guides and reports with layered QC, 7-lock anti-drift anchors, and offline fallback.", "skill_id": "skill_2053083113738002432" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 109, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-12T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "AI Agent Product Design Handbook", "subtitle_en": "From zero to one — use AI to write a 5-chapter, 30,000-word product design handbook", "description_en": "Want to systematically document product methodology, but dread writing 30,000 words? This case shows how FBS-BookWriter goes from a one-line topic through web research, outline generation, chapter-by-chapter writing, and multi-layer review to produce a 5-chapter \"AI Agent Product Design Handbook.\"\n\nThe handbook covers core Agent product design elements: from user intent understanding and context management to tool orchestration and multi-agent collaboration, each chapter supported with real-world examples.\n\nBest for: Product managers, technical leads, anyone who needs to write long-form structured content.", "tags_en": [ "Book Writing", "Handbook", "AI Agent", "Product Design", "Long-form Writing" ], "is_recommended": false, "recommended_order": 0 }, { "id": "ancient-sweet-romance-drama", "title": "古风甜宠短剧", "title_en": "Ancient Costume Sweet Romance Drama", "subtitle": "用身份悬念和高密度心动瞬间完成一场雨夜初遇", "subtitle_en": "A rainy first meeting built on hidden identities and concentrated romantic beats", "description": "古风甜宠容易只剩造型展示，缺少人设反差、身份悬念和有效的情感推进。\n\n案例让药铺少女在夜街救下受伤的黑甲将军；他表面冷硬却怕苦药，她看似莽撞却认出他的秘密令牌，两人的距离在包扎与试探中迅速拉近。\n\n适合：古装甜宠、角色 CP 预告、短剧投流片和以高密度情绪点吸引观众的内容。", "description_en": "Costume romance can become a wardrobe showcase when it lacks character contrast, hidden identity, and emotional movement.\n\nA medicine-shop girl rescues an injured general in black armor on a rainy street. His fear of bitter medicine and her recognition of his secret token bring them rapidly closer.\n\nBest for costume romance, couple trailers, drama acquisition clips, and content built around concentrated emotional beats.", "categories": [ "creation" ], "tags": [ "古风甜宠", "雨夜初遇", "身份悬念" ], "tags_en": [ "costume romance", "rainy encounter", "hidden identity" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「动漫短剧创作」专家调用 Flova 的「古风甜宠短剧」Skill，制作一支约 30 秒、16:9 的雨夜初遇短片《长街救下冷面将军》。药铺少女在收摊后的古城长街发现一名受伤的黑甲男子，冒雨扶他进屋并替他包扎；男子表面冷硬，却因怕苦药露出不自然的迟疑。少女在他腰间看见皇城禁军密令，但装作不知，笑着说“药不苦，秘密才苦”。请先呈现完整剧本让我审核，再确认男女主形象、朝代氛围、画幅、背景音乐与字幕偏好；随后依次确认角色图、故事板和空场景图，再生成镜头。保持高冷与怕苦、莽撞与细腻两组反差，至少安排包扎触手、近距离对视和令牌试探三个心动/悬念点，并保证服装、伤口、令牌和雨势连续。", "prompt_en": "Ask the “Flova Animation and Short Drama Director” expert to invoke Flova’s “Ancient Costume Sweet Romance Drama” Skill and create a 16:9 rainy-night first-encounter short of about 30 seconds titled Rescuing the Cold General on the Night Street. After closing her medicine shop, a young woman finds an injured man in black armor on an old city street, helps him inside through the rain, and bandages him. He appears cold and severe but hesitates awkwardly because he hates bitter medicine. She notices a secret imperial-guard token at his waist, pretends not to recognize it, and smiles: “Medicine is not bitter. Secrets are.” Present the complete script for approval first, then confirm the leads’ appearance, period atmosphere, aspect ratio, music, and subtitle preferences. Obtain approval for character images, storyboard, and empty location images in sequence before generating shots. Preserve the contrasts of coldness versus fear of medicine and recklessness versus sensitivity. Include three clear romantic or suspense beats: hands touching during bandaging, a close gaze, and a test involving the token. Keep costumes, wound, token, and rain continuous.", "skills": [], "experts": [ { "id": "flova-anime-short-drama-director", "name": "动漫短剧创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及连续短剧成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "6afa1d59a44f", "expert_id": "ex_jh9eA3mhVTje" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-limitboard-morning-brief", "title": "涨停晨会简报", "title_en": "Morning Limit-Up Brief", "subtitle": "把收盘后的涨跌停线索整理成一份晨会可直接翻阅的编辑部版面。", "subtitle_en": "A newsroom-style morning brief that turns limit-up signals into a ready-to-use meeting handout.", "description": "收盘后最头疼的，不是知道今天有多少涨停，而是怎么把这些线索在第二天晨会上讲清楚。这个案例把涨停家数、连板高度、主线板块、封单集中度和全量清单整理成一份编辑部版式的晨会简报，领导和研究员都能直接翻阅。\n\n你能看到一份真正可交付的双语 HTML：首屏先给结论，正文再展开指数、板块、连板梯队和全量涨停明细，适合在内部晨会、复盘邮件或投研群里直接转发。\n\n适合：卖方/买方晨会、短线复盘团队、投资经理助理、需要做市场日报的运营同学。", "description_en": "The hardest part after the close is not counting limit-ups, but explaining the session clearly at the next morning meeting. This case turns limit-up breadth, ladder height, leading sectors, seal concentration, and the full stock list into a newsroom-style brief that both portfolio managers and researchers can scan quickly.\n\nIt demonstrates a delivery-grade bilingual HTML report: conclusions first on the opening screen, then index moves, sector leadership, ladder details, and the complete limit-up list for easy reuse in internal meetings, recap emails, or research channels.\n\nBest for: sell-side or buy-side morning meetings, short-term recap desks, PM assistants, and operators who produce daily market notes.", "categories": [ "finance", "data", "search" ], "tags": [ "涨停复盘", "晨会简报", "板块主线", "多语HTML" ], "tags_en": [ "limit-up recap", "morning brief", "sector leadership", "bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于最新一个有完整收盘数据的 A 股交易日，生成一份适合晨会直接使用的涨跌停复盘 HTML。用真实查询到的涨停/跌停、连板梯队、板型分布、主线板块、封单集中度和全量涨停明细，先给出 3-5 条结论，再展开数据。整体做成明亮、纸感明显、像晨报头版与内页拼版的编辑部风格；单文件 HTML、资源全内联、可离线打开。请在页脚注明数据日期与来源口径，并保留“仅供复盘参考，不构成投资建议”的声明。", "prompt_en": "Using the latest A-share trading day with complete closing data, generate a limit-up/down recap HTML suitable for a morning meeting. Query real data for limit-ups, limit-downs, consecutive-board ladders, board distribution, leading sectors, seal concentration, and the full limit-up list. Lead with 3-5 conclusions, then expand into the evidence. Keep the page bright, editorial, and clearly newspaper-like, with the feel of a front page plus inner spreads; deliver a single self-contained HTML file that works offline. State the data date and source notes in the footer, and keep the disclaimer that the page is for recap purposes only and not investment advice.", "artifacts": [ { "file": "output.html", "label": "双语交付 HTML 报告" } ], "skills": [ { "id": "a-share-limitboard-report", "name": "A股涨跌停日报", "description": "A股涨跌停日报，收盘后汇总涨停/跌停、连板梯队、板型分布、主线板块、炸板与全量列表。", "skill_id": "skill_2076627892488605696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T21:22:40.164728+08:00", "updated_at": "2026-07-30T16:43:23.113221+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-sentiment-cycle-atlas", "title": "A股情绪周期图谱", "title_en": "A-share Sentiment Cycle Atlas", "subtitle": "把赚钱效应、连板高度和主线扩散放到一张图里", "subtitle_en": "Put profitability breadth, ladder height and theme diffusion onto one canvas", "description": "做情绪复盘最怕两件事：一是看见满屏红盘就误以为主升来了，二是只盯涨停数量，却忽略了主线是否成型、量能是否跟上。结果往往是最热的时候冲进去，第二天才发现只是修复日。\n\n这个案例演示了如何把 A 股收盘后的真实行情、连板天梯、板块涨幅与龙虎榜资金整合成一份情绪周期图谱：先给情绪温度计，再拆赚钱效应、主线扩散、机构净买入和空仓检验，最后落到次日策略提醒。\n\n适合：偏短线的活跃投资者、每天要做盘后复盘的交易员、希望快速判断“能不能追高”的市场观察者。", "description_en": "Two mistakes dominate sentiment reviews: reading a sea of green as an automatic breakout, and focusing only on limit-up counts while ignoring whether a real mainline exists and whether volume actually confirmed it. That is how many traders chase the hottest tape only to find out it was merely a repair day.\n\nThis case shows how to turn real A-share closing data, ladder rankings, theme moves and Dragon-Tiger capital flows into a sentiment-cycle atlas: start with a thermometer, then break down profitability breadth, theme diffusion, institutional net buying and the empty-hand test before ending with next-session strategy notes.\n\nIdeal for active short-term investors, traders who do post-close reviews every day, and market observers who want a fast answer to whether the tape is still chaseable.", "categories": [ "finance", "data" ], "tags": [ "A股复盘", "情绪周期", "连板梯队", "主线题材", "双语HTML" ], "tags_en": [ "A-share Review", "Sentiment Cycle", "Limit-up Ladder", "Mainline Theme", "Bilingual HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "情绪周期分析 HTML报告" } ], "prompt": "请基于指定交易日做一份 A 股情绪周期分析单文件 HTML。先判断当天是否为交易日；若休市直接说明。若有交易数据，请围绕赚钱效应、连板高度、主线题材、机构净买入和空仓检验，输出情绪温度计、主线辨识、样本梯队、风险提示与次日策略提醒。所有资源内联、可离线打开，数据使用真实公开行情并注明统计日期。", "prompt_en": "Create a single-file HTML A-share sentiment-cycle analysis for a specified trade date. First confirm whether the date is a trading session; if the market is closed, state that directly. If data exists, focus on profitability breadth, ladder height, mainline themes, institutional net buying and the empty-hand test, and output a sentiment thermometer, mainline reading, representative ladder names, risk notes and next-session strategy reminders. Fully self-contained for offline use, and based on real public market data with the statistics date shown clearly.", "skills": [ { "id": "a-share-daily-review", "name": "A股每日复盘", "description": "A股每日收盘后全维度复盘报告生成，支持五种分析风格，输出自包含HTML报告。", "skill_id": "skill_2076628582742429696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:12:45+08:00", "updated_at": "2026-07-27T20:12:45+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "cashflow-model-dashboard", "title": "现金与商业模型检查台", "title_en": "Cashflow & Business Model Checkpoint", "subtitle": "收入增长前，先看结构能否让公司活下去", "subtitle_en": "Before Chasing Revenue, Test Whether the Structure Can Sustain You", "categories": [ "business", "education" ], "tags": [ "现金流", "商业模型", "资源约束", "创业财务" ], "tags_en": [ "Cashflow", "Business Model", "Resource Constraints", "Startup Finance" ], "description": "这是一家企业培训服务创业公司的经营检查台：收入在增长，但每新增客户都伴随新的课程定制与交付协调，现金边界只剩 5 个月。案例不把订单增长等同于结构改善，而是先把财务放回经营仪表盘的位置。\n\n检查台围绕现金、交付、定价与复用三层展开。它把商业模型视为对时间、现金与创始人精力的约束，要求先辨别新增订单带来的是能力复用，还是同步放大人力、管理与交付压力。\n\n产物进一步给出暂停与继续验证的信号，并以具体反问帮助创始人回到自己的位置：先稳定生存结构，再决定是否扩大收入。", "description_en": "This checkpoint examines a corporate-training startup whose revenue is rising while every additional client brings new course customization and delivery coordination, with only 5 months of cash coverage. It does not equate order growth with structural improvement; it first restores finance to its role as the operating dashboard.\n\nThe checkpoint reviews three layers: cash, delivery, and pricing with reuse. It treats the business model as a constraint on time, cash, and founder attention, asking whether additional orders create reusable capability or simultaneously amplify people, management, and delivery pressure.\n\nThe artifact then presents signals to pause or continue validating, and uses specific questions to return the founder to the right position: stabilize the survival structure before deciding whether to expand revenue.", "prompt": "一家提供企业培训服务的创业公司，收入持续上升，但每增加一个客户都需要新增课程定制与交付协调；账上现金只够 5 个月。销售希望以更低价格换取更多订单，创始人犹豫是否继续扩张。请不要把这理解为单纯的销售或执行问题，而要诊断真正的顺序问题：财务应作为经营仪表盘，商业模型应约束有限的时间、现金与创始人精力，规模未必能解决结构性亏损。围绕现金、交付、定价与复用三层做检查，说明哪些信号意味着应暂停并校准，哪些信号仅支持继续验证。不要虚构外部事实或新增数值。每个模块都用一句核心判断开头，给出具体解释，并以 1–2 个贴近创始人处境的反问收尾，帮助其先稳定生存结构，再决定是否扩大收入。", "prompt_en": "A startup that provides corporate training services is seeing revenue rise, yet every additional client requires new course customization and delivery coordination. Cash on hand covers only 5 months. Sales wants to trade lower prices for more orders, while the founder is unsure whether to keep expanding. Do not treat this as only a sales or execution problem. Diagnose the underlying sequencing problem: finance should act as the operating dashboard, and the business model should constrain limited time, cash, and founder attention; scale may not solve structural losses. Review three layers—cash, delivery, and pricing with reuse—and explain which signals call for pausing and recalibration and which signals only support continued validation. Do not invent external facts or additional numbers. Begin every module with one core judgment, follow it with concrete explanation, and close with one or two questions that fit the founder’s situation, helping them stabilize the survival structure before deciding whether to expand revenue.", "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "中英双语现金与商业模型检查台" } ], "skills": [ { "id": "the-entrepreneurship-handbook", "name": "创业可以学", "description": "服务创业者和管理者，解答创业/商业/管理问题，引发深度思考", "skill_id": "skill_2053083248287678464" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 97, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-07-27T16:30:16+08:00", "updated_at": "2026-07-27T16:30:16+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "content-asset-multiplatform-pack", "title": "内容资产四平台分发包", "title_en": "Four-Platform Content Asset Pack", "subtitle": "同一核心观点，写成四种平台原生表达", "subtitle_en": "One core idea, rewritten in four platform-native voices", "description": "很多内容团队把“一稿多发”理解成改字数：公众号删短一点发小红书，再把开头读出来发抖音，结果每个平台都不像原生内容。\n\n这个案例以“小团队如何停止追热点、开始积累内容资产”为原稿，使用一键改稿多平台分发 Skill 生成视频号、小红书、抖音和公众号四套完整版本。交付包含核心观点拆解、每平台三组标题、正文、标签、配图比例、分镜与发布前合规自检。\n\n适合：新媒体运营、品牌内容团队、个人 IP、需要降低跨平台分发成本的小型营销团队。", "description_en": "Many content teams interpret multi-platform publishing as changing length: shorten the newsletter for Xiaohongshu, then read the opening aloud for Douyin. The result feels native nowhere.\n\nThis case starts with an original article about helping a small team stop chasing trends and build reusable content assets. The Multi-Platform Content Rewrite Skill produces complete versions for WeChat Channels, Xiaohongshu, Douyin, and WeChat Official Accounts, including core-message extraction, three title options per platform, body copy, tags, image ratios, shot guidance, and a compliance checklist.\n\nIdeal for: social-media operators, brand content teams, personal brands, and small marketing teams reducing cross-platform distribution effort.", "categories": [ "creation" ], "tags": [ "内容运营", "多平台分发", "一稿多写", "平台适配" ], "tags_en": [ "content operations", "multi-platform publishing", "content repurposing", "platform adaptation" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "四平台内容分发报告" } ], "prompt": "请把我提供的原稿改写成视频号、小红书、抖音和公众号四个平台可直接发布的版本。主题是：三人内容团队不应每天追热点，而应把选题、论据、案例和用户问题沉淀成可复用的内容资产。原稿包含一个匿名案例：团队连续 30 天日更但复用率很低，后来建立选题母表、论据卡、案例库和用户问题库，每周只新增两个核心观点，并围绕同一观点制作长文、清单、短视频和社交讨论版本。受众是 3—10 人的新媒体与品牌内容团队，调性理性、务实，不夸大效果。四个平台的核心观点必须一致，但标题、钩子、结构、CTA 和配图建议要原生化；小红书与抖音禁止微信号、私信和诱导互动话术。", "prompt_en": "Rewrite the supplied source article into ready-to-publish versions for WeChat Channels, Xiaohongshu, Douyin, and WeChat Official Accounts. The topic is that a three-person content team should stop chasing daily trends and instead turn topics, evidence, cases, and customer questions into reusable content assets. The source includes an anonymous case: after 30 days of daily posting with little reuse, the team built a topic master sheet, evidence cards, a case library, and a customer-question library. Each week it added only two core ideas, then turned each into a long article, checklist, short video, and social discussion version. The audience is social-media and brand-content teams of 3–10 people. Keep the tone rational and practical without exaggerated results. Preserve the same core thesis across platforms while making titles, hooks, structure, CTA, and image guidance native to each platform. Xiaohongshu and Douyin must not include WeChat IDs, private-message prompts, or engagement bait.", "skills": [ { "id": "lingyi-one-click-rewrite-multi-publish", "name": "一键改稿多平台分发", "description": "【零一数科·出品】把一篇原稿按四平台调性差异化改写为可直接发布的分版本，并做合规自检。", "slug": "lingyi-one-click-rewrite-multi-publish" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-29T08:00:00Z", "updated_at": "2026-08-31T16:31:11+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "efunds-weekly-hotspot", "title": "热点周报", "title_en": "Weekly Hotspot Report", "subtitle": "速览市场行情与热门资讯", "subtitle_en": "Market moves and trending news at a glance", "description": "告别手工周报耗时难题！一键抓取「易方达基金」公众号本周推文，智能生成 TOP5 市场热点；行情素材取自官方资讯栏目，热点附带原文引文与跳转链接，自动关联基金产品，一键导出 HTML 周报。\n适配理财顾问、投资者等需高效产出周度市场简报的人群。", "description_en": "Tired of spending hours on manual weekly reports? One click pulls this week's posts from the E Fund official WeChat account and intelligently generates a TOP5 of market hotspots. Market material is drawn from official information columns; every hotspot comes with a verbatim quote and a jump link, auto-matched to the relevant fund products, and exports an HTML weekly report in one click.\nBuilt for wealth advisors, investors, and anyone who needs to produce a weekly market briefing efficiently.", "categories": [ "finance" ], "tags": [ "热点周报", "投资理财", "市场周报", "基金", "易方达" ], "tags_en": [ "Weekly Hotspots", "Investment", "Market Report", "Fund", "E Fund" ], "artifact_type": "html", "preview": "efunds-weekly-hotspot.html", "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "用「易方达热点梳理」技能出一份本周热点周报 HTML：抓「易方达基金」「易方达指数通」两官号本周文章，归并 TOP5 热点（全球指数配图取自原文），每个热点配原文引文＋可点击来源＋关联易方达基金；行情概览用官方周更栏目；过敏感词，数据真实可溯源。", "prompt_en": "Use the \"易方达热点梳理\" skill to produce this week's hotspot weekly report as an HTML page: pull this week's posts from both E Fund official accounts, cluster a TOP5 of hotspots (global index chart from the original article), attach verbatim quotes with clickable sources and matched E Fund products; use E Fund's official weekly columns for the market overview; run the sensitive-word check and keep all data real and traceable.", "skills": [ { "id": "efunds-hotspot-combing", "name": "易方达热点梳理", "description": "易方达基金「热点梳理」总入口：基于 efunds-mcp 检索「易方达基金」「易方达指数通」公众号公开文章，覆盖事件热点梳理/近期热点速览/热点周报三类成稿，AI 只做资讯与事实梳理，市场观点一律原文引用。", "skill_id": "skill_2092287909491437568" } ], "experts": [], "mcps": [ { "id": "efunds", "name": "易方达基金", "description": "易方达基金公众号文章检索与正文读取（search_wx_articles / get_wx_article_detail）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T11:12:00+08:00", "updated_at": "2026-08-31T17:07:35+08:00", "artifacts": [ { "file": "efunds-weekly-hotspot.html", "label": "本周热点周报（HTML 聚合报告）" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "health-report-family-guardian", "title": "远程帮父母看化验单", "title_en": "Read Your Parents' Lab Results Remotely", "subtitle": "父母微信发来一张化验单照片，隔着几百公里也能说明白", "subtitle_en": "A photo of a lab report from your parents — explain it clearly from hundreds of miles away", "description": "在外地工作，父母去医院查完血，随手拍张化验单照片发微信过来，配一句「你看看有没有事」。照片歪的、有反光、术语一个不认识，只能干着急，要么就是全家一起在网上乱搜，越搜越慌。\n\n这个案例演示把一张手机拍的血常规/生化化验单照片交给 AI：先从照片里把指标逐项识别出来，标出偏高偏低，然后按「哪些其实不用担心 / 哪些建议复查 / 哪些要尽快去医院」三档分级，用能直接转发给父母看的口语化说法解释清楚，并给出一段可以直接发到家庭群的简明结论，附带该挂什么科、需要带哪些资料。\n\n适合：异地照护父母的子女、家庭健康管理场景、社区与基层医疗随访、老年人健康服务机构、企业员工家庭健康福利。", "description_en": "You work in another city. Your parents get a blood test, snap a photo of the lab report, and send it over WeChat with \"take a look, is anything wrong?\" The photo is crooked and glary, the terms are unfamiliar, and the whole family ends up googling and panicking.\n\nThis case shows how to hand a phone photo of a CBC/biochemistry lab report to AI: it recognizes each indicator from the image, flags highs and lows, then sorts findings into three tiers — nothing to worry about / worth a recheck / see a doctor soon — explaining each in conversational wording you can forward straight to your parents. It also produces a short conclusion ready to paste into the family chat, plus which department to book and what documents to bring.\n\nGood for: adult children caring for parents remotely, family health management, community and primary-care follow-up, elderly care service providers, and corporate family health benefits.", "categories": [ "education", "productivity" ], "tags": [ "化验单", "家庭健康", "异地照护", "指标异常", "科普解读" ], "tags_en": [ "lab report", "family health", "remote caregiving", "abnormal indicators", "plain-language explanation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我父母刚做完血检，微信发来结果：血红蛋白 103 g/L（参考 115–150）、血清铁偏低、铁蛋白 8 ng/mL（参考 15–200）、血肌酐 156 μmol/L（参考 44–106）、eGFR 42（参考 >90），血脂和肝功能正常。我在外地，想弄明白哪些要紧、怎么跟他们解释、要不要催他们去医院。", "prompt_en": "My parents just had blood work and sent me the results: hemoglobin 103 g/L (ref 115–150), low serum iron, ferritin 8 ng/mL (ref 15–200), creatinine 156 μmol/L (ref 44–106), eGFR 42 (ref >90); lipids and liver function normal. I live far away and want to know what's urgent, how to explain it to them, and whether to push them to see a doctor.", "skills": [ { "id": "tencent-medical-report-interpreter", "name": "腾讯健康-AI就医助手-医疗报告解读", "description": "对接 ADP LLMReportInterpretation API，解读检验单/检查单/体检报告，识别异常值并用通俗语言解析医学术语", "slug": "tencent-medical-report-interpreter" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 95, "usage_count": 0, "created_at": "2026-08-22T16:09:15Z", "updated_at": "2026-08-22T16:36:30Z", "artifacts": [ { "file": "output.html", "label": "父母化验单解读与家庭沟通版" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "industry-long-short-signal", "title": "行业多空风向标", "title_en": "Industry Long/Short Signal", "subtitle": "按有效事件窗口拆解行业方向，把核心信号、雷达因子和来源放在一页", "subtitle_en": "Split an industry view by valid event windows with core signals, six-factor bars, and reviewable sources.", "description": "同一个行业在 3 日、5 日、7 日和 20 日窗口里的反应可能不同。只给一句“看多/看空”，用户很难知道该相信哪个验证窗口；缺少长期样本时，更不能为了版式硬凑长期结论。\n\n这个案例演示如何用同舟股市投研专家，把行业要闻、行情数据、产业图谱、供需位置、风险舆情、同舟观点和历史事件统计整理成一页多空风向标，只展示有有效样本的事件窗口，并给出方向、核心信号、失败条件和可复核来源。六维因子用横向进度条和明细卡呈现，方便用户看出强弱结构和每个分数的依据。\n\n适合：金融内容团队、投研分析师、投顾、产品经理，以及想快速理解行业风向但不想看长报告的用户。", "description_en": "The same industry can behave differently across 3-day, 5-day, 7-day, and 20-day event windows. A single bullish or bearish sentence does not tell users which window to trust; when long-horizon samples are missing, the page should not force a long-term conclusion just for layout symmetry.\n\nThis case shows how Tongzhou Equity Research Expert turns industry news, market data, industry graph context, supply-demand position, risk sentiment, Tongzhou viewpoints, and historical event statistics into a one-page long/short signal board, showing only valid event windows with direction, core signals, failure conditions, and reviewable sources. Six factors are shown as horizontal bars with detail cards so users can see both the structure and the evidence behind each score.\n\nSuitable for: financial content teams, equity researchers, advisors, product managers, and users who want to understand sector direction without reading a long report.", "categories": [ "finance", "data", "search" ], "tags": [ "行业多空", "数据看板", "六维因子", "HTML" ], "tags_en": [ "Sector signal", "Dashboard", "Six-factor bars", "HTML" ], "artifact_type": "html", "playbook_type": "industry-windvane", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "行业多空风向标" } ], "prompt": "请调用同舟股市投研专家，围绕「{行业或主题，例如：锂电池行业}」生成一页单文件 HTML《行业多空风向标》。\n\n先完成同舟 MCP Gateway 认证；如果没有 API Key 或认证失败，只提示用户去控制台申请/绑定，不要继续编造内容。请动态解析用户输入的行业或主题，不要写死示例行业、行业编号或历史样例参数。\n\n请优先调用同舟 MCP 获取可复核材料，包括行业要闻、公告或研报、行情/估值、行业图谱公开因子、同舟观点、历史相似事件统计和相关市场数据。页面使用专家内置的「行业多空风向标」HTML playbook 生成；具体版式、证据解读、六维因子、情景矩阵、新闻情绪、源头复核、数据口径和安全边界以该 playbook/reference 为准，不要临时发明新模板。 保留源头复核入口；没有文章级链接的数据只标注来源类型，不做假链接；页面导航不要使用 href=\"#section-id\"；统计数据有则展示真实样本，没有则标注样本不足。若 60 日或其他长期窗口没有有效事件样本，请从期限拆解、综合判断和情景矩阵中跳过，只在数据口径中说明缺口，不要用宽基指数或示例数字替代。页面只展示用户能理解的证据强弱说明，不输出模型化拆分。\n\n输出一个 HTML 文件即可。允许源头复核外链，但不要加载外部脚本、CSS、图片或字体；不承诺收益，不读取个人持仓或交易历史。", "prompt_en": "Use Tongzhou Equity Research Expert to create a single-file HTML Industry Long/Short Signal page for \"{industry or theme, e.g. lithium battery industry}\".\n\nFirst complete the Tongzhou MCP Gateway credential check. If the API key is missing or invalid, only ask the user to apply or bind a key in the console and do not invent content. Dynamically resolve the industry or theme from the user's input; do not hard-code sample sectors, industry IDs, or historical demo parameters.\n\nUse Tongzhou MCP to collect reviewable evidence first, including industry news, announcements or reports, market/valuation data, public industry-graph factors, Tongzhou viewpoints, historical similar-event statistics, and related market data. Generate the page with the expert's built-in Industry Long/Short Signal HTML playbook. Layout, evidence interpretation, six-factor display, scenario matrix, news sentiment, source review, methodology, and safety boundaries should follow that playbook/reference; do not invent a new template in the prompt. Keep source-review links when available; data without article-level links should only show source type; do not use href=\"#section-id\" navigation; show real statistics when available and mark insufficient samples when they are not. If 60-day or other long-horizon windows have no valid event samples, skip them from the horizon cards, overall view, and scenario matrix, and explain the gap only in methodology. Do not substitute broad-index numbers or demo values. Use user-facing evidence-strength explanations rather than model-like decomposition.\n\nOutput one HTML file only. Source-review external links are allowed, but do not load external scripts, CSS, images, or fonts. Do not promise returns. Do not read personal holdings or trade history.", "skills": [], "experts": [], "mcps": [ { "id": "public-mcp-gateway", "name": "同舟公开 MCP Gateway", "description": "受控暴露 fin-data-query、doc-search、fin-graph、same-boat 等公开投研能力。" } ], "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-07-03T00:00:00+08:00", "updated_at": "2026-07-27T11:27:39+08:00", "source_links_supported": true, "is_recommended": false, "recommended_order": 0 }, { "id": "kdocs-sheet-governance-studio", "title": "表格治理控制台", "title_en": "Sheet Governance Studio", "subtitle": "把销售跟进表变成持续可维护的规则面板", "subtitle_en": "Turn a shared sales tracker into a governed operating surface", "description": "销售跟进表最难的不是第一次搭出来，而是团队一起录入后很快失控：阶段叫法不统一、手机号和日期格式混乱、关键字段经常空着，月底统计时总要重新洗表。\n\n这个案例把金山文档的表格读取、批量更新、条件格式、数据校验和区域保护能力串成一条治理路径：先识别现有问题，再统一字段结构，给枚举列补下拉规则，为异常值加高亮提醒，并在需要时锁定表头和公式区域。页面同时保留规则设计、前后对比和复用 prompt，便于团队直接照着落地。\n\n适合：销售运营、CRM 管理员、区域销售负责人，以及所有需要多人维护共享表格、又希望把规则长期固化下来的业务团队。", "description_en": "The hardest part of a sales follow-up sheet is not building the first version. It starts to drift after the whole team edits it together: stage names diverge, phone and date formats become inconsistent, key fields are left blank, and month-end reporting turns into another cleanup cycle.\n\nThis case connects Kdocs sheet reading, batch updates, conditional formatting, data validation, and protection ranges into one governance flow: identify the current issues first, standardize the field structure, add dropdown rules to enum columns, highlight exceptions, and lock headers or formula areas when needed. The page also preserves the rule design, before/after comparisons, and a reusable prompt so teams can apply the same pattern directly.\n\nBest for: sales operations, CRM owners, regional sales leads, and any business team that maintains shared sheets together and wants the rules to remain stable over time.", "categories": [ "office", "productivity" ], "tags": [ "销售运营", "表格治理", "数据校验", "异常高亮", "区域保护" ], "tags_en": [ "Sales Ops", "Sheet Governance", "Data Validation", "Exception Highlighting", "Protection Ranges" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "external_url_en": "", "prompt": "请把我给你的销售跟进表整理成一张可持续维护的规范表。先定位目标文件并读取工作表信息与当前数据，识别重复字段、空白必填项和不一致值；再统一字段名与列顺序，补齐核心字段。对阶段、来源、区域等枚举列先检查现有数据校验，再创建或更新下拉规则；对重复客户名、空白下次跟进日期、超 7 天未跟进和超阈值金额等异常，先检查现有条件格式，再创建或更新高亮规则。若需要锁定表头或公式区域，请先给出建议范围并在我确认后执行。", "prompt_en": "Please turn the sales follow-up sheet I provide into a governed workspace that can be maintained over time. Start by locating the target file and reading the worksheet structure and current data, then identify duplicate fields, blank required values, and inconsistent entries. Next, standardize the field names and column order and complete the core columns. For enum columns such as stage, source, and region, check existing data validations first, then create or update dropdown rules. For issues such as duplicate customer names, blank next follow-up dates, no follow-up for more than seven days, and amounts above the threshold, check existing conditional formats first, then create or update highlight rules. If header rows or formula ranges should be locked, propose the ranges first and apply them only after I confirm.", "artifacts": [ { "file": "output.html", "label": "双语表格治理控制台" }, { "file": "source_snapshot.json", "label": "表格治理流程快照" } ], "skills": [ { "id": "kdocs", "name": "金山文档|WPS云文档", "description": "金山文档官方 Skill。对话即操作——知识一键存入、碎片内容整理、接龙转表格、文档转 Markdown、表格美化、收发表生成，全在一句话内完成。", "skill_id": "skill_2053082745828995072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T15:16:14+08:00", "updated_at": "2026-07-30T17:16:12+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "learning-progress-card", "title": "学习进度仪表板", "title_en": "Learning Progress Dashboard", "subtitle": "输入学习数据，生成精美的课程进度展示卡片", "subtitle_en": "Input learning data, generate elegant course progress dashboard cards", "description": "在线教育平台的学员后台、企业培训系统的进度看板、个人学习计划的可视化——都需要直观地展示学习进度和成就。但默认的进度条太简陋，学员看了没有成就感，也无法一目了然地掌握整体学习情况。\n\n这个案例演示了输入一个学习者的课程数据（在学课程、进度、成绩、学习时长、成就），AI 自动生成一个精美的学习进度仪表板 HTML 页面：居中卡片组布局（max-width 640px）、顶部总览统计（课程数/总时长/连续天数/证书数）、在学课程进度卡片（含进度环/课程名/进度百分比）、成就徽章区域。紫色学习主题，数据驱动的视觉设计。\n\n适合：在线教育平台、企业培训系统、个人学习计划管理、教育类产品 UI 参考。", "description_en": "Online education platform dashboards, corporate training progress boards, personal learning plan visualization — all need intuitive progress and achievement displays. But default progress bars are too plain — learners feel no sense of accomplishment and can't grasp their overall learning status at a glance.\n\nThis case demonstrates inputting a learner's course data (active courses, progress, grades, study hours, achievements), with AI generating an elegant learning progress dashboard HTML page: centered card group layout (max-width 640px), top overview stats (courses/total hours/streak days/certificates), active course progress cards (progress ring/course name/percentage), achievement badge area. Purple learning theme, data-driven visual design.\n\nBest for: Online education platforms, corporate training systems, personal learning plan management, education product UI reference.", "categories": [ "education", "data" ], "tags": [ "学习进度", "教育平台", "仪表盘", "数据可视化" ], "tags_en": [ "Learning Progress", "Education Platform", "Dashboard", "Data Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个学习进度仪表板页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，现代精美，展示学习者的总览数据、在学课程进度和成就徽章。\n\n请用下面这样的数据（可替换为你自己的）：学习者与平台、总览（在学/已完成课程数、总时长、连续天数、证书数）、在学课程（名称/进度/课时/最近学习/预计完成）、成就徽章。\n...", "prompt_en": "Generate a learning-progress dashboard page (single-file HTML, inline with no external dependencies), UI text in Simplified Chinese, modern and polished, showing the learner's overview stats, in-progress course progress, and achievement badges.\n\nUse data like the following (replace with your own): learner and platform; overview (in-progress/completed course counts, total hours, streak days, certificates); in-progress courses (name/progress/lessons/last studied/expected completion); achievement badges.\n...", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "学习进度卡片" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "lexiang-longform-to-kb", "title": "长文一键长成知识库", "title_en": "Turn a Long Doc into a Knowledge Base in One Click", "subtitle": "腾讯云知结构化导入：一篇杂乱长文 → 在线文档 + 多层级目录树", "subtitle_en": "Tencent Yunzhi structured import: a messy long doc into an online doc plus a multi-level tree", "description": "写好一篇长文档想沉淀进团队知识库，却得手动新建一堆页面、复制粘贴、逐个调排版，再一层层搭目录——又累又乱，一不小心标题层级和表格还全乱套。\n\n这个案例把一篇结构复杂的《企业级 AI 知识问答系统建设指南》一键导入腾讯云知：云知自动把 Markdown 转成保留完整排版（标题层级、表格、有序 / 无序列表）的在线文档，还能按章节拆解成 folder + 子页面的多层级目录树，直接形成团队知识库骨架；最后生成一份暗色科技感的成果展示页，含完整文档链接和拆解目录树。\n\n适合：知识运营、技术 writer、团队负责人、培训师等需要把长文沉淀成结构化知识库的人。", "description_en": "You've finished a long document and want to file it into the team knowledge base, but you end up creating a pile of pages by hand, copy-pasting, fixing formatting one by one, and building the folder tree layer by layer. It's tedious and messy, and heading levels or tables easily break along the way.\n\nThis case imports a complex 'Enterprise AI Q&A System Build Guide' into Tencent Yunzhi in one click: Yunzhi turns the Markdown into an online doc that preserves the full layout (heading levels, tables, ordered / unordered lists), then splits it by chapter into a multi-level tree of a folder plus sub-pages, forming the skeleton of a team knowledge base. Finally it generates a dark, tech-styled result page with the full-doc link and the split tree.\n\nGood for: knowledge ops, technical writers, team leads, trainers and anyone who needs to turn long docs into a structured knowledge base.", "categories": [ "productivity", "education" ], "tags": [ "文档导入", "知识库搭建", "结构化", "Markdown" ], "tags_en": [ "Doc Import", "Knowledge Base", "Structured", "Markdown" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "结构化导入成果展示" } ], "prompt": "帮我把下面这篇《企业级 AI 知识问答系统建设指南》长文用腾讯云知（乐享）一键导入知识库。\n\n[粘贴你的长文档]\n\n要求：\n1. 用云知真实执行导入（entry_import_content），把 Markdown 原文自动转成保留完整排版的在线文档——标题层级、表格、有序 / 无序列表全部保留，产出真实的在线文档页面链接和 entry_id。\n2. 再按章节把这篇长文拆解成一棵多层级目录树：新建一个 folder「知识问答系统建设（拆解目录）」，下挂若干子页面（如 01 背景与目标 / 02 整体架构 / 03 实施步骤），每个子页面都是真实创建的乐享页面，带真实链接。\n3. 最后生成一份单文件 HTML 成果展示页：暗色科技感、专业排版；顶部标题区放目标知识库名与链接；中间用「导入前 → 导入后」分栏对比体现一键转化；成果一展示完整在线文档卡片（标题 / entry_id / 可点击链接）；成果二用带层级缩进和连接线的树形结构可视化拆解目录树，每个节点显示名称和可点击链接；底部给价值总结。\n\n所有页面链接、entry_id 必须来自云知真实导入结果，严禁编造。样式方向：深色背景 + 浅色文字，科技蓝 / 紫点缀，图标用内联 SVG，不用任何外部资源。", "prompt_en": "Please import the following long document, 'Enterprise AI Q&A System Build Guide', into a knowledge base with Tencent Yunzhi (Lexiang) in one click.\n\n[paste your long document]\n\nRequirements:\n1. Actually run the import via Yunzhi (entry_import_content), turning the Markdown into an online doc that preserves the full layout — heading levels, tables, ordered / unordered lists — and produce a real online-doc page link and entry_id.\n2. Then split the long doc by chapter into a multi-level tree: create a folder 'Q&A System Build (Split Directory)' with several sub-pages (e.g. 01 Background & Goals / 02 Overall Architecture / 03 Implementation Steps), each a real Lexiang page with a real link.\n3. Finally generate a single-file HTML result page: dark, tech aesthetic, professional layout; a header with the target KB name and link; a 'before → after' two-column comparison showing the one-click transform; result one shows a full online-doc card (title / entry_id / clickable link); result two visualizes the split tree with indentation and connector lines, each node showing its name and clickable link; a value summary at the bottom.\n\nAll page links and entry_ids must come from the real Yunzhi import result — never fabricate them. Style direction: dark background with light text, tech blue / purple accents, inline SVG icons, no external resources.", "skills": [ { "id": "tencent-yunzhi", "name": "腾讯云知", "description": "腾讯云知（乐享）知识库专用操作——搜索、阅读、创建、编辑、上传", "skill_id": "skill_2053083480194342912" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-07-27T11:30:00+08:00", "updated_at": "2026-07-27T11:30:00+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "medication-safety-polypharmacy", "title": "老人多重用药风险排查", "title_en": "Polypharmacy Risk Screening for Seniors", "subtitle": "华法林、阿司匹林等慢病药叠加，新发不适一键归因", "subtitle_en": "Trace new symptoms to overlapping chronic medications like warfarin and aspirin", "description": "家里老人长期吃好几种慢病药，最近突然刷牙出血、身上淤青，到底是药吃多了还是别的问题，心里没底。\n\n这个案例演示了把「用药清单 + 新发症状」交给 AI 临床助手，它自动解析华法林、美托洛尔、二甲双胍、阿司匹林等药物，识别抗凝 / 抗血小板叠加的出血风险，并区分需立即就医的危险信号与可观察的轻微症状。\n\n适合：慢病老人家属、照顾老人的子女、多重用药患者、社区健康管理员、需要药物不良反应科普的用户。", "description_en": "An elderly family member takes several chronic medications and recently started bleeding gums and bruising. Is it the drugs, or something else? Hard to know.\n\nThis case shows how to hand a full medication list plus new symptoms to the AI clinical assistant, which parses drugs like warfarin, metoprolol, metformin and aspirin, flags the bleeding risk from overlapping anticoagulant / antiplatelet effects, and separates urgent warning signs from mild, watchable symptoms.\n\nFor: family members caring for elderly patients, children of aging parents, polypharmacy patients, community health workers, and anyone wanting plain-language guidance on drug side effects.", "categories": [ "agent", "education" ], "tags": [ "用药安全", "多重用药", "药物不良反应", "老年用药", "AI临床助手" ], "tags_en": [ "Medication Safety", "Polypharmacy", "Adverse Drug Reactions", "Geriatric Medication", "AI Clinical Assistant" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "老人多重用药风险排查演示（HTML，含中英双语）" } ], "prompt": "请以「腾讯健康-AI临床助手-用药安全」的能力，帮我排查一位老人的多重用药风险：\n\n我爸 72 岁，有高血压、房颤、糖尿病，每天服用华法林、美托洛尔、二甲双胍、阿司匹林。最近刷牙老出血，身上也有淤青。请判断：\n1. 出血倾向最可能由哪些药物叠加引起（重点：抗凝药 + 抗血小板药叠加）；\n2. 哪些症状是需立即就医的危险信号（黑便、呕血、血尿、持续出血、意识改变、胸痛），哪些是可观察的轻微症状；\n3. 下一步建议（就医查 PT/INR、就诊时带完整用药清单），并说明为什么不能擅自停用抗凝药。\n\n请严格区分「紧急征象」与「可观察症状」；不擅自建议停用或调整抗凝 / 抗血小板、降糖等慢病维持用药；不给出具体剂量调整数值；末尾建议尽快就医查凝血功能。", "prompt_en": "Using the medication-safety capabilities of the Tencent Health AI Clinical Assistant, help me screen polypharmacy risk for an elderly patient:\n\nMy dad is 72, with hypertension, atrial fibrillation and diabetes. He takes warfarin, metoprolol, metformin and aspirin daily. Lately his gums bleed and he bruises easily. Please assess:\n1. Which drug combination most likely explains the bleeding tendency (focus: anticoagulant + antiplatelet stacking);\n2. Which symptoms are danger signs needing immediate care (black stools, hematemesis, hematuria, persistent bleeding, altered consciousness, chest pain) versus mild watchable symptoms;\n3. Next steps (see a doctor for PT/INR, bring the full medication list) and explain why anticoagulants should not be stopped on your own.\n\nStrictly separate emergency signs from watchable symptoms. Do not recommend stopping or adjusting anticoagulant/antiplatelet or chronic glucose-lowering drugs on your own. Do not give specific dose-adjustment numbers. End with a recommendation to see a doctor promptly to check coagulation.", "skills": [ { "id": "medication-advisor", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "slug": "medication-advisor" } ], "experts": [ { "id": "chief-physician", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_0JLBYdNE0Hxk" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T11:01:00+08:00", "updated_at": "2026-08-27T00:01:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-hpc-expert-case-2", "title": "执行流程并生成分析结果报告", "title_en": "Execute Pipeline and Generate Analysis Report", "subtitle": "集群上跑起来，跑完拿到一份漂亮的 RNA-seq 结果报告", "subtitle_en": "Submit to cluster, run it, and get a polished RNA-seq analysis report", "description": "代码写好了，但怎么在真实集群上跑起来、跑完还能拿到一份可读的报告？这个案例演示了如何衔接上一步脚本，以普通用户身份投递 SLURM 作业，每 30 秒轮询状态到完成，并自动固化一份含 KPI 卡、reads 长度分布图、GC 曲线、转录本双轴柱状图与定量精度散点图（Pearson r 评分）的「RNA-seq 分析结果报告」HTML。\n\n适合：已完成流程代码设计、需要在 HPC 集群上实际运行 RNA-seq 并获取可视化结果报告的生信用户。", "description_en": "Got the scripts ready, but how do you actually run them on a real cluster and get a readable report? This case demonstrates how to follow up from the previous step, submit a SLURM job as a regular user, poll status every 30 seconds until completion, and automatically produce an 'RNA-seq Analysis Report' HTML with KPI cards, read length distribution, GC curve, transcript dual-axis bar chart, and a quantification accuracy scatter plot with Pearson r score.\n\nSuitable for: bioinformatics users who have finished pipeline code design and need to actually run RNA-seq on an HPC cluster and obtain visual result reports.", "categories": [ "dev" ], "tags": [ "HPC集群", "RNA-seq", "SLURM作业提交", "结果报告", "定量分析" ], "tags_en": [ "HPC Cluster", "RNA-seq", "SLURM Job Submission", "Result Report", "Quantification Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "RNA-seq 分析结果报告" } ], "prompt": "请把上一步生成的 RNA-seq 流程跑起来并输出结果报告。\n\n先自动发现资源：列出运行中集群，选第一个 SLURM 集群，确认 compute 队列与工作目录，确认脚本已就绪。然后以普通用户身份（先确认集群用户，不要用 root）把 run_rnaseq.sh 投递到 compute 队列，每 30 秒轮询作业状态直到完成（用 sacct 确认退出码为 0），取回 results.json 和生成的报告。\n\n最终固化输出一份「RNA-seq 分析结果报告」HTML，风格专业美观、自包含内联 SVG，包含：渐变蓝色 Hero + 顶部 KPI 卡（reads 总数/平均读长/平均 GC%/N 占比）+ 流程执行概览（数据模拟→QC→定量→精度评估四阶段+耗时）+ reads 长度分布直方图 + GC 含量分布曲线 + 转录本表达量双轴柱状图（reads 数 vs TPM）+ 定量精度散点图（设定丰度 vs 实测 TPM，含理想对角线和 Pearson r 环形评分）+ 定量结果明细表（长度/设定丰度/实测 reads/TPM/偏差，偏差着色）。", "prompt_en": "Please run the RNA-seq pipeline generated in the previous step and output a result report.\n\nFirst auto-discover resources: list running clusters, select the first SLURM cluster, confirm the compute queue and working directory, confirm scripts are ready. Then submit run_rnaseq.sh to the compute queue as a regular user (confirm cluster username first, do not use root), poll job status every 30 seconds until completion (use sacct to confirm exit code 0), retrieve results.json and the generated report.\n\nFinally produce an 'RNA-seq Analysis Report' HTML, professional and self-contained with inline SVG, including: gradient blue Hero + top KPI cards (total reads/average read length/average GC%/N ratio) + pipeline execution overview (data simulation→QC→quantification→accuracy evaluation, 4 stages with timing) + read length distribution histogram + GC content curve + transcript expression dual-axis bar chart (read count vs TPM) + quantification accuracy scatter plot (set abundance vs measured TPM, with ideal diagonal and Pearson r ring score) + quantification detail table (length/set abundance/measured reads/TPM/deviation, deviation color-coded).", "skills": [], "experts": [ { "id": "omics-hpc-expert", "name": "腾讯组学HPC集群运维与作业管理专家", "description": "远程组学HPC一站式运维，管控节点、队列、存储全生命周期，适配SLURM/SGE调度，自然语言交互降低门槛", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_oVOV90lFvBz5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-ori-expert-case-2", "title": "从头设计（De Novo Design）→ 结构验证", "title_en": "De Novo Protein Design → Structure Validation", "subtitle": "ORI 从头生成蛋白质候选序列，自动级联 USMFold 结构预测一体化呈现", "subtitle_en": "ORI de novo protein sequence generation with auto-cascaded USMFold structure prediction in one unified report", "description": "从头设计出满足功能约束的蛋白质序列并验证结构，历来是蛋白设计师最费力的环节。\n\n这个案例演示了通过 ORI 专家的 Generate Protein 子应用，一键执行蛋白质从头设计（目标：耐 60℃ 双功能酶），生成候选序列排名表，并自动级联 USMFold 对最优候选进行结构预测——候选序列和 3D 结构一体化呈现在 HTML 报告中，可交互、可下载 PDB 深度操作。\n\n适合：工业酶工程师、蛋白质设计研究员、合成生物学从业者。", "description_en": "Designing a protein sequence meeting functional constraints from scratch and then validating its structure has always been the most demanding step for protein designers.\n\nThis case demonstrates how to use the ORI expert's Generate Protein sub-application to run de novo protein design in one click (target: a thermostable bifunctional enzyme for 60°C), generate a ranked candidate sequence table, and automatically cascade USMFold to predict the top candidate's structure — sequences and 3D structure unified in a single interactive HTML report, with downloadable PDB for deep analysis.\n\nIdeal for: industrial enzyme engineers, protein design researchers, and synthetic biologists.", "categories": [ "data", "education" ], "tags": [ "蛋白质从头设计", "ORI", "序列生成", "结构验证", "USMFold级联" ], "tags_en": [ "De Novo Protein Design", "ORI", "Sequence Generation", "Structure Validation", "Cascaded USMFold" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "蛋白从头设计 + 结构预测一体化 HTML 报告" } ], "prompt": "Prompt 2 ｜ 从头设计（De Novo Design）→ 结构验证\n\n我是工业酶方向的蛋白设计师，目标是耐 60℃ 的双功能酶（几丁质酶 + 溶菌酶活性）。请使用 ORI 公共应用的 generate-protein 子流程，完全按默认参数（默认设计 prompt、模型与资源配置），执行一次蛋白质从头设计，并完成分析：\n\n1. 提交前给我任务确认单（子流程｜设计意图｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：候选序列表（排名｜名称｜长度｜score），渲染为 HTML 报告页；并将 score 最优候选自动级联 usmfold-predict（默认参数），其预测结构同样嵌入报告内 3D 区域（3Dmol.js，PDB 数据内嵌）：**卡通 + 半透明分子表面叠加**、按 pLDDT rwb 着色、**默认自动旋转**，配 3 个控制按钮（自动旋转 / 隐藏表面 / 复位视角）+ 鼠标交互提示，报告内直接可交互；另附 PDB 文件供 Mol* 深度操作。", "prompt_en": "Prompt 2 | De Novo Design → Structure Validation\n\nI am a protein designer in the industrial enzyme field, targeting a thermostable bifunctional enzyme (chitinase + lysozyme activity) that can withstand 60°C. Please use the ORI public application's generate-protein sub-workflow, following all default parameters (default design prompt, model, and resource configuration), and run a de novo protein design, then complete the analysis:\n\n1. Before submission, provide me with a task confirmation sheet (sub-workflow | design intent | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and fix suggestions based on the error knowledge base;\n3. Upon completion, output: candidate sequence table (rank | name | length | score), rendered as an HTML report page; then automatically cascade usmfold-predict (default parameters) on the top-scoring candidate, with its predicted structure embedded in the report's 3D area (3Dmol.js, PDB data embedded): **cartoon + semi-transparent surface overlay**, colored by pLDDT rwb, **auto-rotation enabled by default**, with 3 control buttons (auto-rotate / hide surface / reset view) + mouse interaction tips, directly interactive within the report; also provide the PDB file for deep analysis in Mol*.", "skills": [], "experts": [ { "id": "omics-ori-expert", "name": "ORI蛋白设计专家", "description": "腾讯ORI蛋白设计专家，基于ORI蛋白质设计系统，提供从头序列设计、USMFold结构预测、溶解性/热稳定性/信号肽预测五大能力，并支持3D结构可视化", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_HrdscGoeTkwq" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T07:58:16Z", "updated_at": "2026-08-30T11:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-scprotein-expert-case-2", "title": "细胞表示嵌入（Stage2）", "title_en": "Cell Representation Embedding (Stage2)", "subtitle": "从 Stage1 结果直接学习细胞嵌入，t-SNE 分群一目了然", "subtitle_en": "Learn cell embeddings directly from Stage1 results — t-SNE clusters at a glance", "description": "完成了肽段质控，下一步怎么把单细胞数据转化成可分群的细胞表征？这个案例演示了如何衔接 Stage1 输出，用 scPROTEIN Stage2 学习细胞低维嵌入，生成 PCA→t-SNE 细胞分群散点图（KMeans 着色）、Silhouette vs k 曲线与最优聚类数推荐，为后续细胞类型注释提供直观依据。\n\n适合：已完成 scPROTEIN Stage1 分析、需要进一步做细胞分群与类型表征的蛋白质组学研究人员。", "description_en": "Done with peptide QC — how do you turn single-cell data into cluster-ready cell representations? This case demonstrates how to follow up from Stage1 output, use scPROTEIN Stage2 to learn low-dimensional cell embeddings, and produce a PCA→t-SNE cell cluster scatter plot (KMeans colored), Silhouette vs k curve, and optimal cluster count recommendation — providing intuitive guidance for downstream cell type annotation.\n\nSuitable for: proteomics researchers who have completed scPROTEIN Stage1 analysis and need to proceed with cell clustering and type representation.", "categories": [ "data" ], "tags": [ "单细胞蛋白组", "细胞类型注释", "t-SNE", "scPROTEIN", "聚类分析" ], "tags_en": [ "Single-cell Proteomics", "Cell Type Annotation", "t-SNE", "scPROTEIN", "Clustering Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "细胞嵌入分群交互报告" } ], "prompt": "我已完成 Stage1 肽级不确定性定量，现需学习细胞嵌入用于分群注释。请使用 scPROTEIN 应用，按默认参数执行 Stage2 细胞嵌入。\n\n1. 提交前确认 Stage1 衔接输出可读、数据规模与参数摘要，我确认后提交；\n2. 跟踪任务至完成，失败给根因诊断；\n3. 结果可视化：PCA→t-SNE 细胞分群散点图（KMeans 着色）、Silhouette vs k 曲线与最优 k、PCA 方差解释图，HTML 交互报告在内置浏览器打开。", "prompt_en": "I have completed Stage1 peptide-level uncertainty quantification and now need to learn cell embeddings for cluster annotation. Please use the scPROTEIN application to run Stage2 cell embedding with default parameters.\n\n1. Confirm that Stage1 output is readable, along with data scale and parameter summary before submission; submit after my confirmation;\n2. Track the task to completion; provide root cause diagnosis on failure;\n3. Visualize results: PCA→t-SNE cell cluster scatter plot (KMeans colored), Silhouette vs k curve with optimal k, PCA variance explained plot — open the HTML interactive report in the built-in browser.", "skills": [], "experts": [ { "id": "omics-scprotein-expert", "name": "腾讯scPROTEIN单细胞蛋白组建模专家", "description": "腾讯 scPROTEIN表征模型，实现对单细胞蛋白组的多肽不确定性估计，去除批次效应噪声以及单细胞蛋白组的细胞类型注释", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_SlvtyVitPzY8" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-unipert-expert-case-2", "title": "化合物 MOA 聚类与作用机制分析", "title_en": "Compound MOA Clustering and Mechanism of Action Analysis", "subtitle": "SMILES 输入，余弦聚类判定药物作用机制，支持重定向与联合用药筛选", "subtitle_en": "SMILES input, cosine clustering for MOA determination, supports drug repurposing and combination screening", "description": "手头有一批化合物，想知道它们的作用机制是否相似、能否做药物重定向？这个案例演示了如何用 UniPert compounds 子应用将化合物 SMILES 编码为统一扰动表征，基于余弦相似度做层次聚类与 MOA 判定，输出含聚类热图+树状图、共享激活维度对比图与已知药理交叉验证表的 ECharts 交互报告。\n\n适合：从事药物发现、MOA 分析、联合用药筛选的研究人员与制药公司研发团队。", "description_en": "Got a batch of compounds and want to know if their mechanisms of action are similar or if drug repurposing is possible? This case demonstrates how to use the UniPert compounds sub-application to encode compound SMILES into unified perturbation representations, perform hierarchical clustering and MOA determination via cosine similarity, and produce an ECharts interactive report with clustering heatmap + dendrogram, shared activation dimension comparison, and known pharmacology cross-validation table.\n\nSuitable for: researchers and pharmaceutical R&D teams working on drug discovery, MOA analysis, and combination therapy screening.", "categories": [ "data" ], "tags": [ "化合物MOA", "药物重定向", "UniPert", "扰动表征", "聚类分析" ], "tags_en": [ "Compound MOA", "Drug Repurposing", "UniPert", "Perturbation Representation", "Clustering Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "化合物 MOA 聚类分析报告" } ], "prompt": "我需要通过统一扰动表征完成化合物的 MOA 聚类与作用机制分析。请使用 UniPert-WDL 合集的 compounds 子应用，按默认输入模板（compounds.csv，SMILES 输入，无需我提供数据）执行化合物表征编码，再基于余弦相似度做层次聚类与 MOA 判定。\n\n1. 提交前确认子应用、同名检查结果与参数摘要，我确认后提交；\n2. 跟踪任务至完成，失败给根因诊断（debug 取证 + 错误知识库匹配 + 修复建议）；\n3. 结果固化为一份可复用的 ECharts 交互式 HTML 报告（如 unipert_moa_report.html），在内置浏览器打开预览，须包含：概览卡片（化合物数 / 表征维度 / MOA 类别数 / 聚类对相似度 / Top-20 重叠维度数 / 运行时长）、聚类热图 + 树状图（含同类阈值线 0.30 与判定结论面板）、Top 激活维度对比图（共享激活维度琥珀色高亮）、已知药理交叉验证表（模型判定 vs 已知 MOA + 一致性标签）、跨模态参照图（内源 vs 外源分离）、结论与应用价值（药物重定向 / 联合用药筛选 / MOA 注释）。", "prompt_en": "I need to perform compound MOA clustering and mechanism of action analysis through unified perturbation representations. Please use the compounds sub-application from the UniPert-WDL collection, with the default input template (compounds.csv, SMILES input, no need for me to provide data) to execute compound representation encoding, then perform hierarchical clustering and MOA determination based on cosine similarity.\n\n1. Confirm the sub-application, duplicate name check result, and parameter summary before submission; submit after my confirmation;\n2. Track the task to completion; provide root cause diagnosis on failure (debug evidence + error knowledge base matching + fix suggestions);\n3. Consolidate results into a reusable ECharts interactive HTML report (e.g., unipert_moa_report.html), open preview in built-in browser, must include: overview cards (compound count / representation dimensions / MOA category count / cluster pair similarity / Top-20 overlapping dimensions / runtime), clustering heatmap + dendrogram (with same-class threshold line 0.30 and determination conclusion panel), Top activation dimension comparison chart (shared activation dimensions highlighted in amber), known pharmacology cross-validation table (model determination vs known MOA + consistency labels), cross-modal reference chart (endogenous vs exogenous separation), conclusions and application value (drug repurposing / combination drug screening / MOA annotation).", "skills": [], "experts": [ { "id": "omics-unipert-expert", "name": "腾讯UniPert虚拟细胞专家", "description": "基于腾讯UniPert模型，构建虚拟细胞，实现对于内外源扰动源统一表征，并且实现基因和药物扰动预测，包括细胞特异性药物响应", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_w7OMyXyVUKQh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "palace-museum-porcelain-kiln-distribution", "title": "故宫馆藏瓷器装饰带与窑口分布", "title_en": "Palace Museum Porcelain: Decoration Bands & Kiln Distribution", "subtitle": "377,440 件馆藏瓷器的窑口占比与彩带/装饰带构成", "subtitle_en": "Kiln distribution & decoration bands among 377,400 porcelain items", "description": "故宫博物院是中国收藏瓷器数量最大的博物馆，但研究者面对 20 余个窑口、近 38 万件文物时，常常苦于如何把'装饰带（彩带）'维度量化展示。这个案例演示了如何借助腾讯探元文物库多维检索，配合故宫官方公开数据，生成以柱状图为主、饼图与横向柱状图为辅的数据看板，让'一超多强'的窑口格局、清代各朝烧造量、青花/粉彩等装饰工艺占比跃然纸上。\n\n适合：陶瓷史研究者、博物馆策展与陶瓷研究馆员、文化遗产数字化工程师、明清艺术史与皇家工艺研究者、博物馆数据可视化爱好者。", "description_en": "While the Palace Museum is the world's largest porcelain holder (377,440 items across 20+ kilns), researchers often struggle to quantify the 'decoration band' dimension. This case shows how to leverage the Tencent Tanyuan relic database MCP for multi-dimensional queries, combined with the Palace's public data, to build a data dashboard with a bar chart (kiln distribution), a pie chart (decoration share), and a horizontal bar chart (Qing-era production). The result is a clear visualization of the 'one-super-multiple-strong' kiln pattern.\n\nBest for: ceramic history researchers, museum curators & ceramic-collection staff, cultural heritage digitalization engineers, Ming-Qing art historians, and museum data visualization enthusiasts.", "categories": [ "data" ], "tags": [ "故宫博物院", "瓷器", "窑口", "装饰带", "数据看板" ], "tags_en": [ "Palace Museum", "Porcelain", "Kilns", "Decoration Bands", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 数据看板报告（内置中英 i18n）" } ], "prompt": "请按「腾讯探元学术研究-量化分析」专家的工作流，研究「故宫博物院馆藏瓷器的装饰带（彩带）与窑口分布」。\n\n研究范围：\n- 窑口维度：景德镇御窑、龙泉窑、耀州窑、磁州窑、钧窑、越窑、官/汝/哥窑、定窑共 8 个对比维度\n- 装饰带维度（彩带）：青花、粉彩/珐琅彩、颜色釉、五彩/斗彩、釉里红、其他彩瓷（数据按公开文献相对估算）\n- 清代各朝烧造量：乾隆、光绪、道光、康熙、雍正、嘉庆、同治、咸丰、宣统\n\n交付要求（强制）：\n1. 探元文物库 MCP 多维检索：按 category=瓷器 维度分批查询（2—4 批次），统计窑口、装饰、年代分布\n2. WebSearch 1—2 次补充故宫官方公开数据（藏品总目、装饰工艺命名规律等）\n3. 图表 ≥ 3 张：1 张柱状图（窑口占比）+ 1 张饼图（装饰工艺）+ 1 张横向柱状图（清代各朝烧造量），数据全部基于真实出处并标注口径\n4. 四格式报告：Markdown 母版 → HTML → DOCX → PDF\n5. 每个数据章节'先图后文'：量化结论须可回溯到具体数据点；标注数据来源与统计口径；不含'建议'\n6. 关于探元稀疏数据：故宫博物院条目在探元文物库中索引覆盖不足（实测搜'故宫博物院'返回 0 行），采用'探元 + 公开文献双链路'如实标注的方法", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Quantitative Analysis' expert workflow to research 'the decoration bands and kiln distribution of porcelain in the Palace Museum collection'.\n\nResearch scope:\n- Kiln dimension: 8 comparison kilns (Jingdezhen Imperial, Longquan, Yaozhou, Cizhou, Jun, Yue, Guan/Ru/Ge, Ding)\n- Decoration band dimension: blue-and-white, famille rose/enamel, monochrome glazes, wucai/doucai, underglaze red, others\n- Qing-dynasty production: Qianlong / Guangxu / Daoguang / Kangxi / Yongzheng / Jiaqing / Tongzhi / Xianfeng / Xuantong\n\nDelivery requirements (mandatory):\n1. Multi-dimensional Tanyuan MCP queries (2-4 batches) by category=porcelain\n2. 1-2 WebSearch passes for Palace Museum public data\n3. ≥3 charts: bar (kiln) + pie (decoration) + horizontal bar (Qing production)\n4. Four-format report: Markdown master → HTML → DOCX → PDF\n5. Each chapter 'figure-first then text'; quantitative conclusions traceable to data points; data source & statistical calibre labeled; no 'recommendations'\n6. On data sparsity: Tanyuan's Palace Museum coverage is limited (search '故宫博物院' returned 0 rows); using 'Tanyuan + public literature' dual-track method with explicit annotation", "skills": [], "experts": [ { "id": "tanyuan-deep-research-quantitative", "name": "文博可视化分析专家", "description": "以腾讯探元文物数据库为核心信息源，按用户维度分批多次查询，交付图表先行的数据看板报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "13bdc1132eb4", "expert_id": "ex_WaXheY0OvJSu" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:30:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-infographic-chenghua-chicken-cup", "title": "明成化斗彩鸡缸杯工艺图解", "title_en": "Craft Diagram of the Chenghua Doucai Chicken Cup", "subtitle": "斗彩巅峰 · 一图看懂二次烧成的「青花斗彩」", "subtitle_en": "Peak of Doucai — single diagram explaining the two-fire「underglaze-blue vs overglaze-enamel」process", "description": "跟社教讲解员说「斗彩」时，对面八成是懵的——「釉下青花」到底是哪一笔？「釉上填彩」何时上？「二次烧成」靠什么控？ 这个案例用 tanyuan-relic-infographic-expert 把工艺卡在「鸡缸杯」这件明星器上：什么是斗彩、为什么成品率低、「姹紫」为什么只成化有、2014 年玫茵堂拍出 2.8 亿港元背后是什么——一张可贴公众号、可贴展板的 HTML 长图。\n\n适合：博物馆瓷器展厅的工艺讲解页、景德镇御窑厂相关课件、瓷器鉴定自媒体拍品档案、青花/五彩对比教学页、文博类公众号封面。", "description_en": "When you say「doucai」to a museum docent, they almost always look blank — which stroke is「underglaze blue」? When does the「overglaze enamel」go on? What controls the「double firing」? This case uses tanyuan-relic-infographic-expert to lock the craft onto one star-piece, the Chicken Cup: what doucai actually is, why the yield is low, why the unique「chazi (magenta)」color belongs only to Chenghua, and what's behind the 2014 Meiyintang sale at HK$281M — a single HTML infographic ready for WeChat publishing or exhibition panels.\n\nSuitable for: porcelain-craft explanation pages in museum galleries, classroom courseware on the Jingdezhen imperial kiln, appraisal-self-media catalogs for porcelain lots, teaching sheets on underglaze-blue vs wucai contrast, and WeChat public-account covers in cultural heritage.", "categories": [ "creation", "education" ], "tags": [ "文物图解", "斗彩", "成化", "景德镇御窑", "工艺图解" ], "tags_en": [ "relic infographic", "doucai", "Chenghua", "Jingdezhen imperial kiln", "craft diagram" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请调用 tanyuan-relic-infographic-expert 专家，为「明成化斗彩鸡缸杯」生成一张单文件 HTML 工艺图解长图。\n\n要求：\n1. 数据真实：使用 tanyuan-wb-mcp-server 的 search_relics（查询「明成化斗彩鸡缸杯」）与 search_movable_relics（查询「斗彩鸡缸杯」）真实检索——本次会话已确认 MCP 可用。检索返回的真实字段包括：御窑元华堂复刻记录（5 名技师分绘山石/花草/公鸡/母鸡/小鸡 + 向元华大师领衔 + 釉下青花勾线 + 700-800℃ 低温二次烧成 + 顶级高岭土胎质「糯米胎」+ 鸡群寓意「专宠一人、三生万物」+ 2014 苏富比 2.8 亿港元拍卖原品著录）；粉彩鸡缸杯对照件（清乾隆官窑，口径 8cm 高 6.9cm，从斗彩到粉彩工艺演进）。\n2. 文物本体图：探元检索已返回 3 张复刻品授权图（static.tanyuan.qq.com 域名，已授权直接用），原品本体图按专家 SKILL.md §5.2 走 Wikimedia Commons（File:Percival David Collection DSCF3631 10.jpg，CC BY-SA 4.0，作者 Johnbod，大英博物馆 Percival David Collection Room 95 馆藏），下载 base64 内联；aspect-ratio 等于真实图片比例（如 1200/655）。\n3. 视觉：使用简约版模板，配色采用「朱红·朱砂宫墙」（更契合瓷器气质）。\n4. 结构：标题区（明·成化 / 斗彩巅峰）→ HERO 顶图（原品真实照片）→ 器物特征 → 斗彩工艺详解（釉下青花 + 釉上多彩 + 二次烧成 + 色料来源 + 分绘制度 + 工序稀缺性）→ 章节内插图（探元复刻品授权图，标注「复刻品」）→ 名号由来与历史意义（含明·沈德符《万历野获编》「成窑酒杯，每对值十万金」原典 + 2014 苏富比 2.8124 亿港元成交 + 存世稀少十余件）→ 同类对比（斗彩 vs 五彩 4 维度）→ 研究与鉴定要点（姹紫仅成化独有）→ 底部四栏 → 页脚来源标注。\n5. 硬约束：单文件 HTML、CSS/图片全部内联、不得 https:// 外链、不得 href=\"#\" 空锚点、不得 emoji；所有数值与日期逐处标源。", "prompt_en": "Use the tanyuan-relic-infographic-expert to produce a single-file HTML craft-infographic for the「Chenghua Doucai Chicken Cup」.\n\nRequirements:\n1. Real data: query via tanyuan-wb-mcp-server (search_relics「明成化斗彩鸡缸杯」+ search_movable_relics「斗彩鸡缸杯」) — MCP confirmed reachable this session. Real fields returned include: Imperial Kiln Yuanhuatang replica record (5 artisans drawing rocks/flowers/rooster/hen/chicks + master Xiang Yuanhua leading + underglaze-blue outlining + 700-800°C low-temperature double firing + top Jingdezhen kaolin「sticky-rice」body + chicken flock symbolizing「exclusive favor, three-generation multiplication」 + 2014 Sotheby's HK$281M record); Famille-rose chicken-cup reference (Qianlong imperial kiln, mouth 8cm, height 6.9cm, showing doucai→fencai evolution).\n2. Relic photo: TanYuan returns 3 replica authorized images (static.tanyuan.qq.com domain, free to use); the original piece follows expert SKILL.md §5.2 → Wikimedia Commons (File:Percival David Collection DSCF3631 10.jpg, CC BY-SA 4.0, photographer Johnbod, British Museum Percival David Collection Room 95), downloaded + base64 inlined; aspect-ratio equals actual image (e.g. 1200/655).\n3. Visual: «concise edition» template with「zhuhong (cinnabar-wall)」 theme (better fits porcelain mood).\n4. Structure: title block (Ming · Chenghua / peak of doucai) → HERO (original real photo) → physical features → craft details (underglaze blue + overglaze enamels + double firing + colorant sourcing + 5-artisan division + scarcity) → inline figure (TanYuan authorized replica image, tagged「replica」) → naming & historical significance (with Ming Shen Defu's《Wanli Yehuobian》「paired Chenghua cups at 100,000 taels」quote + 2014 Sotheby's HK$281.24M + ~10 surviving pieces) → class comparison (doucai vs wucai, 4 dimensions) → research & appraisal tips (chazi color exclusive to Chenghua) → bottom four-cell bar → footer sources.\n5. Hard constraints: single HTML, all CSS & images inlined, no external https://, no href=\"#\" anchors, no emoji; every auction record and museum count source-tagged.", "skills": [], "experts": [ { "id": "tanyuan-relic-infographic", "name": "文物图解专家", "description": "基于探元文物库+知识库+MCP检索的单文件文物图解专家，使用简约版/典藏版模板输出可展陈、可传播的 HTML 长图", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_OPJE3pWwnB1j" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:06:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "明成化斗彩鸡缸杯工艺图解" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "sancai-duck-ewer-collect-card", "title": "三彩鸭形水注文创 · 小卡", "title_en": "Sancai Duck Ewer Cultural Product · Collectible Card", "subtitle": "温暖手绘插画风的东方收藏小卡", "subtitle_en": "A warm hand-drawn Oriental collectible card", "description": "想给文创品配一张能收藏、有温度的小卡，实拍图又太硬，缺手作出版物的味道。\n\n案例把实拍卡包产品转绘为深茶褐轮廓、水彩颗粒质感的手绘插画，居中悬浮暖象牙白圆角卡片，顶部手写标题「春江水暖，好事游来～」，点缀祥云、星芒与朱砂小印，底部「2026」，呈现轻复古、东方手工出版物气质的 3:4 收藏小卡。\n\n适合：文创品牌运营、插画/平面设计师、博物馆文创团队、会员礼品开发方", "description_en": "This case redraws the photographed card-holder into a hand-drawn illustration with deep tea-brown outlines and watercolor grain, floating a warm ivory rounded card center-stage, a handwritten title on top, auspicious clouds, star sparkles and a cinnabar seal, with '2026' at the base — a lightly vintage, Oriental hand-published 3:4 collectible card.\n\nGood for: Cultural brand ops, Illustrators & designers, Museum creative teams, Membership gift developers", "categories": [ "creation" ], "tags": [ "收藏小卡", "手绘插画", "东方文创", "3:4竖版" ], "tags_en": [ "Collectible Card", "Hand-Drawn", "Oriental Creative", "3:4 Portrait" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据这张文创品图片，将实拍产品转绘成温暖、细腻、略带复古感的手绘插画，并设计成一张3:4竖版东方文创收藏小卡。准确保留原产品的横向卡包造型、长宽比例、浅棕色皮革材质、卡槽结构、边缘缝线、鸭形浮雕、羽毛细节和底部深蓝色水波纹，不得改变产品品类、结构、朝向、颜色或纹样，不得将其变成真实鸭子、古代文物或其他商品；只将照片质感转化为带深茶褐色轮廓、柔和平面明暗、低饱和暖色、水彩水粉颗粒和纸张肌理的精致手绘插画。\n\n画面中央悬浮一张完整的暖象牙白圆角纸质卡片，背景为低饱和暖灰褐色，带柔和自然投影和充足留白；卡面设置纤细茶褐色双线边框与简洁中式几何角花。顶部居中准确排版标题「春江水暖，好事游来～」，使用自然亲和的中文手写字体；中部完整展示横向文创品插画，主体不裁切，周围疏朗点缀浅茶色线描祥云、少量四角星芒和一个无文字的朱砂红色小印记；下方加入淡墨灰色远山与水面纹样，底部居中排版「2026」，搭配纤细短线和小菱形装饰。整体呈现温暖克制、轻复古、具有东方手工出版物气质的收藏小卡，避免乱码、多余文字、品牌Logo、二维码、水印、繁复国潮装饰、厚重做旧、过度卡通和强烈3D效果。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-promo-expert", "name": "文创宣传视觉设计专家", "description": "将定稿文创设计转化为海报、明信片、收藏小卡、易拉宝等宣传物料", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c6229d08eb8d", "expert_id": "ex_zaYhWoDF5zrp" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-position-quick", "title": "职位快速检索：浙江事业编会计岗", "title_en": "Quick Position Search: Zhejiang Accounting Posts", "subtitle": "不设画像直查职位库，快速摸底某类岗位的体量与分布", "subtitle_en": "Query the position database directly to size up a role family fast", "description": "有时你还没想好自己报什么，只想先知道「这类岗有多少、都在哪」。\n\n本案例按「事业单位 + 浙江 + 会计」三个条件直查职位库，命中 88 个岗位（本页 10 条即覆盖 8 份不同公告），逐条列出单位、地区、招录人数、状态与来源公告，并如实说明本接口不返回进面分与报考条件明细。\n\n适合：处于摸底阶段的考生、对比多地岗位供给的考生、做择业咨询的老师。", "description_en": "Sometimes you haven't decided what to apply for — you just want to know how many such posts exist and where.\n\nThis case queries the database directly with \"public institution + Zhejiang + accounting\", hitting 88 positions (the 10 shown span 8 different announcements), listing unit, region, headcount, status and source announcement for each, and honestly noting that this interface returns neither historical cutoff scores nor detailed eligibility requirements.\n\nFor: candidates at the scoping stage, those comparing supply across regions, and counsellors.", "categories": [ "search" ], "tags": [ "事业单位", "职位检索", "浙江", "会计" ], "tags_en": [ "public institution", "position search", "Zhejiang", "accounting" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "浙江事业编会计岗检索结果" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查一下浙江的事业单位岗位里，名称含「会计」的有哪些在招。要求：1) 直接按考试类型=事业单位、地区=浙江、关键词=会计检索职位库并给出总数；2) 展示前 10 条：岗位、单位、地区、招录人数、状态、来源公告；3) 统计这批结果来自几份公告、状态分布如何；4) 接口没有的信息（进面分、报考条件）如实说明没有，不要推测。", "prompt_en": "Find Zhejiang public-institution positions whose titles contain \"accounting\". Requirements: 1) query the database with exam type=public institution, region=Zhejiang, keyword=accounting and give the total; 2) show the top 10 with position, unit, region, headcount, status and source announcement; 3) report how many distinct announcements the results span and the status distribution; 4) for information the interface lacks (cutoff scores, eligibility details), state so honestly without guessing.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "支持按考试类型/地区/关键词快速检索职位库", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "solo-dev-account-launch", "title": "一个人做产品，怎么拿到第一批用户", "title_en": "Getting Your First Users as a Solo Founder", "subtitle": "产品刚上线没用户没案例，给一份能直接执行的 30 天起号方案", "subtitle_en": "No users, no case studies — get a 30-day launch plan you can execute today", "description": "一人公司 / 独立开发者用 AI 编程做出产品后，最常卡在「内容从哪开始」——没有用户案例、没有数据截图、没有客户证言，常规起号方法全都用不上。这个案例基于你的产品、目标用户和每周可投入时间，输出赛道选择、账号定位卡、万粉以内的对标账号清单、能直接落笔的前 10 条选题（含排序逻辑）、按周拆解的 30 天路线图，并明确列出现阶段不该做的事。所有选题围绕「做产品的过程本身」设计，不需要任何用户案例即可执行。", "description_en": "Solo founders and indie developers who ship a product with AI coding tools get stuck at the same place: where does content even start? No user case studies, no data screenshots, no customer testimonials — every conventional launch playbook is off the table. Based on your product, target users and weekly time budget, this case outputs a niche recommendation, a positioning card, a benchmark account list capped at sub-10K followers, the first 10 ready-to-write topics with the reasoning behind their order, a 30-day roadmap broken down by week, and an explicit list of what you should NOT do yet. Every topic is designed around the act of building the product itself, so none of them require a single user case study.", "categories": [ "content" ], "tags": [ "起号规划", "独立开发", "冷启动", "内容日历", "一人公司" ], "tags_en": [ "Launch Planning", "Indie Dev", "Cold Start", "Content Calendar", "Solo Founder" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "一人公司 30 天起号方案" } ], "prompt": "你是社媒内容起号规划师。我是一人公司（独立开发者），刚用 AI 编程做出了自己的产品并已上线，现在需要通过做内容拿到第一批用户。\n\n我的情况：\n- 产品是什么：{产品一句话介绍}\n- 解决谁的什么问题：{目标用户和他们的痛点}\n- 目前状态：{上线多久、有多少用户}\n- 我每周能投入做内容的时间：{小时数}\n- 我更愿意出镜还是只写字：{出镜 / 只写字 / 都可以}\n\n请给我一份完整的起号方案，必须包含这五部分：\n\n1. 赛道选择建议：列 3 个候选内容赛道，每个写推荐理由和风险提示，最后给明确推荐结论（不要让我自己选）\n2. 账号定位卡：人设一句话、目标受众、内容形态、差异化角度\n3. 对标账号清单：3-5 个，说明粉丝量级和具体可学的点。优先给万粉以内、能看出起量路径的，不要只给头部大号\n4. 前 10 条内容清单：给出顺序和每条的具体选题（不是方向，是能直接落笔的标题），并说明这个顺序为什么这样排\n5. 30 天路线图：按周拆，写清每周选题方向、内容形式、目标、要观察哪个数据\n\n三条硬要求：\n\n- 我没有用户案例、没有数据截图、没有客户证言。不要给需要这些素材才能做的选题，请围绕「做产品的过程本身」设计内容\n- 严格按我的产能来排。如果我每周只有 3 小时，就不要给需要日更的方案，也不要让我同时做多个平台\n- 明确告诉我哪些事现在不该做（比如 0 粉投流、为凑数日更），并说明原因\n\n输出为 HTML 卡片，单文件、CSS 内联，可直接双击打开。", "prompt_en": "You are a social media launch strategist. I run a one-person company (indie developer). I just shipped my own product built with AI coding tools and it's live. Now I need content to bring in my first users.\n\nMy situation:\n- What the product is: {one-line product description}\n- Whose problem it solves: {target users and their pain points}\n- Current status: {how long it's been live, how many users}\n- Hours per week I can spend on content: {number of hours}\n- Whether I'd rather appear on camera or just write: {on camera / writing only / either}\n\nPlease give me a complete launch plan covering these five parts:\n\n1. Niche recommendation: list 3 candidate content niches, with the rationale and the risk for each, then give me a clear final recommendation (don't make me pick)\n2. Positioning card: one-line persona, target audience, content format, differentiation angle\n3. Benchmark account list: 3-5 accounts, with follower tier and the specific thing worth learning from each. Prioritize accounts under 10K followers whose growth path is still visible — don't just give me the top players\n4. First 10 posts: give the order and the specific topic for each (not a direction — an actual title I can start writing), and explain why they're sequenced this way\n5. 30-day roadmap: broken down by week, with each week's topic direction, content format, goal, and the one metric to watch\n\nThree hard requirements:\n\n- I have no user case studies, no data screenshots, no customer testimonials. Don't give me topics that require those — design the content around the act of building the product itself\n- Stick strictly to my capacity. If I only have 3 hours a week, don't hand me a daily-posting plan, and don't put me on multiple platforms at once\n- Tell me explicitly what I should NOT do right now (e.g. paid promotion at 0 followers, daily posting just to hit a number), and why\n\nOutput as an HTML card: single file, inlined CSS, openable with a double click.", "skills": [], "experts": [ { "id": "account-launch-expert", "name": "启号·营销通起号规划师", "description": "从 0 粉到能跑起来的账号：定赛道、定位、对标与 30 天路线图。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "abab585e0c1e", "expert_id": "ex_AB1EkOhHLzbY" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T09:00:00Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-lit-search-case-2", "title": "黑色素瘤靶向治疗证据检索", "title_en": "Targeted Therapy Evidence Retrieval for Melanoma", "subtitle": "5年BRAF V600E黑色素瘤临床证据，RCT到真实世界一网打尽", "subtitle_en": "5-year BRAF V600E melanoma clinical evidence from RCTs to real-world studies in one report", "description": "肿瘤治疗方案选择需要同时参考 RCT、系统综述、真实世界研究与在研试验，但按 PICO 框架手动分层检索各类证据需要检索多个数据库、逐一筛选，极度耗时。腾讯生命科学文献检索专家按场景自动联查 PubMed、ClinicalTrials 等专业来源，分层去重，并输出带 OS/PFS 比较的结构化证据表，附完整 PMID、DOI、NCT 溯源。本案例以 BRAF V600E 阳性黑色素瘤 BRAF/MEK 抑制剂治疗的近 5 年临床证据为示例，产出分层证据 HTML 报告，便于快速比较不同研究设计的疗效结论。适合肿瘤科医生、医学证据评估人员及临床研究设计者使用。", "description_en": "Selecting oncology treatment regimens requires simultaneously reviewing RCTs, systematic reviews, real-world evidence, and ongoing trials — but manually stratifying these across multiple databases by PICO framework is extremely time-consuming. The Tencent Life Sciences Literature Search Expert automatically federates PubMed, ClinicalTrials, and other authoritative sources by scenario, deduplicates results, and outputs a structured evidence table with OS/PFS comparisons and full PMID/DOI/NCT traceability. This case demonstrates 5-year clinical evidence retrieval for BRAF/MEK inhibitor treatment in BRAF V600E-positive melanoma, producing a stratified evidence HTML report for quick comparison across study designs. Ideal for oncologists, medical evidence evaluators, and clinical trial designers.", "categories": [ "search", "education" ], "tags": [ "临床证据", "循证医学", "文献检索", "黑色素瘤", "BRAF" ], "tags_en": [ "Clinical Evidence", "Evidence-Based Medicine", "Literature Search", "Melanoma", "BRAF" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "tencent-lit-search-output-2.html", "artifacts": [ { "file": "tencent-lit-search-output-2.html", "label": "BRAF V600E 黑色素瘤分层证据报告" } ], "prompt": "检索近 5 年 BRAF V600E 突变阳性黑色素瘤接受 BRAF/MEK 抑制剂治疗的临床证据。按 PICO 区分 RCT、系统综述、真实世界研究和在研试验，比较 OS/PFS，输出带 PMID、DOI、NCT 的分层证据表。", "prompt_en": "Retrieve clinical evidence from the past 5 years for BRAF V600E-positive melanoma treated with BRAF/MEK inhibitors. Distinguish RCTs, systematic reviews, real-world studies, and ongoing trials by PICO. Compare OS/PFS and output a stratified evidence table with PMID, DOI, and NCT.", "experts": [ { "id": "tencent-lit-search", "name": "生命科学文献检索", "description": "生命科学证据检索专家。覆盖药研、临床、生信及遗传与罕见病，按场景联查专业来源，结果分层去重可溯源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "843d1587f82b", "expert_id": "ex_DXuzV9trsD3U" } ], "skills": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 110, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-24T16:32:00+08:00", "updated_at": "2026-08-25T12:18:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "bakery-customer-persona-insight", "title": "连锁烘焙用户画像与人群洞察", "title_en": "Bakery Customer Persona and Crowd Insight", "subtitle": "从 4.2 万会员数据到四类人群动作回路", "subtitle_en": "From 42,000 member records to four audience action loops", "description": "会员数据分散在门店、小程序和企微时，团队常常只有“年轻女性喜欢甜品”这类无法圈选、无法行动的模糊认知。\n\n这个案例以某城市连锁烘焙品牌的匿名数据为输入，通过用户画像与人群洞察 Skill 完成六维就绪度评估、CDP 六类标签体系、四个量化分群、四张 Persona 卡、人群机会矩阵和洞察到动作的闭环，同时交付结构化 Markdown 与可视化白皮书 HTML。\n\n适合：连锁零售会员运营、CRM 团队、私域负责人、需要将用户数据转成可执行策略的增长团队。", "description_en": "When member data is split across stores, mini programs, and WeCom, teams often end up with vague statements such as “young women like desserts”—insights that cannot be selected or acted upon.\n\nUsing anonymized data from a city bakery chain, this case applies the User Persona and Crowd Insight Skill to produce a six-dimension readiness assessment, six-domain CDP tag system, four measurable segments, four persona cards, an opportunity matrix, and insight-to-action loops. It delivers both structured Markdown and a visual HTML white paper.\n\nIdeal for: chain-retail membership operators, CRM teams, private-domain leaders, and growth teams turning customer data into executable strategy.", "categories": [ "business" ], "tags": [ "用户画像", "人群洞察", "CDP标签", "连锁零售" ], "tags_en": [ "customer persona", "crowd insight", "CDP tags", "chain retail" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "用户画像白皮书" } ], "prompt": "请为某城市连锁烘焙品牌构建用户画像与人群洞察。品牌有约 4.2 万名会员、18 个月有效订单和部分小程序浏览数据；门店会员与小程序 OneID 匹配率约 72%，企微好友尚未统一映射。近 180 天有效购买会员约 1.8 万，整体 90 天复购率约 27%，早餐时段订单占 34%，社群运营没有统一标签。现有标签只有城市、注册渠道、累计消费和最近购买时间，尚无正式 RFM 分群、Persona、预测模型或标签治理机制。核心目标是提高早餐复购、识别高价值家庭客、降低 90 天以上沉睡会员占比，并让营销和内容团队能直接圈选人群。所有品牌和人物均须匿名，不臆造未提供的真实规模或模型准确率。", "prompt_en": "Build customer personas and crowd insights for an anonymized city bakery chain. It has about 42,000 members, 18 months of valid order history, and partial mini-program browsing data. Store-member and mini-program OneID matching is roughly 72%, while WeCom contacts are not yet mapped. About 18,000 members purchased in the last 180 days, the overall 90-day repeat-purchase rate is about 27%, breakfast orders represent 34%, and community operations lack unified tags. Current tags cover only city, registration source, cumulative spend, and last-purchase time; there are no formal RFM segments, personas, predictive models, or tag-governance mechanisms. The goals are to increase breakfast repeat purchases, identify high-value family customers, reduce the share of members dormant for more than 90 days, and enable direct audience selection by marketing and content teams. Keep all brands and people anonymous, and do not invent unprovided scale or model-accuracy figures.", "skills": [ { "id": "lingyi-user-persona-and-crowd-insight", "name": "用户画像与人群洞察", "description": "【零一数科·出品】盘点画像就绪度，设计 CDP 标签、分群、Persona 与洞察到动作回路。", "slug": "lingyi-user-persona-and-crowd-insight" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 111, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-29T08:00:00Z", "updated_at": "2026-08-31T16:31:11+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "bronze-vessel-evolution-timeline", "title": "中国古代青铜器类型与用途演变", "title_en": "Chinese Bronze Vessels: Type & Function Evolution", "subtitle": "2000 年时间线下从二里头到秦汉的礼器—实用器转向", "subtitle_en": "Ritual-to-utilitarian shift from Erlitou to Qin-Han across 2000 years", "description": "青铜器研究最怕就是把'礼器—实用器'转向讲清楚。这个案例演示如何从腾讯探元文物库的青铜器样本出发，按六大器类（食器/酒器/乐器/兵器/水器/车马器）与六朝（夏—商—西周—春秋—战国—秦汉）分批量化统计，配合殷墟 4380 件兵器、二里头爵、列鼎制度等公开学术坐标，呈现一张'范铸法→失蜡法、神权→日用'的可量化时间线。\n\n适合：中国古代史与先秦史研究者、青铜器鉴定与收藏家、博物馆陈列与策展人、考古与文化遗产保护从业者、艺术史与美术专业师生。", "description_en": "Bronze scholarship often struggles to make the ritual-to-utilitarian shift legible. This case shows how to mine 10 Tanyuan MCP entries of category=bronze (Shang, Western Zhou, Warring States, Han) plus the Yinxu 4380-weapon benchmark, Erlitou Jue discovery, and Western Zhou's tripod-cauldron ranking system to produce a quantified 2000-year timeline that traces the shift from block-mold casting → lost-wax casting, from divine kingship to daily life.\n\nBest for: ancient Chinese historians & pre-Qin scholars, bronze collectors & appraisers, museum exhibition designers, archaeology & cultural heritage practitioners, and art-history students.", "categories": [ "data" ], "tags": [ "青铜器", "礼器", "二里头", "列鼎制度", "数据看板" ], "tags_en": [ "Bronze Vessels", "Ritual Vessels", "Erlitou", "Tripod Ranking System", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 数据看板报告（内置中英 i18n）" } ], "prompt": "请按「腾讯探元学术研究-量化分析」专家的工作流，研究「中国古代青铜器的类型与用途演变」。\n\n研究范围：\n- 类型维度：六大器类——食器（鼎/簋/鬲/豆）、酒器（爵/角/斝/尊/罍/卣/觥）、乐器（编钟/铙/镈）、兵器（戈/矛/剑/戟/钺）、水器（盘/匜/鉴）、车马器（銮铃/当卢）\n- 朝代维度：夏（二里头）→ 商 → 西周 → 春秋 → 战国 → 秦汉\n- 用途维度：礼器 / 实用器 / 兵器 / 乐器\n\n交付要求（强制）：\n1. 探元文物库 MCP 以 category=青铜器 维度分批查询（2—4 批次），按朝代/类型/用途分组统计\n2. WebSearch 1—2 次补充学术背景（如后母戊鼎尺寸、殷墟兵器数量、列鼎制度等）\n3. 图表 ≥ 3 张：1 张柱状图（类型分布）+ 1 张堆叠柱状图（朝代×用途演变）+ 1 张手绘 SVG 时间线（朝代节点 + 工艺演化方向）\n4. 四格式报告：Markdown 母版 → HTML → DOCX → PDF\n5. 每个数据章节'先图后文'；量化结论须可回溯到具体数据点；标注数据来源与统计口径；不含'建议'\n6. 关注三组核心对比：礼器 vs 实用器的朝代演变、范铸法 vs 失蜡法的工艺转向、青铜器用途的'神权—王权—日用'三段叙事", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Quantitative Analysis' expert workflow to research 'Chinese bronze vessel types and function evolution'.\n\nResearch scope:\n- Type dimension: 6 categories — food vessels (ding/gui/li/dou), wine vessels (jue/jue/jia/zun/lei/you/gong), music (bianzhong/nao/bo), weapons (ge/mao/jian/ji/yue), water vessels (pan/yi/jian), chariot & harness\n- Dynasty dimension: Xia (Erlitou) → Shang → Western Zhou → Spring & Autumn → Warring States → Qin-Han\n- Function dimension: ritual / utilitarian / weapon / music\n\nDelivery requirements (mandatory):\n1. Multi-batch Tanyuan MCP queries (2-4 batches) by category=bronze, grouped by dynasty/type/function\n2. 1-2 WebSearch passes for academic context (Houmuwu Ding size, Yinxu weapon count, tripod ranking system)\n3. ≥3 charts: bar (types) + stacked bar (dynasty×function) + hand-drawn SVG timeline (dynasty nodes + craft direction)\n4. Four-format report: Markdown master → HTML → DOCX → PDF\n5. Each chapter 'figure-first then text'; quantitative conclusions traceable to data points; labeled source & calibre; no 'recommendations'\n6. Focus on: ritual vs utilitarian dynasty evolution, block-mold vs lost-wax craft shift, and the three-act narrative of divine kingship → royal authority → daily life", "skills": [], "experts": [ { "id": "tanyuan-deep-research-quantitative", "name": "文博可视化分析专家", "description": "以腾讯探元文物数据库为核心信息源，按用户维度分批多次查询，交付图表先行的数据看板报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "13bdc1132eb4", "expert_id": "ex_WaXheY0OvJSu" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 111, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:30:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-quick-tax-national", "title": "职位快搜：国考税务系统报录比透视", "title_en": "Quick Search: Competition Ratios Across the National Tax System", "subtitle": "同一个税务局的 30 个岗位，报录比从 23:1 到 1180:1", "subtitle_en": "Thirty posts at one tax bureau, with ratios from 23:1 to 1180:1", "description": "都说国考税务岗好考，可同一个税务局里，有的岗位 23 个人抢 1 个，有的 1180 个人抢 1 个——只看系统名字根本挑不出岗。\n\n本案例按「国考 + 招录单位关键词=税务局」直查职位库，命中 11,789 个岗位，本页 30 条全部落在北京市税务局各城区，薪资区间几乎一致。逐条列出报录比与往年进面分，做报录比分档统计，并指出报录比与分数线并不同步这一常见误区。\n\n适合：锁定某个系统报考的考生、需要在同系统内二次挑岗的人。", "description_en": "People say tax posts are easy to get into — yet within a single tax bureau, some posts draw 23 applicants per opening and others 1,180.\n\nThis case queries the database with \"national exam + unit keyword = tax bureau\", hitting 11,789 posts. The 30 shown all sit in Beijing districts with near-identical salary bands. It lists each post's competition ratio and historical cutoff score, bands the ratios, and debunks the assumption that a higher ratio means a higher score line.\n\nFor: candidates targeting a specific system who need to pick among its posts.", "categories": [ "search" ], "tags": [ "国考", "税务系统", "报录比", "职位检索" ], "tags_en": [ "national exam", "tax system", "competition ratio", "position search" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "国考税务系统 30 个岗位报录比透视", "label_en": "Competition ratios across 30 national tax-system posts" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查国考里招录单位含「税务局」的岗位。要求：1) 真实检索并给出命中总数；2) 取 30 条结果，按报录比升序列出岗位、所在区、招录人数、报录比、往年进面最低分、薪资区间；3) 指出本批报录比的最大值与最小值差多少倍，并说明这些岗位是否属于同一单位、薪资是否一致；4) 把报录比分成 50:1 以下 / 50–100 / 100–300 / 300–1000 / 1000 以上五档做分布统计；5) 用本批真实数据说明报录比与进面分是否同步，不要给结论性推测。", "prompt_en": "Find national-exam posts whose recruiting unit contains \"tax bureau\". Requirements: 1) run the query and report the total; 2) take 30 results and list them by ascending competition ratio with post, district, headcount, ratio, historical cutoff score and salary band; 3) state how many times larger the highest ratio is than the lowest, and whether these posts share the same unit and salary band; 4) band the ratios into under 50:1 / 50–100 / 100–300 / 300–1000 / over 1000 and give the distribution; 5) use the actual data to show whether ratio and cutoff score move together — no speculation.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 111, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "smart-page-insight-ai-page-dau", "title": "AI Page 产品数据周报 · 金字塔", "title_en": "AI Page Weekly Data Review · Pyramid", "subtitle": "用金字塔·数据版把 DAU 复盘做成结论先行的可交互 HTML", "subtitle_en": "Turn a weekly DAU review into a conclusion-first interactive HTML with Pyramid Data narrative", "description": "产品数据复盘最怕什么？怕 Leader 翻 5 页数据图才知道「到底涨了还是跌了」。\n\n这个案例基于腾讯文档「智能页面」（Smart Page）技能，用「金字塔·数据版」叙事（Minto Pyramid，结论先于数字）+ 脉冲皮肤（Pulse 深色数据仪表盘），把一份 AI Page 本周 DAU 复盘做成单文件可交互 HTML。首屏直接给结论卡（headline + tone），紧跟 4 个核心 KPI 象限（DAU 峰值 / 分享率 / 文档内生成占比 / 新场景日新增），每个带周内 delta 和置信度标签。中段是 DAU 7 天趋势图（Chart.js，支持 launch / peak 节点 annotation），下面 2 组 breakdown（场景拆解 + 入口分布），再接 5 条 insight 卡（每条观察 + 归因 + severity）和 4 条带 owner / ETA 的下一步行动。整体动效含滚动揭示、数字 CountUp、Sticky 章节编号，同时降级 prefers-reduced-motion。示例中所有敏感数字已模糊化处理。\n\n适合：\n- 每周 / 每月负责产品数据复盘的数据 PM、增长 PM、运营负责人\n- KPI 汇报、Review 会、OKR 对齐、向上同步 DAU / 留存 / 转化等核心指标\n- 受够了 Leader 问「所以结论是什么」的同学\n- 希望把数据洞察做成结论先行、可滚动可查、可分享传播的人", "description_en": "The biggest pain of product data reviews? Your leader flipping 5 chart pages just to figure out \"did it grow or drop?\"\n\nThis case is built on Tencent Docs Smart Page. Using the Pyramid Data narrative (Minto Pyramid — conclusion before numbers) + Pulse skin (dark data dashboard), an AI Page weekly DAU review is packaged into a single-file interactive HTML. The first fold shows the headline verdict (with tone), followed by 4 core KPIs (Peak DAU / Share Rate / Doc-embedded Ratio / New Scene Daily Adds), each with a weekly delta and confidence badge. Mid-section: a 7-day DAU trend chart (Chart.js, with launch / peak annotations), two breakdown groups (scene split + entry split), then 5 insight cards (observation + reason + severity) and 4 next-step actions with owner and ETA. Animations include scroll reveal, CountUp for numbers, sticky section index, and a graceful prefers-reduced-motion fallback. All sensitive numbers in the demo are obfuscated.\n\nBest for:\n- Data PMs, growth PMs, and ops leads running weekly / monthly data reviews\n- KPI reporting, review meetings, OKR syncs, and upward reporting on DAU / retention / conversion\n- Anyone tired of being asked \"so what's the conclusion?\"\n- People who want their data insights to lead with verdict, be scrollable, and be shareable", "categories": [ "office", "data" ], "tags": [ "数据周报", "金字塔原理", "DAU 复盘", "可交互 HTML", "Smart Page" ], "tags_en": [ "Weekly Data Report", "Pyramid Principle", "DAU Review", "Interactive HTML", "Smart Page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI Page 本周数据复盘（可交互 HTML）" } ], "prompt": "请使用「腾讯文档智能页面 · Smart Page」技能，帮我生成一份产品数据周报的可交互 HTML 汇报网页。\n\n参数：\n- 场景：insight（数据洞察）\n- 叙事范式：pyramid-data（金字塔·数据版，结论先于数字）\n- 皮肤：pulse（脉冲，深色数据仪表盘）\n- 受众：直属 Leader\n\n内容框架（金字塔结构，结论先行）：\n1. Headline：一句话结论\n2. KPI 四象限：4 个核心指标，各带数值、周内 delta、置信度\n3. 趋势图：1-3 张，可标注关键节点\n4. Breakdown：1-3 组维度拆解\n5. Insights：3-5 条洞察，每条含观察 + 归因 + 严重度\n6. Actions：2-4 条下一步，含 owner 与 ETA\n\n第一步：（提供本周真实数据，敏感数字可模糊化保留量级）。\n...\n\n输出要求：\n- 单文件 HTML，CSS / JS / 字体 / 图片全部内联\n- 首屏不滚动即可看到 headline + 4 个 KPI\n- 趋势图用 Chart.js 渲染\n- 具备滚动揭示、数字 CountUp、Sticky 章节编号、进度条，并支持 prefers-reduced-motion 降级\n- 最后发布到腾讯文档，返回可在线访问的链接", "prompt_en": "Please use the \"Tencent Docs Smart Page\" skill to generate an interactive HTML report page for a product data weekly review.\n\nParameters:\n- Scenario: insight (data insight)\n- Narrative: pyramid-data (conclusion before numbers)\n- Skin: pulse (dark data dashboard)\n- Audience: my direct leader\n\nContent framework (pyramid structure, conclusion first):\n1. Headline: a one-sentence conclusion\n2. KPI four-quadrant: 4 core metrics, each with value, weekly delta, and confidence\n3. Trend charts: 1-3, with annotations for key points\n4. Breakdown: 1-3 dimension splits\n5. Insights: 3-5, each with observation + reason + severity\n6. Actions: 2-4 next steps with owner and ETA\n\nStep 1: (provide this week's real data; sensitive numbers can be blurred while keeping the order of magnitude).\n...\n\nOutput requirements:\n- Single-file HTML, with all CSS / JS / fonts / images inlined\n- The headline + 4 KPIs must be visible on the first screen without scrolling\n- Render trend charts with Chart.js\n- Include scroll reveal, number CountUp, sticky section numbering, a progress bar, and a prefers-reduced-motion fallback\n- Finally publish to Tencent Docs and return an online-accessible link", "skills": [ { "id": "smart-page", "name": "腾讯文档智能页面", "description": "下一代可交互在线汇报网页生成工具，覆盖汇报 / 述职 / 周报 / 分享 / 项目复盘 / 数据看板等场景，生成即获永久在线链接", "skill_id": "skill_2053082927661453312" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 111, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-05-09T11:15:00+08:00", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "solo-dev-viral-teardown", "title": "竞品那条爆了，拆开它换成我的", "title_en": "That Competitor Post Went Viral — Take It Apart and Make It Yours", "subtitle": "抽出可复制的结构填自己的经历，并说清哪段你根本学不了", "subtitle_en": "Extract the reusable structure, fill it with your own story, and see exactly which part you can't copy", "description": "看到同类产品的内容火了想借鉴，直接照抄往往没用——因为爆款的流量里有一半来自你拿不到的东西（时机、已有粉丝基础）。这个案例把一条爆款拆成钩子、论证、CTA 三段，逐段说明手法并给可复制性分级，明确区分「可复制的结构」「不可复制的时机」「不可复制的资产」，然后抽出模板骨架、给出用你自己产品和经历填充的选题变体，最后评估换到你的账号上会衰减多少、怎么补。不做逐句改写和洗稿：引用原文仅截 ≤15 字短语用于说明句式，选题变体全部基于你自己的素材重写。", "description_en": "When a competitor's post takes off, copying it rarely works — half of that reach came from things you can't get: timing, and an existing follower base. This case breaks a viral post into hook, argument and CTA, explains the technique behind each part and grades it for reusability, explicitly separating the copyable structure from the uncopyable timing and the uncopyable assets. It then extracts a template skeleton, produces topic variants filled with your own product and experience, and estimates how much performance will decay on your account and how to compensate. No paraphrasing or content laundering: original text is quoted only as phrases of 15 characters or fewer to illustrate sentence patterns, and every topic variant is rewritten from your own material.", "categories": [ "content" ], "tags": [ "爆款拆解", "结构提取", "选题变体", "独立开发", "对标分析" ], "tags_en": [ "Viral Teardown", "Structure Extraction", "Topic Variants", "Indie Dev", "Benchmark Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "爆款拆解与选题变体报告" } ], "prompt": "你是社媒爆款分析师。我是一人公司（独立开发者），做了自己的产品在做内容推广。我看到同类产品的一条内容火了，想学它的做法。\n\n我的情况：\n- 我的产品：{产品一句话介绍}\n- 我的目标用户：{目标用户是谁}\n- 我看到的那条爆款：{把标题和大致内容粘贴进来，或描述它讲了什么}\n- 它的数据表现：{点赞/收藏/评论量级，不知道可以写不清楚}\n- 发布平台：{小红书 / 抖音 / 公众号 / 视频号 / 其他}\n\n请帮我拆解它，输出这五部分：\n\n1. 它凭什么火：分开讲钩子、论证、CTA 三段各做对了什么，不要笼统说「内容好」\n2. 可复制性分级：把它的做法分成三类 —— 可复制的结构、不可复制的时机（比如蹭上了热点）、不可复制的资产（比如作者已有粉丝基础）。诚实告诉我哪些我学不了\n3. 可复用模板：抽出骨架，标明哪些位置填我自己的内容\n4. 给我的选题变体：3-5 个，用同样的结构但换成我的产品和经历，每个附一句话说明对应原爆款的哪个手法\n5. 换到我这里会衰减多少：它的受众和我的可能不同，说明哪些效果会打折、为什么\n\n三条硬要求：\n\n- 不要帮我洗稿或逐句改写。我要的是结构，不是它的内容\n- 引用原文只截关键短语（不超过 15 字）用于说明手法，不要整段复制\n- 选题变体必须用我自己的产品和经历来填，缺什么素材直接指出来，不要用原文内容凑\n\n输出为 HTML 卡片，单文件、CSS 内联，可直接双击打开。", "prompt_en": "You are a viral content analyst. I run a one-person company (indie developer) and I'm doing content to promote my own product. I saw a post from a similar product go viral and I want to learn how they did it.\n\nMy situation:\n- My product: {one-line product description}\n- My target users: {who the target users are}\n- The viral post I saw: {paste the title and roughly what it said, or describe what it covered}\n- How it performed: {likes / saves / comments, write \"not sure\" if you don't know}\n- Platform: {Xiaohongshu / Douyin / WeChat Official Account / Channels / other}\n\nPlease take it apart and give me these five parts:\n\n1. Why it worked: address the hook, the argument and the CTA separately, saying what each one got right — don't just tell me \"the content was good\"\n2. Reusability grading: sort what they did into three buckets — copyable structure, uncopyable timing (e.g. they rode a trend), and uncopyable assets (e.g. the author already had followers). Be honest about what I can't replicate\n3. Reusable template: extract the skeleton and mark which slots I fill with my own material\n4. Topic variants for me: 3-5 of them, same structure but swapped to my product and my experience, each with one line noting which technique from the original it maps to\n5. How much it will decay for me: their audience may differ from mine — spell out which effects will be discounted and why\n\nThree hard requirements:\n\n- Don't launder or paraphrase it line by line. I want the structure, not their content\n- Quote the original only as key phrases (15 characters or fewer) to illustrate technique — never copy whole passages\n- Topic variants must be filled with my own product and experience. If material is missing, point it out — don't pad with the original's content\n\nOutput as an HTML card: single file, inlined CSS, openable with a double click.", "skills": [], "experts": [ { "id": "viral-analysis-expert", "name": "包火·营销通爆款分析师", "description": "把火了的内容拆开：凭什么火、哪些能复制、哪些学不了。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "dd477ddeb3c9", "expert_id": "ex_cq0nuGKgPkcX" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 111, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T09:05:00Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "haina-iphone17-analysis", "title": "iPhone 17 深度分析", "title_en": "iPhone 17 Deep Analysis", "subtitle": "用海纳购物管家全面解析 iPhone 17 真实体验", "subtitle_en": "Comprehensive iPhone 17 review with Haina Shopping Assistant", "description": "想买新手机却被网上两极分化的评价搞得很纠结？吹的捧上天，黑的一文不值，根本看不到真实客观的使用感受。这个案例演示了如何把「iPhone 17 怎么样」这类问题交给海纳购物管家，自动搜索全网真实评测文章，从性能、影像、续航、信号、充电速度到长期使用的真实优缺点，全面分析后给出明确的购买建议。\n\n案例中展示了海纳购物管家的「商品总结」模式完整流程：意图解析 → 查询改写 → 全网内容搜索 → 按模板生成结构化分析报告。报告涵盖商品概览、核心性能、优势劣势、用户评价、适用场景、竞品对比、购买建议和避坑提示，帮助消费者理性判断是否值得入手。\n\n适合：正在纠结换手机的消费者、想要快速了解新品真实口碑的数码爱好者、需要为他人推荐手机的亲友、追求理性消费决策的购物达人。", "description_en": "Torn between polarized online reviews when buying a new phone? This case demonstrates how to hand off questions like \"How is iPhone 17?\" to the Haina Shopping Assistant, which automatically searches real reviews across the web and generates a comprehensive analysis covering performance, camera, battery life, signal strength, charging speed, and long-term pros and cons.\n\nThe case showcases the complete \"Product Summary\" workflow: intent parsing → query rewriting → web content search → structured analysis report generation. The report covers product overview, core performance, advantages and disadvantages, user reviews, applicable scenarios, competitor comparison, purchase recommendations, and anti-pitfall tips.\n\nTarget audience: Consumers hesitating about phone upgrades, tech enthusiasts seeking authentic product feedback, friends helping others choose phones, and savvy shoppers who value rational purchase decisions.", "categories": [ "search", "business" ], "tags": [ "手机选购", "iPhone", "商品分析", "数码评测", "购物决策" ], "tags_en": [ "Phone Shopping", "iPhone", "Product Analysis", "Tech Review", "Shopping Decision" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "iPhone 17 深度分析报告", "label_en": "iPhone 17 Deep Analysis Report" } ], "prompt": "请使用海纳购物管家的「商品总结」模式，帮我全面分析以下商品：\n\n商品名称：[填写你想了解的具体商品名称，如 iPhone 17、戴森吹风机等]\n\n要求：\n1. 自动识别「商品总结」意图，搜索全网真实评测和使用体验文章\n2. 按照商品总结模板输出完整分析报告，包含：商品概览、核心性能、卖点分析、优势与劣势、用户评价、适用场景、竞品对比、购买建议、避坑提示\n3. 所有信息必须基于搜索结果，不编造数据\n4. 没有商品搜索结果时不输出价格和购买链接\n5. 语言通俗易懂，聚焦购买决策关键信息\n\n示例：\n商品名称：iPhone 17", "prompt_en": "Please use the \"Product Summary\" mode of the Haina Shopping Assistant to comprehensively analyze the following product:\n\nProduct name: [Fill in the specific product you want to learn about, such as iPhone 17, Dyson hair dryer, etc.]\n\nRequirements:\n1. Automatically recognize the \"Product Summary\" intent and search the web for real review and user-experience articles\n2. Output a complete analysis report following the product summary template, including: product overview, core performance, selling-point analysis, pros and cons, user reviews, applicable scenarios, competitor comparison, purchase recommendations, and pitfall warnings\n3. All information must be based on search results; do not fabricate data\n4. When there are no product search results, do not output prices or purchase links\n5. Use plain, easy-to-understand language, focusing on the key information for purchase decisions\n\nExample:\nProduct name: iPhone 17", "skills": [ { "id": "haina-shopping-assistant", "name": "海纳购物管家", "description": "值得买开发的海纳AI购物管家，支持商品推荐、商品总结、商品对比、优惠好价四种消费决策模式", "skill_id": "skill_2061810811874226176" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 112, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-06-02T11:27:51Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-quick-sichuan-secretary", "title": "职位快搜：四川县域文秘岗全景", "title_en": "Quick Search: Clerical Posts Across Sichuan Counties", "subtitle": "10 个岗位、20 个名额，一页看完某个岗位方向在一个省的真实供给", "subtitle_en": "Ten posts, twenty openings — a whole province's supply for one role, on one page", "description": "想报文秘岗，但不知道全省到底有几个坑、都在哪、待遇差多少——大多数人只能翻公告一份份数。\n\n本案例按「公务员 + 四川 + 关键词=文秘」直查，命中 10 个岗位共 20 个名额，全部落在区县一级。逐条列出单位、地市、区县、招录人数、薪资区间与状态，做地市分布与薪资档位分析，并如实说明这批岗位没有报录比和进面分数据。\n\n适合：锁定某个岗位方向的考生、需要判断跨市备考必要性的人。", "description_en": "If you want a clerical post, you need to know how many exist province-wide, where they are, and how pay differs — most people count them announcement by announcement.\n\nThis case queries \"civil servant + Sichuan + keyword 文秘\", hitting 10 posts with 20 openings, all at county level. It lists unit, city, county, headcount, salary band and status for each, analyses geographic and pay distribution, and states honestly that none of these posts carry ratio or cutoff-score data.\n\nFor: candidates fixed on one role type who need to judge whether to prepare across cities.", "categories": [ "search" ], "tags": [ "四川", "文秘岗", "县域", "职位检索" ], "tags_en": [ "Sichuan", "clerical post", "county level", "position search" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "四川县域文秘岗全景清单", "label_en": "Full view of clerical posts in Sichuan counties" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查四川省公务员岗位里职位名称含「文秘」的岗位。要求：1) 真实检索并给出总数；2) 全部列出：岗位、招录单位、地级市、区县、招录人数、薪资区间、报名状态；3) 统计招录名额合计、覆盖了几个地级市、各市分别几个岗位；4) 分析薪资区间的档位分布，指出最高与最低相差多少；5) 如果这批岗位没有报录比或往年进面分数据，如实说明，不要估算。", "prompt_en": "Find Sichuan civil-service posts whose title contains \"文秘\" (clerical). Requirements: 1) run the query and give the total; 2) list them all with post, unit, prefecture-level city, county, headcount, salary band and application status; 3) report the total openings, how many cities are covered and how many posts fall in each; 4) analyse the salary bands and state the gap between highest and lowest; 5) if these posts lack competition-ratio or cutoff-score data, say so honestly rather than estimating.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 112, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "mogao-caves-mural-distribution", "title": "莫高窟壁画朝代与类别分布", "title_en": "Mogao Caves Mural: Dynasty & Category Distribution", "subtitle": "492 个有壁画洞窟 × 7 大题材的朝代与类别占比", "subtitle_en": "492 mural caves × 7 subject categories: dynasty & category share", "description": "敦煌壁画研究最深的需求就是把'朝代 × 类别'两个维度同时量化。这个案例演示如何从敦煌研究院公开学术统计出发，配合朝代饼图、壁画面积柱状图、题材饼图三张图，把'唐代独占 56.7%'与'经变画 ≈ 38%'两个核心事实打到屏幕上——哪怕探元文化遗产库此次返回结构性错误，公开数据也能撑起完整的量化叙事。\n\n适合：敦煌学与丝绸之路研究者、佛教艺术史研究者、石窟保护与博物馆数字化工程师、文化遗产保育与文旅研究者、考古与图像学专业学生。", "description_en": "Mural scholarship needs both dimensions—dynasty × category—at the same time. This case shows how to build a 3-chart dashboard (dynasty pie, area bar, category pie) from publicly-available Dunhuang Research Academy statistics. Even when the Tanyuan cultural-heritage MCP returned a structural error on `datasourceType=1`, this study surfaces the two central findings: Tang dynasty monopolizes 56.7% of caves and sutra illustrations account for ~38% of all murals.\n\nBest for: Dunhuang & Silk Road researchers, Buddhist art historians, cave-conservation digital engineers, cultural-heritage conservation & tourism students.", "categories": [ "data" ], "tags": [ "莫高窟", "敦煌", "壁画", "朝代分布", "数据看板" ], "tags_en": [ "Mogao Caves", "Dunhuang", "Murals", "Dynasty Distribution", "Data Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 数据看板报告（内置中英 i18n）" } ], "prompt": "请按「腾讯探元学术研究-量化分析」专家的工作流，研究「敦煌莫高窟壁画的朝代分布与题材类别占比」。\n\n研究范围：\n- 朝代维度：十六国/北凉、北魏、西魏、北周、隋、唐（含吐蕃）、五代、北宋、西夏、元 — 共 10 个朝代\n- 数据基础：735 个洞窟 / 492 个有壁画洞窟 / 4.5 万余平方米壁画 / 2,400 余身彩塑\n- 类别维度：佛教尊像画、佛经故事画、佛教史迹画、经变画、神怪画、供养人画像、装饰图案 — 共 7 大类\n\n交付要求（强制）：\n1. 探元文化遗产库（datasourceType=1）按朝代/类别分批查询（2—4 批次）\n2. WebSearch 1—2 次补充学术统计（敦煌研究院张伯元原表统计；中国国家文物局公开口径）\n3. 图表 ≥ 3 张：1 张饼图（朝代分布）+ 1 张柱状图（各朝壁画面积）+ 1 张饼图（题材类别占比）\n4. 四格式报告：Markdown 母版 → HTML → DOCX → PDF\n5. 每个数据章节'先图后文'；量化结论须可回溯到具体数据点；标注数据来源与统计口径；不含'建议'\n6. 关于接口失败：探元文化遗产库（datasourceType=1）此次返回'Error while extracting response for type [byte[]]'结构性错误，按 SOP '数据库暂未返回该维度数据'降级采用学界公开统计口径（如实标注，不伪造数据）", "prompt_en": "Use the 'Tencent Tanyuan Academic Research - Quantitative Analysis' expert workflow to research 'Dunhuang Mogao Cave murals: dynasty & subject category distribution'.\n\nResearch scope:\n- Dynasty dimension: 10 periods — Sixteen-Kingdoms/Northern Liang, Northern Wei, Western Wei, Northern Zhou, Sui, Tang (incl. Tubo), Five Dynasties, Northern Song, Western Xia, Yuan\n- Data baseline: 735 caves / 492 mural-bearing caves / 45,000+ sq.m of murals / 2,400+ painted sculptures\n- Category dimension: 7 classes — buddha-iconography, jataka, Buddhist hagiography, sutra illustration, deity imagery, donor portraits, decorative patterns\n\nDelivery requirements (mandatory):\n1. Tanyuan cultural-heritage MCP (datasourceType=1) batches by dynasty/category\n2. 1-2 WebSearch passes for academic context (Zhang Boyuan's Dunhuang archive table; National Cultural Heritage Administration statistics)\n3. ≥3 charts: pie (dynasty) + bar (mural area by dynasty) + pie (subject categories)\n4. Four-format report: Markdown master → HTML → DOCX → PDF\n5. Each chapter 'figure-first then text'; quantitative conclusions traceable to data points; labeled source & calibre; no 'recommendations'\n6. On API failure: Tanyuan cultural-heritage DB returned 'Error while extracting response for type [byte[]]' structural error → fallback to public Dune Research Academy statistics with explicit annotation", "skills": [], "experts": [ { "id": "tanyuan-deep-research-quantitative", "name": "文博可视化分析专家", "description": "以腾讯探元文物数据库为核心信息源，按用户维度分批多次查询，交付图表先行的数据看板报告。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "13bdc1132eb4", "expert_id": "ex_WaXheY0OvJSu" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 112, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T17:30:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "solo-dev-style-training", "title": "一眼看出是 AI 写的，养成个人专属风格", "title_en": "Everyone Can Tell It's AI-Written — Build a Voice That's Yours", "subtitle": "没写过东西也能定风格，五个问题聊出来再逐项挑掉 AI 味", "subtitle_en": "Five questions to define your voice even with nothing written yet, then strip the AI tells one by one", "description": "用 AI 编程做产品的人，做内容也习惯让 AI 生成，结果发出来一眼假——而个人产品恰恰靠人味建立信任。这个案例针对「还没写过内容、没有样本可提取」的情况，用五问访谈直接聊出初始风格档案（语气人称、开场习惯、论证方式、可信度来源、风格禁区），然后逐项指出你那段内容的 AI 味具体在哪个词哪个句式、为什么显得假、怎么改，给出前后对照改写。特别处理开发者的语气平衡问题：技术味不能全删，全删会连可信度一起删掉，给出该去该留的判断标准。缺真实细节的地方明确指出，不代为编造。", "description_en": "People who build products with AI coding tools tend to let AI write their content too — and it reads fake on arrival, which is fatal when a personal product runs on trust. Built for the case where you have written nothing and have no samples to mine, this case runs a five-question interview to produce an initial voice profile (tone and person, opening habits, how you argue, where credibility comes from, and your hard no-go zones), then points at the exact words and sentence patterns that make your draft read like AI, why each one reads fake, and how to fix it, with a before-and-after rewrite. It handles the developer's tone-balance problem specifically: technical voice must not be stripped entirely, because stripping it takes your credibility with it — so you get clear criteria for what to cut and what to keep. Where real detail is missing it says so instead of inventing it.", "categories": [ "content" ], "tags": [ "去 AI 味", "风格养成", "五问访谈", "独立开发", "改稿" ], "tags_en": [ "De-AI Writing", "Voice Building", "Five-Question Interview", "Indie Dev", "Rewriting" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "个人风格档案与去 AI 味改写" } ], "prompt": "你是内容风格教练。我是一人公司（独立开发者），做了自己的产品在做内容推广。我发的内容一看就是 AI 写的，想改成自己说话的样子。\n\n我的情况：\n- 我的产品：{产品一句话介绍}\n- 我之前写过内容吗：{写过几篇，可以贴 1-2 篇 / 没写过，都是 AI 生成的}\n- 我不想变成哪种账号：{你反感的表达方式或账号类型}\n- 需要改的这段内容：{把你觉得 AI 味重的那段贴进来}\n\n请帮我做三件事：\n\n1. 先问我五个问题定风格：如果我没有作品可参考，用访谈的方式问我五个问题（一次问完，不要来回聊），从我的回答里提炼出初始风格档案：语气人称、开场习惯、论证方式、可信度来源、以及我不能说的话\n2. 逐项指出这段的 AI 味在哪：不要笼统说「太生硬」，要指到具体的词和句式，说明为什么这样写会显得假，怎么改\n3. 按我的风格重写一版：给出前后对照，并说明每处为什么这么改\n\n三条硬要求：\n\n- 技术味不要全删。我是开发者，保留一部分技术表达反而可信；请明确告诉我哪些该去、哪些该留\n- 改写必须用我自己的真实经历补细节。如果我提供的信息不够，直接告诉我缺什么，不要自己编具体数字和场景\n- 最后给我一份可以复用的风格档案，下次我说「按我的风格写」就能直接用\n\n输出为 HTML 卡片，单文件、CSS 内联，可直接双击打开。", "prompt_en": "You are a content style coach. I run a one-person company (indie developer) and I'm doing content to promote my own product. What I post obviously reads as AI-written and I want it to sound like me.\n\nMy situation:\n- My product: {one-line product description}\n- Have I written content before: {written a few pieces, can paste 1-2 / never written any, all AI-generated}\n- The kind of account I don't want to become: {styles or account types you dislike}\n- The piece that needs fixing: {paste the passage that reads most like AI}\n\nPlease do three things:\n\n1. Ask me five questions first to define my voice: if I have no writing to reference, interview me with five questions (all at once — no back and forth) and distill an initial voice profile from my answers: tone and person, opening habits, how I argue, where my credibility comes from, and the things I must never say\n2. Point out exactly where this piece reads like AI: don't just say \"too stiff\" — point at the specific words and sentence patterns, explain why they read fake, and how to fix them\n3. Rewrite it in my voice: show it before and after, and explain why each change was made\n\nThree hard requirements:\n\n- Don't strip out all the technical voice. I'm a developer — keeping some technical language actually makes me credible; tell me explicitly what to cut and what to keep\n- The rewrite must fill in detail from my real experience. If what I gave you isn't enough, tell me what's missing — don't invent specific numbers or scenes\n- End with a reusable voice profile, so next time I can just say \"write it in my voice\" and use it directly\n\nOutput as an HTML card: single file, inlined CSS, openable with a double click.", "skills": [], "experts": [ { "id": "style-training-expert", "name": "赵写·营销通内容风格教练", "description": "把你自己怎么说话变成可复用的风格档案，写出来像本人不像 AI。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "cccafd5caf3d", "expert_id": "ex_eNI9E1wUgwWK" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 112, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-25T09:10:00Z", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "breakfast-cup-pricing-promo", "title": "早餐杯定价与促销决策", "title_en": "Breakfast Cup Pricing and Promotion Decision", "subtitle": "从成本地板到促销保本增量的确定性测算", "subtitle_en": "Deterministic modeling from price floor to promotion break-even lift", "description": "新品定价不能只看竞品，也不能只把成本乘一个毛利率；平台费、退货损耗、支付意愿和促销让利会同时改变答案。\n\n这个案例以七杯装燕麦谷物早餐杯为匿名商品，通过商品定价与促销组合 Skill 的确定性脚本测算贡献毛利保本价、完全成本保本价、目标毛利价、支付意愿区间、建议价格点和三种促销机制的保本增量与风险，最终生成可离线打开的 HTML 决策报告。\n\n适合：新品负责人、电商运营、商品经理、经营分析和需要在活动前核算毛利底线的团队。", "description_en": "New-product pricing cannot rely only on competitors or cost-plus margins. Platform fees, return losses, willingness to pay, and promotional discounts all change the answer.\n\nUsing an anonymized seven-cup oat breakfast product, this case applies the Product Pricing and Promotion Strategy Skill's deterministic engine to calculate contribution and full-cost break-even prices, target-margin price, willingness-to-pay range, recommended price points, and the break-even lift and risk of three promotion mechanisms. The final deliverable is a self-contained offline HTML decision report.\n\nIdeal for: new-product owners, e-commerce operators, merchandise managers, business analysts, and teams checking margin floors before campaigns.", "categories": [ "business" ], "tags": [ "商品定价", "促销测算", "毛利分析", "电商运营" ], "tags_en": [ "product pricing", "promotion modeling", "margin analysis", "e-commerce operations" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "定价与促销决策报告" } ], "prompt": "请为七杯装燕麦谷物早餐杯制定平台电商新品定价与促销组合。已确认输入：单位变动成本 31.60 元，其中货品 22 元、包装 2.60 元、快递 7 元；平台佣金 5%、支付手续费 0.6%；退货率约 4%，货值回收率 80%，逆向物流 6 元；月固定投入 3 万元、月销量预期 2000 件。当前测试价 79 元，竞品价格为 69 元和 89 元，市场价格带按 49—109 元参考，目标贡献毛利率不低于 40%。单点支付意愿依次为太便宜 49 元、划算 59 元、有点贵 79 元、太贵 99 元；Gabor-Granger 的五个案例运营估计为 59/69/79/89/99 元对应购买意愿 68%/55%/40%/25%/14%，月度潜在可触达客群按 15000 人、陈述意愿折减系数按 55%，均标 C 级并需真实调研校准。经济价值估计为便利性增益 8 元、品牌认知弱势 3 元、价值捕获率 40%，同样为 C 级。没有历史销量序列，弹性只能用 C 级基准 -1.8。对比三种促销：满 159 减 20 按活动客单 176 元测算；赠送成本 6 元、感知标价 29 元的摇摇杯，预计销量增量 18%（已确认的运营估计）；新客满 69 减 10 券按测算客单 79 元、预计核销率 25% 计算。活动客单、感知标价和核销率均为 B/C 级运营估计，执行前需用真实数据复核。所有金额与 ROI 必须由官方脚本测算，并保留证据等级、假设与价格合规提示。", "prompt_en": "Create an e-commerce new-product pricing and promotion strategy for a seven-cup oat breakfast product. Confirmed inputs: unit variable cost RMB 31.60, comprising RMB 22 product cost, RMB 2.60 packaging, and RMB 7 shipping; 5% platform commission and 0.6% payment fee; approximately 4% returns, 80% product-cost recovery, and RMB 6 reverse logistics; RMB 30,000 monthly fixed investment and 2,000 expected monthly units. The current test price is RMB 79, competitors are RMB 69 and RMB 89, the reference market band is RMB 49–109, and the target contribution margin is at least 40%. A single-point willingness-to-pay estimate is RMB 49 too cheap, RMB 59 good value, RMB 79 expensive but considered, and RMB 99 too expensive. Case planning estimates for Gabor-Granger are prices RMB 59/69/79/89/99 with purchase intent of 68%/55%/40%/25%/14%, a monthly reachable audience of 15,000, and a stated-intent discount factor of 55%; grade all of these C and require calibration with real research. The economic-value assumptions are RMB 8 convenience uplift, RMB 3 brand-awareness disadvantage, and a 40% capture ratio, also grade C. There is no historical sales series, so elasticity must remain a grade-C benchmark of -1.8. Compare three promotions: RMB 20 off RMB 159 modeled with an RMB 176 promotional basket; a shaker gift costing RMB 6 with RMB 29 perceived value and an operator-confirmed estimated unit lift of 18%; and a RMB 10 new-customer coupon with an RMB 69 threshold, RMB 79 modeled basket, and 25% estimated redemption. Treat the promotional basket, perceived value, and redemption rate as grade-B/C operating estimates that require real-data validation before execution. All financial figures and ROI must be calculated by the official script, with evidence grades, assumptions, and pricing-compliance warnings retained.", "skills": [ { "id": "lingyi-pricing-promo-strategy", "name": "商品定价+促销组合", "description": "【零一数科·出品】用确定性测算引擎生成定价与促销决策报告。", "skill_id": "skill_2093645368646750208" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 113, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-29T08:00:00Z", "updated_at": "2026-08-31T16:31:11+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "marketing-brand-messaging-v2", "title": "品牌 Messaging 手册", "title_en": "Brand Messaging Playbook", "subtitle": "从品牌审计到分受众话术治理的一体化品牌手册", "subtitle_en": "An integrated brand playbook from messaging audit to audience-specific talk tracks", "description": "很多成长型 B2B 品牌并不是没有内容，而是官网、公众号、销售话术、PR 表达各说各话——亮点很多，却没有被组织成一套一致、可复用、可被销售和传播放大的品牌语言。本案例演示营销战役团队如何为典型成长型 B2B 科技 / 企业服务品牌完成一次品牌 Messaging 重塑：从竞品与标杆扫描、品牌问题诊断、Message House 搭建、Proof Points 体系、分受众话术库，到官网 / 公众号 / 销售 / 媒体 / 路演模板，最终形成 Markdown 完整手册 + HTML 品牌展示版。\n\n适合：品牌负责人、市场负责人、创始人、内容团队、销售 enablement 团队、B2B 增长团队", "description_en": "Many growing B2B brands do not suffer from a lack of content, but from fragmented language across the website, social channels, sales decks, and PR: too many good points, with no single system organizing them into reusable brand messaging. This case shows how the Marketing Campaign Team rebuilds a Messaging v2.0 system for a typical growth-stage B2B technology or enterprise-service brand: from competitor and benchmark scans, messaging diagnosis, Message House design, and proof-point architecture, to audience-specific talk tracks and templates for the website, official content, sales, media, and investor narratives. Outputs both a full Markdown handbook and an HTML brand-manual version.\n\nBest for: Brand Leads, Marketing Leads, Founders, Content Teams, Sales Enablement Teams, B2B Growth Teams", "categories": [ "business", "creation", "agent" ], "tags": [ "品牌审计", "Messaging", "B2B 品牌", "话术体系", "品牌手册" ], "tags_en": [ "brand audit", "messaging", "B2B brand", "talk tracks", "brand handbook" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "品牌手册 HTML 展示版" }, { "file": "output.md", "label": "品牌手册 Markdown 完整版" } ], "prompt": "请完成一次品牌审核与 Messaging 体系重塑，适用于典型成长型 B2B 科技 / 企业服务品牌：\n1. 扫描 2-3 个竞品 / 标杆品牌的定位与 messaging 结构\n2. 审计官网 / 公众号 / 销售话术中的 messaging 一致性问题\n3. 提出新的 Messaging 框架（定位句、三大支柱、痛点映射、Proof Points）\n4. 分受众输出话术库（C 端 / B 端 / 投资人 / 媒体）\n5. 补充官网 Hero、About、公众号文末、销售开场、媒体采访、投资人路演等标准文案模板\n\n输出《品牌 Messaging 体系 v2.0》Markdown 完整手册 + HTML 品牌手册版本。", "prompt_en": "Please complete a brand audit and messaging system rebuild, applicable to a typical growth-stage B2B technology / enterprise-service brand:\n1. Scan the positioning and messaging structure of 2-3 competitors / benchmark brands\n2. Audit messaging consistency issues across the website / official content channels / sales talk tracks\n3. Propose a new messaging framework (positioning statement, three pillars, pain-point mapping, proof points)\n4. Output audience-specific talk tracks (consumer / business / investor / media)\n5. Add standard copy templates for the website hero, About, end-of-article on official content channels, sales openers, media interviews, investor pitches, etc.\n\nOutput a complete \"Brand Messaging System v2.0\" Markdown handbook + an HTML brand-handbook version.", "skills": [], "experts": [ { "id": "MarketingCampaignTeam", "name": "营销战役团队", "description": "由营销总监领导的 4 人营销专家团队：内容创作者（博客/邮件/社媒/品牌声音）、活动策划师（战役策略/受众/渠道/预算）、SEO 专家（技术审计/内容优化/效果分析）和品牌分析师（竞品定位/品牌审核）。覆盖完整营销生命周期。", "expertType": "team", "source": "expert-marketplace", "name_en": "Marketing Campaign Team", "description_en": "A team of 4 marketing specialists led by a Marketing Director: Content Creator (blog posts/email/social media/brand voice), Campaign Planner (campaign strategy/audience/channels/budget), SEO Expert (technical audit/content optimization/performance analytics), and Brand Analyst (competitive positioning/brand review). Covers the full marketing lifecycle.", "expert_id": "ex_wg8SWSt1T33S" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 113, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:59:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "user-insight-marketing-ton", "title": "用营销通自己当案例：洞察报告长什么样", "title_en": "Eating Our Own Dog Food: A User Insight Report for Marketing Connect Itself", "subtitle": "以营销通（思研 + 金手指双平台）自身为演示对象——微信聊天、社群、评论、思研访谈四源语料，交叉验证产出洞察报告，看看自己的用户洞察闭环长什么样", "subtitle_en": "Using Marketing Connect itself as the demo product — four sources of corpus cross-validated into an insight report, showing the full insight loop on our own platform", "description": "营销通有两个平台：思研做调研、金手指做投放。那用营销通自己当产品，跑一遍「用户洞察报告」会得到什么？这个案例给出了答案：以营销通为演示对象，虚构了一套早期用户的反馈语料（微信聊天、一人公司社群、应用商店评论、思研 AI 访谈四源 12 条），演示技能如何交叉验证出「起点缺失」「小预算诉求」「报告后断档」三个多源结论，并产出人群线索、话题钩子和可直接引用的用户原话。给想搞懂营销通定位的内部同学，也给想看洞察报告实际效果的用户——你自己的产品，照着 prompt 里的语料格式替换即可。\n\n**关于演示语料**：Prompt 里内嵌的营销通用户反馈为虚构演示语料（4 个来源 12 条），分析框架与交叉验证逻辑是真实的；照着格式替换成你产品的真实反馈即可。", "description_en": "Marketing Connect has two platforms: SiYan for research and JinShouZhi for ad delivery. What happens if we run the User Insight Report skill on Marketing Connect itself? This case shows exactly that: a fictitious corpus of early-user feedback (WeChat chats, one-person-company communities, app-store reviews, and SiYan AI interviews — 12 items across 4 sources), cross-validated into three multi-source conclusions — missing starting point, small-budget demand, and the post-report gap — plus audience leads, topic hooks, and quotable user voices.\n\n**About demo corpus**: The embedded feedback is fictitious; the analysis framework and cross-validation logic are real. Replace with your own product's real feedback in the same format.", "categories": [ "business" ], "tags": [ "用户洞察", "营销通", "产品决策", "吃自己的狗粮", "产品定位" ], "tags_en": [ "user insight", "Marketing Connect", "product decisions", "dogfooding", "positioning" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "营销通用户洞察报告（自身演示）" } ], "prompt": "请把下面散落各处的用户反馈整合成一份完整的用户洞察报告（单文件 HTML，一次输出，不要反问我）。\n\n【我的产品】\n营销通 —— 面向小商家、一人公司与独立开发者的一站式营销平台，包含思研（AI 调研平台：发一个访谈链接即可回收真实用户反馈）和金手指（AI 广告投放系统：素材策略与投放咨询）。\n（替换为你的产品：{产品名 + 一句话介绍}）\n\n【语料一：微信聊天】（示例语料，替换为你的真实记录）\n对话 A（朋友，一人公司，卖 Notion 模板）：\n「上个月模板卖了两万三，全是靠社群自发。我试了下金手指，给的素材策略挺全，但我连广告账户都没有，卡在开户那一步就放弃了」「最想要的是你直接告诉我第一周花多少钱、投什么词」\n对话 B（朋友，淘宝手作店主）：\n「我不懂什么人群包定向，我就想知道每天 50 块能不能投，投了有没有人来问我」\n\n【语料二：社群/朋友圈反馈】（示例语料，替换为你的）\n「试了思研，发了 20 个链接回来 14 份，比我自己一个个约人聊快多了。问题是报告出来我不知道下一步干嘛」\n「金手指的素材策略 PDF 很专业，但对我这种连图都不会修的人，能不能直接把图也生成好」\n「从验证需求到投广告整个链路是通的，就是入口藏得有点深，我找金手指找了半天」\n\n【语料三：应用商店/评论区】（示例语料，替换为你的）\n评论 1（5 星）：「思研访谈是真省事，发到两个群里当天就回收了十几份，免费版额度够用。」\n评论 2（3 星）：「素材策略给了十几页，看不完。我只想知道先投哪一条。」\n评论 3：「什么时候出手机版？在外面想看投放数据只能等回家。」\n\n【语料四：思研 AI 访谈】（示例语料，替换为你的；暂无可整段删除）\n受访者 1（vibe coding 开发者，做了个 AI 记账工具）：\n「我用思研访谈了 10 个用户，发现根本没人需要自动记账，大家要的是月底能给我看花了多少。这一轮帮我省了三个月白做功能」「广告我肯定投，但我预算一个月就 500，怕全打水漂」\n受访者 2（一人公司，做知识付费）：\n「内容我自己能写，就是不知道发什么平台、什么时间发。要是能根据洞察直接排个内容日历我就付费」\n\n【分析要求】\n1. 严格区分「用户原话」与「推测」——原话单独列出并标注来源；推测必须标注为待验证\n2. 多源交叉验证——同一问题至少两个来源重复出现才标高频；单源问题标注「仅单一来源」；口头付费意愿标注「口头报价」；矛盾点（如\"策略要专业完整\" vs \"我只要先投哪条\"）单独列出并指出需要产品决策\n3. 报告包含六节：核心结论（≤3 条，标来源数）· 洞察明细（表格，含状态标签）· 用户原话（按主题分组，可直接引用，标来源）· 人群线索（人群×特征×话题钩子×卖点角度——这是给内容与广告的创意输入）· 当前未知 · 下一步建议（编号，按顺序做）\n4. 输出单文件 HTML：无外部资源引用，原话与推测要有视觉区分（不同底色/标签）\n5. 如果我的语料太少不足以分析，直接在报告「当前未知」节说明，并给我 3-5 个建议用思研（营销通专区的 AI 调研平台，发访谈链接即可回收反馈）补采的具体问题", "prompt_en": "Please integrate the scattered user feedback below into a complete user insight report as a single-file HTML, in one pass — do not ask me follow-up questions.\n\n[My product]\nMarketing Connect — a one-stop marketing platform for small merchants, one-person companies, and indie developers. It includes SiYan (AI research platform: send one interview link to collect real user feedback) and JinShouZhi (AI ad system: creative strategy and delivery consulting).\n(Replace with yours: {product name + one-liner})\n\n[Source 1: WeChat chats] (sample corpus — replace with yours)\nChat A (friend, one-person company selling Notion templates): \"Made 23k CNY last month, all organic from communities. Tried JinShouZhi — the creative strategy is thorough, but I don't even have an ad account. I gave up at account setup.\" \"What I want: just tell me how much to spend in week one and which keywords to bid.\"\nChat B (friend, handmade Taobao shop owner): \"I don't understand audience targeting. I just want to know: can I spend 50 CNY a day, and will people actually message me?\"\n\n[Source 2: Community feedback] (sample — replace with yours)\n\"Tried SiYan: sent 20 links, got 14 back — way faster than scheduling interviews myself. Problem: after the report, I don't know what to do next.\"\n\"JinShouZhi's strategy PDF is professional, but I can't even edit images — can't it just generate them for me?\"\n\"The full chain from validating demand to running ads works — but the entrance is buried. Took me forever to find JinShouZhi.\"\n\n[Source 3: App-store reviews] (sample — replace with yours)\nReview 1 (5 stars): \"SiYan interviews are effortless — posted to two group chats and collected a dozen responses same day. Free tier is enough.\"\nReview 2 (3 stars): \"The creative strategy is a dozen pages — can't finish it. I just want to know which one to run first.\"\nReview 3: \"When is the mobile version coming? Can't check ad performance until I'm home.\"\n\n[Source 4: SiYan AI interviews] (sample — replace with yours; delete if none)\nInterviewee 1 (vibe-coding developer, built an AI bookkeeping tool): \"I interviewed 10 users with SiYan and found nobody wants auto-bookkeeping — they want a month-end spend summary. That round saved me three months of building the wrong feature.\" \"I'll definitely run ads, but my budget is 500 CNY a month — afraid it all goes down the drain.\"\nInterviewee 2 (one-person company, knowledge products): \"I can write content myself — I just don't know which platform or what time to post. If it could turn insights into a content calendar, I'd pay.\"\n\n[Analysis requirements]\n1. Strictly separate verbatim quotes from speculation — quotes listed separately with sources; speculation marked as to-be-verified\n2. Cross-validate across sources — an issue counts as high-frequency only if it repeats in 2+ sources; single-source findings marked as such; stated willingness-to-pay marked as verbal offer; contradictions listed separately as product decisions\n3. Six sections: key conclusions (max 3, with source counts) / insight details (table with status tags) / user quotes (grouped by theme, citable) / audience leads (audience x traits x topic hooks x selling angles — creative input for content and ads) / current unknowns / next steps (numbered, in order)\n4. Single-file HTML: no external resources; visual distinction between quotes and speculation\n5. If corpus is insufficient, say so in the Current Unknowns section and give 3-5 specific follow-up questions to collect via SiYan", "skills": [], "experts": [ { "id": "user-insight-expert", "name": "言真·营销通用户洞察专家", "description": "把聊天记录、访谈、评论变成能开工的洞察：这些人是谁、真正卡在哪、下一步该去找什么样的人。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1b5ad88e42da", "expert_id": "ex_s9tpnTzewHOY" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 113, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-24T19:30:00+08:00", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "labor-contract-dispute-analysis", "title": "劳动合同纠纷法律风险分析", "title_en": "Labor Contract Dispute Legal Risk Analysis", "subtitle": "员工被以「不胜任工作」辞退的法律风险与维权路径分析", "subtitle_en": "Legal risk and rights protection analysis for employee dismissed for 'incompetence'", "description": "面对突然被公司辞退的困境，不知道自己的权益如何保障？本案例演示如何运用系统化法律推理，从管辖权、事实认定、法律争点、适用法律、风险评估到行动建议六个维度，全面分析劳动合同解除的法律风险。适合职场人士、HR从业者、法律专业学生及需要了解劳动法维权路径的人群。", "description_en": "Facing sudden dismissal and unsure about your rights? This case demonstrates systematic legal reasoning across six dimensions — jurisdiction, facts, issues, applicable law, risk assessment, and action recommendations — to comprehensively analyze legal risks in labor contract termination. Suitable for professionals, HR practitioners, law students, and anyone needing to understand labor law protection paths.", "categories": [ "education" ], "tags": [ "劳动法", "合同纠纷", "法律分析", "维权" ], "tags_en": [ "Labor Law", "Contract Dispute", "Legal Analysis", "Rights Protection" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "法律分析报告" } ], "prompt": "你是一位资深劳动法律师。公司以「不胜任工作」为由解除劳动合同，员工认为考核不合理。请从管辖权、事实认定、法律争点、适用法律、风险评估、行动建议六个维度进行系统分析。\n\n背景：员工在北京某互联网公司工作3年，最近一次绩效考核为C（不合格），公司依据内部规章制度解除劳动合同，未支付经济补偿金。\n\n输出要求：以 Markdown 格式输出完整法律分析报告，包含表格对比、风险矩阵和分点行动建议，不少于 2000 字。", "prompt_en": "You are a senior labor law attorney. A company terminated an employment contract on the grounds of \"inability to perform the job,\" and the employee believes the performance evaluation was unreasonable. Please conduct a systematic analysis across six dimensions: jurisdiction, fact-finding, legal points of dispute, applicable law, risk assessment, and recommended actions.\n\nBackground: The employee worked at an internet company in Beijing for 3 years, and the most recent performance evaluation was a C (unsatisfactory). The company terminated the employment contract based on internal rules and regulations, without paying severance.\n\nOutput requirements: Output a complete legal analysis report in Markdown format, including comparison tables, a risk matrix, and itemized action recommendations, no fewer than 2,000 words.", "skills": [ { "id": "legal-logic-analysis", "name": "Legal Logic Analysis", "description": "系统化法律推理分析（管辖权、事实、争点、适用法律、风险评估）", "skill_id": "skill_2053082264957227008" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 114, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-05-08T12:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "user-insight-multi-source", "title": "散落的用户反馈，整合成洞察报告", "title_en": "Turn Scattered User Feedback into an Insight Report", "subtitle": "微信聊天、社群反馈、应用商店评论、思研 AI 访谈——多源语料交叉验证，产出可直接用于内容与广告素材的用户洞察", "subtitle_en": "WeChat chats, community feedback, app-store reviews, and SiYan AI interviews — cross-validate multi-source corpus into user insights ready for content and ad creatives", "description": "产品做出来了，用户的反馈散落在微信聊天、社群、应用商店评论和访谈记录里，看不过来也串不起来？这个案例演示「用户洞察报告」技能如何把多源语料交叉验证，整合成一份结构化洞察报告：区分用户原话与推测、标注每个结论的来源、指出哪些问题在多个渠道重复出现。报告末尾直接给出人群线索、话题方向和可引用的用户原话，作为后续内容创作和广告素材的创意输入。没有语料？报告会先建议用思研（AI 调研平台）发一个访谈链接，低成本回收真实反馈后再分析。适合刚做出产品、想搞懂谁需要它、为什么需要的独立开发者和小团队。\n\n**关于演示语料**：Prompt 里内嵌了一套完整的示例语料（4 个来源，虚构演示用），照着格式替换成你的真实反馈即可；专家一次直接产出完整 HTML 报告，不会反问。", "description_en": "You shipped your product, but user feedback is scattered across WeChat chats, communities, app-store reviews, and interview notes — hard to read, harder to connect. This case shows how the User Insight Report skill cross-validates multi-source corpus into one structured insight report: separating verbatim quotes from speculation, labeling the source of every conclusion, and highlighting issues that repeat across channels. The report ends with audience leads, topic directions, and quotable user voices — creative inputs for your next content and ad creatives. No corpus yet? The report will first suggest running a SiYan AI interview to collect real feedback at low cost. Ideal for indie developers and small teams who just built a product and want to understand who needs it and why.\n\n**About demo corpus**: The shown output.html uses fictitious placeholder corpus (10 items across 4 sources). After copying this prompt, replace with your real multi-source feedback to get the same analysis.", "categories": [ "business" ], "tags": [ "用户洞察", "多源整合", "思研", "创意输入", "独立开发" ], "tags_en": [ "user insight", "multi-source", "SiYan", "creative input", "indie dev" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "多源用户反馈洞察报告" } ], "prompt": "请把下面散落各处的用户反馈整合成一份完整的用户洞察报告（单文件 HTML，一次输出，不要反问我）。\n\n【我的产品】\n简单清单 —— 给个人和小团队的轻量任务管理工具，核心是\"打开就能用\"：每人每天聚焦最重要的三件事，小团队版支持任务分配与完成状态一眼可见。\n（替换为你的产品：{产品名 + 一句话介绍}）\n\n【语料一：微信聊天】（示例语料，替换为你的真实记录）\n对话 A（朋友，自由职业设计师）：\n「上次你说那个清单工具我用了两天，最大的问题是我懒得打开它。要是能微信里直接提醒我今天的三个事就好了」「别的都挺好，就是手机上打开还要等加载，等一下我就不想看了」\n对话 B（前同事，3 人小团队负责人）：\n「我们组现在任务全在群里喊，谁做完了也不说。你那个团队版能不能让人勾一下就显示谁没做？」「价格别太贵，我们这种小团队一年人均两三百以内可以接受」\n\n【语料二：社群/朋友圈反馈】（示例语料，替换为你的）\n「试了一下，跟 Notion 比最大的优势就是不用配置，打开就能记。但团队功能还比较初级，只能标记完成，看不到任务流转。」\n「这类工具最大的坑是坚持不下来。我自己做过的清单 App 死于没人第二天再打开。」\n\n【语料三：应用商店/评论区】（示例语料，替换为你的）\n评论 1（5 星）：「终于有不用学习的清单了。试用过滴答、Todoist 都太复杂，这个 30 秒上手不是吹的。」\n评论 2（3 星）：「功能太简单了，连重复任务都没有。当作轻量临时清单可以，主力工具不行。」\n评论 3：「请问什么时候出安卓平板适配？」\n\n【语料四：思研 AI 访谈】（示例语料，替换为你的；暂无可整段删除）\n受访者 1（独立开发者）：「我不是不知道要做计划，是每天打开工具这个动作本身就有阻力。要是早上微信推给我，我点一下确认，我可能真能坚持。」「付费的话一年一百以内我肯定付。」\n受访者 2（全职妈妈，手工副业）：「我的单子散在微信、小红书私信和一个本子上，经常漏发货。我就想看见今天必须做的三件事，手机上顺手点掉就行。」「别给我推协作功能，我一个人用，看到团队功能反而觉得不是给我做的。」\n\n【分析要求】\n1. 严格区分「用户原话」与「推测」——原话单独列出并标注来源；推测必须标注为待验证\n2. 多源交叉验证——同一问题至少两个来源重复出现才标高频；单源问题标注「仅单一来源」\n3. 报告包含六节：核心结论（≤3 条，标来源）· 洞察明细（表格，含状态标签）· 用户原话（可直接引用，标来源）· 人群线索（人群×特征×话题钩子×卖点角度——这是给内容与广告的创意输入）· 当前未知 · 下一步建议（编号）\n4. 输出单文件 HTML：无外部资源引用，原话与推测要有视觉区分（不同底色/标签）\n5. 如果我的语料太少不足以分析，直接在报告「当前未知」节说明，并给我 3-5 个建议用思研（营销通专区的 AI 调研平台，发访谈链接即可回收反馈）补采的具体问题", "prompt_en": "Please integrate the scattered user feedback below into a complete user insight report as a single-file HTML, in one pass — do not ask me follow-up questions.\n\n[My product]\nSimple List — a lightweight task manager for individuals and small teams. Core idea: \"just open and use\" — three most important things per day; the team version shows task assignment and completion at a glance.\n(Replace with yours: {product name + one-liner})\n\n[Source 1: WeChat chats] (sample corpus — replace with yours)\nChat A (friend, freelance designer): \"Used it for two days. Biggest problem: I'm too lazy to open it. If it could remind me of today's three things right in WeChat, that'd be great.\" \"Everything else is fine, but mobile loading is slow — once I wait, I don't want to look.\"\nChat B (ex-colleague, 3-person team lead): \"Our team shouts tasks in group chat, no one says when done. Can your team version show who hasn't done with one tick?\" \"Keep price under 200-300 CNY per person per year.\"\n\n[Source 2: Community feedback] (sample — replace with yours)\n\"Tried it. Biggest advantage over Notion: zero setup, open and write. But team features are basic — only completion marks, no task flow.\"\n\"The biggest pit of such tools is retention. My own to-do app died because nobody opened it on day two.\"\n\n[Source 3: App-store reviews] (sample — replace with yours)\nReview 1 (5 stars): \"Finally a to-do list that needs no learning. Tried TickTick and Todoist — too complex. 30-second onboarding is real.\"\nReview 2 (3 stars): \"Too simple — no recurring tasks. Fine as a light temporary list, not a main tool.\"\nReview 3: \"When will Android tablet support land?\"\n\n[Source 4: SiYan AI interviews] (sample — replace with yours; delete if none)\nInterviewee 1 (indie developer): \"It's not that I don't know I should plan — the act of opening the tool itself is the friction. If it pushed to my WeChat in the morning and I just tap to confirm, I could actually stick with it.\" \"Under 100 CNY a year, I'd pay.\"\nInterviewee 2 (full-time mom, handcraft side business): \"My orders scatter across WeChat, Xiaohongshu DMs and a notebook — I miss shipments. I just want to see today's three must-dos and tick them off on my phone.\" \"Don't push collaboration features — I use it alone; seeing team features makes me feel it's not for me.\"\n\n[Analysis requirements]\n1. Strictly separate \"verbatim quotes\" from \"speculation\" — quotes listed separately with sources; all speculation marked to-be-validated\n2. Cross-validate across sources — flag high-frequency only if repeated in 2+ sources; single-source findings labeled \"single source only\"\n3. Report contains six sections: core conclusions (max 3, with sources) · insight details (table with status tags) · quotable user voices (with sources) · audience leads (audience × traits × topic hooks × sellable angles — creative inputs for content and ads) · current unknowns · numbered next steps\n4. Output a single-file HTML with no external resources; quotes and speculation visually distinct (different backgrounds/tags)\n5. If my corpus is too thin, state so in \"current unknowns\" and give 3-5 specific questions to collect via SiYan (the Marketing Hub AI research platform — send one interview link to collect feedback)", "skills": [], "experts": [ { "id": "user-insight-expert", "name": "言真·营销通用户洞察专家", "description": "把聊天记录、访谈、评论变成能开工的洞察：这些人是谁、真正卡在哪、下一步该去找什么样的人。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1b5ad88e42da", "expert_id": "ex_s9tpnTzewHOY" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 114, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-24T18:40:00+08:00", "updated_at": "2026-08-25T20:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "wechat-store-30day-launch", "title": "微信小店 30 天冷启动计划", "title_en": "30-Day WeChat Store Launch Plan", "subtitle": "私域破零、推客分发与素材测试的执行界面", "subtitle_en": "An execution interface for private-domain launch, affiliates, and creative testing", "description": "微信小店新店最常见的问题不是“没有流量技巧”，而是没有先跑通关联账号、支付履约、私域破零和首批成交数据。\n\n这个案例为一个早餐食品新店设计 30 天冷启动方案，通过微信小店运营 Skill 的测算、违禁词扫描和 HTML 渲染链路，交付包含目标倒推、三条增长路径、四周甘特图、逐周动作、责任人、完成标准、风险项和开店自检的交互式页面。\n\n适合：微信小店新商家、视频号电商运营、私域团队和需要把冷启动任务拆到周与责任人的小型团队。", "description_en": "The most common new-store problem on WeChat is not a lack of traffic tricks, but failure to establish linked accounts, payment and fulfillment, private-domain first orders, and initial transaction signals.\n\nThis case creates a 30-day launch plan for a breakfast-food store. Using the WeChat Store Ops Skill's calculator, compliance scanner, and HTML renderer, it delivers an interactive interface covering target decomposition, three growth paths, a four-week Gantt, weekly actions, owners, completion criteria, risk flags, and a store-launch checklist.\n\nIdeal for: new WeChat Store merchants, WeChat Channels commerce operators, private-domain teams, and small teams assigning launch work by week and owner.", "categories": [ "business" ], "tags": [ "微信小店", "冷启动", "私域破零", "电商增长" ], "tags_en": [ "WeChat Store", "cold start", "private-domain first sales", "e-commerce growth" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "30 天冷启动交互方案" } ], "prompt": "请为主营七杯装燕麦谷物早餐杯的微信小店制定 30 天冷启动方案。店铺处于新店未破零阶段，客单价约 79 元；现有触点包括约 8000 名公众号粉丝、5 个社群共约 900 人、企微好友约 1500 人，尚未开通视频号。30 天目标支付 GMV 为 5 万元，可投预算 1 万元，团队包括老板、运营、客服和仓储各一名。GMV 倒推采用案例规划假设 UV=8000、转化率=2.5%、客单价=79 元；69 元测试到手价的毛利测算采用货品及履约成本 31.60 元、技术服务费 3%、推客佣金 10%、退货损耗 2 元、运费已含在 31.60 元成本中。以上均不是历史实绩，真实执行前必须以后台费率、实际佣金和履约成本重跑官方脚本。方案必须先跑通关联视频号、运费与退货模板、有效电话、商户号和 1 元测试单，再依次安排私域预售破零、推客素材包、首场直播和小额素材投放。所有 GMV 倒推和毛利数字使用官方脚本；每个时间轴动作必须有责任人、完成标准和必要风险提示；最终交付自包含 HTML。", "prompt_en": "Create a 30-day launch plan for a new WeChat Store selling a seven-cup oat breakfast product. The store has not yet made its first sale, with an average order value of about RMB 79. Existing touchpoints include roughly 8,000 Official Account followers, five communities totaling about 900 people, and around 1,500 WeCom contacts; there is no WeChat Channels account yet. The 30-day paid-GMV target is RMB 50,000 with a RMB 10,000 test budget. The team consists of one owner, one operator, one customer-service person, and one warehouse person. Use case-planning assumptions of 8,000 UV, a 2.5% conversion rate, and RMB 79 AOV for GMV decomposition. For the RMB 69 delivered-price margin model, use RMB 31.60 product-and-fulfillment cost, a 3% technical service fee, 10% affiliate commission, RMB 2 return loss, and treat shipping as already included in the RMB 31.60 cost. These are planning assumptions, not historical performance; rerun the official script with actual backend rates, commissions, and fulfillment cost before execution. The plan must first establish the linked Channels account, shipping and return templates, valid contact numbers, merchant account, and a RMB 1 test order, then sequence private-domain presale for first orders, an affiliate asset pack, the first livestream, and small-budget creative testing. Use the official scripts for all GMV decomposition and margin figures. Every timeline action must include an owner, completion criterion, and necessary risk warning. Deliver a self-contained HTML report.", "skills": [ { "id": "lingyi-wechat-store-ops", "name": "微信小店运营方案", "description": "【零一数科·出品】把小店运营诉求转成带责任人、时间点和判定标准的可执行 HTML 方案。", "skill_id": "skill_2091129604447596544" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 114, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-29T08:00:00Z", "updated_at": "2026-08-31T16:31:11+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-competitor-monitor", "title": "Agent 竞品网站变更监控", "title_en": "Agent Competitor Website Change Monitoring", "subtitle": "多 Agent 协作监控竞品定价、功能与内容变更并生成可视化仪表盘", "subtitle_en": "Multi-Agent Workflow for Monitoring Competitor Pricing, Features & Content Changes with Visual Dashboard", "description": "竞品悄悄调整了定价策略、上线了新功能、或在博客中暗示了产品方向——等你从行业群聊中偶然得知，竞争窗口期可能已经过去了。手动定期检查竞品网站不仅耗时，而且很容易遗漏细微但关键的变化。\n\n这个案例演示了一个多 Agent 竞品监控工作流：采集 Agent 模拟抓取 Notion、Coda、Craft 三个竞品的定价页面、功能页面和博客更新日志；对比 Agent 将本次抓取结果与历史数据对比，识别价格调整、新增功能、下线功能和文案变化；分析 Agent 评估每项变更的影响等级（高/中/低）并生成竞品动态摘要；报告 Agent 生成包含变更时间线、详情卡片、影响标签和应对建议的可视化仪表盘。\n\n适合：产品经理、竞品分析师、市场研究员、战略规划团队、SaaS 创业公司的业务负责人。", "description_en": "Competitors quietly adjust pricing, launch new features, or hint at product direction in blog posts — by the time you hear about it casually, the competitive window may have passed. Manually checking competitor websites is time-consuming and easy to miss subtle but critical changes.\n\nThis case demonstrates a multi-agent competitor monitoring workflow: a Collection Agent simulates scraping Notion, Coda, and Craft's pricing pages, feature pages, and changelogs; a Comparison Agent identifies changes against historical data; an Analysis Agent evaluates impact levels; a Report Agent generates a visual dashboard with change timelines, detail cards, impact badges, and recommended actions.\n\nIdeal for: product managers, competitive analysts, market researchers, strategy teams, and SaaS startup leads.", "categories": [ "agent", "search", "business" ], "tags": [ "竞品分析", "网站监控", "多Agent协作", "商业智能" ], "tags_en": [ "Competitive Analysis", "Website Monitoring", "Multi-Agent", "Business Intelligence" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我需要一个 Agent 工作流来监控 3 个竞品的网站变更。竞品：Notion、Coda、Craft。请设计流程：1) 采集 Agent 模拟抓取三个竞品的定价页面、功能页面和博客/更新日志；2) 对比 Agent 将本次抓取结果与上次（模拟的历史数据）对比，识别变更项（价格调整/新功能/下线功能/文案变化）；3) 分析 Agent 评估变更的影响等级（高/中/低）并生成竞品动态摘要；4) 报告 Agent 生成可视化的竞品监控仪表盘。产出一个 HTML 仪表盘，包含三个竞品的变更时间线、变更详情卡片、影响等级标签和应对建议。", "prompt_en": "I need an Agent workflow to monitor changes on the websites of 3 competitors. Competitors: Notion, Coda, Craft. Please design the workflow: 1) A Collection Agent simulates scraping each competitor's pricing page, features page, and blog/changelog; 2) A Comparison Agent compares the latest scrape with the previous one (simulated historical data) and identifies changes (price adjustments / new features / discontinued features / copy changes); 3) An Analysis Agent assesses the impact level of each change (high/medium/low) and generates a competitor activity summary; 4) A Reporting Agent generates a visual competitor monitoring dashboard. Produce an HTML dashboard containing a change timeline for the three competitors, change detail cards, impact-level tags, and recommended responses.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 115, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "竞品监控仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "marketing-seo-audit-xxx", "title": "B2B 官网 SEO 审计", "title_en": "B2B Website SEO Audit", "subtitle": "技术诊断到 90 天优化路线图的一体化审计案例", "subtitle_en": "An end-to-end SEO audit from technical diagnosis to a 90-day optimization roadmap", "description": "很多 B2B 官网的问题不是“没做 SEO”，而是技术底座、栏目结构、关键词覆盖和转化路径彼此脱节——首页讲品牌、博客讲知识、产品页讲功能，但搜索流量进来之后并没有被有效导向商机页面。本案例演示营销战役团队如何为典型 B2B 官网 xxx.com 做完整 SEO 审计：梳理技术 SEO 风险、内容 SEO 短板、竞品内容差距、关键词机会地图、TOP 20 内容优化优先级，以及 90 天执行计划与 KPI 汇报机制。最终输出 Markdown 审计完整版 + HTML 可视化报告。\n\n适合：市场负责人、增长负责人、SEO 经理、内容负责人、B2B 官网运营、企业服务创始人", "description_en": "The real problem with many B2B websites is not a lack of SEO effort, but disconnected foundations: technical setup, site architecture, keyword coverage, and conversion paths all operate in silos. The homepage talks brand, the blog talks education, the product pages list features—yet incoming search traffic is not effectively guided toward pipeline-generating pages. This case shows how the Marketing Campaign Team conducts a full SEO audit for a typical B2B website, xxx.com: covering technical SEO risks, content SEO gaps, competitor content benchmarking, a keyword opportunity map, a prioritized top-20 content optimization list, and a 90-day execution plan with KPI reporting cadence. Outputs both a full Markdown audit and an HTML visualization-ready report.\n\nBest for: Marketing Leads, Growth Leads, SEO Managers, Content Leads, B2B Website Operators, B2B SaaS Founders", "categories": [ "business", "search", "agent" ], "tags": [ "SEO 审计", "B2B 官网", "关键词策略", "内容差距", "90 天计划" ], "tags_en": [ "SEO audit", "B2B website", "keyword strategy", "content gap", "90-day plan" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "SEO 审计 HTML 可视化版" }, { "file": "output.md", "label": "SEO 审计 Markdown 完整版" } ], "prompt": "我们有一个 B2B 官网（xxx.com），流量 6 个月没增长。请做完整 SEO 审计：\n- 技术 SEO（爬取/索引/速度/结构化数据）\n- 内容 SEO（现有内容表现、关键词覆盖、竞品对比）\n- 外链现状\n- 关键词机会地图\n- 现有 TOP 20 文章优化建议\n- 90 天 SEO 行动计划\n\n输出《官网 SEO 全面审计报告》Markdown + HTML 可视化（关键词机会热力图、内容差距矩阵）。", "prompt_en": "We have a B2B corporate website (xxx.com), and traffic hasn't grown in 6 months. Please do a complete SEO audit:\n- Technical SEO (crawling/indexing/speed/structured data)\n- Content SEO (performance of existing content, keyword coverage, competitor comparison)\n- Backlink status\n- A keyword opportunity map\n- Optimization recommendations for the current TOP 20 articles\n- A 90-day SEO action plan\n\nOutput a \"Comprehensive Website SEO Audit Report\" in Markdown + an HTML visualization (a keyword opportunity heatmap and a content gap matrix).", "skills": [], "experts": [ { "id": "MarketingCampaignTeam", "name": "营销战役团队", "description": "由营销总监领导的 4 人营销专家团队：内容创作者（博客/邮件/社媒/品牌声音）、活动策划师（战役策略/受众/渠道/预算）、SEO 专家（技术审计/内容优化/效果分析）和品牌分析师（竞品定位/品牌审核）。覆盖完整营销生命周期。", "expertType": "team", "source": "expert-marketplace", "name_en": "Marketing Campaign Team", "description_en": "A team of 4 marketing specialists led by a Marketing Director: Content Creator (blog posts/email/social media/brand voice), Campaign Planner (campaign strategy/audience/channels/budget), SEO Expert (technical audit/content optimization/performance analytics), and Brand Analyst (competitive positioning/brand review). Covers the full marketing lifecycle.", "expert_id": "ex_wg8SWSt1T33S" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 116, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:52:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-git-weekly-digest", "title": "Agent 自动汇总 Git 周报", "title_en": "Agent Automated Git Weekly Digest", "subtitle": "多 Agent 协作采集、分类、统计 Git 数据并生成可视化周报", "subtitle_en": "Multi-Agent Git Data Collection, Classification, Statistics & Visual Dashboard", "description": "每周五下午花两小时手动汇总团队周报？从 Git log 里数 commit、从 GitHub 上查 PR、从 Issue 列表里统计进度——重复劳动不说，还容易遗漏和出错？monorepo 项目模块多、人员多，手工整理根本理不清楚。\n\n本案例展示了一套多 Agent 协作的 Git 周报自动化工作流：采集 Agent 拉取本周所有 commit log、PR 合并记录和 Issue 关闭情况，分类 Agent 将变更按模块分类（组件库/业务页面/基础设施/文档）并标注 PR 类型（feature/fix/refactor/chore），统计 Agent 计算每人代码量、PR 数、Review 数和平均 PR 合并时间，呈现 Agent 生成包含数据可视化的 HTML 周报仪表盘。从原始 Git 数据到专业周报页面，全程自动完成。\n\n适合：前端/全栈团队 Leader、项目经理、Scrum Master、需要定期产出团队产出报告的技术管理者。", "description_en": "Spending hours every Friday manually compiling team weekly reports? Counting commits from git log, checking PRs on GitHub, tallying issue progress — tedious, error-prone, and never complete?\n\nThis case demonstrates a multi-Agent automated Git weekly report workflow: Collection Agent pulls commit logs, merged PRs, and closed issues, Classification Agent categorizes changes by module and PR type, Statistics Agent calculates per-developer metrics, Presentation Agent generates an HTML dashboard with data visualization.\n\nIdeal for: frontend/full-stack team leads, project managers, Scrum Masters, and tech managers needing regular team output reports.", "categories": [ "agent", "dev", "office" ], "tags": [ "Git 周报", "团队统计", "可视化报表", "Agent 自动化" ], "tags_en": [ "Git weekly report", "team statistics", "visual dashboard", "agent automation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请设计一个 Agent 工作流自动生成团队 Git 周报，产出一个 HTML 周报仪表盘（单文件 HTML，内联无外部依赖），界面文字用简体中文（PR/Commit 等术语可保留英文）。\n\n场景：5 人前端团队 monorepo。流程：\n1. 采集 Agent：拉取本周 commit、PR 合并、Issue 关闭（模拟数据）\n2. 分类 Agent：按模块（组件库/业务页面/基础设施/文档）分类，标注 PR 类型（feature/fix/refactor/chore）\n3. 统计 Agent：每人代码量、PR 数、Review 数、平均 PR 合并时间\n4. 呈现 Agent：生成可视化周报", "prompt_en": "Design an Agent workflow that automatically generates a team Git weekly report, producing an HTML weekly dashboard (single-file HTML, inlined with no external dependencies). The interface text should be in Simplified Chinese (terms like PR/Commit may remain in English).\n\nScenario: a 5-person frontend team monorepo. Workflow:\n1. Collection Agent: pull this week's commits, merged PRs, and closed issues (mock data)\n2. Classification Agent: categorize by module (component library / business pages / infrastructure / docs) and label PR type (feature/fix/refactor/chore)\n3. Statistics Agent: per-person lines of code, PR count, review count, and average PR merge time\n4. Presentation Agent: generate the visual weekly report", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 118, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "Git 周报仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "agent-auto-test-report", "title": "Agent 自动化测试报告生成", "title_en": "Agent Automated Test Report Generation", "subtitle": "多 Agent 协作完成测试执行、覆盖率分析与报告生成", "subtitle_en": "Multi-Agent Collaboration for Test Execution, Coverage Analysis & Report Generation", "description": "测试写了不少，但覆盖率报告只是一堆数字？失败用例需要手动排查根因？每次发版前花半天时间整理测试报告？Node.js 电商项目功能越来越多，测试套件越来越庞大，团队急需一套自动化方案来系统化管理测试质量。\n\n本案例展示了一套完整的多 Agent 协作测试工作流：分析 Agent 扫描项目结构识别测试文件与覆盖率目标，执行 Agent 运行 Jest 测试套件收集通过/失败/跳过的用例详情，覆盖率 Agent 分析各模块覆盖率并标注低覆盖率热点，报告 Agent 生成结构化测试报告——包含覆盖率矩阵、失败用例根因分析和具体修复建议。最终产出完整的 TypeScript 测试代码和 JSON 格式测试报告。\n\n适合：全栈开发工程师、QA 工程师、技术团队 Leader、希望提升测试自动化水平的 Node.js 团队。", "description_en": "Tests written but coverage reports are just meaningless numbers? Failed cases need manual root cause analysis? Spending half a day preparing test reports before every release?\n\nThis case demonstrates a complete multi-Agent collaborative testing workflow: Analysis Agent scans project structure and identifies test targets, Execution Agent runs Jest test suites and collects pass/fail/skip details, Coverage Agent analyzes per-module coverage and flags low-coverage hotspots, Report Agent generates structured test reports with coverage matrices, failure root cause analysis, and fix recommendations.\n\nIdeal for: full-stack developers, QA engineers, tech leads, and Node.js teams seeking better test automation.", "categories": [ "agent", "dev" ], "tags": [ "自动化测试", "Agent 协作", "测试报告", "覆盖率分析" ], "tags_en": [ "automated testing", "agent collaboration", "test report", "coverage analysis" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.ts", "prompt": "我需要一个 Agent 工作流来自动执行测试并生成报告。场景：一个 Node.js 电商项目。流程：1) 分析 Agent 扫描项目结构，识别测试文件和覆盖率目标；2) 执行 Agent 运行 Jest 测试套件，收集通过/失败/跳过的用例；3) 覆盖率 Agent 分析各模块覆盖率，标注低覆盖率热点；4) 报告 Agent 生成结构化测试报告（含覆盖率矩阵、失败用例根因分析、修复建议）。输出一个完整的 TypeScript 测试文件 + JSON 格式测试报告。", "prompt_en": "I need an Agent workflow that automatically runs tests and generates a report. Scenario: a Node.js e-commerce project. Workflow: 1) An Analysis Agent scans the project structure, identifying test files and coverage targets; 2) An Execution Agent runs the Jest test suite, collecting passed/failed/skipped cases; 3) A Coverage Agent analyzes per-module coverage and flags low-coverage hotspots; 4) A Report Agent generates a structured test report (including a coverage matrix, root-cause analysis of failed cases, and fix recommendations). Output one complete TypeScript test file plus a test report in JSON format.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 119, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.ts", "label": "自动化测试代码" }, { "file": "output.json", "label": "测试报告" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-limitboard-sentiment-cockpit", "title": "涨停情绪驾驶舱", "title_en": "Limit-Up Sentiment Cockpit", "subtitle": "用高密度卡片和信号面板呈现封板强度、情绪温度与高标接力。", "subtitle_en": "A high-density signal board for sealing strength, sentiment temperature, and ladder continuation.", "description": "盘后看涨跌停，最怕只有清单没有判断：到底是普涨修复、局部抱团，还是高标接力重新启动？这个案例把封板率、涨跌比、连板高度、封单榜和热点扩散整合成一个情绪驾驶舱，适合快速判断市场温度。\n\n你能直接看到一份高密度双语 HTML 看板：大盘温度、指数脉冲、连板塔、热点簇和封单集中区都在一屏到两屏内说清楚，便于交易团队盘后复盘或盘前快速预热。\n\n适合：短线交易员、情绪周期研究者、盘后复盘作者、需要快速抓强弱切换的人。", "description_en": "When reviewing limit-ups after the close, a raw list is never enough: was the market broadening, crowding into one pocket, or reigniting ladder continuation? This case packages seal rate, breadth ratio, ladder height, seal leaders, and hotspot diffusion into a sentiment cockpit built for fast judgment.\n\nThe output is a dense bilingual HTML dashboard where market temperature, index pulse, ladder structure, hotspot clusters, and seal concentration are all visible within one or two screens, making it useful for post-close debriefs or pre-open preparation.\n\nBest for: short-term traders, sentiment-cycle researchers, recap writers, and anyone tracking strength rotation.", "categories": [ "finance", "data", "search" ], "tags": [ "市场情绪", "驾驶舱", "封板率", "交易观察" ], "tags_en": [ "market sentiment", "cockpit", "seal rate", "trading monitor" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于最新一个有完整收盘数据的 A 股交易日，生成一份“涨停情绪驾驶舱”HTML，看清当日市场温度。用真实查询得到的涨停家数、封板未遂、涨跌比、连板高度、封单榜、热点板块与代表个股，突出强修复、分歧还是退潮的判断。页面需要高信息密度但保持明亮易读，风格偏轻科技、信号墙和斜切操作面板，适合交易团队盘后复盘或盘前快速查看；单文件 HTML、资源全内联、可离线打开。请保留数据日期、来源说明与风险声明。", "prompt_en": "Using the latest A-share trading day with complete closing data, generate a “Limit-Up Sentiment Cockpit” HTML page that reveals the day’s market temperature. Query real data for total limit-ups, failed seals, the up/down ratio, ladder height, seal leaders, hotspot sectors, and representative names, then judge whether the session reflected broad repair, divergence, or fading momentum. Keep the page bright but high-density, with a light-tech signal-wall feel and angular operation panels, suitable for a trading team’s post-close debrief or pre-open check. Deliver a single self-contained HTML file that works offline, and include the data date, source notes, and risk disclaimer.", "artifacts": [ { "file": "output.html", "label": "双语交付 HTML 报告" } ], "skills": [ { "id": "a-share-limitboard-report", "name": "A股涨跌停日报", "description": "A股涨跌停日报，收盘后汇总涨停/跌停、连板梯队、板型分布、主线板块、炸板与全量列表。", "skill_id": "skill_2076627892488605696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T21:22:40.164728+08:00", "updated_at": "2026-07-30T16:43:23.113221+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-shortline-game-desk", "title": "A股短线博弈台", "title_en": "A-share Shortline Game Desk", "subtitle": "先定周期，再看高低位与空仓检验", "subtitle_en": "Define the cycle first, then read high-vs-low setup and the empty-hand test", "description": "短线复盘最容易掉进一个坑：只看涨停数量，却看不清到底是谁在定高度、谁只是跟风补涨、谁又会成为第二天最先掉队的中位风险区。很多所谓“龙头战法”，问题都出在节奏而不是名字。\n\n这个案例演示了如何用 A 股收盘后的真实连板、封单、资金流和龙虎榜数据，做一份短线博弈复盘：先定周期，再区分绝对高位和低位试错，接着看换手、人气和空仓检验，最后输出次日纪律清单。\n\n适合：做短线题材与连板观察的人、关注龙头与低位试错节奏的交易者、需要把“今天到底能不能打板”讲清楚的复盘场景。", "description_en": "Shortline reviews often fall into the same trap: people count limit-ups without seeing who actually sets the ceiling, who is merely a sympathy follower, and which mid-ladder names are most likely to fail first the next day. In many so-called leader strategies, the real problem is timing rather than the chosen ticker.\n\nThis case shows how to turn real A-share ladder, seal-volume, flow and Dragon-Tiger data into a shortline game review: define the cycle first, separate absolute highs from low-base probes, then study turnover, popularity and the empty-hand test before ending with a disciplined next-session checklist.\n\nIdeal for traders tracking shortline themes and ladder names, anyone focused on timing high-level leaders versus low-base probes, and post-close review scenarios that need a precise answer to whether the day was truly playable.", "categories": [ "finance", "data" ], "tags": [ "A股复盘", "短线博弈", "连板", "龙虎榜", "交易纪律" ], "tags_en": [ "A-share Review", "Shortline Game", "Consecutive Boards", "Dragon-Tiger List", "Trading Discipline" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "短线博弈复盘 HTML报告" } ], "prompt": "请基于指定交易日做一份 A 股短线博弈复盘单文件 HTML。先判断当天是否为交易日；若休市直接说明。若有交易数据，请围绕情绪周期、绝对高位、低位试错、封单强度、主力资金、人气锚点和空仓检验，输出周期判断、样本梯队、龙虎榜观察、次日纪律与风险提示。所有资源内联、可离线打开，数据使用真实公开行情并标明统计日期。", "prompt_en": "Create a single-file HTML A-share shortline game review for a specified trade date. First confirm whether the day is a trading session; if the market is closed, say so directly. If data exists, focus on the sentiment cycle, absolute highs, low-base probes, seal strength, main-force flow, popularity anchors and the empty-hand test, and output a cycle judgment, representative ladder names, Dragon-Tiger observations, next-session discipline and risk notes. Fully self-contained for offline use, and based on real public market data with the statistics date shown clearly.", "skills": [ { "id": "a-share-daily-review", "name": "A股每日复盘", "description": "A股每日收盘后全维度复盘报告生成，支持五种分析风格，输出自包含HTML报告。", "skill_id": "skill_2076628582742429696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T20:12:45+08:00", "updated_at": "2026-07-27T20:12:45+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "founder-free-execution-system", "title": "创始人不在场的执行系统", "title_en": "An Execution System That Runs Without the Founder", "subtitle": "把反复拍板变成团队可复用的判断", "subtitle_en": "Turn Repeated Approvals into Reusable Team Judgement", "categories": [ "business", "education" ], "tags": [ "执行系统", "创始人瓶颈", "判断标准", "组织能力" ], "tags_en": [ "Execution System", "Founder Bottleneck", "Decision Standards", "Organizational Capability" ], "description": "一家 15 人内容服务创业公司里，创始人外出两天，项目排期、客户承诺和优先级便全部卡住。团队并非不愿意推进，而是在等待创始人逐项解释什么更重要、什么可以承诺、什么才算做对。\n\n本案例不把问题误诊为执行力不足，也不建议用更多例会和逐项审批加码。它回到更前面的判断：系统是否已经把确认过的判断、对错标准与上报边界固定进团队共同可见的环境。\n\n通过三项自检，创始人可以检查判断是否固定、成功标准是否清楚、升级边界是否明确，并把“我不在就停摆”的信号转化为建设可重复执行环境的起点。", "description_en": "At a 15-person content-services startup, the founder leaves for two days and project schedules, client commitments, and priorities all stall. The team is not unwilling to proceed; it is waiting for the founder to explain, item by item, what matters most, what can be promised, and what counts as doing it right.\n\nThis case does not misdiagnose the issue as weak execution, nor does it prescribe more meetings and item-by-item approvals. It returns to the earlier judgement: has the system fixed confirmed judgement, standards for right and wrong, and escalation boundaries in an environment visible to the whole team?\n\nThrough three self-checks, founders can test whether judgement is fixed, success standards are clear, and escalation boundaries are explicit—then turn the signal of “everything stops when I am away” into a starting point for building a repeatable operating environment.", "prompt": "你是一家 15 人内容服务创业公司的创始人。你外出两天后，项目排期、客户承诺和优先级全部卡住，团队说“老板不拍板就不知道怎么做”。你原本想增加每日例会和逐项审批。请不要把问题简单归为团队执行力不足，也不要给出空泛的管理技巧。请先诊断真正卡住的是什么：哪些已确认的判断仍只留在创始人脑中，团队缺少哪些共同的对错标准，以及什么情况下应该自行判断、什么情况下必须上报。然后围绕“创始人不在场时，事情是否还能以认可方式继续发生”设计一个最小的执行系统自检框架，包含判断是否固定、成功标准是否清楚、升级边界是否明确三项。最后把“盯人”的反应替换成“建系统”的动作，并用 1–2 个具体反问收尾，帮助创始人回到自己的位置重新看问题。不得虚构企业数据。", "prompt_en": "You are the founder of a 15-person content-services startup. After you leave for two days, project schedules, client commitments, and priorities all stall. The team says, “Without the boss deciding, we do not know what to do.” Your first instinct is to add daily meetings and item-by-item approvals. Do not reduce this to a simple lack of team execution, and do not offer generic management tips. First diagnose what is actually stuck: which confirmed judgements still exist only in the founder’s mind, which shared standards for right and wrong the team lacks, and which situations should be decided independently versus escalated. Then design a minimal execution-system self-check around this question: when the founder is absent, can work still proceed in a way the founder recognizes? Include three checks: whether judgement is fixed, whether success standards are clear, and whether escalation boundaries are explicit. Finally, replace the instinct to watch people with actions that build a system, and close with one or two concrete questions that bring the founder back to their own position. Do not invent company data.", "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "中英双语执行系统自检台" } ], "skills": [ { "id": "the-entrepreneurship-handbook", "name": "创业可以学", "description": "服务创业者和管理者，解答创业/商业/管理问题，引发深度思考", "skill_id": "skill_2053083248287678464" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 98, "usage_count": 0, "scenario": "", "scenario_use": "", "task_mode": "", "created_at": "2026-07-27T16:30:16+08:00", "updated_at": "2026-07-27T16:30:16+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "is_recommended": false, "recommended_order": 0 }, { "id": "glass-substrate-tgv-chain-report", "title": "玻璃基板TGV产业链", "title_en": "Glass Substrate TGV Chain", "subtitle": "用技术迭代溯源法，从 AI 算力升级反推玻璃基板、TGV 与关键设备材料卡点。", "subtitle_en": "Use technology-iteration tracing to work backward from AI-compute upgrades toward glass substrate, TGV, and the key equipment-material bottlenecks.", "description": "先进封装继续往高密度和高带宽走时，玻璃基板 / TGV 被市场反复提起，但真正有价值的不是“新材料”四个字，而是谁能把成孔、填铜、RDL 和良率稳定做出来。这个案例围绕玻璃基板 TGV 产业链，演示如何从技术迭代一路反推到工艺和设备卡点。\n\n报告会结合真实公开资料，拆出玻璃原片、激光成孔、金属化填铜、检测与封装验证等关键节点，再把这些瓶颈映射到 A 股公司，并说明哪些仍停留在验证阶段。\n\n适合：想理解先进封装新方向的人；需要从技术升级反推产业受益链的投研人员；要评估 TGV 公司映射真实性的产业分析师。", "description_en": "As advanced packaging keeps moving toward higher density and bandwidth, glass substrate / TGV is mentioned repeatedly, but the real value is not the phrase “new material” — it is who can actually stabilize via formation, copper filling, RDL, and yield. This case shows how to trace the glass-substrate TGV chain backward from technology iteration to process and equipment chokepoints.\n\nUsing real public materials, the report breaks the chain into glass sheet, laser via formation, metallization copper filling, inspection, and packaging-validation nodes, then maps those bottlenecks to A-share companies while explaining which names still sit at the verification stage.\n\nBest for: users trying to understand new advanced-packaging directions; researchers tracing beneficiaries from technology upgrades; industry analysts assessing whether TGV mappings are real.", "categories": [ "finance", "data", "business" ], "tags": [ "玻璃基板", "TGV", "先进封装", "AI算力", "产业链分析" ], "tags_en": [ "Glass Substrate", "TGV", "Advanced Packaging", "AI Compute", "Industry Chain Analysis" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "完整分析报告" } ], "skills": [ { "id": "industry-chain-analysis", "name": "产业链分析", "description": "产业链/赛道深度研究，支持全景拆解与瓶颈环节、金矿雷达、Serenity五维瓶颈验证。", "skill_id": "skill_2076627978773680128" } ], "experts": [], "mcps": [], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "请围绕玻璃基板 / TGV 先进封装产业链做一份深度分析。先用真实查询获取技术路线、关键工艺、设备材料、公司披露与验证进度，并注明来源和日期；再按“技术迭代溯源 → 卡点评分 → A股映射 → 风险与跟踪”的路径输出一份单文件、可离线打开的 HTML 报告，正文同步标注证据等级并区分重点线索与观察线索。", "prompt_en": "Create a deep-dive report on the glass substrate / TGV advanced-packaging chain. First use real queries to gather the technology roadmap, key processes, equipment/material links, company disclosures, and validation progress with explicit sources and dates; then deliver a single-file offline HTML report following the path of technology-iteration tracing → bottleneck scoring → A-share mapping → risks and tracking, while marking evidence grades and separating priority leads from watch-list leads in the body.", "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-07-28T09:00:00Z", "updated_at": "2026-07-28T09:40:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "health-report-chronic-followup", "title": "慢病复查报告趋势对比", "title_en": "Track Chronic Disease Indicators Across Rechecks", "subtitle": "多次复查报告一起看，指标是变好还是变差一目了然", "subtitle_en": "Compare multiple recheck reports at once — see instantly whether indicators improved or worsened", "description": "高血压、糖尿病、脂肪肝这类慢病，最关键的不是单次数值，而是趋势。可现实是几次复查单夹在不同文件夹里，上次血糖多少、药调整后有没有效果，全靠回忆，医生问起来也说不清。\n\n这个案例演示把同一个人多次（3 次以上）的复查报告一起交给 AI：先逐份提取指标建立统一口径的对照表，再按时间轴计算每个关键指标的变化方向和幅度，标出「持续改善 / 基本稳定 / 持续恶化 / 波动明显」，指出哪些指标已回到参考范围内、哪些仍在恶化需要重点干预，并整理出一份可以直接带去门诊、方便医生快速判断的随访摘要。\n\n适合：高血压/糖尿病/脂肪肝等慢病长期管理者、术后与用药后复查人群、家庭健康档案整理、基层医疗与慢病随访管理团队、体检中心的持续健康管理服务。", "description_en": "For chronic conditions like hypertension, diabetes, or fatty liver, what matters is the trend, not a single reading. In practice, recheck reports sit in different folders — what was last time's glucose? Did the medication adjustment work? — and it all comes down to memory, which fails when the doctor asks.\n\nThis case shows how to hand AI several (3+) recheck reports from the same person: it extracts indicators from each report into one consistent comparison table, computes the direction and magnitude of change for each key indicator along a timeline, labels them as steadily improving / broadly stable / steadily worsening / notably fluctuating, points out which indicators have returned to the reference range and which still need intervention, and compiles a follow-up summary you can bring straight to the clinic for quick physician review.\n\nGood for: people managing hypertension / diabetes / fatty liver long term, post-op and post-medication recheck patients, family health record keeping, primary-care and chronic disease follow-up teams, and ongoing health management services at checkup centers.", "categories": [ "education", "data", "productivity" ], "tags": [ "慢病管理", "复查对比", "指标趋势", "随访摘要", "健康档案" ], "tags_en": [ "chronic care", "recheck comparison", "indicator trends", "follow-up summary", "health records" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "我有 2 型糖尿病，在吃二甲双胍。四次复查的糖化血红蛋白依次是 8.6%、7.9%、7.1%、6.5%（参考 4.0–6.0），同期高密度脂蛋白是 0.88、0.91、0.89、0.92 mmol/L（参考 >1.04）。这两年到底控制得怎么样，下次门诊该重点跟医生聊什么？", "prompt_en": "I have type 2 diabetes and take metformin. My HbA1c across four follow-ups was 8.6%, 7.9%, 7.1%, 6.5% (ref 4.0–6.0), while HDL-C over the same period was 0.88, 0.91, 0.89, 0.92 mmol/L (ref >1.04). How well have I actually been controlling this, and what should I focus on with my doctor at the next visit?", "skills": [ { "id": "tencent-medical-report-interpreter", "name": "腾讯健康-AI就医助手-医疗报告解读", "description": "对接 ADP LLMReportInterpretation API，解读检验单/检查单/体检报告，识别异常值并用通俗语言解析医学术语", "slug": "tencent-medical-report-interpreter" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-08-22T16:09:15Z", "updated_at": "2026-08-22T16:36:30Z", "artifacts": [ { "file": "output.html", "label": "慢病四次复查趋势对比与随访摘要" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "kdocs-knowledge-handoff-atlas", "title": "知识交接 Atlas", "title_en": "Knowledge Handoff Atlas", "subtitle": "把散落资料折叠成新人能直接走通的知识地图", "subtitle_en": "Fold scattered materials into one onboarding map a newcomer can actually follow", "description": "项目资料散落在网页收藏、会议纪要、碎片笔记和已有云文档里时，新同学接手最难的不是“资料不够”，而是不知道先看什么、哪些是最新版本、哪些内容要继续维护。\n\n这个案例展示如何用金山文档的知识库、导入云文档、创建条目和智能文档整理能力，把分散材料收拢成一个交接入口：先建知识视图，再按项目背景、决策记录、交付资产、运行手册和 FAQ 五层结构归档，同时补出 30 分钟上手路径、交接清单和常见问题。\n\n适合：项目经理、产品运营、研发负责人、交付经理，以及任何需要把历史资料沉淀成团队入口、降低新人接手成本的人。", "description_en": "When project materials are scattered across web saves, meeting notes, loose notes, and existing cloud docs, the hardest part for a newcomer is not a lack of material—it is not knowing where to start, which version is current, and what should keep evolving.\n\nThis case shows how to use Kdocs knowledge views, cloud-doc import, item creation, and smart-document organization to turn scattered materials into one handoff entry. It creates a knowledge view first, then arranges content into five layers—project context, decision records, delivery assets, runbooks, and FAQ—while adding a 30-minute onboarding path, a handoff checklist, and common questions.\n\nBest for: project managers, product operators, engineering leads, delivery managers, and anyone who wants to turn historical materials into a durable team entry and reduce onboarding friction.", "categories": [ "office", "productivity" ], "tags": [ "知识库", "新人交接", "项目资料整理", "知识视图", "文档归档" ], "tags_en": [ "Knowledge Base", "Newcomer Handoff", "Project Cleanup", "Knowledge View", "Document Archiving" ], "artifact_type": "html", "external_url": "", "cover_image": "cover.png", "cover_mode": "landscape", "cover_image_en": "cover-en.png", "scenario_cover": "", "scenario_cover_en": "", "preview": "output.html", "preview_en": "output.html", "external_url_en": "", "prompt": "把我的网页收藏、会议纪要、零散笔记和已有云文档收拢到一个项目知识库里，并整理成新人交接入口。先定位或创建知识库，再按项目背景、决策记录、交付资产、运行手册和常见问题建立结构；网页内容存入知识库，已有云文档直接导入，零散文本整理成智能文档。最后给出知识库首页结构、30 分钟上手路径、交接清单和 FAQ。", "prompt_en": "Consolidate my web saves, meeting notes, loose notes, and existing cloud docs into one project knowledge base, then organize it as a newcomer handoff entry. First locate or create the knowledge base, then build a structure for project context, decision records, delivery assets, runbooks, and FAQ; save web content into the knowledge base, import existing cloud docs directly, and turn loose text into smart documents. Finally, provide the home structure, a 30-minute onboarding path, a handoff checklist, and FAQ.", "artifacts": [ { "file": "output.html", "label": "双语知识交接地图" }, { "file": "source_snapshot.json", "label": "知识库流程与结构快照" } ], "skills": [ { "id": "kdocs", "name": "金山文档|WPS云文档", "description": "金山文档官方 Skill。对话即操作——知识一键存入、碎片内容整理、接龙转表格、文档转 Markdown、表格美化、收发表生成，全在一句话内完成。", "skill_id": "skill_2053082745828995072" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-28T15:16:14+08:00", "updated_at": "2026-07-30T17:16:12+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "lexiang-kb-health-check", "title": "给知识库做一次全面体检", "title_en": "Run a Full Health Check on Your Knowledge Base", "subtitle": "云知批量遍历所有页面，逐篇诊断过时/缺结构，输出可视化体检报告", "subtitle_en": "Traverse every page with Tencent Yunzhi, diagnose stale & unstructured docs, and get a visual health report", "description": "知识库越攒越大，文档越堆越多，可没人说得清哪些内容早就过时了、哪些是当年随手记的没结构流水账，真要用的时候翻起来一团糟，找半天还不敢信。\n\n这个案例让腾讯云知批量遍历你知识库下的所有页面，逐篇诊断内容是否过时、是否缺乏结构、是否重复，给每篇打一个健康分，最后汇总成一份带总体健康分和整改优先级清单的可视化体检报告（暗色科技感单文件 HTML），每条结论都附真实页面链接可回溯。\n\n适合：知识库管理员、知识运营、团队负责人、文档 owner，以及任何想定期给知识库做治理的人。", "description_en": "Your knowledge base keeps growing, but nobody can tell which docs went stale long ago or which ones are just unstructured brain-dumps. When you actually need something, it's a mess to dig through and you can't even trust what you find.\n\nThis case lets Tencent Yunzhi batch-traverse every page under your knowledge base, diagnosing each doc for staleness, missing structure, and duplication, scoring each one on health, then rolling it all up into a visual health report with an overall score and a prioritized fix-list (a dark, techy single-file HTML). Every finding links back to the real page.\n\nGood for: knowledge base admins, knowledge ops, team leads, doc owners, and anyone who wants to govern their KB regularly.", "categories": [ "productivity", "agent" ], "tags": [ "知识治理", "内容体检", "知识库维护", "批量诊断" ], "tags_en": [ "Knowledge Governance", "Content Health Check", "KB Maintenance", "Batch Diagnosis" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "知识库健康度体检报告" } ], "prompt": "用腾讯云知帮我给知识库做一次全面体检。请批量遍历我这个知识库 [知识库链接或ID] 下的所有页面，逐篇诊断内容是否过时、是否缺乏结构、是否重复，给每篇打一个健康分，最后汇总成一份带总体健康分和整改优先级清单的可视化体检报告。\n\n要求：\n1. 遍历和读取全部走云知（乐享）的真实能力——用 entry_list_children 拿到父节点下所有页面，逐篇读取内容后再诊断，禁止编造任何页面名、评分或结论。\n2. 每篇给出：诊断结论（健康 / 内容过时 / 缺乏结构）、结构分、具体问题清单、以及需要整改的给出整改建议。\n3. 每条诊断都要附上该页面的真实链接，方便我点进去核对。\n4. 汇总总览：总页面数、健康数、过时数、缺结构数、总体健康分，再给一份按优先级排序的整改清单。\n5. 产物做成一份暗色科技感、带仪表盘感的单文件 HTML 体检报告；健康用绿、过时用橙、缺结构用红。样式你自己把握，重点是数据真实、可回溯。", "prompt_en": "Use Tencent Yunzhi to run a full health check on my knowledge base. Batch-traverse every page under my KB [KB link or ID], diagnose each doc for staleness, missing structure, and duplication, give each a health score, then roll it up into a visual health report with an overall score and a prioritized fix-list.\n\nRequirements:\n1. All traversal and reading must use Yunzhi's real capabilities — use entry_list_children to get every page under the parent node, read each one before diagnosing, and never fabricate page names, scores, or conclusions.\n2. For each page, give: a diagnosis (Healthy / Outdated / Unstructured), a structure score, a concrete issue list, and a fix suggestion where needed.\n3. Attach the real page link to every finding so I can click through and verify.\n4. Summary overview: total pages, healthy count, outdated count, unstructured count, overall health score, plus a fix-list sorted by priority.\n5. Deliver a dark, techy, dashboard-style single-file HTML health report; green for healthy, orange for outdated, red for unstructured. Style is up to you — what matters is the data being real and traceable.", "skills": [ { "id": "tencent-yunzhi", "name": "腾讯云知", "description": "腾讯云知（乐享）知识库专用操作——搜索、阅读、创建、编辑、上传", "skill_id": "skill_2053083480194342912" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-07-27T11:30:00+08:00", "updated_at": "2026-07-27T11:30:00+08:00", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "marketing-tax-ai-campaign-retrospective-draft", "title": "税务小智战役复盘（草稿）", "title_en": "Tax AI Campaign Review Draft", "subtitle": "新品发布 8 周效果评估与结构优化建议的草稿案例", "subtitle_en": "A draft retrospective of an 8-week launch campaign with optimization recommendations", "description": "很多营销团队在新品发布后能拿到一批数据，却迟迟产不出一份能真正指导下一轮投放和预算调整的复盘报告：数字散在各处，渠道好坏说不清，内容问题和投放问题也很难放到同一张图里。本案例封装的是“税务小智”新品发布战役的复盘草稿版，已包含核心漏斗、ROI/CAC/LTV、渠道对比、预算重分配和 30/60/90 天优化动作，并保留定性样本与汇报结构的待补充区，适合先做管理层预览、结构评审与后续补完。\n\n适合：市场负责人、增长负责人、投放经理、内容负责人、B2B SaaS 创始人、复盘汇报撰写者", "description_en": "After a product launch, many marketing teams end up with plenty of campaign data but no retrospective that actually guides the next round of spend and optimization. Metrics are scattered, channel performance is hard to compare, and content issues rarely connect cleanly with business outcomes. This case packages a draft retrospective for the 'Tax Small Genius' launch campaign, already covering the funnel, ROI/CAC/LTV, channel comparison, budget reallocation, and 30/60/90-day optimization actions, while clearly preserving placeholder sections for qualitative sample review and reporting structure. It is useful for leadership preview, structural review, and later completion.\n\nBest for: Marketing Leads, Growth Leads, Performance Managers, Content Leads, B2B SaaS Founders, retrospective report writers", "categories": [ "business", "data", "agent" ], "tags": [ "营销复盘", "ROI 分析", "渠道对比", "预算优化", "草稿版" ], "tags_en": [ "campaign retrospective", "ROI analysis", "channel comparison", "budget optimization", "draft" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "复盘报告 HTML 草稿版" }, { "file": "output.md", "label": "复盘报告 Markdown 草稿版" } ], "prompt": "请基于一个已运行 8 周的新品发布营销战役，完成正式复盘报告。场景：面向中小企业的 AI 财税 SaaS“税务小智”。要求包括：\n1. 漏斗回顾（曝光、点击、线索、SQL、付费、收入）\n2. ROI / CAC / LTV / 回本周期分析\n3. 渠道对比（至少覆盖搜索广告、公众号、短视频、私域、销售跟进/邮件）\n4. 预算重分配建议\n5. 30/60/90 天优化动作\n6. 复盘报告需保留对内容样本、问题分级、汇报节奏的扩展空间\n\n输出 Markdown 复盘报告 + HTML 展示版。当前案例为草稿封装版。", "prompt_en": "Please complete a formal retrospective report based on a product launch marketing campaign that has been running for 8 weeks. Scenario: an AI tax & finance SaaS for small and medium-sized enterprises called \"Tax Small Genius.\" Requirements include:\n1. Funnel review (impressions, clicks, leads, SQLs, paid, revenue)\n2. ROI / CAC / LTV / payback period analysis\n3. Channel comparison (covering at least search ads, WeChat Official Accounts, short video, private domain, and sales follow-up/email)\n4. Budget reallocation recommendations\n5. 30/60/90-day optimization actions\n6. The retrospective report should leave room for expansion on content samples, issue severity grading, and reporting cadence\n\nOutput a Markdown retrospective report + an HTML presentation version. The current case is a draft package version.", "skills": [], "experts": [ { "id": "MarketingCampaignTeam", "name": "营销战役团队", "description": "由营销总监领导的 4 人营销专家团队：内容创作者（博客/邮件/社媒/品牌声音）、活动策划师（战役策略/受众/渠道/预算）、SEO 专家（技术审计/内容优化/效果分析）和品牌分析师（竞品定位/品牌审核）。覆盖完整营销生命周期。", "expertType": "team", "source": "expert-marketplace", "name_en": "Marketing Campaign Team", "description_en": "A team of 4 marketing specialists led by a Marketing Director: Content Creator (blog posts/email/social media/brand voice), Campaign Planner (campaign strategy/audience/channels/budget), SEO Expert (technical audit/content optimization/performance analytics), and Brand Analyst (competitive positioning/brand review). Covers the full marketing lifecycle.", "expert_id": "ex_wg8SWSt1T33S" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:59:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "medication-safety-herb-interaction", "title": "中西药保健品联用排查", "title_en": "Herb-Drug & Supplement Interaction Check", "subtitle": "华法林拟加用鱼油 / 银杏 / 三七粉，出血风险 + 禁忌组合 + INR 监测 + 宣教一键评估", "subtitle_en": "Planning to add fish oil, ginkgo and panax to warfarin — bleeding risk, contraindications, INR monitoring and counseling in one", "description": "患者长期服用华法林，正打算自行加用鱼油、银杏叶、三七粉——这个组合安不安全、有没有禁忌、该怎么监测？\n\n这个案例演示了把「拟加用的保健品清单」交给 AI 临床助手，它逐项评估鱼油、银杏叶、三七粉的抗凝 / 抗血小板活性与华法林的叠加出血风险，判定哪些组合需避免、哪些需谨慎，并给出 INR 监测方案与患者宣教要点。\n\n适合：长期服用华法林等抗凝药的患者、拟自行加服保健品的中老年人群、需要相互作用与用药安全科普的用户、临床药师带教。", "description_en": "A patient on long-term warfarin is about to self-add fish oil, ginkgo leaf and panax powder — is this combination safe, are any of them contraindicated, and how should it be monitored?\n\nThis case shows how to hand a planned supplement list to the AI clinical assistant, which itemizes the anticoagulant / antiplatelet activity of fish oil, ginkgo and panax against warfarin's bleeding risk, judges which combinations to avoid and which to use cautiously, and lays out an INR monitoring plan plus patient counseling points.\n\nFor: patients on warfarin and other anticoagulants, seniors planning to self-add supplements, anyone needing interaction and medication-safety guidance, and clinical pharmacists for teaching.", "categories": [ "agent", "education" ], "tags": [ "用药安全", "药物相互作用", "保健品", "中药", "AI临床助手" ], "tags_en": [ "Medication Safety", "Drug Interactions", "Supplements", "Traditional Chinese Medicine", "AI Clinical Assistant" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "中西药保健品联用排查演示（HTML，含中英双语）" } ], "prompt": "请以「腾讯健康-AI临床助手-用药安全」的能力，评估一位患者的华法林与保健品联用风险：\n\n患者长期服用华法林，近期拟自行加用鱼油、银杏叶保健品及三七粉。请评估：\n1. 联用后的出血风险（重点：鱼油、银杏叶、三七粉的抗血小板 / 抗凝活性与华法林「叠加强化」，可能推高 INR）；\n2. 是否存在禁忌或需避免的组合（逐项判定：哪些组合应避免、哪些需谨慎、哪些可谨慎监测）；\n3. 相应的 INR 监测建议（基线 INR、加服后复查节点、目标范围）；\n4. 患者宣教建议（出血预警征象、就医时机、用药依从性、不擅自停药）。\n\n请客观说明相互作用机制，不武断否定中药 / 保健品价值；不擅自建议停用华法林等抗凝药（停药有血栓风险）；出血相关风险应主动强调、不等用户追问；最终建议以医生 / 药师意见为准。", "prompt_en": "Using the medication-safety capabilities of the Tencent Health AI Clinical Assistant, evaluate a patient's risk of combining warfarin with supplements:/n/nThe patient has been on warfarin long-term and now plans to self-add fish oil, ginkgo leaf supplements and panax powder. Please assess:/n1. The bleeding risk after combining them (focus: the antiplatelet / anticoagulant activity of fish oil, ginkgo and panax stacks with warfarin, possibly raising INR);\n2. Whether any combination is contraindicated or should be avoided (itemize: which to avoid, which to use cautiously, which to use with careful monitoring);\n3. INR monitoring recommendations (baseline INR, follow-up timing after adding, target range);\n4. Patient counseling (bleeding warning signs, when to seek care, medication adherence, not stopping on their own).\n\nObjectively explain the interaction mechanism. Do not flatly dismiss the value of traditional Chinese medicines / supplements. Do not recommend stopping warfarin or other anticoagulants on your own (stopping carries thrombotic risk). Flag the bleeding risk proactively rather than waiting for the user to ask. End with a recommendation to defer to a doctor / pharmacist.", "skills": [ { "id": "medication-advisor", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "slug": "medication-advisor" } ], "experts": [ { "id": "chief-physician", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_0JLBYdNE0Hxk" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T11:01:00+08:00", "updated_at": "2026-09-05T11:11:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-hpc-expert-case-3", "title": "任务跟踪 + 集群运维监控大盘", "title_en": "Job Tracking + Cluster Ops Monitoring Dashboard", "subtitle": "作业生命周期与集群运维状态，一张大盘全掌握", "subtitle_en": "Job lifecycle and cluster ops status — everything on one dashboard", "description": "提交了作业不知道跑到哪了，集群资源用得怎样也不清楚？这个案例演示了如何全程跟踪作业从 PENDING 到 COMPLETED 的每个阶段，同时巡检集群节点健康、存储用量与弹性伸缩事件，输出一张含作业甘特图、节点弹性伸缩曲线、资源环形仪表与扩缩容事件日志的「运维监控大盘」HTML，作业与集群状态一图尽览。\n\n适合：需要同时掌握作业进度与集群运维全局状态的 HPC 运维人员、生信工程师，以及想了解 SLURM 弹性伸缩机制的用户。", "description_en": "Submitted a job but have no idea how it's going, or what the cluster resource usage looks like? This case demonstrates how to track a job through every stage from PENDING to COMPLETED, while simultaneously inspecting cluster node health, storage usage, and elastic scaling events, producing an 'Ops Monitoring Dashboard' HTML with a job Gantt chart, node elastic scaling curve, resource ring gauges, and a scaling event log — full visibility in one view.\n\nSuitable for: HPC ops engineers and bioinformatics engineers who need simultaneous visibility into job progress and cluster-wide status, and users who want to understand SLURM elastic scaling mechanics.", "categories": [ "dev" ], "tags": [ "HPC集群", "运维监控", "SLURM", "弹性伸缩", "作业跟踪" ], "tags_en": [ "HPC Cluster", "Ops Monitoring", "SLURM", "Elastic Scaling", "Job Tracking" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "集群运维监控大盘" } ], "prompt": "请生成一份任务跟踪 + 集群运维监控大盘。\n\n先列出所有运行中集群，选第一个 SLURM 集群，用 squeue/sacct 定位最近提交的作业作为跟踪目标。跟踪该作业全生命周期（每 30 秒轮询 PENDING→RUNNING→COMPLETED 各阶段时间戳，观察 sinfo 弹性扩容与缩容释放，作业完成后用 sacct 取最终状态/退出码/耗时/内存/节点）。同时巡检集群运维状态：节点健康（sinfo）、存储用量（df -h）、内存水位（free -h）、扩缩容事件、最近 24h 失败作业与 OOM。\n\n最终输出一份「运维监控大盘」HTML，风格专业美观、运维深色青主题、自包含内联 SVG，包含：顶部运维 KPI 卡（作业最终状态/峰值活跃节点/集群可用率/峰值 CPU/告警数）+ 作业生命周期甘特图（PENDING/RUNNING/COMPLETED 分段耗时）+ 作业状态看板 + 节点弹性伸缩曲线（时间-节点数，叠加作业阶段背景带与扩容/缩容触发点）+ 集群资源环形仪表（CPU/内存/存储）+ 集群与队列健康列表 + 扩缩容事件日志表 + 运维事件时间线。只读巡检，无需确认。", "prompt_en": "Please generate a job tracking + cluster operations monitoring dashboard.\n\nFirst list all running clusters, select the first SLURM cluster, use squeue/sacct to locate the most recently submitted job as the tracking target. Track the job's full lifecycle (poll every 30 seconds for PENDING→RUNNING→COMPLETED stage timestamps, observe sinfo elastic scale-out and scale-in releases, use sacct after completion to get final status/exit code/duration/memory/nodes). Simultaneously inspect cluster ops status: node health (sinfo), storage usage (df -h), memory level (free -h), scaling events, failed jobs and OOM in the last 24h.\n\nFinally output an 'Ops Monitoring Dashboard' HTML, professional and self-contained with dark teal ops theme and inline SVG, including: top ops KPI cards (final job status/peak active nodes/cluster availability/peak CPU/alert count) + job lifecycle Gantt chart (PENDING/RUNNING/COMPLETED segment timing) + job status board + node elastic scaling curve (time vs node count, with job stage background bands and scale-out/in trigger points) + cluster resource ring gauges (CPU/memory/storage) + cluster and queue health list + scaling event log table + ops event timeline. Read-only inspection, no confirmation needed.", "skills": [], "experts": [ { "id": "omics-hpc-expert", "name": "腾讯组学HPC集群运维与作业管理专家", "description": "远程组学HPC一站式运维，管控节点、队列、存储全生命周期，适配SLURM/SGE调度，自然语言交互降低门槛", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_oVOV90lFvBz5" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-29T17:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-ori-expert-case-3", "title": "热稳定性评估（Thermostability）· 设计验证", "title_en": "Thermostability Assessment · Design Validation", "subtitle": "ORI 预测蛋白 Tm 值，60℃ 耐温门槛判定 + HTML 验证报告一键生成", "subtitle_en": "ORI predicts protein Tm values with 60°C threshold judgment and one-click HTML validation report", "description": "发酵工艺验证最怕到最后才发现蛋白不耐温，白白浪费整条设计流程。\n\n这个案例演示了通过 ORI 专家的 Thermostability 子应用，对示例序列执行热稳定性评估，输出每条序列的预测 Tm（℃）与 60℃ 门槛判定，渲染为达标/不达标一目了然的 HTML 验证报告（含差值标注）。**进阶用法：可衔接 omics-ori-expert-case-2，直接验证从头设计最优候选序列的耐温性。**\n\n适合：发酵工艺工程师、蛋白热稳定性优化研究员、工业生物技术从业者。", "description_en": "The worst outcome in fermentation process validation is discovering a protein's thermal tolerance falls short only at the very end, wasting the entire design pipeline.\n\nThis case demonstrates how to use the ORI expert's Thermostability sub-application to assess a sequence set, output predicted Tm (°C) and 60°C threshold judgments, and render a clear pass/fail HTML validation report with margin annotations. **Advanced usage: pair with omics-ori-expert-case-2 to directly validate the top de novo designed candidate's thermal tolerance.**\n\nIdeal for: fermentation process engineers, protein thermostability optimization researchers, and industrial biotechnology practitioners.", "categories": [ "data", "education" ], "tags": [ "热稳定性预测", "ORI", "Tm预测", "耐温性验证", "蛋白质性质评估" ], "tags_en": [ "Thermostability Prediction", "ORI", "Tm Prediction", "Thermal Tolerance Validation", "Protein Property Assessment" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "热稳定性评估 HTML 验证报告" } ], "prompt": "Prompt 3 ｜ 热稳定性评估（Thermostability）· 设计验证\n\n我是发酵工艺方向的工程师，需要确认蛋白耐温性是否满足 60℃ 工艺要求。请使用 ORI 公共应用的 thermostability 子流程，完全按默认参数（默认示例输入、模型与资源配置），执行一次热稳定性评估，并完成分析：\n\n1. 提交前给我任务确认单（子流程｜输入摘要｜关键参数｜预计时长），确认后提交；\n2. 跟踪任务直至完成，失败则按错误知识库给出根因与修复建议；\n3. 完成后输出：预测 Tm 结果表（序列｜预测 Tm(℃)｜60℃ 门槛判定），并渲染为 HTML 验证报告（达标/不达标 + 差值标注）；进阶用法：紧接 Prompt 2 时，改用其最优候选序列作输入（模型与资源仍默认），验证\"耐 60℃\"设计目标是否达成。", "prompt_en": "Prompt 3 | Thermostability Assessment · Design Validation\n\nI am an engineer in the fermentation process field and need to confirm whether a protein's thermal tolerance meets the 60°C process requirement. Please use the ORI public application's thermostability sub-workflow, following all default parameters (default example input, model, and resource configuration), and run a thermostability assessment, then complete the analysis:\n\n1. Before submission, provide me with a task confirmation sheet (sub-workflow | input summary | key parameters | estimated duration), then submit after confirmation;\n2. Track the task until completion; if it fails, provide root cause analysis and fix suggestions based on the error knowledge base;\n3. Upon completion, output: predicted Tm result table (sequence | predicted Tm(°C) | 60°C threshold judgment), and render an HTML validation report (pass/fail + margin annotation); advanced usage: when following Prompt 2, switch to using its top candidate sequence as input (model and resources still default) to verify whether the '60°C-tolerant' design goal has been achieved.", "skills": [], "experts": [ { "id": "omics-ori-expert", "name": "ORI蛋白设计专家", "description": "腾讯ORI蛋白设计专家，基于ORI蛋白质设计系统，提供从头序列设计、USMFold结构预测、溶解性/热稳定性/信号肽预测五大能力，并支持3D结构可视化", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_HrdscGoeTkwq" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T07:58:16Z", "updated_at": "2026-08-27T13:25:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-scprotein-expert-case-3", "title": "端到端全流程（Stage1 → Stage2）", "title_en": "End-to-End Full Pipeline (Stage1 → Stage2)", "subtitle": "一次运行完整演示 scPROTEIN 两阶段，交付完整可视化报告", "subtitle_en": "One run, full scPROTEIN two-stage demo with a complete visual report", "description": "想一次性跑完 scPROTEIN 全流程，从原始数据到细胞分群结果？这个案例演示了从默认公共数据集出发，依次跑 Stage1 不确定性定量和 Stage2 细胞嵌入，Stage1 完成校验后自动衔接，最终交付一份融合概览卡片、t-SNE 分群图、不确定性分布、Top20 复核表与热图的 ECharts 交互式完整报告。\n\n适合：想完整体验 scPROTEIN 端到端分析流程、快速获得全量可视化结果的蛋白质组学研究人员与平台评测用户。", "description_en": "Want to run the full scPROTEIN pipeline in one go — from raw data to cell clustering results? This case demonstrates starting from the default public dataset, running Stage1 uncertainty quantification and Stage2 cell embedding sequentially (Stage2 auto-starts after Stage1 validation), and delivering a complete ECharts interactive report combining overview cards, t-SNE cluster plots, uncertainty distribution, Top20 review table, and heatmap.\n\nSuitable for: proteomics researchers and platform evaluators who want to experience the full scPROTEIN end-to-end analysis and quickly obtain comprehensive visual results.", "categories": [ "data" ], "tags": [ "单细胞蛋白组", "端到端分析", "scPROTEIN", "细胞分群", "批次效应去除" ], "tags_en": [ "Single-cell Proteomics", "End-to-End Analysis", "scPROTEIN", "Cell Clustering", "Batch Effect Removal" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "scPROTEIN 端到端完整分析报告" } ], "prompt": "我需完整演示 scPROTEIN 端到端能力：从默认公共数据集出发，依次执行 Stage1（不确定性定量）与 Stage2（细胞嵌入），交付可视化报告。两阶段均用默认参数。\n\n1. 提交前确认两个子应用与参数摘要，我确认后提交；\n2. Stage1 完成并校验输出后，再衔接提交 Stage2，任一失败给根因诊断；\n3. 结果固化为一份可复用的 ECharts 交互式 HTML 报告文件（如 scprotein_visual_report.html），在内置浏览器打开预览，须包含：概览卡片（肽段数/细胞数/嵌入维度/最优 k/silhouette/低置信占比）、t-SNE 细胞分群散点图（KMeans 着色 + Silhouette vs k 曲线）、不确定性分布直方图（P90 阈值线）、Top20 复核表、高方差肽段热图。", "prompt_en": "I need to fully demonstrate scPROTEIN end-to-end capability: starting from the default public dataset, run Stage1 (uncertainty quantification) and Stage2 (cell embedding) sequentially, and deliver a visualization report. Both stages use default parameters.\n\n1. Confirm both sub-applications and parameter summary before submission; submit after my confirmation;\n2. After Stage1 completes and output is validated, then submit Stage2; provide root cause diagnosis for any failure;\n3. Consolidate results into a reusable ECharts interactive HTML report file (e.g., scprotein_visual_report.html), open preview in built-in browser, must include: overview cards (peptide count/cell count/embedding dimension/optimal k/silhouette/low-confidence ratio), t-SNE cell cluster scatter plot (KMeans colored + Silhouette vs k curve), uncertainty distribution histogram (P90 threshold line), Top20 review table, high-variance peptide heatmap.", "skills": [], "experts": [ { "id": "omics-scprotein-expert", "name": "腾讯scPROTEIN单细胞蛋白组建模专家", "description": "腾讯 scPROTEIN表征模型，实现对单细胞蛋白组的多肽不确定性估计，去除批次效应噪声以及单细胞蛋白组的细胞类型注释", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_SlvtyVitPzY8" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-25T18:35:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "omics-unipert-expert-case-3", "title": "蛋白多层分类", "title_en": "Protein Multi-level Classification", "subtitle": "FASTA 序列输入，三层分类层级揭示蛋白扰动特性", "subtitle_en": "FASTA sequence input — three-level classification hierarchy reveals protein perturbation characteristics", "description": "拿到蛋白序列，想知道它们在扰动表征空间里属于哪个类别、与哪些扰动最相似？这个案例演示了如何用 UniPert genes 子应用将蛋白 FASTA 序列编码为统一表征，完成扰动模态层 → 表征邻域层 → 激活特征层三层多级分类，输出含三层分类层级图、表征指纹热图、邻域相似度对比图与分类依据汇总的 ECharts 交互报告。\n\n适合：需要对蛋白质进行功能分类、理解蛋白质在扰动表征空间中定位的计算生物学与药物靶点研究人员。", "description_en": "Got protein sequences and want to know which category they belong to in the perturbation representation space and which perturbations they most resemble? This case demonstrates how to use the UniPert genes sub-application to encode protein FASTA sequences into unified representations, complete three-level multi-tier classification (perturbation modality layer → representation neighborhood layer → activation feature layer), and produce an ECharts interactive report with a three-level classification hierarchy diagram, representation fingerprint heatmap, neighborhood similarity comparison, and classification basis summary.\n\nSuitable for: computational biology and drug target researchers who need to functionally classify proteins and understand their positioning in the perturbation representation space.", "categories": [ "data" ], "tags": [ "蛋白多层分类", "基因扰动预测", "UniPert", "虚拟细胞", "表征编码" ], "tags_en": [ "Protein Multi-level Classification", "Gene Perturbation Prediction", "UniPert", "Virtual Cell", "Representation Encoding" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "蛋白多层分类分析报告" } ], "prompt": "我需要基于 UniPert 统一表征完成蛋白多层分类。请使用 UniPert-WDL 合集的 genes 子应用，按默认输入模板（protein.fasta，无需我提供数据）执行蛋白表征编码。\n\n1. 提交前确认子应用、同名检查结果与参数摘要，我确认后提交；\n2. 跟踪任务至完成，失败给根因诊断（debug 取证 + 错误知识库匹配 + 修复建议）；\n3. 结果固化为一份可复用的 ECharts 交互式 HTML 报告（如 unipert_protein_report.html），在内置浏览器打开预览，须包含：概览卡片（蛋白数 / 表征维度 / 激活维度数 / 最近邻扰动 / 邻域相似度 / 层级判定结果）、三层分类层级图（扰动模态层 → 表征邻域层 → 激活特征层，附各层依据）、表征指纹热图（峰值维度标注）、邻域相似度对比图、分类依据汇总与局限说明。", "prompt_en": "I need to complete protein multi-level classification based on UniPert unified representations. Please use the genes sub-application from the UniPert-WDL collection, with the default input template (protein.fasta, no need for me to provide data) to execute protein representation encoding.\n\n1. Confirm the sub-application, duplicate name check result, and parameter summary before submission; submit after my confirmation;\n2. Track the task to completion; provide root cause diagnosis on failure (debug evidence + error knowledge base matching + fix suggestions);\n3. Consolidate results into a reusable ECharts interactive HTML report (e.g., unipert_protein_report.html), open preview in built-in browser, must include: overview cards (protein count / representation dimensions / active dimensions / nearest neighbor perturbation / neighborhood similarity / hierarchical classification results), three-level classification hierarchy diagram (perturbation modality layer → representation neighborhood layer → activation feature layer, with rationale for each level), representation fingerprint heatmap (peak dimensions annotated), neighborhood similarity comparison chart, classification basis summary and limitations.", "skills": [], "experts": [ { "id": "omics-unipert-expert", "name": "腾讯UniPert虚拟细胞专家", "description": "基于腾讯UniPert模型，构建虚拟细胞，实现对于内外源扰动源统一表征，并且实现基因和药物扰动预测，包括细胞特异性药物响应", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_w7OMyXyVUKQh" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-25T11:28:00+08:00", "updated_at": "2026-08-30T11:15:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "painted-pottery-ding-rollup-banner", "title": "彩陶鼎文创 · 易拉宝", "title_en": "Painted Pottery Ding Cultural Product · Roll-Up Banner", "subtitle": "基于大汶口文化彩陶鼎文创品的宣传易拉宝", "subtitle_en": "A monument-like Oriental cultural roll-up banner", "description": "展会要一张 80×200 的易拉宝，主体、标题、说明、收卷安全区全要排明白，做起来最头疼。\n\n案例以大汶口文化彩陶鼎文创摆件为唯一主体，2:5 超长竖构图、对称纪念碑式排版：象牙米白宣纸背景，顶部小标签、黑色主标题「大汶口文化彩陶鼎」、主题词「古器新生」、朱红小印、放大浅底暗纹、产品说明与底部山形收束，底部预留收卷安全区，输出克制清雅的当代东方博物馆展陈易拉宝。\n\n适合：文创展陈策划、平面/展览设计师、博物馆市场团队、文化活动主办方", "description_en": "This case makes a Dawenkou painted-pottery Ding creative object the sole subject in a 2:5 ultra-tall symmetric monument layout: ivory rice-paper backdrop, a top label, black main title, the theme phrase, a cinnabar seal, an enlarged faint watermark motif, product notes and a mountain-shaped base, with a safe roll-up zone reserved — a restrained, elegant contemporary museum banner.\n\nGood for: Exhibition planners, Graphic & exhibition designers, Museum marketing teams, Cultural event organizers", "categories": [ "creation" ], "tags": [ "易拉宝", "展陈设计", "东方编辑", "2:5竖版" ], "tags_en": [ "Roll-Up Banner", "Exhibition Design", "Oriental Editorial", "2:5 Portrait" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请根据用户上传的文创品图片，生成一张80×200cm、2:5超长竖构图的东方文化展陈易拉宝平面设计。产品是画面唯一核心主体，必须严格保持其原有品类、正面朝向、整体轮廓、长宽比例、色彩、木质质感、旋涡纹样、部件数量和三足结构，不得重新设计、替换、拉伸、压扁或增加细节。\n\n画面采用稳定、对称、具有纪念碑感的纵向中心构图。背景为温暖柔和的象牙米白色宣纸，带有轻微纸张纤维、自然颗粒与克制的旧印刷质感。顶部居中放置小号深青绿色文字“文创摆件 / CULTURAL OBJECT”，中文使用精致宋体，英文使用增加字距的纤细衬线字体，下方设置一条短而细的青绿色分隔线。\n\n画面上部居中排布黑色主标题“大汶口文化彩陶鼎”，使用高对比现代宋体，字号醒目、字距紧凑；标题下方排布主题词“古器新生”，字号略小，使用低饱和深青绿色。主题词下方居中点缀一枚小型朱红色抽象印记，只作为色彩锚点，不使用真实机构印章或复杂文字。\n\n提取产品表面的旋涡纹样，将其大幅放大并以约5%–10%的透明度铺陈在标题与产品之间，形成左右延伸、局部越出画面的浅米灰色背景暗纹，不得抢夺主体。文创品完整置于画面中下部并严格水平居中，约占画面高度的38%–45%、宽度的60%–72%，保留清晰自然的木质纹理，底部加入边缘柔和的接地阴影，不添加基座、道具或复杂场景。\n\n产品下方居中排布两行说明文字：“取意于大汶口文化彩陶鼎，将史前彩陶的旋涡纹样与三足器形，转译为温润木质摆件。”使用深灰黑色小号宋体，行距疏朗，文本宽度不超过画面的70%；文字下方设置一条细短的低饱和青绿色横线。\n\n画面最底部使用2–3层低饱和灰青绿色山峦与河流状抽象平面色块横向延展，边缘自然、层次柔和并带轻微版画颗粒，用于稳定画面重心。底部8%–12%区域不放置任何重要文字、Logo或二维码，作为易拉宝收卷安全区。\n\n整体色彩限定为象牙米白、炭黑、深青绿、陶土红棕和少量朱红，呈现克制、温润、清雅的当代东方博物馆编辑设计审美。确保“顶部小标签—主标题—古器新生—产品—说明文字—山形收束”形成明确的纵向阅读顺序。输出干净的易拉宝平面设计稿，不展示支架、展厅、墙面透视或样机场景；不复制产品，不添加未经提供的Logo，不虚构尺寸、编号、馆藏信息或历史结论，避免厚重边框、祥云堆叠、金色装饰、泛滥书法、错误文字和水印。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-promo-expert", "name": "文创宣传视觉设计专家", "description": "将定稿文创设计转化为海报、明信片、收藏小卡、易拉宝等宣传物料", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "c6229d08eb8d", "expert_id": "ex_zaYhWoDF5zrp" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "作品展示页（含产物图与提示词）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-20T16:58:13+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "product-promo-short", "title": "商品宣传短片", "title_en": "Product Promo Short", "subtitle": "用岩石、木火与干花呈现琥珀香水的沉静气味", "subtitle_en": "Rock, firewood, and dried flowers express the calm scent of an amber fragrance", "description": "单张产品图很难直接变成有层次的商业短片，产品外观还容易在生成中走样。\n\n案例围绕一只黑盖琥珀香水瓶，用湿润森林、黑色岩石、木材火星和干花四组环境建立气味联想，最后回到干净的品牌定帧。\n\n适合：香水、护肤、酒饮、家居香氛和需要竖屏种草素材的消费品牌。", "description_en": "A single product image rarely becomes a layered commercial on its own, and the product can deform during generation.\n\nAn amber fragrance bottle is placed among a damp forest, black rock, wood sparks, smoke, and dried flowers before returning to a clean brand lockup.\n\nBest for fragrance, skincare, beverages, home scent, and consumer brands that need vertical social ads.", "categories": [ "creation" ], "tags": [ "商品广告", "香水", "竖屏短片" ], "tags_en": [ "product ad", "fragrance", "vertical short" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「商品宣传短片」Skill，根据我上传的黑盖琥珀色香水瓶产品图，制作一支约 17 秒、9:16 的商业短片《琥珀香水的森林余烬》。必须保持瓶身比例、棕色玻璃、黑色瓶盖与标签排版不变。镜头顺序为：湿润森林中的瓶身轮廓；香水瓶立于黑色岩石、侧面暖光扫过；木材火星和烟雾掠过瓶身；干花与琥珀液体的微距；最后在纯黑背景中定格产品与品牌名。先确认规格和故事板，再依次确认产品环境图、氛围/功能镜头和音频方案；每阶段批准后再进入下一步。运镜以缓慢推进、微距滑动和一次利落转场为主，配低频环境音乐、火星与玻璃轻响，不新增未经提供的功效宣称。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Product Promo Short” Skill and use my uploaded product image of an amber fragrance bottle with a black cap to create a 9:16 commercial titled Forest Embers for an Amber Fragrance, about 17 seconds long. Preserve the bottle proportions, brown glass, black cap, and label layout exactly. Use this shot order: the bottle silhouette in a damp forest; the bottle standing on black rock as warm side light sweeps across it; wood sparks and smoke passing the bottle; macro views of dried flowers and amber liquid; and a final product-and-brand lockup on pure black. Confirm the specification and storyboard first, then obtain approval for the product-in-environment images, atmospheric or functional shots, and audio plan in sequence. Use slow push-ins, macro slides, and one crisp transition, with low-frequency ambient music, sparks, and delicate glass sounds. Do not add product claims that I did not provide.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-compare-ru-vs-guan", "title": "汝窑 vs 官窑（宋代青瓷巅峰一对）", "title_en": "Ru Ware vs Guan Ware: Two Peaks of Song Celadon", "subtitle": "玛瑙入釉的天青孤品 vs 紫口铁足的皇家御窑：釉色/工艺/存世量三维对峙", "subtitle_en": "Ru ware's agate-glazed sky-blue singletons vs Guan ware's imperial 'purple rim, iron foot' — glaze, craft, and survival compared", "description": "「汝官哥钧定」里汝窑排名第一、官窑紧随其后，但每次讲起两者的区别，要么停留在「都很美、都是青色」，要么直接被「汝窑只有 67 件」的传说堵住。这个案例演示了如何用腾讯探元文博知识库对两个窑口**分窑口粒度**分别检索，用粒度对齐的方式补齐「**釉色、工艺、存世量**」三个用户点名维度的硬数据，并扩展到胎体、装烧、器型、开片四个证据维度，最后给到一份能拿到博物院里就能直接上手辨识的对比清单。\n\n适合：宋代青瓷爱好者、博物馆观众（尤其是台北故宫 / 上博 / 国博常客）、文博自媒体编辑、文物拍卖入门者、艺术院校陶瓷史方向师生。", "description_en": "When fans talk about the 'Wu Da Ming Yao' (five great Song kilns), Ru ranks first and Guan second, but most comparisons either stop at 'both are blue and beautiful' or get blocked by the half-legend that 'only 67 Ru pieces survive.' This case shows how the TanYuan knowledge base uses **kilo-level granularity** to query each kiln separately, hardens the three user-named dimensions (glaze, craft, surviving quantity) with published evidence, and extends to body, firing setup, vessel shape, and crackle — finishing with a hands-on comparison list you can take straight into a museum gallery. Best for: Song celadon enthusiasts, museum visitors (especially NPM/Shanghai Museum/National Museum regulars), cultural-heritage editors, newcomers to relic auctions, and ceramics-history students.", "categories": [ "business" ], "tags": [ "汝窑", "官窑", "宋代青瓷", "五大名窑", "存世量" ], "tags_en": [ "Ru Ware", "Guan Ware", "Song Celadon", "Five Great Kilns", "Surviving Quantity" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Markdown 对比报告（一句话结论 + 对比表 + 快速辨识要点 + 代表器物 + 来源）" } ], "prompt": "请按「腾讯探元文物对比分析师」专家的工作流，对比分析**汝窑**与**官窑**在**釉色、工艺、存世量**三个用户指定维度，以及胎体、装烧方式、器型用途、开片纹饰、特征性标识等由证据决定的补充维度上的差异。\n\n研究范围：\n- 对象 A：汝窑（北宋晚期，约 1086—1106 年；河南宝丰清凉寺村）\n- 对象 B：官窑（含北宋官窑汴京 / 南宋官窑杭州修内司与郊坛下）\n\n用户**必答维度**（必须排在前三行且必须出现）：釉色、工艺、存世量\n补充维度候选（资料撑住才上，不全部列）：\n- 年代与产地\n- 胎体与材质\n- 装烧方式（芝麻钉 / 垫饼支钉）\n- 器型与用途\n- 开片纹饰（蟹爪纹 vs 金丝铁线）\n- 特征性标识（紫口铁足 / 奉华款等后刻）\n\n交付要求（强制）：\n1. 粒度识别为「窑口 / 品类」，**严格只走探元知识库（mcp__tanyuan-wb-mcp-server__search_movable_relics）分别查两窑口**，分别 query: \"汝窑\"、\"官窑\"、\"南宋官窑\"；维度不足时分别补精简子 query 如 \"汝窑 釉色\"、\"官窑 开片\"、\"汝窑 玛瑙\"、\"官窑 紫口铁足\"\n2. **绝不能用一次 query 问「汝窑和官窑的区别」**——必须分两次检索\n3. **存世量、口径必须走联网核对**，区分台北故宫 21 件、故宫博物院、上海博物馆、大维德基金会等不同机构统计，保留不同口径而非编造伪精确数字\n4. 输出严格遵循专家骨架：**一句话结论 → 对比表（用户点名维度最前）→ 快速辨识要点（3—5 条可操作动作）→ 代表器物（每方 2—4 件附馆藏）→ 来源（探元知识库 + 联网补充出处）**\n5. 空白格写「公开资料未见明确记录」，**不留空、不编造**\n6. **不向用户暴露内部运行信息**——禁止出现 query、检索、工具名、requestId、条数机制词", "prompt_en": "Use the 'TanYuan Cultural Relic Comparison Analyst' expert workflow to comparatively analyse the differences between **Ru ware** and **Guan ware** along three user-named dimensions (glaze, craftsmanship, surviving quantity) and any evidence-supported supplementary dimensions (body material, firing setup, vessel shape & use, crackle pattern, signature marks like 'purple-rim/iron-foot').\n\nScope:\n- Object A: Ru ware (Northern Song late period, ~1086–1106 CE; Qingliangsi kiln site, Baofeng, Henan)\n- Object B: Guan ware (includes Northern Song Guan at Bianjing / Southern Song Guan at Xiuneisi & Jiaotanxia in Hangzhou)\n\nMandatory user-named dimensions (must occupy the first three rows of the table): glaze, craftsmanship, surviving quantity.\n\nEvidence-driven supplementary dimensions (only if the retrieved material actually supports them — never pad): date and provenance, body and material, firing setup (sesame-spur supports / pad-and-spur), vessel shape and use, crackle pattern (crab-claw vs gold-and-iron-wire), signature marks (purple-rim/iron-foot, post-cut 'Fenghua' inscriptions).\n\nDelivery requirements (mandatory):\n1. Granularity is 'kiln / category' — query the TanYuan knowledge base **only** via `mcp__tanyuan-wb-mcp-server__search_movable_relics`, once each for 'Ru ware', 'Guan ware', 'Southern Song Guan ware'. Supplement missing dimensions with focused sub-queries like 'Ru ware glaze', 'Guan ware crackle', 'Ru ware agate', 'Guan ware purple-rim iron-foot'.\n2. **Never** combine objects in a single query ('difference between Ru and Guan'). Always query separately.\n3. **Surviving quantity and canonical figures must come from public web verification** — distinguish between NPM Taipei 21 pieces, Palace Museum Beijing, Shanghai Museum, David Foundation etc., keep multiple calibres, never fabricate pseudo-exact numbers.\n4. Output must follow the expert skeleton strictly: **one-line conclusion → comparison table (user-named dimensions first) → 3–5 actionable quick-identification tips → 2–4 representative pieces per side with current custodian → sources (TanYuan knowledge base + web-verified supplements)**.\n5. Empty cells write 'no public record found', **never leave blanks, never fabricate**.\n6. **Never expose internal mechanics** — no queries, no search log, no tool names, no requestId, no result-count mechanism words.", "skills": [], "experts": [ { "id": "tanyuan-relic-compare-expert", "name": "文物对比专家", "description": "把两个或多个文博对象在同一套维度下讲清楚异同：一句话结论 + 对比表 + 快速辨识要点 + 代表器物。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1ce329117e4c", "expert_id": "ex_9FcVuwD0NDKB" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-14T19:04:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-infographic-changxin-lantern", "title": "长信宫灯结构拆解科普图", "title_en": "Exploded Diagram of the Changxin Palace Lantern", "subtitle": "千年环保灯 · 展厅级结构拆解与烟道走向图", "subtitle_en": "Millennium eco-lamp — gallery-grade exploded view with smoke-flow annotation", "description": "讲一件汉代灯具，「环保设计」四个字往往一笔带过；但凡想让观众真的「看见」烟怎么走、水怎么容、光怎么调，没有结构图基本是空谈。这个案例用 tanyuan-relic-infographic-expert 调出「典藏版」模板（暗金主题，适合深色展陈背景），把长信宫灯拆成头部、躯壳、左袖、右臂（烟道）、灯盘灯罩、底盘六大分件，附 HERO 封面、三栏拼版、烟道走向 SVG 标注、五节点时间轴——一张可直接挂展厅或公众号头条的单文件 HTML。\n\n适合：博物馆汉代专题展常设讲解屏、文物环保设计课引入案例、青铜灯具研究课论文配图、青少儿博物馆手册栏目、科技与文物跨学科课件。", "description_en": "Explaining a Han dynasty lamp with the four characters「eco design」rarely lands — to actually show visitors how the smoke travels, how water absorbs soot, and how light is adjusted, you need structural diagrams, not words. This case uses tanyuan-relic-infographic-expert's「premium edition」template (dark-gold theme, suited to deep exhibition backdrops) to break down the Changxin Palace Lantern into six major parts: head, torso, left sleeve, right arm (the smoke duct), shade-and-tray, and base — with a HERO cover, three-column split layout, SVG smoke-flow annotations, and a five-node timeline — a single HTML ready to hang in the gallery or top a public-account post.\n\nSuitable for: permanent interpretation screens at Han-era special exhibitions, opening cases for museum eco-design lessons, illustrations for bronze-lamp research papers, junior-museum handbook columns, and STEAM courseware bridging craft and conservation.", "categories": [ "creation", "education" ], "tags": [ "文物图解", "结构拆解", "长信宫灯", "汉代灯具", "典藏版" ], "tags_en": [ "relic infographic", "exploded view", "Changxin Palace Lantern", "Han dynasty lamp", "premium edition" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请调用 tanyuan-relic-infographic-expert 专家，使用「典藏版（进阶）」模板，为「西汉长信宫灯」生成一张含「结构拆解」的单文件 HTML 科普图解长图。\n\n要求：\n1. 数据真实：使用 tanyuan-wb-mcp-server 的 search_relics（查询「长信宫灯」）与 search_movable_relics（查询「长信宫灯」）真实检索——本次会话已确认 MCP 可用。检索返回的真实字段包括：通高 48cm / 宫女高 44.5cm / 重量 15.85kg / 材质青铜 / 铭文「长信尚浴」「阳信家」9 处共 65 字 / 现藏河北省博物院 / 制作年代约公元前 151 年（西汉景帝至武帝时期）/ 出土背景（1968.8 河北满城窦绾墓，灯盘内残留碳化动物油脂）/ 流转链（阳信侯刘揭 → 吴楚七国之乱朝廷没收 → 长信宫窦太后御用 → 赐予窦绾随葬）/ 工艺（分铸套接 + 鎏金 + 头身右臂灯座灯盘灯罩六分件可拆装 + 右臂虹管导烟 + 体内中空容水）/ 学术价值（铭文为汉代宫廷制度第一手资料 + 写实主义艺术代表 + 环保设计理念在世界科技史具重要意义）。\n2. 文物本体图：探元检索已返回 2 张授权图（miniapp-static.tanyuan.qq.com 域名，已授权直接用）—— 一张正面照（1222×1406）+ 一张抠图版（540×643，适合中部展示）；下载 base64 内联；aspect-ratio 等于真实图片比例。\n3. 视觉：用典藏版模板（cyan / gold / red 三主题之一），HERO 封面采用「左文字 + 右真实照片」横向布局，避免我之前画的火柴人 SVG——按专家 SKILL.md「文物本体绝不让 AI 自由重画」铁律执行。\n4. 结构：HERO 封面（文字 + 探元真实照片）→ 三栏拼版（左=基础信息 8 字段 + 造型叙事；中=探元抠图版授权图 + 特征要点 5 条；右=环保巧思 4 项 + 铭文流转链）→ 烟道走向 SVG 教学辅助图（标注 ① 灯盘点燃 ② 烟雾上行 ③ 烟尘溶水，虚线箭头）→ 六大分件拆解 SVG（头部/中空躯壳/左臂持盘/右臂烟道虹管/灯盘+灯罩/底盘，每件编号+SVG+名称）→ 五节点时间轴（前 151 制作 / 前 154 后没收归长信宫 / 前 113 入葬 / 1968 出土 / 现藏河北省博）→ 页脚诗 + 免责声明。\n5. 硬约束：单文件 HTML、CSS/图片/SVG 全部内联、不得 https:// 外链、不得 href=\"#\" 空锚点、不得 emoji；中部为探元抠图版授权图，非我自画 SVG；拆解 SVG 与烟道 SVG 明确标注「教学辅助示意图，非真实文物照片或拆解模型」。", "prompt_en": "Use the tanyuan-relic-infographic-expert's「premium edition」 template to produce a single-file HTML infographic for the「Changxin Palace Lantern」 with explicit structural breakdown.\n\nRequirements:\n1. Real data: query via tanyuan-wb-mcp-server (search_relics「长信宫灯」 + search_movable_relics「长信宫灯」) — MCP confirmed reachable this session. Real fields returned include: total height 48cm / maid height 44.5cm / weight 15.85kg / material bronze / inscription「Changxin Shangyu」「Yangxin Jia」9 places totaling 65 characters / current keeper Hebei Museum / made around 151 BCE (Western Han, reigns of Jingdi to Wudi) / excavation context (1968.8 Mancheng Douwan tomb, lamp tray contained carbonized animal-fat residue) / circulation chain (Yangxin Marquis Liu Jie → imperial confiscation after Seven States Rebellion → Changxin Palace Empress Dou → bestowed on Douwan for burial) / craft (separate casting + gilding + six-part modular head-body-right arm-base-tray-shade + right-arm smoke duct + hollow torso water reservoir) / academic value (inscriptions as first-hand evidence for Han palace institutions + realism art representative + environmental design significant in world history of technology).\n2. Relic photo: TanYuan returns 2 authorized images (miniapp-static.tanyuan.qq.com domain, free to use) — a front view (1222×1406) + a cut-out version (540×643, suited for central display); download + base64 inline; aspect-ratio equals actual image.\n3. Visual: «premium edition» template (one of cyan / gold / red themes); HERO uses「left text + right real photo」 horizontal layout — DO NOT hand-draw stick-figure SVGs of the maid; strictly follow expert SKILL.md「the relic body must never be re-drawn by AI」 rule.\n4. Structure: HERO cover (text + TanYuan real photo) → three-column layout (left: 8-field basic info + form narrative; center: TanYuan cut-out authorized image + 5-item feature highlights; right: 4 eco-design points + inscription circulation chain) → smoke-flow SVG teaching-auxiliary diagram (steps ① lamp tray ignition ② smoke rising ③ soot dissolving in water, dashed arrows) → six-part exploded-view SVG (head / hollow torso / left arm holding tray / right-arm-as-smoke-duct / shade+tray / base — each with number + SVG + name) → five-node timeline (151 BCE making / post-154 BCE confiscation to Changxin Palace / 113 BCE burial / 1968 excavation / current keeper Hebei Museum) → footer poem + disclaimer.\n5. Hard constraints: single HTML, all CSS/images/SVG inlined, no external https://, no href=\"#\" anchors, no emoji; the central image is the TanYuan cut-out authorized image (NOT a self-drawn SVG); the exploded-cell SVG and the smoke-flow SVG must be explicitly marked as「teaching-auxiliary schematic, not a real artifact photo or disassembly model」.", "skills": [], "experts": [ { "id": "tanyuan-relic-infographic", "name": "文物图解专家", "description": "基于探元文物库+知识库+MCP检索的单文件文物图解专家，使用简约版/典藏版模板输出可展陈、可传播的 HTML 长图", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_OPJE3pWwnB1j" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-14T18:06:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "artifacts": [ { "file": "output.html", "label": "长信宫灯结构拆解科普图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "scene-announcement", "title": "公告汇总：浙江事业编本周截止", "title_en": "Notice Digest: Zhejiang Deadlines This Week", "subtitle": "把一个省即将截止的招考公告排成倒计时清单", "subtitle_en": "Turn a province's closing recruitment notices into a countdown list", "description": "考编最亏的不是没考上，是压根没看到公告。公告散在几十个官网入口，刷到时报名通道常已关闭。\n\n本案例实时检索浙江省 2026 年事业单位公告 185 条，按截止时间升序取最紧急的一批，逐条列报名起止、招录人数、编制情况与联考/单招属性，再按紧急程度给出处理顺序。\n\n适合：同时盯多省公告的考生、做每周招考播报的老师、帮孩子盯报名节点的家长。", "description_en": "The worst loss in exam season isn't failing — it's never seeing the announcement. Notices scatter across dozens of official sites and often close before you find them.\n\nThis case live-searches 185 Zhejiang public-institution notices for 2026, sorts by closing time, and lists registration windows, headcounts, establishment status and joint/standalone exam type, then ranks what to handle first.\n\nFor: candidates tracking multiple provinces, teachers publishing weekly briefings, and parents watching deadlines.", "categories": [ "search" ], "tags": [ "事业单位", "招考公告", "报名截止", "浙江" ], "tags_en": [ "public institution", "recruitment notice", "deadline", "Zhejiang" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "浙江事业编截止倒计时清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我盯浙江省的事业单位招考。请实时检索：1) 范围=浙江省、2026 年、事业单位，状态取「报名中」和「即将截止」，按截止时间由近到远；2) 每条列公告标题、报名起止、招录人数、职位数、地区、有无编制、联考还是单招；3) 数据必须来自官方公告实时检索；4) 最后按「今天/明天/还剩几天」分档，告诉我先处理哪条。", "prompt_en": "Track public-institution recruitment in Zhejiang for me. Live search: 1) scope=Zhejiang, 2026, public institutions, status \"open\" and \"closing soon\", sorted by nearest deadline; 2) list each notice's title, registration window, headcount, post count, region, establishment status, and joint vs standalone; 3) data must come from a live search of official notices; 4) bucket by \"today / tomorrow / N days left\" and tell me what to handle first.", "skills": [ { "id": "gaodun-notice-search", "name": "公告快速查询", "description": "按省份、类型、年份、状态检索官方招考公告并汇总关键要点", "skill_id": "skill_2089642798870765568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "stereopy-single-sample-sop", "title": "Stereopy 单样本空间聚类 SOP", "title_en": "Stereopy Single-Sample Spatial Clustering SOP", "subtitle": "从 GEF 读入到聚类注释的标准下游分析——可复制粘贴的最小可运行流水线", "subtitle_en": "Standard downstream from GEF to clustering & annotation — copy-paste-ready runnable pipeline", "description": "拿到 SAW 跑出来的 .tissue.gef 或 .cellbin.gef 之后怎么往下走？文档散落在 readthedocs 各篇教程，GEF 读写、bin_type 区别、raw_checkpoint 什么时候打、Seurat 怎么转、SingleR 注释要不要 GPU——全靠自己拼。这个案例把立体Seq专家内置的 stereopy-docs 官方文档浓缩成一份「拿到 GEF 就能跑」的单样本 SOP：覆盖环境准备（Python 3.8 + pip/conda）、GEF 读写速查（square bin / cell bin 对比 + IO 速查表）、标准流水线（read → cal_qc → filter → raw_checkpoint → normalize → HVG → PCA → neighbors → UMAP → Leiden → find_marker_genes）、QC 过滤前后分布、空间散点 + UMAP 必看原则、marker 基因解读 + SingleR 注释、Scanpy/Seurat 互转、结果保存。6 张代表性示意图全部内嵌，单文件自包含可直接当团队 SOP。适合：空间组学方向研究生、需要把 SAW 产出矩阵接上下游分析的算法团队、刚上手 Stereopy 想跳过踩坑的工程师。", "description_en": "What's next after SAW gives you a .tissue.gef or .cellbin.gef? The docs are scattered across readthedocs tutorials — GEF I/O, bin_type differences, when to call raw_checkpoint, how to convert to Seurat, whether SingleR needs GPU — all of it is a stitching job. This case distills the `stereopy-docs` skill into a 'pick up a GEF and just run' single-sample SOP covering environment setup (Python 3.8 + pip/conda), GEF I/O cheatsheet (square vs cell bin comparison + I/O table), the standard pipeline (read → cal_qc → filter → raw_checkpoint → normalize → HVG → PCA → neighbors → UMAP → Leiden → find_marker_genes), pre/post-filter QC distributions, the space-scatter + UMAP viewing principle, marker gene interpretation + SingleR annotation, Scanpy/Seurat interop, and result saving. 6 representative figures all embedded in a self-contained file. Suitable for: spatial-omics graduate students, algorithm teams bridging SAW outputs to downstream, and engineers new to Stereopy who want to skip the pitfalls.", "categories": [ "data", "education" ], "tags": [ "Stereopy", "GEF", "Leiden 聚类", "SingleR 注释", "空间可视化" ], "tags_en": [ "Stereopy", "GEF", "Leiden clustering", "SingleR annotation", "spatial visualization" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Stereopy 单样本空间聚类 SOP（自包含 HTML）" } ], "prompt": "你是 Stereo-seq 空间转录组分析专家。请基于立体Seq专家内置的 `stereopy-docs` 官方 skill 文档与教程，生成一份 Stereopy 单样本空间聚类 SOP。\n\n要求：\n1. 覆盖九个章节：环境准备（conda Python 3.8 + pip install stereopy + 验证 import）、GEF 数据读入（square bin / cell bin 概念对比 + IO 速查表 + 完整 read 代码）、标准分析流水线（read → cal_qc → filter → raw_checkpoint → normalize → HVG → PCA → neighbors → UMAP → Leiden → find_marker_genes 完整可运行代码）、QC 过滤前后分布、聚类结果可视化（空间散点 + UMAP 双视图原则）、marker 基因解读（find_marker_genes 输出表 + SingleR 注释完整代码）、格式转换与互操作（Scanpy / Seurat 风味 AnnData 互转表）、结果保存（write_h5ad / write_mid_gef / update_gef）、避坑清单（9 条 raw_checkpoint / filter / bin_size 等典型坑）。\n2. 每章必须包含：目标说明、可运行 Python 代码（带注释）、API 速查、产出文件清单。\n3. 用 matplotlib 预先生成 6 张代表性结果图（square vs cell bin 概念对比、分析流水线管线图、QC 过滤前后小提琴图、空间聚类散点、UMAP 嵌入、marker × cluster 热图）。\n4. 将所有图片以 base64 data URI 内联嵌入，产出单文件自包含 HTML（无外链）。\n5. 末尾附图表索引与数据来源声明。\n\n约束：\n- 输出单文件 HTML（output.html）：所有 CSS 内联、图片 base64 内嵌、Noto Sans SC 子集按文档字符内联嵌入、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航。\n- 所有 API、参数、典型流程均直接来自 stereopy-docs 官方文档原文（SquareBin_Clustering.ipynb、CellBin_Clustering.ipynb、SingleR.ipynb、basic-workflow.md、annotation.md），不要虚构函数签名。\n- 中文版，代码标识符保持英文生态惯例。\n- 文档可直接作为团队培训 SOP 使用。", "prompt_en": "You are a Stereo-seq spatial transcriptomics analysis expert. Using the `stereopy-docs` official skill documents and tutorials built into the stereo-seq expert, generate a Stereopy single-sample spatial clustering SOP.\n\nRequirements:\n1. Cover nine chapters: environment setup (conda Python 3.8 + pip install stereopy + import verification), GEF data loading (square vs cell bin concept comparison + I/O cheatsheet + complete read code), the standard analysis pipeline (read → cal_qc → filter → raw_checkpoint → normalize → HVG → PCA → neighbors → UMAP → Leiden → find_marker_genes with runnable code), pre/post-filter QC distributions, clustering result visualization (spatial scatter + UMAP dual-view principle), marker gene interpretation (find_marker_genes output table + SingleR annotation complete code), format conversion & interop (Scanpy / Seurat flavor AnnData table), result saving (write_h5ad / write_mid_gef / update_gef), and a 9-item pitfall checklist (raw_checkpoint / filter / bin_size typical issues).\n2. Each chapter must include: goal, runnable Python code (commented), API cheatsheet, output file list.\n3. Pre-generate 6 representative figures with matplotlib (square vs cell bin concept comparison, analysis pipeline diagram, pre/post-filter QC violin plot, spatial clustering scatter, UMAP embedding, marker × cluster heatmap).\n4. Inline all figures as base64 data URIs and produce a single-file self-contained report (no external links).\n5. Append a figure index and a data-source disclosure.\n\nConstraints:\n- Output a single-file HTML (output.html): all CSS inlined, figures base64-embedded, Noto Sans SC subset inlined by document-character filtering, no external resource references, content visible by default, no hash-anchor navigation.\n- All APIs, parameters, and typical flows come directly from the official stereopy-docs (SquareBin_Clustering.ipynb, CellBin_Clustering.ipynb, SingleR.ipynb, basic-workflow.md, annotation.md) — do not fabricate function signatures.\n- Chinese language; code identifiers follow English ecosystem conventions.\n- The document must be usable directly as team training SOP.", "skills": [], "experts": [ { "id": "stereo-seq-expert", "name": "Stereo-seq空间转录组分析专家", "description": "Stereo-seq 空间转录组全流程分析（SAW 投递 / Stereopy 下游 / Spateo 分割·对齐·3D），内嵌 saw-st-v8 / dcs-cli / saw-st-v8-makeref / saw-st-v8-realign / saw-st-v8-reanalyze-lasso / saw-st-v8-gef2gem / stereopy-docs / spateo / stereo-seq-technology-principles 共 9 个 skill", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_b2vdl3DaOp7A" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 93, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-29T16:00:00+08:00", "updated_at": "2026-08-29T18:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tencent-lit-search-case-3", "title": "变异致病性证据链检索", "title_en": "Evidence Chain Retrieval for Variant Pathogenicity", "subtitle": "CFTR变异与囊性纤维化：从病例到分子机制的深度证据链", "subtitle_en": "CFTR variant and cystic fibrosis: deep evidence chain from case reports to molecular mechanisms", "description": "遗传病变异致病性评估要求系统覆盖病例报道、遗传共分离、分子功能实验及冲突证据，但变异位点命名规则多样（cDNA/蛋白/转录本歧义），手工梳理极易遗漏关键证据且无法有效标注缺口。腾讯生命科学文献检索专家能自动识别转录本歧义并标记，联查 ClinVar、OMIM、PubMed 等多源数据库，输出分层证据链并清晰标注证据缺口。本案例以 CFTR c.1521_1523del (p.Phe508del) 与囊性纤维化的关联检索为示例，产出覆盖病例/家系、表型、遗传共分离、蛋白机制及冲突证据的深度 HTML 报告。适合遗传咨询师、罕见病研究员及临床基因组学团队使用。", "description_en": "Variant pathogenicity assessment for genetic diseases requires systematic coverage of case reports, genetic co-segregation, functional molecular data, and conflicting evidence — but variant nomenclature inconsistencies (cDNA/protein/transcript ambiguity) make manual curation error-prone and prone to evidence gaps. The Tencent Life Sciences Literature Search Expert automatically detects and flags transcript ambiguities, federates ClinVar, OMIM, PubMed, and other sources, and outputs a stratified evidence chain with clearly annotated gaps. This case demonstrates retrieval of the CFTR c.1521_1523del (p.Phe508del)–cystic fibrosis association, producing a deep HTML report covering case/family data, phenotypes, co-segregation, protein mechanisms, and conflicting evidence. Ideal for genetic counselors, rare disease researchers, and clinical genomics teams.", "categories": [ "search", "education" ], "tags": [ "变异致病性", "遗传病", "罕见病", "文献检索", "CFTR" ], "tags_en": [ "Variant Pathogenicity", "Genetic Disease", "Rare Disease", "Literature Search", "CFTR" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "tencent-lit-search-output-3.html", "artifacts": [ { "file": "tencent-lit-search-output-3.html", "label": "CFTR 变异致病性深度证据报告" } ], "prompt": "Prompt 3 ｜遗传病变异致病性证据检索\n\nCFTR c.1521_1523del (p.Phe508del) 与囊性纤维化有何关联？请做深度检索，覆盖病例与家系、汗液氯离子/胰腺/肺表型、常染色体隐性遗传与共分离证据、蛋白折叠与膜转运缺陷机制及冲突证据；若转录本不明确，先标记歧义，并说明证据缺口。", "prompt_en": "Prompt 3 | Variant Pathogenicity Evidence Retrieval for Rare Disease\n\nWhat is the association between CFTR c.1521_1523del (p.Phe508del) and cystic fibrosis? Perform a deep search covering case and family reports, sweat chloride/pancreatic/pulmonary phenotypes, autosomal recessive inheritance and co-segregation evidence, protein folding and membrane transport defect mechanisms, and conflicting evidence; if transcript identity is ambiguous, flag it and describe the evidence gaps.", "experts": [ { "id": "tencent-lit-search", "name": "生命科学文献检索", "description": "生命科学证据检索专家。覆盖药研、临床、生信及遗传与罕见病，按场景联查专业来源，结果分层去重可溯源。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "843d1587f82b", "expert_id": "ex_DXuzV9trsD3U" } ], "skills": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 120, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-24T16:32:00+08:00", "updated_at": "2026-08-25T12:18:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-workflow-visualizer", "title": "Agent 协作流程可视化", "title_en": "AI Agent Workflow Visualizer", "subtitle": "交互式节点图展示多 Agent 协作流水线", "subtitle_en": "Interactive node graph for multi-agent collaboration pipelines", "description": "想理解多 Agent 协作到底怎么运作，但文字描述总是抽象难懂？这个交互式可视化器用动画节点图展示 Agent 协作全流程——Team Lead 分配任务、Researcher/Analyst/Writer 并行执行、Reviewer 质审反馈，连线动画模拟数据流转，点击节点查看详情面板。让 AI Agent 的协作编排一目了然。适合 AI 产品经理、开发者、技术管理者理解和展示 Agent 架构。", "description_en": "Struggling to explain how multi-agent collaboration works beyond abstract text? This interactive visualizer uses animated node graphs to show the entire Agent workflow — Team Lead dispatches tasks, Researcher/Analyst/Writer execute in parallel, Reviewer provides quality feedback. Animated connections simulate data flow, click any node for details. Perfect for AI product managers, developers, and tech leads demonstrating Agent architectures.", "categories": [ "agent", "dev" ], "tags": [ "Agent 协作", "流程可视化", "节点图", "多 Agent", "工作流" ], "tags_en": [ "Agent Collaboration", "Workflow Visualization", "Node Graph", "Multi-Agent", "Workflow" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Agent 协作流程可视化器" } ], "cover_image": "cover.png", "cover_mode": "landscape", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" }, { "id": "impeccable", "name": "高品质前端设计", "description": "高品质 UI/UX 设计工具集，避免泛 AI 审美", "skill_id": "skill_2053082862415904768" } ], "experts": [], "prompt": "请生成一个 AI Agent 协作流程可视化器单页（单文件 HTML，内联无外部依赖），暗色主题，用 SVG 绘制交互式节点图，展示「智能研报生成」多 Agent 流水线。\n\n流程：团队协调者 → 并行分发到「信息采集 / 数据分析 / 内容撰写」三个 Agent → 质量审核 Agent → 最终交付物；审核不通过时回退到撰写 Agent。\n\n交互：播放按钮触发节点逐个点亮 + 连线流动动画；点击节点弹出详情面板（角色描述/输入输出/工具列表）；底部展示流水线统计（节点数/并行分支/预估耗时/Token 用量/质量评分）。", "prompt_en": "Please create a single-page AI Agent collaboration workflow visualizer (single-file HTML, inline with no external dependencies), with a dark theme, using SVG to draw an interactive node graph that shows the multi-agent pipeline for \"Intelligent Research Report Generation.\"\n\nWorkflow: Team Coordinator → parallel dispatch to three agents (\"Information Gathering / Data Analysis / Content Writing\") → Quality Review Agent → final deliverable; if review fails, fall back to the Writing Agent.\n\nInteraction: a play button triggers the nodes lighting up one by one plus a flowing animation along the connecting lines; clicking a node pops up a detail panel (role description / inputs and outputs / tool list); the bottom displays pipeline statistics (node count / parallel branches / estimated duration / token usage / quality score).", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "badge": "official", "avatar": "" }, "source": "official", "quality_score": 93, "created_at": "2026-05-03T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 121, "featured_order": 0, "usage_count": 0, "is_featured": false, "mcps": [], "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-notice-teacher-multi", "title": "公告检索：三省教师招聘在招清单", "title_en": "Announcement Search: Live Teacher Recruitment Across Three Provinces", "subtitle": "一次拉齐山东、河南、四川的在招教师公告，按规模与编制排序", "subtitle_en": "Pull live teacher postings from three provinces, ranked by scale and tenure track", "description": "教师岗公告发得又碎又快，一个县一份、一所学校一份，逐省翻很容易错过刚发布的。\n\n本案例一次传入三个省份 + 状态=报名中检索，命中 279 份在招公告，取最新 20 份，按招录人数排序列出地区、编制情况、岗位数与报名时间，统计省份分布与有编/无编比例，并给出一套「怎么用这份清单」的操作顺序。\n\n适合：跨省找教师岗的考生、需要每天扫新公告的备考者。", "description_en": "Teacher postings are fragmented and fast-moving — one per county, one per school — and checking province by province means missing fresh ones.\n\nThis case passes three provinces plus status \"open\" in a single query, hitting 279 live announcements. The 20 newest are ranked by headcount with region, tenure-track status, post count and application window, plus a province breakdown and a step-by-step way to work the list.\n\nFor: candidates searching across provinces, and anyone scanning new postings daily.", "categories": [ "search" ], "tags": [ "教师招聘", "多省检索", "在招公告", "编制" ], "tags_en": [ "teacher recruitment", "multi-province", "live postings", "tenure track" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "三省教师招聘在招公告清单", "label_en": "Live teacher recruitment across three provinces" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我一次性查山东、河南、四川三个省正在报名的教师招聘公告。要求：1) 真实检索并给出命中总数；2) 取最新 20 份，按招录人数从多到少列出公告标题、地区、是否有编制、招录人数、岗位数、报名时间；3) 统计这 20 份的省份分布和有编制/无编制/未标注各几份；4) 说明教师岗公告与省考大公告在颗粒度上的区别；5) 给出一套使用这份清单的操作顺序（先看什么、再看什么）。", "prompt_en": "Query live teacher-recruitment announcements across Shandong, Henan and Sichuan in one go. Requirements: 1) run the search and give the total; 2) take the 20 newest, ranked by headcount, listing title, region, tenure-track status, headcount, post count and application window; 3) break down the 20 by province and by tenure-track/non-tenure/unlabelled; 4) explain how teacher postings differ in granularity from large provincial-exam announcements; 5) give a practical order of operations for working through the list.", "skills": [ { "id": "gaodun-notice-search", "name": "公告快速查询", "description": "按考试类型、地区、招考状态检索招考公告", "skill_id": "skill_2089642798870765568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 121, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-compare-shang-vs-zhou-bronze", "title": "商代青铜器 vs 周代青铜器", "title_en": "Shang Bronze vs Zhou Bronze: From Divine Kingship to Ritual Order", "subtitle": "饕餮神权之器 vs 列鼎礼制之器：纹饰·礼制·铭文三轴跨越 800 年", "subtitle_en": "From Taotie-fuelled Shang ritual vessels of divine authority to ranked-ding Western Zhou vessels of Confucian order — three axes: pattern, ritual system, inscription", "description": "把一件青铜器摆你面前，如何一眼判断它是商代的、周代的、还是中间过渡的？这是博物馆讲解员、青铜器收藏者、艺术史学生最常被问到、也最容易被答错的问题。这个案例演示了如何用腾讯探元文博知识库按**年代体系粒度**分别检索两个时代的青铜主题，让对比的维度（纹饰主线、礼制用途、铸造工艺、铭文风格、器型与足型、出土分布）自然浮现，给出一份「**不用查文献就能上展柜讲解**」的三轴对照清单。\n\n适合：青铜器爱好者与收藏者、博物馆讲解员与策展人、中国古代史与先秦史研究者、艺术史与美术专业师生、考古与文化遗产保护从业者。", "description_en": "Put a bronze vessel in front of you — can you tell whether it's Shang, Zhou, or transitional? This is the question museum docents, bronze collectors, and art-history students get asked most often and answer wrong most often. This case shows how the TanYuan knowledge base uses **chronological-system granularity** to query each era's bronze topic independently so the comparison dimensions (pattern keyline, ritual use, casting technique, inscription style, vessel and foot shapes, excavation distribution) emerge naturally — producing a three-axis checklist you can use at a museum vitrine without consulting references. Best for: bronze collectors & enthusiasts, museum docents & curators, ancient Chinese and pre-Qin historians, art-history students, archaeology and heritage practitioners.", "categories": [ "business" ], "tags": [ "青铜器", "商代", "周代", "列鼎制度", "饕餮纹" ], "tags_en": [ "Chinese Bronze", "Shang Dynasty", "Zhou Dynasty", "Tripod Ranking System", "Taotie Pattern" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Markdown 对比报告（一句话结论 + 对比表 + 快速辨识要点 + 代表器物 + 来源）" } ], "prompt": "请按「腾讯探元文物对比分析师」专家的工作流，对比分析**商代青铜器**与**周代青铜器**在纹饰主题、铸造工艺、礼制用途（列鼎制度）、器型与器类、胎体与铭文、器型风格、出土分布上的差异。\n\n研究范围：\n- 对象 A：商代青铜器（约公元前 16 世纪—前 11 世纪，含商前期与商晚期）\n- 对象 B：周代青铜器（公元前 11 世纪—前 256 年，含西周早/中/晚期、春秋、战国）\n\n用户未点名维度，按本次实际能检到的资料决定补充维度（建议但不强制）：\n- 纹饰主题（饕餮纹 / 夔龙纹 / 鸟纹 vs 窃曲纹 / 波曲纹 / 重环纹 / 瓦楞纹 / 鸟纹）\n- 铸造工艺（范铸法 vs 失蜡法的出现时机）\n- 礼制用途（重酒礼器 vs 列鼎制度、九鼎八簋）\n- 器型风格（庄严神秘 vs 质朴规整；柱足 / 锥足 / 扁足 vs 蹄足）\n- 铭文风格（族徽日名 vs 长篇册命、玉箸体）\n- 出土分布（安阳殷墟 / 郑州商城 vs 宝鸡周原 / 琉璃河 / 曲村）\n\n交付要求（强制）：\n1. 粒度识别为「年代体系」，**只用探元知识库（mcp__tanyuan-wb-mcp-server__search_movable_relics）分别检索两个时代主题**，query: \"商代青铜器\"、\"西周青铜器\"、\"列鼎制度\"、\"饕餮纹\"、\"窃曲纹\"。**代表器物（后母戊鼎、四羊方尊、子龙鼎、毛公鼎、散氏盘、大盂鼎、利簋、何尊）必须走联网核对**\n2. **绝不能用一次 query 问「商周青铜器的区别」**，必须分次检索\n3. **search_relics 仅在举例证时使用**，不得用其返回的「地方馆藏品」充当「代表作清单」\n4. 输出严格遵循专家骨架：**一句话结论 → 对比表（每行同一件事）→ 快速辨识要点（3—5 条可操作动作）→ 代表器物（每方 3—6 件附馆藏与时代）→ 来源（探元知识库 + 中国国家博物馆 chnmuseum.cn + 联网补充）**\n5. 空白格写「公开资料未见明确记录」，**不留空、不编造**\n6. **不向用户暴露内部运行信息**——禁止出现 query、检索、工具名、requestId、条数机制词", "prompt_en": "Use the 'TanYuan Cultural Relic Comparison Analyst' expert workflow to comparatively analyse the differences between **Shang bronzes** and **Zhou bronzes** across pattern keyline, casting technique, ritual use (the tripod ranking system), vessel shape and category, body and inscription style, style and aesthetic, and excavation distribution.\n\nScope:\n- Object A: Shang bronzes (~16th–11th c. BCE)\n- Object B: Zhou bronzes (11th c.–256 BCE, covering Western Zhou early/middle/late, Spring & Autumn, Warring States)\n\nThe user has not named dimensions — let evidence decide (recommended but not required):\n- Pattern keyline (Taotie / Kui-dragon / bird vs 窃曲 qūqū / 波曲 bōqū / double-ring / tile-ridge / long-tailed bird)\n- Casting (piece-mould vs emergence of lost-wax)\n- Ritual use (wine-vessel emphasis vs tripod-ranking system, nine ding + eight gui)\n- Vessel shape (solemn and mysterious vs austere and orderly; post-foot columnar/conical/hoof)\n- Inscription (clan emblem + day-name vs long commissioning inscriptions, yùzhù body)\n- Excavation distribution (Yinxu Anyang / Zhengzhou Shang vs Zhouyuan Baoji / Liulihe / Qucun)\n\nDelivery requirements (mandatory):\n1. Granularity is 'chronological system' — query the TanYuan knowledge base **only** via `mcp__tanyuan-wb-mcp-server__search_movable_relics`, once each for 'Shang bronze', 'Western Zhou bronze', 'tripod ranking system', 'Taotie pattern', 'qūqū pattern'. **Representative pieces (Houmuwu Ding, Four-Goat Square Zun, Zilong Ding, Maogong Ding, Sanzhong Pan, Dayu Ding, Li Gui, He Zun) MUST be cross-verified via public web**.\n2. **Never** combine eras in a single query ('difference between Shang and Zhou bronzes'). Query separately.\n3. **search_relics only when pointing to a specific piece**; never use its regional-catalog hits as a 'representative list'.\n4. Output follows the expert skeleton strictly: **one-line conclusion → comparison table (every row one topic) → 3–5 actionable quick-identification tips → 3–6 representative pieces per side with custodian and date → sources (TanYuan + chnmuseum.cn + web supplements)**.\n5. Empty cells write 'no public record found', **never leave blanks, never fabricate**.\n6. **Never expose internal mechanics** — no queries, no search log, no tool names, no requestId, no result-count mechanism words.", "skills": [], "experts": [ { "id": "tanyuan-relic-compare-expert", "name": "文物对比专家", "description": "把两个或多个文博对象在同一套维度下讲清楚异同：一句话结论 + 对比表 + 快速辨识要点 + 代表器物。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1ce329117e4c", "expert_id": "ex_9FcVuwD0NDKB" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 121, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-08-14T19:04:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-notice-large-scale", "title": "公告检索：在招公告招录人数榜", "title_en": "Announcement Search: Ranking Live Postings by Headcount", "subtitle": "把排序从「最新」换成「招考人数多」，看到的是另一张榜单", "subtitle_en": "Switch the sort from newest to largest, and a different list appears", "description": "每天刷最新公告，会一直看到小规模的单位招聘，而省级联考这种上万人的机会可能翻十几页都看不到。\n\n本案例把排序方式换成「招考人数多」，命中 2,779 份在招公告，取前 20 份排出招录人数榜——榜首是包含 111 份子公告的省级联考专题。案例讲清汇总专题与单份公告的区别，以及两种排序为什么必须都用。\n\n适合：规划全年备考节奏的考生、想抓住大规模招录窗口的人。", "description_en": "Scanning newest-first shows a steady stream of small single-unit postings, while a provincial joint exam recruiting tens of thousands can sit pages deep.\n\nThis case switches the sort to \"largest headcount\", hitting 2,779 live announcements. The top 20 form a headcount ranking led by a provincial joint-exam topic bundling 111 sub-announcements. It explains how summary topics differ from individual postings and why both sort orders are needed.\n\nFor: candidates planning a year-long schedule and those chasing large recruitment windows.", "categories": [ "search" ], "tags": [ "招录规模", "汇总专题", "省级联考", "公告排序" ], "tags_en": [ "recruitment scale", "summary topic", "joint exam", "announcement sorting" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "在招公告招录人数榜", "label_en": "Live announcements ranked by headcount" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我按「招考人数多」的排序查正在报名的公务员和事业单位公告。要求：1) 真实检索并给出总数；2) 取前 20 份，列出公告标题、是汇总专题还是单份公告、招录人数、岗位数、包含几份子公告；3) 说明汇总专题（招录人数带「+」号的那类）和单份公告在使用上有什么区别、各自该怎么用；4) 对比「按最新排序」和「按招考人数排序」分别会看到什么，说明为什么两种都要用；5) 全部数据用真实检索结果，不要举例说明。", "prompt_en": "Search live civil-service and public-institution announcements sorted by largest headcount. Requirements: 1) run the query and give the total; 2) take the top 20, listing title, whether it is a summary topic or a single posting, headcount, post count and number of sub-announcements; 3) explain how summary topics (those with a \"+\" in the headcount) differ from single postings and how to use each; 4) contrast what newest-first versus largest-first sorting surfaces and why both are needed; 5) use real query results throughout, not illustrative examples.", "skills": [ { "id": "gaodun-notice-search", "name": "公告快速查询", "description": "按考试类型、地区、招考状态检索招考公告", "skill_id": "skill_2089642798870765568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 122, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "job-application-tracker", "title": "求职进度追踪表", "title_en": "Job Application Tracker", "subtitle": "输入求职目标，生成系统化的求职进度追踪与分析表格", "subtitle_en": "Input job search goals, generate systematic application tracking and analysis spreadsheet", "description": "求职季投了几十家公司，哪些回复了、哪些进入面试、哪些 offer 在谈——用备忘录记根本管不过来。邮件翻不到、面试时间记混、薪资谈判没有参照系，手忙脚乱中错过心仪的机会。\n\n这个案例演示了输入求职目标信息（目标岗位、城市、薪资期望），AI 自动生成一份完整的求职进度追踪 Excel 表格：投递总览仪表盘（投递数/回复率/面试率/offer率）、30 条投递明细记录（公司/职位/薪资范围/状态/面试轮次）、Top 5 公司面试准备清单、3 个 offer 薪资福利对比分析表。条件格式标注各阶段状态，数据验证确保录入规范。\n\n适合：应届毕业生秋招春招、社招跳槽者、猎头管理候选人进度。", "description_en": "During job hunting season, tracking dozens of applications becomes overwhelming — which companies responded, which progressed to interviews, which offers are being negotiated. Notes apps can't keep up. Missed emails, confused interview schedules, and salary negotiations without reference points lead to lost opportunities.\n\nThis case demonstrates inputting job search goals (target position, city, salary expectations), with AI generating a complete job application tracking Excel spreadsheet: application overview dashboard (applications/response rate/interview rate/offer rate), 30 detailed application records (company/position/salary range/status/interview rounds), top 5 company interview preparation checklist, and 3-offer salary comparison analysis. Conditional formatting highlights stage status, data validation ensures consistent entry.\n\nBest for: Fresh graduates during recruitment season, mid-career job switchers, recruiters managing candidate pipelines.", "categories": [ "data", "office" ], "tags": [ "求职追踪", "面试管理", "Excel", "求职工具", "薪资对比" ], "tags_en": [ "Job Tracking", "Interview Management", "Excel", "Job Search Tool", "Salary Comparison" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "求职进度追踪表" } ], "prompt": "请为一位准备跳槽的 3 年经验前端工程师生成一份求职进度追踪 Excel（output.xlsx）。\n\n求职者：目标岗位高级前端 / 前端 Tech Lead，目标城市北京、上海、深圳、杭州，期望薪资 40-55K，求职周期 2026 年 4-5 月。\n\n请生成以下工作表：\n1. 投递总览：总投递数、回复率、面试率、offer 率，按状态 / 城市 / 渠道分布\n2. 投递明细（约 30 条）：公司、职位、城市、薪资范围、渠道、投递日期、当前状态、面试轮次、下一步等；状态分布要真实，含 3 个 offer\n3. 面试准备清单：进入面试的 Top 5 公司，含面试阶段、形式、需准备的知识点\n4. 薪资对比：对比 3 个 offer 的月薪、年终奖、股票期权、福利、年度总包\n\n要求：状态列用数据验证下拉，关键状态加条件格式，统计数据用公式汇总。输出为 .xlsx 文件。", "prompt_en": "Please generate a job-search progress tracking Excel (output.xlsx) for a front-end engineer with 3 years of experience preparing to switch jobs.\n\nCandidate: target positions Senior Front-end / Front-end Tech Lead, target cities Beijing, Shanghai, Shenzhen, Hangzhou, expected salary 40-55K, job-search period April-May 2026.\n\nPlease generate these worksheets:\n1. Application Overview: total applications, reply rate, interview rate, offer rate, distribution by status / city / channel\n2. Application Detail (about 30 records): company, position, city, salary range, channel, application date, current status, interview round, next action, etc.; the status distribution should be realistic, including 3 offers\n3. Interview Prep Checklist: the Top 5 companies reaching interviews, with stage, format, knowledge points to prepare\n4. Salary Comparison: compare 3 offers on monthly salary, annual bonus, stock/options, benefits, total annual package\n\nRequirements: data-validation dropdown for the status column, conditional formatting on key statuses, formula-based summaries. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 122, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "relic-compare-tangsancai-vs-qingbai", "title": "唐三彩 vs 宋青白瓷", "title_en": "Tang Sancai vs Song Qingbai: Tomb Ware vs Daily-use Porcelain", "subtitle": "低温铅釉多彩陶器 vs 高温单色青白瓷器：陶与瓷、墓与桌、明器与日用之别", "subtitle_en": "Low-fire lead-glazed polychrome pottery vs high-fire monochrome Qingbai porcelain — earthenware vs porcelain, tomb ware vs tableware", "description": "唐三彩是陶器、宋青白瓷是瓷器——这一基本分野很多爱好者都模糊。可一旦上手触摸、对光照射、侧耳倾听，二者差异其实一秒钟就能感知。这个案例演示了如何用腾讯探元文博知识库按「**品类粒度**」分别检索多彩低温釉陶与单色高温青白瓷，让「**工艺类别、釉色、胎体、用途、成型装饰、窑口**」六维对比自然浮现，并把最能上手的「颜色数量 + 敲声 + 照光」三招放在辨识要点首位。\n\n适合：陶瓷史爱好者、博物馆观众（陕西历史博物馆 / 河南博物院 / 景德镇陶瓷馆常客）、文博自媒体编辑、文物拍卖入门者、艺术与考古专业师生。", "description_en": "Tang Sancai is pottery, Song Qingbai is porcelain — even enthusiasts blur this basic distinction. Yet the moment you touch them, hold them to the light, and tap them, the gap shows in a second. This case shows how the TanYuan knowledge base uses **category granularity** to query polychrome low-fire earthenware and monochrome high-fire porcelain independently, surfacing a six-axis comparison (process category, glaze, body, use, shaping/ornament, kiln) — and places the three most hands-on cues (count the colours + listen to the ring + hold to the light) at the top of the quick-identification list. Best for: ceramics-history enthusiasts, museum visitors (Shaanxi History Museum / Henan Museum / Jingdezhen Ceramics Museum regulars), cultural-heritage editors, newcomers to relic auctions, and art/archaeology students.", "categories": [ "business" ], "tags": [ "唐三彩", "青白瓷", "影青", "景德镇", "陶器与瓷器" ], "tags_en": [ "Tang Sancai", "Qingbai Porcelain", "Yingqing", "Jingdezhen", "Earthenware vs Porcelain" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Markdown 对比报告（一句话结论 + 对比表 + 快速辨识要点 + 代表器物 + 来源）" } ], "prompt": "请按「腾讯探元文物对比分析师」专家的工作流，对比分析**唐三彩**与**宋青白瓷**（又称影青）在工艺类别、釉色、胎体、用途、成型与装饰、代表窑口六维上的差异。\n\n研究范围：\n- 对象 A：唐三彩（唐代低温铅釉多彩陶器；以河南巩县窑、陕西铜川窑为代表）\n- 对象 B：宋青白瓷 / 影青（宋代高温单色青白瓷器；以江西景德镇湖田窑为代表）\n\n用户未点名维度，按本次实际能检到的资料决定补充维度（建议但不强制）：\n- 工艺类别（陶器 vs 瓷器；低温铅釉 750—800 °C vs 高温还原焰 1200+ °C）\n- 釉色（黄/绿/白/蓝多彩 vs 青白单色 / 影青）\n- 胎体（白色粘土松脆陶胎 vs 高硅致密瓷胎）\n- 主要用途（随葬明器 vs 日用器与陈设器）\n- 成型与装饰（模制/雕塑/贴花 vs 轮制/刻花/印花）\n- 代表窑口与釉色机理（多金属氧化物呈色 vs 少量铁还原青色调）\n\n交付要求（强制）：\n1. 粒度识别为「品类 / 工艺体系」，**只用探元知识库（mcp__tanyuan-wb-mcp-server__search_movable_relics）分别检索两个品类主题**，query: \"唐三彩\"、\"青白瓷\"、\"景德镇窑\"、\"巩县窑\"。**代表器物（唐三彩骆驼载乐俑、唐三彩女立俑、唐三彩天王俑、宋景德镇窑青白釉刻花梅瓶、宋景德镇窑青白釉注子注碗、宋湖田窑青白釉印花碗）必须走联网核对**\n2. **绝不能用一次 query 问「唐三彩和青白瓷的区别」**，必须分次检索\n3. 输出严格遵循专家骨架：**一句话结论 → 对比表（六维或更多实质维度，每行同一件事）→ 快速辨识要点（3—5 条可操作动作，按「看颜色/听声音/照光」的优先级排序）→ 代表器物（每方 3—5 件附馆藏）→ 来源（探元知识库 + 陕西历史博物馆 + 故宫博物院 + 景德镇陶瓷馆 + 联网补充）**\n4. 空白格写「公开资料未见明确记录」，**不留空、不编造**\n5. **不向用户暴露内部运行信息**——禁止出现 query、检索、工具名、requestId、条数机制词", "prompt_en": "Use the 'TanYuan Cultural Relic Comparison Analyst' expert workflow to comparatively analyse the differences between **Tang Sancai** and **Song Qingbai** (aka Yingqing) along six axes: process category, glaze, body, primary use, shaping and ornament, representative kiln site.\n\nScope:\n- Object A: Tang Sancai (Tang low-fire lead-glazed polychrome pottery; Gongxian kiln Henan, Tongchuan kiln Shaanxi as leading producers)\n- Object B: Song Qingbai / Yingqing (Song high-fire monochrome Qingbai porcelain; Jingdezhen Hutian kiln Jiangxi as the representative producer)\n\nThe user has not named dimensions — let evidence decide (recommended but not required):\n- Process category (earthenware vs porcelain; low-fire lead glaze 750–800 °C vs high-fire reduction 1200+ °C)\n- Glaze (yellow/green/white/blue polychrome vs Qingbai monochrome / Yingqing)\n- Body (white clay porous vs high-silica vitrified porcelain)\n- Primary use (tomb mingqi vs tableware and display ware)\n- Shaping and ornament (moulding / sculptural / appliqué vs wheel-throwing / incised / impressed)\n- Representative kilns and colouring mechanism (poly-metal-oxide coloration vs iron-reduction cyan tonality)\n\nDelivery requirements (mandatory):\n1. Granularity is 'category / craft system' — query the TanYuan knowledge base **only** via `mcp__tanyuan-wb-mcp-server__search_movable_relics`, once each for 'Tang Sancai', 'Qingbai porcelain', 'Jingdezhen kiln', 'Gongxian kiln'. **Representative pieces (Tang Sancai camel-musicians, Tang Sancai standing lady, Tang Sancai heavenly king, Song Jingdezhen Qingbai carved-meiping, Song Jingdezhen Qingbai ewer-and-bowl, Song Hutian Qingbai impressed bowl) MUST be cross-verified via public web**.\n2. **Never** combine categories in a single query ('difference between Tang Sancai and Qingbai'). Query separately.\n3. Output follows the expert skeleton strictly: **one-line conclusion → comparison table (every row one topic) → 3–5 actionable quick-identification tips ordered by 'count colours / listen to the ring / hold to the light' priority → 3–5 representative pieces per side with current custodian → sources (TanYuan + Shaanxi History Museum + Palace Museum + Jingdezhen Ceramics Museum + web supplements)**.\n4. Empty cells write 'no public record found', **never leave blanks, never fabricate**.\n5. **Never expose internal mechanics** — no queries, no search log, no tool names, no requestId, no result-count mechanism words.", "skills": [], "experts": [ { "id": "tanyuan-relic-compare-expert", "name": "文物对比专家", "description": "把两个或多个文博对象在同一套维度下讲清楚异同：一句话结论 + 对比表 + 快速辨识要点 + 代表器物。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "1ce329117e4c", "expert_id": "ex_9FcVuwD0NDKB" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 122, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-14T19:04:00+08:00", "updated_at": "2026-08-31T12:20:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "marketing-tax-ai-launch-campaign", "title": "税务小智发布战役方案", "title_en": "AI Tax SaaS Launch Campaign", "subtitle": "3个月冲刺5000家付费企业的完整新品发布打法", "subtitle_en": "A full launch playbook to win 5,000 paying companies in 3 months", "description": "新品发布最怕的是目标很大、渠道很多、团队很忙，但没有一份真正能落地的整合营销方案——策略、品牌定位、SEO、内容节奏、效果复盘各做各的，最后无法形成合力。本案例演示营销战役团队如何为面向中小企业的 AI 财税 SaaS“税务小智”设计一套完整新品发布战役：从目标/KPI、受众画像、差异化 messaging、预算与时间线，到 SEO 关键词策略、12 周内容日历、16 篇首批核心内容，再到汇报节奏与 ROI 管理，一次性产出 Markdown 完整版 + HTML 内网展示版。\n\n适合：市场负责人、增长负责人、品牌经理、内容团队、B2B SaaS 创始人、渠道运营", "description_en": "Launching a new product usually breaks down when the goal is big, channels are many, and teams are busy—but there is no single integrated plan tying strategy, positioning, SEO, content cadence, and reporting together. This case shows how the Marketing Campaign Team builds a complete launch campaign for 'Tax Small Genius', an AI tax SaaS for SMBs: from goals/KPIs, personas, differentiated messaging, budget and timeline, to SEO keyword strategy, a 12-week content calendar, 16 launch-ready content assets, and a reporting rhythm with ROI management. Outputs both a full Markdown plan and an HTML intranet-ready version.\n\nBest for: Marketing Leads, Growth Leads, Brand Managers, Content Teams, B2B SaaS Founders, Channel Operators", "categories": [ "business", "creation", "agent" ], "tags": [ "新品发布", "整合营销", "B2B SaaS", "SEO策略", "内容日历" ], "tags_en": [ "product launch", "integrated marketing", "B2B SaaS", "SEO strategy", "content calendar" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "新品发布战役 HTML 展示版" }, { "file": "output.md", "label": "新品发布战役 Markdown 完整版" } ], "prompt": "下个月要发布新产品——面向中小企业的 AI 财税 SaaS \"税务小智\"，目标 3 个月内 5000 家付费企业。请出完整营销战役方案：\n1. 战役策略（目标/受众画像/核心 messaging/渠道组合/预算建议/时间线）\n2. 品牌差异化定位（对标 2-3 个竞品做 messaging 对比）\n3. SEO 关键词策略与内容优化建议\n4. 12 周内容日历 + 首批核心内容（公众号长文 3 篇、短视频脚本 3 个、邮件序列 5 封、小红书 5 篇）\n5. 效果评估指标与汇报节奏\n\n输出 Markdown 完整战役方案 + HTML 版本（可直接挂在公司内网）。", "prompt_en": "Next month we're launching a new product—an AI tax & finance SaaS for small and medium businesses called \"Tax Xiaozhi\"—with a goal of 5,000 paying companies within 3 months. Please produce a complete marketing campaign plan:\n1. Campaign strategy (goals/audience personas/core messaging/channel mix/budget recommendations/timeline)\n2. Brand differentiation positioning (compare messaging against 2-3 competitors)\n3. SEO keyword strategy and content optimization suggestions\n4. A 12-week content calendar + the first batch of core content (3 long-form WeChat Official Account articles, 3 short-video scripts, a 5-email sequence, 5 Xiaohongshu posts)\n5. Performance metrics and reporting cadence\n\nOutput the complete campaign plan in Markdown + an HTML version (ready to host on the company intranet).", "skills": [], "experts": [ { "id": "MarketingCampaignTeam", "name": "营销战役团队", "description": "由营销总监领导的 4 人营销专家团队：内容创作者（博客/邮件/社媒/品牌声音）、活动策划师（战役策略/受众/渠道/预算）、SEO 专家（技术审计/内容优化/效果分析）和品牌分析师（竞品定位/品牌审核）。覆盖完整营销生命周期。", "expertType": "team", "source": "expert-marketplace", "name_en": "Marketing Campaign Team", "description_en": "A team of 4 marketing specialists led by a Marketing Director: Content Creator (blog posts/email/social media/brand voice), Campaign Planner (campaign strategy/audience/channels/budget), SEO Expert (technical audit/content optimization/performance analytics), and Brand Analyst (competitive positioning/brand review). Covers the full marketing lifecycle.", "expert_id": "ex_wg8SWSt1T33S" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 123, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:31:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "annual-conference-plan", "title": "年会策划方案", "title_en": "Annual Conference Plan", "subtitle": "500人年度盛典全案，从主题到执行一键生成 Word 文档", "subtitle_en": "Complete Word document for a 500-person annual gala — theme to execution", "description": "年会策划千头万绪，预算怎么分、节目怎么排、场地怎么选，每年都是一场噩梦。这个案例展示了如何用 Word 文档生成技能一键输出一份500人年会完整策划案，包含主题设计、预算分配、时间线、节目单、场地布置、供应商清单、应急预案，直接生成可打印签批的正式 Word 文档。\n\n从「共创·共赢·2026」主题出发，方案覆盖30万预算的精细化拆分、T-30天倒计时推进表、3小时活动流程编排、5个工作组的职责矩阵、3家候选酒店对比、6类应急预案。每个板块都给出了可直接落地的表格、清单和执行标准，生成 Word 格式拿来就能用。\n\n适合：HR、行政主管、企业文化负责人、活动策划公司。", "description_en": "Planning an annual conference is overwhelming — how to allocate the budget, arrange performances, choose a venue — it's a nightmare every year. This case demonstrates using the Word Document Generation skill to produce a complete plan for a 500-person annual gala as a formal Word document, covering theme design, budget allocation, timeline, program schedule, venue setup, vendor list, and contingency plans.\n\nStarting from the theme \"Co-Create · Win Together · 2026\", the plan covers detailed breakdown of a 300K CNY budget, a T-30 day countdown schedule, a 3-hour event flow, responsibility matrix for 5 working groups, comparison of 3 candidate hotels, and 6 types of contingency plans. Every section provides ready-to-use tables, checklists, and execution standards in a printable Word format.\n\nBest for: HR managers, administrative supervisors, corporate culture leads, and event planning agencies.", "categories": [ "business", "office" ], "tags": [ "年会策划", "活动方案", "企业文化", "预算管理", "执行手册" ], "tags_en": [ "Annual Gala", "Event Planning", "Corporate Culture", "Budget Management", "Execution Handbook" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "年会策划全案（Word）" } ], "prompt": "请帮我生成一份500人规模的企业年会策划方案 Word 文档（output.docx）。\n\n要求：\n1. 年会主题：「共创·共赢·2026」，需包含主视觉描述、色调方案和 slogan\n2. 总预算30万，需要详细的预算明细表\n3. T-30天倒计时推进表\n4. 3家候选酒店对比表\n5. 3小时完整活动流程（开场→颁奖→表演→互动→晚宴）\n6. 5个工作组（统筹组/节目组/后勤组/技术组/宣传组）职责分工矩阵\n7. 供应商清单和评估标准\n8. 6种常见突发状况应急预案\n9. 文档格式要求：包含封面页、目录、页眉页脚、多级标题编号、统一表格样式\n10. 输出为正式 Word 文档，可直接打印使用", "prompt_en": "Please generate a Word document (output.docx) with a corporate annual-gala planning proposal for a 500-person event.\n\nRequirements:\n1. Gala theme: \"Co-create · Win Together · 2026\", including a description of the key visual, a color scheme, and a slogan\n2. A total budget of ¥300,000, with a detailed budget breakdown table\n3. A T-30 day countdown schedule\n4. A comparison table of 3 candidate hotels\n5. A complete 3-hour event flow (opening → awards → performances → interactive segments → banquet)\n6. A responsibility/role matrix for 5 working groups (Coordination / Program / Logistics / Technical / PR)\n7. A vendor list and evaluation criteria\n8. Contingency plans for 6 common emergencies\n9. Document formatting requirements: include a cover page, table of contents, headers/footers, multi-level heading numbering, and a consistent table style\n10. Output as a formal Word document, ready to print and use directly", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 124, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "3d-data-visualization-screen", "title": "3D 数据可视化大屏", "title_en": "3D Data Visualization Dashboard", "subtitle": "CSS 3D 变换 × 伪等距投影 × 动态数据流，汇报级数据大屏", "subtitle_en": "CSS 3D transforms, isometric projection & animated data flows for executive dashboards", "description": "年终汇报、战略展示最怕数据页面平平无奇——一堆 2D 饼图柱图根本镇不住场。这个案例演示了一个 3D 风格数据可视化大屏：用 CSS 3D 变换实现伪等距投影效果，包含 3D 柱状图、环形指标盘、网络拓扑连线、实时数据流动画，整体视觉效果对标企业展厅级数据大屏。\n\n适合：数据分析师做年度汇报、市场运营做品牌数据展示、技术团队做架构监控大屏。", "description_en": "Annual reports and strategy presentations fall flat when data pages look ordinary — a bunch of 2D pie charts and bar graphs simply cannot impress. This case demonstrates a 3D-style data visualization dashboard: CSS 3D transforms create pseudo-isometric projection effects, featuring 3D bar charts, rotating ring indicators, network topology connections, and real-time data flow animations — delivering enterprise exhibition-grade visual impact.\n\nIdeal for: data analysts preparing annual reports, marketing teams building brand data showcases, and tech teams designing architecture monitoring dashboards.", "categories": [ "data", "creation" ], "tags": [ "3D可视化", "数据大屏", "CSS动画", "数据展示" ], "tags_en": [ "3D Visualization", "Data Dashboard", "CSS Animation", "Data Presentation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "3D数据可视化大屏页面" } ], "prompt": "请生成一个「智慧城市运营监控中心」3D 数据可视化大屏页面（单文件 HTML，内联无外部依赖），暗色主题、伪 3D 立体感、带数据动画。\n\n用 CSS 3D 变换营造立体投影效果，包含 3D 柱状图、环形指标盘、数据节点连线动画、关键指标发光脉冲、数字计数动画。布局：顶部 KPI 面板 + 中间 3D 图表 + 底部趋势线。监控指标：人口流量、能耗指数、交通密度、空气质量。", "prompt_en": "Please create a \"Smart City Operations Monitoring Center\" 3D data visualization dashboard page (single-file HTML, fully inline with no external dependencies), with a dark theme, a pseudo-3D stereoscopic look, and animated data.\n\nUse CSS 3D transforms to create a stereoscopic projection effect, including a 3D bar chart, ring-shaped gauge dials, animated connecting lines between data nodes, glowing pulse effects on key metrics, and number count-up animations. Layout: a top KPI panel + 3D charts in the middle + a trend line at the bottom. Monitored metrics: population flow, energy consumption index, traffic density, and air quality.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "全栈前端开发，融合高端 UI 设计与动画效果", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 125, "is_featured": false, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-05-02T14:37:43Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "multi-step-form-wizard", "title": "多步骤智能表单向导", "title_en": "Multi-Step Smart Form Wizard", "subtitle": "条件分支 × 实时校验 × 步骤导航，复杂表单也能丝滑填写", "subtitle_en": "Conditional branching, live validation & step navigation for complex forms", "description": "企业内部系统里的长表单最劝退——几十个字段堆一页，填到一半不知道还剩多少，提交才报错又得从头来。这个案例演示了一个企业级多步骤表单向导：带进度条的步骤导航、字段间条件分支（选了 A 才出现 B 组）、每步实时校验、最终提交前数据摘要确认，整体交互体验对标主流 SaaS 产品。\n\n适合：前端工程师学习表单最佳实践、产品设计师做中后台原型、企业内部系统开发团队。", "description_en": "Long forms in enterprise systems are painful — dozens of fields crammed on one page, no sense of progress halfway through, and errors only appear after submission forcing you to start over. This case demonstrates a production-grade multi-step form wizard: step navigation with a progress bar, conditional branching between fields (selecting A reveals group B), real-time validation per step, and a data summary confirmation before final submission. The overall interaction quality matches mainstream SaaS products.\n\nIdeal for: frontend engineers learning form best practices, product designers prototyping back-office systems, and enterprise internal system development teams.", "categories": [ "dev", "creation" ], "tags": [ "表单设计", "前端交互", "企业级组件", "用户体验" ], "tags_en": [ "Form Design", "Frontend Interaction", "Enterprise Components", "User Experience" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "多步骤表单向导页面" } ], "prompt": "请创建一个企业员工入职信息采集的多步骤表单向导页面（单文件 HTML，内联无外部依赖），现代简洁、响应式。\n\n要求：顶部进度条（4 步：基本信息→岗位详情→IT 设备需求→信息确认）；字段条件分支（如工作模式选「远程办公」显示远程相关字段，选「驻场」显示工位偏好）；每步实时校验必填项并提示；支持上一步/下一步；最后一步汇总所有数据供确认。", "prompt_en": "Please create a multi-step form wizard page for collecting new-employee onboarding information (single-file HTML, inlined with no external dependencies), modern and clean, responsive.\n\nRequirements: a progress bar at the top (4 steps: Basic Info → Position Details → IT Equipment Needs → Confirmation); conditional field branching (e.g., selecting \"Remote\" as the work mode shows remote-related fields, while \"On-site\" shows workstation preferences); real-time validation of required fields at each step with prompts; support for previous/next step; and a final step that summarizes all data for confirmation.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "全栈前端开发，融合高端 UI 设计与动画效果", "skill_id": "skill_2052764137292271616" }, { "id": "impeccable", "name": "高品质前端设计", "description": "高品质 UI/UX 设计工具集，避免泛 AI 审美", "skill_id": "skill_2053082862415904768" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 126, "is_featured": false, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-05-02T14:36:46Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "cicd-pipeline-diagram", "title": "CI/CD 流水线架构图", "title_en": "CI/CD Pipeline Architecture Diagram", "subtitle": "一图看懂从代码提交到生产部署的全流程", "subtitle_en": "Visualize the full journey from code commit to production deployment", "description": "每次跟新同事解释 CI/CD 流程，光靠嘴说总是讲不清楚——代码提交后到底经过哪些阶段、哪里会卡住、哪里需要人工审批？\n\n这个案例用 Excalidraw 画了一张完整的 CI/CD 流水线架构图，覆盖代码提交、自动化测试、构建打包、镜像推送、Staging 部署、人工审批、生产发布 7 个关键阶段，每个节点标注了真实的工具和命令。\n\n适合：DevOps 工程师、技术团队 Leader、需要做技术分享的开发者、正在搭建 CI/CD 体系的团队", "description_en": "Tired of explaining CI/CD pipelines verbally to new team members? It's hard to convey which stages code passes through, where bottlenecks occur, and where manual approvals are needed.\n\nThis case uses Excalidraw to create a complete CI/CD pipeline architecture diagram covering 7 key stages: code commit, automated testing, build & package, image push, staging deployment, manual approval, and production release — each node annotated with real tools and commands.\n\nIdeal for: DevOps engineers, tech team leads, developers preparing tech talks, and teams building CI/CD systems", "categories": [ "dev", "creation" ], "tags": [ "CI/CD", "流水线", "架构图", "DevOps", "自动化部署" ], "tags_en": [ "CI/CD", "Pipeline", "Architecture Diagram", "DevOps", "Automated Deployment" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "prompt": "请用 Excalidraw 画一张 CI/CD 流水线架构图。\n\n要求：\n1. 从左到右的横向流程布局\n2. 包含 7 个阶段，用不同语义色区分：代码提交、自动化测试、构建打包、镜像推送、Staging 部署、人工审批（菱形）、生产发布\n3. 每个节点下方用小字标注真实工具名\n4. 「人工审批」节点加分支：通过走生产发布，拒绝回退到代码提交\n5. 加一个深色背景的证据块，展示关键的 GitHub Actions workflow YAML 片段\n6. 顶部加标题 \"CI/CD Pipeline Architecture\"\n7. 风格干净专业，使用 Excalidraw 标准配色", "prompt_en": "Please use Excalidraw to draw a CI/CD pipeline architecture diagram.\n\nRequirements:\n1. Left-to-right horizontal flow layout\n2. Include 7 stages, distinguished by semantic colors: code commit, automated testing, build & package, image push, staging deployment, manual approval (diamond), production release\n3. Label the real tool name in small text under each node\n4. At the \"Manual Approval\" node, add a branch: if approved, proceed to production release; if rejected, fall back to code commit\n5. Add a dark-background evidence block showing a key GitHub Actions workflow YAML snippet\n6. Add the title \"CI/CD Pipeline Architecture\" at the top\n7. Clean, professional style using the standard Excalidraw palette", "skills": [ { "id": "excalidraw-diagram", "name": "Excalidraw 绘图", "description": "手绘风格架构图与流程图生成", "skill_id": "skill_2053082275194277888" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 127, "is_featured": false, "featured_order": 0, "quality_score": 82, "source": "official", "usage_count": 0, "created_at": "2026-04-15T08:23:23Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.png", "label": "CI/CD 流水线架构图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "annual-goal-dashboard", "title": "个人年度目标看板", "title_en": "Personal Annual Goals Dashboard", "subtitle": "交互式年度目标追踪面板，进度可视化 + 暗色主题切换", "subtitle_en": "Interactive annual goal tracker with visual progress, milestones, and dark/light theme toggle", "description": "年初立的 flag 到年底还记得几个？健身卡办了没去几次，读书计划停在第三本，存款目标年中就放弃了。问题不是缺乏意志力，而是缺少一个随时可见的进度看板——看不到进展，就感受不到动力，最终不了了之。\n\n这个案例生成一个精美的交互式年度目标看板 HTML 页面：4 大目标类别（健康、职业、学习、财务），每类 3 个具体目标，配有动态进度条、里程碑检查点、连续打卡计数器。页面顶部是整体完成度环形图和月度趋势折线图，底部轮播激励名言。支持暗色/亮色主题一键切换，纯前端实现无需后端。\n\n适合：希望量化管理个人成长的职场人、年度规划爱好者、OKR 实践者、想要一个酷炫个人仪表盘的效率达人。", "description_en": "How many New Year's resolutions survive until December? Gym membership barely used, reading plan stalled at book three, savings goal abandoned by June. The problem isn't willpower — it's the lack of a visible progress dashboard. When you can't see progress, you can't feel momentum, and goals fade away.\n\nThis case generates a beautiful interactive annual goals dashboard HTML page: 4 goal categories (health, career, learning, finance), each with 3 specific goals featuring animated progress bars, milestone checkpoints, and streak counters. Top section shows an overall completion ring chart and monthly trend line graph, bottom rotates motivational quotes. Supports dark/light theme toggle, pure frontend with no backend required.\n\nBest for: Professionals who want to quantify personal growth, annual planning enthusiasts, OKR practitioners, productivity fans who want a sleek personal dashboard.", "categories": [ "office", "creation" ], "tags": [ "年度目标", "个人成长", "数据可视化", "交互式看板", "HTML" ], "tags_en": [ "Annual Goals", "Personal Growth", "Data Visualization", "Interactive Dashboard", "HTML" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个个人年度目标看板页面（单文件 HTML，内联无外部依赖），现代精美、支持明暗主题切换。\n\n展示整体完成度（综合百分比 + 月度趋势）和各类目标的进度与里程碑，让人一眼看到全年达成情况。每个目标显示进度、里程碑节点和状态（进行中 / 已完成 / 落后），不同类别用不同颜色区分。\n\n我的年度目标（粘贴你的目标，按类别组织）：\n...", "prompt_en": "Generate a personal annual goals dashboard page (single-file HTML, inlined with no external dependencies), modern and polished, with light/dark theme switching.\n\nShow the overall completion rate (composite percentage + monthly trend) and the progress and milestones of each goal category, so the full-year status is visible at a glance. Each goal should show its progress, milestone checkpoints, and status (in progress / completed / behind); distinguish categories with different colors.\n\nMy annual goals (paste your goals, organized by category):\n...", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "Full-stack frontend development with premium UI", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 128, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "个人年度目标看板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ad-campaign-performance", "title": "广告投放效果分析表", "title_en": "Ad Campaign Performance Analysis", "subtitle": "多渠道广告投放数据追踪与 ROI 优化分析", "subtitle_en": "Multi-channel advertising data tracking and ROI optimization analysis", "description": "广告预算花出去了，效果却说不清——这是很多市场团队的痛点。不同渠道的数据口径不一、对比困难，预算分配靠「感觉」而非数据驱动，导致投放效率低下、ROI 难以优化。\n\n本案例演示一份覆盖微信朋友圈、抖音信息流、百度 SEM、知乎效果广告、小红书薯条、B站花火六大主流渠道的 Q2 广告投放效果分析表。每个渠道按月拆解预算/花费/曝光/点击/转化等核心指标，并附带汇总对比 Sheet，包含渠道排名、预算分配优化建议和环比趋势分析，帮助团队用数据驱动下一期投放策略。\n\n适合数字营销经理、增长团队负责人、CMO 及代理商投放优化师。无论是季度复盘还是预算申请汇报，这份表格都能直接作为决策依据。", "description_en": "Advertising budgets are hard to optimize without unified cross-channel analytics. This case creates a Q2 ad performance analysis covering 6 major Chinese advertising channels with monthly breakdowns, ROI comparisons, and budget optimization recommendations.", "categories": [ "business", "data" ], "tags": [ "广告投放", "ROI分析", "数据分析", "渠道对比" ], "tags_en": [ "Ad Campaign", "ROI Analysis", "Data Analytics", "Channel Comparison" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "prompt": "请制作一份 2026 年 Q2 多渠道广告投放效果分析表。渠道覆盖：微信朋友圈广告、抖音信息流、百度 SEM、知乎效果广告、小红书薯条推广、B站花火。每个渠道包含：月度预算、实际花费、曝光量、点击量、CTR、CPC、转化数、CVR、CPA、ROI。另附汇总对比Sheet（渠道排名、预算分配优化建议、环比趋势）。数据要真实合理，反映不同渠道的典型表现差异。", "prompt_en": "Please create a multi-channel ad campaign performance analysis spreadsheet for Q2 2026. Channels covered: WeChat Moments ads, Douyin in-feed ads, Baidu SEM, Zhihu performance ads, Xiaohongshu (RED) \"French Fries\" promotion, and Bilibili Sparkle (Huahuo). For each channel, include: monthly budget, actual spend, impressions, clicks, CTR, CPC, conversions, CVR, CPA, and ROI. Also add a summary comparison Sheet (channel ranking, budget allocation optimization recommendations, and month-over-month trends). The data should be realistic and reasonable, reflecting the typical performance differences across the various channels.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel spreadsheet processing", "slug": "minimax-xlsx" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 129, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.md", "label": "广告投放效果分析表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ashare-limitboard-sector-atlas", "title": "涨停热点数据地图", "title_en": "Limit-Up Sector Atlas", "subtitle": "把连板梯队、主线板块与封单热区整理成一张叙事感更强的数据地图。", "subtitle_en": "A narrative data atlas that maps ladders, hotspot sectors, and seal clusters into one story.", "description": "做涨跌停复盘时，经常会遇到一个问题：数据很多，但难以把板块、连板、封单和市场宽度放进同一条叙事链。这个案例把涨停数据整理成一张热点数据地图，用更强的故事结构去讲清楚当天资金最认哪条线。\n\n产物是一份双语交付级 HTML，包含主题节点、连板路径、热点板块、封单热区和全量清单，既能作为复盘材料，也能拿来做团队内分享或案例展示。\n\n适合：题材研究、市场主线跟踪、内容型投研输出、需要把数据讲成故事的人。", "description_en": "A common pain point in limit-up reviews is that the data exists, yet sectors, ladders, seal clusters, and market breadth rarely form one coherent story. This case turns the dataset into a sector atlas so you can explain which themes the market truly paid for.\n\nThe output is a bilingual delivery-grade HTML page with theme nodes, ladder routes, hotspot boards, seal clusters, and the full list, making it suitable for both recap work and polished internal sharing.\n\nBest for: theme researchers, market leadership trackers, content-style research output, and anyone who needs data to tell a story.", "categories": [ "finance", "data", "search" ], "tags": [ "热点地图", "数据叙事", "题材轮动", "交付案例" ], "tags_en": [ "hotspot atlas", "data storytelling", "theme rotation", "delivery case" ], "artifact_type": "html", "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "preview": "output.html", "preview_en": "output.html", "prompt": "请基于最新一个有完整收盘数据的 A 股交易日，生成一份“涨停热点数据地图”HTML，把连板梯队、主线板块、封单热区和市场宽度组织成一条清晰的复盘叙事。使用真实查询得到的涨跌停、板块排行、代表个股与全量涨停列表，突出当天最强的主题扩散和高标结构。页面风格要求明亮、像展板式的数据地图与章节化故事页，信息丰富且适合展示；单文件 HTML、资源全内联、可离线打开。请在页脚注明数据日期、来源口径和投资建议免责声明。", "prompt_en": "Using the latest A-share trading day with complete closing data, generate a “Limit-Up Sector Atlas” HTML page that organizes ladders, leading sectors, seal clusters, and market breadth into a clear recap narrative. Query real data for limit-ups and limit-downs, sector rankings, representative names, and the full limit-up list, with emphasis on theme diffusion and the structure of leadership. Keep the page bright, exhibition-like, and story-led, more like a chaptered data map than a generic dashboard; deliver a single self-contained HTML file that works offline. State the data date, source notes, and the investment disclaimer in the footer.", "artifacts": [ { "file": "output.html", "label": "双语交付 HTML 报告" } ], "skills": [ { "id": "a-share-limitboard-report", "name": "A股涨跌停日报", "description": "A股涨跌停日报，收盘后汇总涨停/跌停、连板梯队、板型分布、主线板块、炸板与全量列表。", "skill_id": "skill_2076627892488605696" } ], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-07-27T21:22:40.164728+08:00", "updated_at": "2026-07-30T16:43:23.113221+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "cultural-product-3d-plush", "title": "文创品建模 · 毛绒材质", "title_en": "Cultural Product 3D · Plush Material", "subtitle": "把文创品变成可 360° 旋转的毛绒公仔 3D 模型", "subtitle_en": "Turn a creative product into a 360°-rotatable plush 3D model", "description": "想把文创品做成周边公仔，却很难直观展示立体效果和材质手感。\n\n案例把文创品转译为毛绒公仔化的可交互 3D 模型：短绒布料质感、可见织物纹理与缝线、填充饱满有肉感，网页端可自由旋转缩放查看，单文件离线可开。\n\n适合：文创衍生开发、周边/潮玩设计师、博物馆文创团队、IP 授权方", "description_en": "This case turns a creative product into an interactive plush 3D model — short-pile fabric texture, visible weave and stitching, plump and huggable — freely rotate and zoom in the browser, single-file and offline-ready.\n\nGood for: Merch developers, Plush & toy designers, Museum creative teams, IP licensors", "categories": [ "creation" ], "tags": [ "3D建模", "毛绒材质", "可交互模型", "文创衍生" ], "tags_en": [ "3D Modeling", "Plush Material", "Interactive Model", "Cultural Merch" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据这个文创品生成的毛绒公仔化产品摄影，全身以柔软短绒布料呈现，可见织物纹理、细微缝线与刺绣走线，填充饱满、有肉感，带哑光绒面质感；这只鸭子有一段纤细的 S 形弯颈，头回转搭在肩背上，圆形黑豆眼、扁平圆喙，双翼收拢、原本雕刻的羽毛层次以凸起绣线还原，尾部上翘尖起，前方是一只张开的蹼足，整体端坐在一体成型的波浪形底座上；鸭身为温暖的沙琥珀色、米褐与琥珀米色并带柔和棕色羽纹阴影，底座则保留青花瓷质感，深靛蓝笔触水波纹叠印在米白底色之上，所有颜色都严格对齐原文物；画面为单体居中、四分之三侧视，纯白无缝背景，柔和均匀的漫射棚光配以淡淡接触阴影，对焦锐利、细节丰富，整体是高品质毛绒周边与收藏玩具的渲染风格，可爱、治愈且富有商业质感。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-3d-expert", "name": "文创品3D建模专家", "description": "将文创品转译为可交互的 3D 模型效果，支持毛绒、硅胶、金属珐琅、亚克力等材质渲染", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_P14ik4i7Utla" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "可交互 3D 模型（单文件内联）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-22T15:31:07+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "medication-safety-bleeding-risk", "title": "抗凝药出血风险与用药过量排查", "title_en": "Anticoagulant Bleeding Risk & Overdose Check", "subtitle": "华法林 + 阿司匹林双抗凝，牙龈出血、皮下瘀斑是否用药过量？", "subtitle_en": "Warfarin plus aspirin — do bleeding gums and bruising mean the anticoagulation is too strong?", "description": "72 岁老人，长期吃华法林、阿司匹林等慢病药，最近反复牙龈出血、皮下出现瘀斑——到底是药物叠加的出血风险，还是用药过量了？\n\n这个案例演示了把「老年患者 + 完整用药清单 + 新发出血症状」交给 AI 临床助手，它逐药归因华法林（抗凝）与阿司匹林（抗血小板）的叠加出血风险，判断牙龈出血、皮下瘀斑是否提示抗凝强度 / INR 超标（用药过量可能），并区分需立即就医的危险信号与可观察的轻微症状。\n\n适合：慢病老人及其家属、长期抗凝 / 抗血小板治疗患者、需要出血风险与用药过量科普的用户、社区健康管理员。", "description_en": "A 72-year-old on long-term warfarin and aspirin has recently developed recurrent gum bleeding and subcutaneous bruising — is it an additive bleeding risk from the drug combination, or is the anticoagulation too strong?\n\nThis case shows how to hand an elderly patient plus a full medication list plus new bleeding symptoms to the AI clinical assistant, which attributes bleeding risk to the warfarin (anticoagulant) + aspirin (antiplatelet) combination, judges whether gum bleeding and bruising signal excessive anticoagulation / an elevated INR (possible overdose), and separates urgent danger signs from mild watchable symptoms.\n\nFor: seniors and their families, patients on long-term anticoagulant / antiplatelet therapy, anyone needing plain-language guidance on bleeding risk and overdose, and community health workers.", "categories": [ "agent", "education" ], "tags": [ "用药安全", "出血风险", "用药过量", "华法林", "抗凝药", "AI临床助手" ], "tags_en": [ "Medication Safety", "Bleeding Risk", "Overdose", "Warfarin", "Anticoagulant", "AI Clinical Assistant" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "抗凝药出血风险与用药过量排查演示（HTML，含中英双语）" } ], "prompt": "请以「腾讯健康-AI临床助手-用药安全」的能力，评估一位老年患者的药物相关出血风险与用药过量可能：\n\n72 岁男性，合并高血压、房颤及糖尿病，长期服用华法林、美托洛尔、二甲双胍和阿司匹林，近期反复牙龈出血并出现皮下瘀斑。请评估：\n1. 是否存在药物相关出血风险（重点：华法林「抗凝」与阿司匹林「抗血小板」叠加，共同升高出血风险）；\n2. 是否存在用药过量可能（重点：牙龈出血 + 皮下瘀斑是否提示抗凝强度 / INR 超标，需查凝血功能判断，而非简单判定「药吃多了」）；\n3. 哪些是需立即就医的危险信号（黑便、呕血、血尿、持续出血、头痛意识改变、胸痛），哪些是可观察的轻微症状；\n4. 下一步建议（尽快就医查 PT/INR、就诊带完整用药清单），并说明为什么不能擅自停用抗凝 / 抗血小板药（房颤停药有卒中 / 血栓风险）。\n\n请客观说明出血机制与过量判断依据，不擅自建议停用或调整抗凝 / 抗血小板、降糖等慢病维持用药；不给出具体剂量调整数值；出血相关风险应主动强调；末尾建议尽快就医查凝血功能。", "prompt_en": "Using the medication-safety capabilities of the Tencent Health AI Clinical Assistant, evaluate a patient's drug-related bleeding risk and the possibility of overdose:\n\nA 72-year-old man with hypertension, atrial fibrillation and diabetes has been on warfarin, metoprolol, metformin and aspirin long-term. Recently he has recurrent gum bleeding and subcutaneous bruising. Please assess:\n1. Whether there is a drug-related bleeding risk (focus: warfarin 'anticoagulant' + aspirin 'antiplatelet' stack to raise bleeding risk);\n2. Whether overdose is possible (focus: whether gum bleeding + bruising signals excessive anticoagulation / an elevated INR — needs coagulation testing to judge, rather than simply calling it 'too many drugs');\n3. Which are danger signs needing immediate care (black stools, hematemesis, hematuria, persistent bleeding, headache or altered consciousness, chest pain) versus mild watchable symptoms;\n4. Next steps (see a doctor promptly for PT/INR, bring the full medication list), and explain why anticoagulant / antiplatelet drugs must not be stopped on one's own (stopping raises stroke / thrombotic risk in AF).\n\nObjectively explain the bleeding mechanism and the basis for judging overdose. Do not recommend stopping or adjusting anticoagulant/antiplatelet or chronic glucose-lowering drugs on your own. Do not give specific dose-adjustment numbers. Flag the bleeding risk proactively. End with a recommendation to see a doctor promptly to check coagulation.", "skills": [ { "id": "medication-advisor", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "slug": "medication-advisor" } ], "experts": [ { "id": "chief-physician", "name": "腾讯健康-AI临床助手-用药安全", "description": "用药决策辅助（CDS）：药品信息、处方用法用量、药物相互作用（DDI）、禁忌与特殊人群、不良反应风险与合理用药审方", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_0JLBYdNE0Hxk" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-09-05T11:19:00+08:00", "updated_at": "2026-09-05T11:19:00+08:00", "cover_image": "cover.png", "cover_mode": "landscape", "is_recommended": false, "recommended_order": 0 }, { "id": "multi-grid-video", "title": "多宫格生视频", "title_en": "Multi-Grid Video", "subtitle": "一张多镜头拼版控制四段奢华香氛影像", "subtitle_en": "One multi-cut sheet controls four luxury fragrance movements", "description": "多镜头商品视频常在产品角度、材质和剪辑顺序上失去一致，逐镜返工成本很高。\n\n案例把黑金香水瓶、瓶盖、丝绸烟雾与喷雾粒子组织成带编号的多镜头拼版，再据此生成连续的棚拍商业片。\n\n适合：香水、珠宝、美妆、酒饮等需要高质感多镜头展示的产品内容。", "description_en": "Multi-shot product films often lose consistency in angle, material, and edit order, making per-shot rework expensive.\n\nThis case organizes a black-and-gold bottle, cap, silk, smoke, and spray particles into numbered multi-cut sheets before generating a continuous studio commercial.\n\nBest for fragrance, jewelry, beauty, and beverage products that need premium multi-shot presentation.", "categories": [ "creation" ], "tags": [ "多宫格", "黑金广告", "香氛" ], "tags_en": [ "multi-grid", "black-and-gold ad", "fragrance" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「多宫格生视频」Skill，使用我上传的黑色瓶身、金色瓶盖香水产品图，制作一支约 15 秒、16:9 的棚拍商业短片《黑金香氛四重奏》。四段内容按明确编号推进：1）黑丝绸从瓶身后方掠过，勾出金色边缘；2）瓶盖悬停旋转并与瓶身对齐；3）喷雾在逆光中形成细密金色粒子；4）产品落在黑色镜面台上完成英雄定帧。请先确认全局规格与故事板，并绑定产品资产；随后为每个镜头生成包含内部节拍编号的多镜头拼版，逐批展示让我确认，确认后再从拼版生成视频、设计低沉电子 BGM 与喷雾/金属轻响并完成组装。产品轮廓、Logo、材质和金色调必须始终一致。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Multi-Grid Video” Skill and use my uploaded image of a black fragrance bottle with a gold cap to create a 16:9 studio commercial of about 15 seconds titled Black-and-Gold Fragrance Quartet. Progress through four explicitly numbered movements: 1) black silk passes behind the bottle and reveals its gold edge; 2) the cap hovers, rotates, and aligns with the bottle; 3) backlit spray forms fine golden particles; 4) the product lands on a black mirrored plinth for the hero frame. Confirm the global specification and storyboard and bind the product asset first. Then generate a numbered multi-cut sheet for the internal beats of every shot and show each batch for approval. Once approved, generate video from the sheets, design a deep electronic score with spray and light metallic sounds, and assemble the film. Keep product silhouette, logo, materials, and gold tone consistent throughout.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "notification-center", "title": "通知中心消息列表", "title_en": "Notification Center Message List", "subtitle": "输入消息数据，生成精美的通知中心 UI 页面", "subtitle_en": "Input message data, generate an elegant notification center UI page", "description": "做产品原型、写 PRD、给前端展示设计稿——经常需要一个逼真的通知中心界面作参考或占位。但从零画一个又费时间，Figma 模板里的通知组件跟自己的产品风格对不上。\n\n这个案例演示了输入一组产品通知消息（系统通知、评论提醒、任务更新等），AI 自动生成一个精美的通知中心 UI 页面：居中窄面板（max-width 440px）、顶部标题栏含未读计数和筛选标签、消息列表每条含图标/标题/内容摘要/时间戳、未读消息高亮标记、hover 效果。模拟 iOS/macOS 通知中心的精致感。\n\n适合：产品经理原型参考、前端 UI 组件开发、设计系统 showcase、PRD 配图。", "description_en": "Product prototyping, PRD writing, design spec presentations — often need a realistic notification center UI as reference or placeholder. But building one from scratch takes time, and Figma template notification components don't match your product style.\n\nThis case demonstrates inputting a set of product notifications (system alerts, comment mentions, task updates), with AI generating an elegant notification center UI page: centered narrow panel (max-width 440px), header with unread count and filter tabs, message list with icon/title/content preview/timestamp per item, unread message highlighting, hover effects. Mimics the polished feel of iOS/macOS notification center.\n\nBest for: Product manager prototype reference, frontend UI component development, design system showcase, PRD illustrations.", "categories": [ "creation", "dev" ], "tags": [ "通知中心", "UI组件", "消息列表", "原型参考" ], "tags_en": [ "Notification Center", "UI Component", "Message List", "Prototype Reference" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个通知中心消息列表页面（单文件 HTML，内联无外部依赖），界面文字用简体中文。\n\n要能区分未读/已读、按类型展示不同通知（系统通知 / 评论提醒 / 任务更新 / 审批 / 日程 / 安全提醒等），支持「全部 / 未读 / 已读」筛选，整体精致现代、类 iOS/macOS 通知中心质感。\n\n请造一组示例通知（约 8 条，时间倒序，混合未读与已读），每条含类型、标题、内容摘要和时间戳。", "prompt_en": "Please generate a notification center message list page (single-file HTML, inline with no external dependencies), with interface text in Simplified Chinese.\n\nIt should distinguish unread/read, display notifications by type (system / comment mention / task update / approval / schedule / security alert, etc.), support \"All / Unread / Read\" filtering, and have an overall refined, modern feel reminiscent of the iOS/macOS notification center.\n\nPlease create a set of sample notifications (about 8, in reverse chronological order, mixing unread and read), each with a type, title, content preview, and timestamp.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "通知中心组件" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "scene-score-filter", "title": "进面分查询：浙江低分岗位", "title_en": "Cutoff Scores: Zhejiang's Lower-Bar Positions", "subtitle": "按往年进面分筛出分数线相对友好的保底岗", "subtitle_en": "Filter by historical cutoff to find score-friendly fallback posts", "description": "「够得着」比「喜欢」更重要。行测申论刷到瓶颈时，选个分数线友好的岗比再涨五分现实得多。\n\n本案例按「浙江 + 公务员 + 进面分≤130」实时筛出 551 个岗位，取前 10 条列出单位、地区与往年进面分实值，并计算本页分数区间与均值。\n\n适合：分数处于中游想稳妥上岸的考生、需要给学生定报考梯度的老师。", "description_en": "\"Reachable\" beats \"desirable\". When your mock scores plateau, picking a score-friendly post is more realistic than squeezing out five more points.\n\nThis case live-filters 551 Zhejiang civil-service positions with historical cutoffs ≤130, listing the top 10 with unit, region and actual cutoff values, plus the page's score range and mean.\n\nFor: mid-scoring candidates seeking a safe landing and teachers setting application tiers.", "categories": [ "search" ], "tags": [ "进面分", "公务员", "浙江", "保底岗" ], "tags_en": [ "cutoff score", "civil service", "Zhejiang", "fallback position" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "浙江进面分≤130 岗位清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我在浙江公务员岗位里找往年进面分不超过 130 的。要求：1) 实时查询并给出总数；2) 前 10 条列岗位、单位、地区、往年进面分、招录人数、来源公告；3) 进面分必须是接口返回的历史原值，不要估算；4) 提醒我进面分数据的正确用法。", "prompt_en": "Find Zhejiang civil-service positions with historical interview cutoffs no higher than 130. Requirements: 1) query live and give the total; 2) list the top 10 with position, unit, region, historical cutoff, headcount, source announcement; 3) cutoffs must be the interface's historical values, no estimates; 4) remind me how to use cutoff data correctly.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "按往年进面分区间筛选岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "smart-medical-record-outpatient", "title": "门诊病历口述生成与安全拦截", "title_en": "Outpatient Note from Verbal Dictation with Safety Check", "subtitle": "口述症状自动解析，术语规范化 + 过敏-用药冲突 P0 红线", "subtitle_en": "Parse verbal symptoms, normalize terms, block allergy-drug conflicts", "description": "患者口述「肚子疼两天，有点发烧，拉肚子」，青霉素过敏史却开了阿莫西林——门诊医生每天面对的口述转病历场景，错录一条就可能致命。\n\n这个案例演示了把患者口述交给 AI 临床助手，它自动判别门诊场景、规范化术语（肚子疼→腹痛、发烧→发热、拉肚子→腹泻）、提取主诉与既往史，并在处方环节触发 P0 安全红线：青霉素过敏史与阿莫西林直接冲突，必须人工确认才能放行，绝不静默通过。\n\n适合：门诊医师、急诊预检、基层全科、电子病历录入员、关注用药安全的临床管理者。", "description_en": "A patient says \"stomach pain for two days, a bit feverish, diarrhea\", yet amoxicillin is on the script despite a penicillin allergy — every day in outpatient clinics, a single misrecorded verbal dictation can be fatal.\n\nThis case shows how to hand the patient's verbal account to the AI clinical assistant. It identifies the outpatient scenario, normalizes lay terms (stomach pain → abdominal pain, feverish → fever, diarrhea → loose stools), extracts chief complaint and past history, then triggers a P0 safety redline at prescribing: a penicillin allergy directly conflicts with amoxicillin and requires explicit human confirmation — never silently passes.\n\nFor: outpatient physicians, emergency triage, primary care providers, EMR data entry staff, and clinical managers focused on medication safety.", "categories": [ "agent", "education" ], "tags": [ "智能病例", "门诊病历", "口述解析", "用药安全", "AI临床助手", "腾讯健康" ], "tags_en": [ "Smart Medical Record", "Outpatient Note", "Verbal Parsing", "Medication Safety", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我写一份门诊病历，从我的口述内容里提取结构化字段并做术语规范化。\n\n口述内容：患者女性，35 岁，肚子疼两天，有点发烧，拉肚子；既往有青霉素过敏史；本次开具阿莫西林、蒙脱石散。\n\n要求：\n1. 判别场景为「门诊病历」；\n2. 解析口述，提取主诉、现病史、既往史、用药等信息，口语转标准医学术语（如肚子疼→腹痛、发烧→发热、拉肚子→腹泻）；\n3. 录入处方药品名与用法用量；\n4. 校验 P0 安全红线：检测到「青霉素过敏史」与处方「阿莫西林」冲突时，必须拦截并提示人工确认，不得静默通过；\n5. 对缺失关键字段（如体温数值、腹痛部位）进行渐进式追问，每批不超过 5 个、支持跳过，未作答字段如实标注缺失；\n6. 不做诊断推理、不做治疗建议，不编造临床数据；\n7. 末尾附安全声明，说明本记录仅供医生复核。", "prompt_en": "Draft an outpatient note from my dictation, extracting structured fields and normalizing terminology.\n\nDictation: Patient is female, 35, stomach pain for two days, a bit feverish, diarrhea; known penicillin allergy; prescribed amoxicillin and montmorillonite powder.\n\nRequirements:\n1. Identify the scenario as \"outpatient note\".\n2. Parse the dictation to extract chief complaint, history of present illness, past history and medications; convert lay terms to standard medical terms (stomach pain → abdominal pain, feverish → fever, diarrhea → loose stools).\n3. Record drug names and dosage/administration for the prescription.\n4. Check the P0 safety redline: when a \"penicillin allergy\" conflicts with prescribed \"amoxicillin\", block it and require human confirmation — never pass silently.\n5. Prompt progressively for missing key fields (e.g. temperature value, abdominal pain site), no more than 5 per batch, skippable; honestly mark unanswered fields as missing.\n6. Make no diagnostic inference and no treatment recommendation; never fabricate clinical data.\n7. End with a safety statement noting the record is for clinician review only.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:57:00+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "artifacts": [ { "file": "output.html", "label": "门诊病历口述生成与安全拦截" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "solid-state-electrolyte-chain-report", "title": "固态电池电解质链", "title_en": "Solid-State Electrolyte Chain", "subtitle": "供给约束 + 技术迭代 + 公司映射，识别真卡点与概念跟随。", "subtitle_en": "Supply constraints + technology iteration + company mapping to separate real bottlenecks from concept followers.", "description": "固态电池讲了很多年，真正决定产业节奏的往往不是整车叙事，而是硫化锂前驱体、电解质粉体、设备能力和验证良率能不能跑顺。这个案例围绕硫化物电解质链，拆解上游原料、粉体制备、设备与客户验证的关键约束。\n\n报告会基于真实公开资料，区分哪些公司已经有中试线、工艺积累和客户验证，哪些仍停留在主题映射或远期想象阶段，并补上需要持续跟踪的时间点与风险条件。\n\n适合：关注固态电池产业化进度的人；需要识别电解质链卡点的投研人员；要建立设备与材料验证框架的产业分析师。", "description_en": "Solid-state batteries have been discussed for years, but the real pacing factors are often not the vehicle narrative — they are whether lithium-sulfide precursors, electrolyte powder, equipment capability, and validation yield can actually run smoothly. This case breaks down the sulfide-electrolyte chain into raw materials, powder preparation, equipment, and customer-validation constraints.\n\nGrounded in real public materials, the report separates companies that already have pilot lines, process accumulation, and customer validation from names still living in theme exposure or distant optionality, then adds the key dates and risks worth tracking.\n\nBest for: users following solid-state battery industrialization; researchers identifying bottlenecks in the electrolyte chain; industry analysts building a validation framework for equipment and materials.", "categories": [ "finance", "data", "business" ], "tags": [ "固态电池", "硫化物电解质", "硫化锂", "产业链分析", "公司映射" ], "tags_en": [ "solid-state battery", "sulfide electrolyte", "lithium sulfide", "industry chain analysis", "company mapping" ], "artifact_type": "html", "preview": "output.html", "preview_en": "output.html", "artifacts": [ { "file": "output.html", "label": "完整分析报告" } ], "skills": [ { "id": "industry-chain-analysis", "name": "产业链分析", "description": "产业链/赛道深度研究，支持全景拆解与瓶颈环节、金矿雷达、Serenity五维瓶颈验证。", "skill_id": "skill_2076627978773680128" } ], "experts": [], "mcps": [], "cover_image": "cover.png", "cover_image_en": "cover-en.png", "cover_mode": "landscape", "prompt": "请围绕固态电池硫化物电解质产业链做一份深度分析。先用真实查询核对上游原料、工艺路线、设备能力、公司披露与验证进度，并注明来源和日期；再按“供给约束 → 技术迭代 → 公司映射 → 风险与跟踪”的路径输出一份单文件、可离线打开的 HTML 报告，正文对公司映射同步标注证据等级与验证阶段。", "prompt_en": "Create a deep-dive report on the sulfide solid-state electrolyte chain. First use real queries to verify upstream raw materials, process routes, equipment capability, company disclosures, and validation progress with explicit sources and dates; then deliver a single-file offline HTML report following the path of supply constraints → technology iteration → company mapping → risks and tracking, while marking evidence grades and validation stage for each mapped company.", "scenario": "", "scenario_use": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-07-28T09:33:14Z", "updated_at": "2026-07-28T09:40:41Z", "is_recommended": false, "recommended_order": 0 }, { "id": "spateo-segmentation-3d-reconstruction", "title": "Spateo 细胞分割与 3D 重建实战", "title_en": "Spateo Cell Segmentation & 3D Reconstruction Playbook", "subtitle": "从染色图像到三维器官模型——分割、对齐、点云/表面/体素全流程", "subtitle_en": "From stain images to 3D organ models — segmentation, alignment, point-cloud/surface/voxel pipeline", "description": "细胞分割要么 SAW 出 cellbin.gef 直接用，要么只能从 ssDNA / DAPI / HE 自己切——后者网上代码片段散落，wavelet + watershed 调参就够折腾一天，切完还要切片对齐做 3D 重建。这个案例把立体Seq专家内置的 spateo skill 浓缩成一份「拿到 BGI counts + 染色图就能跑」的全流程指南：覆盖环境准备（spateo-release + [3d] extras + h5py/anndata pin）、染色图像细胞分割 9 步（read_bgi_agg → refine_alignment → mask_nuclei → peaks → watershed → stardist → augment → expand → read_bgi）、分割质量检查与互补策略、多切片 Morpho 对齐（group_pca + morpho_align）、TDR 3D 重建（construct_pc → construct_surface → construct_cells → voxelize → save_model）、4 种 VTK 模型（点云/表面/细胞/体素）用途速查、刚体 vs 非刚体选择。4 张代表性示意图全部内嵌，单文件自包含可直接当团队 SOP。适合：要做细胞分割或多切片 3D 重建的算法团队、做组织发育形态学的研究组、需要把染色与 RNA 配准的研究生。", "description_en": "Cell segmentation either uses SA W's cellbin.gef directly, or you cut cells from ssDNA/DAPI/HE yourself — the latter has scattered code online, wavelet + watershed tuning is a day's fight, and after segmentation you still need slice alignment for 3D reconstruction. This case distills the `spateo` skill into a 'pick up BGI counts + stain image and just run' full-pipeline guide covering environment setup (spateo-release + [3d] extras + h5py/anndata pin), the 9-step stain image cell segmentation (read_bgi_agg → refine_alignment → mask_nuclei → peaks → watershed → stardist → augment → expand → read_bgi), segmentation QC and complementary strategies, multi-slice Morpho alignment (group_pca + morpho_align), TDR 3D reconstruction (construct_pc → construct_surface → construct_cells → voxelize → save_model), the 4 VTK model types (point cloud / surface / cells / voxel) cheatsheet, and rigid vs non-rigid selection. 4 representative figures all embedded in a self-contained file. Suitable for: algorithm teams doing cell segmentation or multi-slice 3D reconstruction, developmental tissue morphology labs, and graduate students needing to register stain with RNA.", "categories": [ "data", "education" ], "tags": [ "Spateo", "细胞分割", "Morpho 对齐", "TDR 3D 重建", "marching_cube" ], "tags_en": [ "Spateo", "cell segmentation", "Morpho alignment", "TDR 3D reconstruction", "marching_cube" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "Spateo 细胞分割与 3D 重建实战（自包含 HTML）" } ], "prompt": "你是 Stereo-seq 空间转录组分析专家。请基于立体Seq专家内置的 `spateo` skill 文档（含 01_stain_segmentation.md / 03_alignment_basic_2d.md / 04_tdr_3d_reconstruction.md），生成一份 Spateo 细胞分割与 3D 重建实战指南。\n\n要求：\n1. 覆盖八个章节：环境准备（pip install spateo-release + [3d] extras + h5py==3.7.0 anndata==0.8.0 pin + MPI 可选）、数据准备（BGI counts + 染色 + 配准前提）、染色图像细胞分割 9 步（read_bgi_agg → refine_alignment → mask_nuclei_from_stain → find_peaks_from_mask → watershed → stardist → augment_labels → expand_labels → read_bgi 完整可运行代码 + 各层可视化代码）、分割质量检查与互补策略表（7 类常见问题）、多切片 Morpho 对齐（group_pca + morpho_align 完整代码 + 刚体 vs 非刚体选择决策表）、TDR 3D 重建（construct_pc → construct_surface → construct_cells → voxelize → save_model 完整代码 + 4 类 VTK 模型用途速查）、完整最小可运行脚本（四阶段串联）、避坑清单（12 条染色配准 / equalize / group_pca 联合 / 刚体优先 3D 等关键坑）。\n2. 每章必须包含：目标说明、可运行 Python 代码（带注释）、API 速查、产出文件清单。\n3. 用 matplotlib 预先生成 4 张代表性结果图（细胞分割 9 步流程图、分割三联图模拟、Morpho 对齐前后 2D 叠加、TDR 3D 重建点云+表面双面板）。\n4. 将所有图片以 base64 data URI 内联嵌入，产出单文件自包含 HTML（无外链）。\n5. 末尾附图表索引与数据来源声明。\n\n约束：\n- 输出单文件 HTML（output.html）：所有 CSS 内联、图片 base64 内嵌、Noto Sans SC 子集按文档字符内联嵌入、无任何外部资源引用、内容默认可见、不使用 hash 锚点导航。\n- 所有 API、参数、典型流程均直接来自 spateo skill 文档原文（01_stain_segmentation.md、03_alignment_basic_2d.md、04_tdr_3d_reconstruction.md），不要虚构 morpho_align / construct_surface 等函数签名。\n- 中文版，代码标识符保持英文生态惯例。\n- 文档可直接作为团队培训 SOP 使用。", "prompt_en": "You are a Stereo-seq spatial transcriptomics analysis expert. Using the `spateo` skill documents built into the stereo-seq expert (including 01_stain_segmentation.md / 03_alignment_basic_2d.md / 04_tdr_3d_reconstruction.md), generate a Spateo cell segmentation & 3D reconstruction playbook.\n\nRequirements:\n1. Cover eight chapters: environment setup (pip install spateo-release + [3d] extras + h5py==3.7.0 anndata==0.8.0 pin + optional MPI), data preparation (BGI counts + stain + registration prerequisite), the 9-step stain image cell segmentation (read_bgi_agg → refine_alignment → mask_nuclei_from_stain → find_peaks_from_mask → watershed → stardist → augment_labels → expand_labels → read_bgi with runnable code + per-layer visualization), segmentation QC and complementary strategies table (7 common issues), multi-slice Morpho alignment (group_pca + morpho_align complete code + rigid vs non-rigid decision table), TDR 3D reconstruction (construct_pc → construct_surface → construct_cells → voxelize → save_model complete code + 4 VTK model-type cheatsheet), complete minimum-runnable script (four-stage pipeline), and a 12-item pitfall checklist (stain registration / equalize / group_pca joint / rigid-first 3D etc.).\n2. Each chapter must include: goal, runnable Python code (commented), API cheatsheet, output file list.\n3. Pre-generate 4 representative figures with matplotlib (9-step segmentation pipeline diagram, segmentation triplet mock, Morpho before/after 2D overlay, TDR 3D point-cloud + surface dual panel).\n4. Inline all figures as base64 data URIs and produce a single-file self-contained report (no external links).\n5. Append a figure index and a data-source disclosure.\n\nConstraints:\n- Output a single-file HTML (output.html): all CSS inlined, figures base64-embedded, Noto Sans SC subset inlined by document-character filtering, no external resource references, content visible by default, no hash-anchor navigation.\n- All APIs, parameters, and typical flows come directly from the spateo skill (01_stain_segmentation.md, 03_alignment_basic_2d.md, 04_tdr_3d_reconstruction.md) — do not fabricate morpho_align / construct_surface signatures.\n- Chinese language; code identifiers follow English ecosystem conventions.\n- The document must be usable directly as team training SOP.", "skills": [], "experts": [ { "id": "stereo-seq-expert", "name": "Stereo-seq空间转录组分析专家", "description": "Stereo-seq 空间转录组全流程分析（SAW 投递 / Stereopy 下游 / Spateo 分割·对齐·3D），内嵌 saw-st-v8 / dcs-cli / saw-st-v8-makeref / saw-st-v8-realign / saw-st-v8-reanalyze-lasso / saw-st-v8-gef2gem / stereopy-docs / spateo / stereo-seq-technology-principles 共 9 个 skill", "expertType": "plugin", "source": "expert-marketplace", "expert_id": "ex_b2vdl3DaOp7A" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 130, "is_featured": false, "featured_order": 0, "quality_score": 96, "cover_image": "cover.png", "cover_mode": "landscape", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-29T16:00:00+08:00", "updated_at": "2026-08-29T18:30:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "brainstorm-mind-map", "title": "头脑风暴思维导图", "title_en": "Brainstorming Mind Map", "subtitle": "输入议题，自动发散生成交互式思维导图", "subtitle_en": "Enter a topic and auto-generate an interactive mind map", "description": "开会讨论新方案时，白板上写的想法越来越多，回头却很难整理成结构化的思路。这个案例演示了如何围绕「提升 SaaS 产品用户月留存率」这一议题，AI 自动从产品体验、用户成功、数据驱动、社区生态、商业策略五大维度展开头脑风暴，生成一个纯 SVG 思维导图，节点层次清晰，各维度用不同颜色区分，无需任何外部依赖。\n\n适合：产品经理、运营负责人、创业者、团队 leader、战略规划人员。", "description_en": "When brainstorming in meetings, ideas pile up on the whiteboard but are hard to organize into structured thinking afterward. This case demonstrates how AI automatically brainstorms around 'improving SaaS product monthly retention rate' across five dimensions — product experience, customer success, data-driven operations, community ecosystem, and business strategy — generating a pure inline SVG mind map with clear hierarchy and color-coded branches, fully self-contained with zero external dependencies.\n\nBest for: Product managers, operations leads, entrepreneurs, team leaders, strategic planners.", "categories": [ "office", "creation" ], "tags": [ "思维导图", "头脑风暴", "留存率", "SaaS" ], "tags_en": [ "Mind Map", "Brainstorming", "Retention", "SaaS" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "交互式思维导图" } ], "prompt": "围绕「如何将 SaaS 产品的用户月留存率从 65% 提升到 80%」生成一个思维导图（单文件 HTML，内联 SVG 实现，无外部依赖、无 JavaScript），暗色背景、中心向四周发散的经典思维导图布局。\n\n覆盖五个一级分支并展开：\n1）产品体验：新手引导、功能发现、性能稳定性、个性化推荐\n2）用户成功：客户成功经理、健康度评分、里程碑庆祝、主动预警\n3）数据驱动：行为漏斗、A/B 实验、流失预测、用户分群\n4）社区生态：用户社区、开发者生态、内容营销、案例库\n5）商业策略：定价优化、增值包、年度合约激励、忠诚度计划", "prompt_en": "Create a mind map around \"How to increase a SaaS product's monthly user retention rate from 65% to 80%\" (single-file HTML, implemented with inline SVG, no external dependencies, no JavaScript), with a dark background and the classic mind-map layout radiating outward from the center.\n\nCover and expand five top-level branches:\n1) Product Experience: onboarding, feature discovery, performance/stability, personalized recommendations\n2) Customer Success: customer success managers, health scoring, milestone celebrations, proactive alerts\n3) Data-Driven: behavioral funnels, A/B experiments, churn prediction, user segmentation\n4) Community Ecosystem: user community, developer ecosystem, content marketing, case-study library\n5) Business Strategy: pricing optimization, add-on packages, annual-contract incentives, loyalty programs", "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表", "description": "使用 Mermaid 语法生成流程图、时序图、甘特图等可视化图表", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "is_featured": false, "sort_order": 131, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-04-23", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-score-national-exam", "title": "进面分筛选：国考 115-135 分段交叉分析", "title_en": "Cutoff-Score Filter: Cross-Analysis of the 115-135 Band", "subtitle": "同一分段内，报录比差 48 倍——分数够得着不等于挤得进去", "subtitle_en": "Within one score band, ratios differ 48-fold", "description": "估完分想找「够得着」的岗位，可只按分数筛出来的岗位，竞争强度天差地别。\n\n本案例按进面分 115–135 筛出 7,331 个国考岗位，本页 30 条全部同时带往年进面分和报录比。按分数升序逐条列出，做分数分段统计，并用真实数据验证「报录比高≠分数线高」，最后给出一套先估分、再看比、再看招录人数的挑岗顺序。\n\n适合：已有稳定估分的考生、需要同时权衡分数与竞争的人。", "description_en": "After estimating your score you want posts within reach — but posts filtered by score alone vary wildly in competition.\n\nThis case filters national-exam posts by a 115–135 cutoff band, hitting 7,331. All 30 shown carry both a historical cutoff score and a competition ratio. They are listed by ascending score with a band breakdown, and the real data is used to test whether a high ratio implies a high score line.\n\nFor: candidates with a stable estimated score who must weigh score against competition.", "categories": [ "search" ], "tags": [ "国考", "进面分", "报录比", "分段筛选" ], "tags_en": [ "national exam", "cutoff score", "competition ratio", "score band" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "国考 115-135 分段进面分与报录比交叉分析", "label_en": "Cutoff score and ratio cross-analysis, 115-135 band" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我国考估分在 120 分左右，帮我找分数够得着的岗位。要求：1) 按往年进面最低分 115–135 分真实筛选国考岗位并给出总数；2) 取 30 条，按进面分升序列出岗位、单位、地区、招录人数、进面最低分、报录比、薪资区间；3) 把进面分按 4 个区间分段统计条数，并分别算出低分段与高分段的平均报录比；4) 用这批真实数据回答「报录比高是不是意味着分数线也高」；5) 给出一套挑岗顺序，并提醒进面分是往年结果、当年岗位设置可能变化。", "prompt_en": "My estimated national-exam score is around 120 — find posts within reach. Requirements: 1) filter posts by a historical cutoff score of 115–135 and give the total; 2) take 30, listed by ascending cutoff, with post, unit, region, headcount, cutoff score, competition ratio and salary band; 3) band the cutoff scores into four ranges with counts, and compute the average ratio for the lower and upper halves; 4) use this real data to answer whether a higher ratio implies a higher score line; 5) give a selection sequence and note that cutoff scores are historical and post configurations may change.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 131, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-code-refactor", "title": "Agent 驱动的代码重构助手", "title_en": "Agent-Driven Code Refactoring Assistant", "subtitle": "多 Agent 协作完成代码审计、方案规划、逐步重构与质量验证", "subtitle_en": "Multi-Agent Code Audit, Refactoring Planning, Step-by-Step Execution & Quality Validation", "description": "遗留代码越来越难维护？单个路由文件膨胀到 500 行，巨型函数、重复逻辑、硬编码配置、缺少错误处理、类型不安全——每次改需求都如履薄冰，新人接手更是无从下手。重构势在必行，但手动重构容易引入回归 Bug，缺乏系统化方法论。\n\n本案例展示了一套多 Agent 协作的代码重构工作流：审计 Agent 全面扫描代码识别所有 code smell 并评估严重程度，规划 Agent 制定重构方案（提取函数、引入类型、配置外部化、统一错误处理中间件），重构 Agent 逐步执行重构并保持每步功能等价，验证 Agent 对比重构前后的代码质量指标。产出生产级别的重构代码，附带关键重构决策说明。\n\n适合：后端开发工程师、全栈工程师、技术 Leader、负责遗留系统维护和代码质量提升的开发团队。", "description_en": "Legacy code becoming unmaintainable? Route files bloated to 500+ lines with god functions, duplicated logic, hardcoded configs, missing error handling, and unsafe types?\n\nThis case demonstrates a multi-Agent code refactoring workflow: Audit Agent scans for code smells and severity assessment, Planning Agent designs refactoring strategy, Refactoring Agent executes step-by-step with functional equivalence, Validation Agent compares before/after quality metrics. Outputs production-quality refactored code with documented decisions.\n\nIdeal for: backend engineers, full-stack developers, tech leads, and teams maintaining legacy systems.", "categories": [ "agent", "dev" ], "tags": [ "代码重构", "Agent 自动化", "代码质量", "TypeScript" ], "tags_en": [ "code refactoring", "agent automation", "code quality", "TypeScript" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.ts", "prompt": "请设计一个 Agent 工作流来分析和重构一段 TypeScript 代码。输入是一个 500 行的 Express.js 路由文件，存在以下问题：巨型函数（单个函数超过100行）、重复代码、硬编码配置、缺少错误处理、类型不安全。流程：1) 审计 Agent 扫描代码，识别所有 code smell 并评估严重程度；2) 规划 Agent 制定重构方案（提取函数、引入类型、配置外部化、统一错误处理中间件）；3) 重构 Agent 逐步执行重构，每步保持功能等价；4) 验证 Agent 对比重构前后的代码质量指标。输出重构后的代码文件。\n\n代码注释和用户可见字符串用简体中文，标识符（变量/函数/类型名）保留英文。", "prompt_en": "Design an Agent workflow to analyze and refactor a piece of TypeScript code. The input is a 500-line Express.js route file with the following problems: god functions (single functions exceeding 100 lines), duplicated code, hardcoded configuration, missing error handling, and unsafe types. Workflow: 1) Audit Agent scans the code, identifies all code smells, and assesses their severity; 2) Planning Agent designs the refactoring plan (extract functions, introduce types, externalize configuration, unify error handling via middleware); 3) Refactoring Agent executes the refactoring step by step, keeping functional equivalence at every step; 4) Validation Agent compares code quality metrics before and after refactoring. Output the refactored code file.\n\nCode comments and user-visible strings should be in Simplified Chinese; keep identifiers (variable/function/type names) in English.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 132, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.ts", "label": "重构后的代码" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-score-jiangsu-band", "title": "进面分筛选：江苏省考 120-135 分段", "title_en": "Cutoff-Score Filter: Jiangsu Provincial Exam, 120-135", "subtitle": "同一座城市、同一份公告里，报录比也能差 33 倍", "subtitle_en": "Same city, same announcement — ratios still differ 33-fold", "description": "都知道省考比国考好考，但好考多少、同一个城市内部差多少，很少有人给出具体数字。\n\n本案例按 120–135 分筛江苏岗位，命中 1,217 个，本页 30 条恰好全部落在苏州，同属一份省考公告。逐条给出报录比、进面分与薪资区间，把中位数与国考同分段对比，并拆到区级与招录单位看集中度。\n\n适合：国省同报的考生、锁定单一城市备考的人。", "description_en": "Everyone knows provincial exams are less competitive than the national one — but by how much, and how much variation sits inside a single city?\n\nThis case filters Jiangsu posts by a 120–135 cutoff, hitting 1,217. All 30 shown fall in Suzhou under one announcement. Each carries a ratio, cutoff and salary band; the median is compared against the national-exam band, then broken down by district and recruiting unit.\n\nFor: candidates sitting both exams, and those committed to one city.", "categories": [ "search" ], "tags": [ "江苏省考", "进面分", "苏州", "分段筛选" ], "tags_en": [ "Jiangsu provincial exam", "cutoff score", "Suzhou", "score band" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "江苏省考 120-135 分段岗位分析", "label_en": "Jiangsu provincial exam, 120-135 band analysis" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查江苏省考里往年进面最低分在 120–135 分之间的岗位。要求：1) 真实筛选并给出总数；2) 取 30 条，按报录比升序列出岗位、单位、地级市、招录人数、报录比、进面最低分、薪资区间；3) 统计这 30 条的地区分布（到区一级）和招录单位集中度；4) 算出本批报录比的中位数（偶数条数取中间两位的均值），与国考查询结果的中位数做量级参照，并说明两次查询的分数区间并不相同；5) 列出分数最高和最低的各三个岗位，分析分数线差异的成因，不要归结为岗位好坏。", "prompt_en": "Find Jiangsu provincial-exam posts with a historical cutoff score between 120 and 135. Requirements: 1) run the filter and give the total; 2) take 30, listed by ascending competition ratio, with post, unit, city, headcount, ratio, cutoff score and salary band; 3) break down the 30 by district and by recruiting unit concentration; 4) compute the median ratio and compare it with the national exam's median for the same band; 5) list the three highest- and three lowest-scoring posts and analyse what drives the difference — not whether the post is 'good'.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 132, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "okr-tracking-dashboard", "title": "团队 OKR 追踪仪表盘", "title_en": "Team OKR Tracking Dashboard", "subtitle": "输入 OKR 数据，生成纯 CSS/SVG 实现的追踪仪表盘", "subtitle_en": "Input OKR data, generate a tracking dashboard using pure CSS/SVG", "description": "每到季度 Review，最头疼的就是把散落在各处的 OKR 进度汇总成一张看得懂的全景图。\n\n这个案例用 frontend-dev 技能生成了一个产品团队 Q2 OKR 追踪仪表盘：4 个 KPI 卡片（DAU / 7 日留存 / NPS / 发版频率）带环比变化指标、各 OKR 目标完成度对比柱状图（目标值 vs 实际值）、负责人维度完成率环形图、底部 OKR 明细表格支持排序。全部纯 CSS/SVG 实现，零外部依赖，打开即用。\n\n适合：产品经理、项目经理、团队 Lead、需要做季度汇报的任何人。", "description_en": "Dreading the quarterly review because OKR progress is scattered across a dozen spreadsheets?\n\nThis case uses the frontend-dev skill to generate a Product Team Q2 OKR tracking dashboard: 4 KPI cards (DAU / 7-day Retention / NPS / Release Frequency) with trend indicators, OKR completion comparison bar chart (target vs actual), owner-dimension completion donut chart, and a sortable detail table. All pure CSS/SVG, zero external dependencies, open and use.\n\nIdeal for: Product managers, project managers, team leads, anyone preparing quarterly reviews.", "categories": [ "data", "office" ], "tags": [ "OKR", "数据看板", "仪表盘", "季度汇报", "CSS图表" ], "tags_en": [ "OKR", "Data Dashboard", "Dashboard", "Quarterly Review", "CSS Charts" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为产品团队生成一个 Q2 OKR 追踪仪表盘（单文件 HTML，内联无外部依赖），界面文字用简体中文。\n\n要直观看到整体 KPI、各 Objective 的完成度，以及负责人维度的达成情况，适合在团队周会上展示。\n\nKPI（本季）：DAU 8.2 万（环比 +15.3%）、7 日留存 42.6%（+2.1%）、NPS 67（+5）、月发版 12 次（-8.3%）。\n\nOKR 完成度：\n- O1 提升用户活跃度：KR1 DAU 破 10 万（实际 82000，82%，陈思远）/ KR2 7 日留存提至 45%（实际 42.6%，94.7%，张芳）/ KR3 日均使用时长 25 分钟（实际 22min，88%，陈思远）\n- O2 提高产品口碑：KR1 NPS 达 70（实际 67，95.7%，李明）/ KR2 App Store 评分 4.7（实际 4.6，97.9%，王浩）\n- O3 加快迭代速度：KR1 月均发版 15 次（实际 12，80%，王浩）/ KR2 线上 P0 故障归零（实际 1，部分达成，李明）/ KR3 自动化测试覆盖率 80%（实际 73%，91.3%，张芳）\n\n负责人完成率：陈思远 85% / 张芳 93% / 李明 96% / 王浩 89%。\n\n完成率高/中/低用不同颜色区分即可。", "prompt_en": "Please generate a Q2 OKR tracking dashboard for the product team (single-file HTML, inline with no external dependencies), with interface text in Simplified Chinese.\n\nIt should give an at-a-glance view of overall KPIs, the completion level of each Objective, and achievement by owner, suitable for presenting at a team weekly meeting.\n\nKPIs (this quarter): DAU 82,000 (MoM +15.3%), 7-day retention 42.6% (+2.1%), NPS 67 (+5), monthly releases 12 (-8.3%).\n\nOKR completion:\n- O1 Improve user engagement: KR1 DAU breaks 100,000 (actual 82,000, 82%, Chen Siyuan) / KR2 raise 7-day retention to 45% (actual 42.6%, 94.7%, Zhang Fang) / KR3 average daily usage 25 min (actual 22min, 88%, Chen Siyuan)\n- O2 Improve product reputation: KR1 NPS reaches 70 (actual 67, 95.7%, Li Ming) / KR2 App Store rating 4.7 (actual 4.6, 97.9%, Wang Hao)\n- O3 Accelerate iteration: KR1 average 15 releases/month (actual 12, 80%, Wang Hao) / KR2 zero online P0 incidents (actual 1, partially achieved, Li Ming) / KR3 automated test coverage 80% (actual 73%, 91.3%, Zhang Fang)\n\nOwner completion rates: Chen Siyuan 85% / Zhang Fang 93% / Li Ming 96% / Wang Hao 89%.\n\nJust distinguish high / medium / low completion rates with different colors.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 133, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "OKR 追踪仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "marketing-campaign-roi", "title": "营销活动 ROI 分析表", "subtitle": "多渠道投放数据一目了然，ROI 自动算清", "description": "做营销最怕什么？钱花出去了，但不知道花在哪个渠道最值。微信、抖音、小红书、SEM、信息流、邮件、KOL、线下活动——八个渠道同时投放，每个渠道的 CPM、CPC、CPA 各不相同，月底汇报时 Excel 拼了十几个表还是算不清楚哪个渠道的 ROI 最高、预算该往哪里倾斜。\n\n这个案例用 AI 一键生成一套完整的营销活动 ROI 分析 Excel：覆盖 8 大主流渠道、6 个月投放数据，自动计算 CPM/CPC/CPA/ROAS 四大核心指标，内置渠道 ROI 横向对比、预算分配优化建议、从曝光到转化的全链路漏斗分析。打开就能看到哪个渠道花钱最值、哪个渠道在浪费预算、下个月的预算该怎么调。\n\n适合：市场营销经理、增长黑客、CMO、投放优化师、数字营销团队、任何需要用数据驱动营销决策的人。", "categories": [ "data", "business" ], "tags": [ "Excel", "营销分析", "ROI", "渠道投放", "ROAS", "预算优化" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "营销活动 ROI 分析表" } ], "prompt": "帮我生成一份专业的营销活动 ROI 分析 Excel 报表。\n\n**业务背景**：我们是一家消费品牌，在 8 个主流渠道投放营销广告，需要分析各渠道投入产出比、优化预算分配。请生成 2025 年下半年 6 个月的合理模拟数据。\n\n**报告结构（多 Sheet）**：\n- Sheet 1 渠道总览：各渠道 KPI 汇总，代表性字段如渠道、投放金额、曝光、点击、转化、收入、CPM、CPC、CPA、ROAS，并给出合计行。\n- Sheet 2 月度明细：6 个月 × 8 渠道逐月数据，体现投放节奏。\n- Sheet 3 ROI 对比：各渠道 ROAS 排名与成本效率对比。\n- Sheet 4 预算分配：当前 vs 建议预算，高 ROAS 增投、低 ROAS 缩减。\n- Sheet 5 转化漏斗：从曝光到成交的全链路漏斗与分环节流失。\n\n格式：专业商务风格，关键指标用条件格式分级，金额带 ¥，首行冻结并加自动筛选。输出为 .xlsx 文件。", "prompt_en": "Help me generate a professional marketing campaign ROI analysis Excel report.\n\n**Business context**: We're a consumer brand running ads across 8 major channels, and we need to analyze each channel's input-output ratio and optimize budget allocation. Please generate reasonable simulated data for 6 months in H2 2025.\n\n**Report structure (multiple sheets)**:\n- Sheet 1 Channel Overview: a KPI summary per channel, with representative fields like channel, spend, impressions, clicks, conversions, revenue, CPM, CPC, CPA, ROAS, plus a total row.\n- Sheet 2 Monthly Detail: month-by-month data for 6 months × 8 channels, reflecting spend rhythms.\n- Sheet 3 ROI Comparison: ROAS ranking and cost-efficiency comparison per channel.\n- Sheet 4 Budget Allocation: current vs. recommended budget — increase high-ROAS channels, cut low-ROAS ones.\n- Sheet 5 Conversion Funnel: a full funnel from impression to conversion with per-stage drop-off.\n\nFormatting: professional business style; use conditional-format tiers for key metrics; amounts with ¥; freeze the top row and add auto-filter. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 134, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Marketing Campaign ROI Analysis", "subtitle_en": "Crystal-clear multi-channel ad performance with automated ROI calculation", "description_en": "The biggest headache in marketing? Money goes out the door, but you can't tell which channel delivers the best bang for the buck. WeChat, Douyin, Xiaohongshu, SEM, Display, Email, KOL, Offline events — eight channels running simultaneously, each with different CPM, CPC, and CPA. Come month-end reporting, you're stitching together a dozen spreadsheets and still can't figure out which channel has the highest ROI or where to shift the budget.\n\nThis case uses AI to generate a complete marketing campaign ROI analysis Excel in one click: covering 8 major channels across 6 months of spend data, with automatic CPM/CPC/CPA/ROAS calculations, built-in cross-channel ROI comparison, budget allocation optimization, and a full-funnel analysis from impressions to conversions. Open it and instantly see which channels are worth every penny, which are burning budget, and how to reallocate spend next month.\n\nBest for: Marketing managers, growth hackers, CMOs, media buyers, digital marketing teams, anyone who needs data-driven marketing decisions.", "tags_en": [ "Excel", "Marketing Analytics", "ROI", "Channel Performance", "ROAS", "Budget Optimization" ], "is_recommended": false, "recommended_order": 0 }, { "id": "product-feature-demo-animation", "title": "产品功能演示动画页", "title_en": "Product Feature Demo Animation Page", "subtitle": "滚动触发动画 × 代码高亮 × UI 模拟，Linear 级产品功能展示", "subtitle_en": "Scroll-triggered animations, code highlights & UI mockups — Linear-quality feature showcase", "description": "产品上新功能想做个漂亮的展示页，写 PPT 太静态，录视频太重，Stripe/Linear 那种「用动画 sell 功能」的效果又不知道怎么实现。这个案例演示了一个 Linear 风格的产品功能展示页：滚动触发渐入动画、代码块打字机效果、UI 界面模拟动画、数据流转动效，整体质感对标顶级 SaaS 产品官网的 Feature Section。\n\n适合：产品市场团队做功能宣传页、创业者做产品展示 Demo、前端工程师学习高级滚动动画。", "description_en": "Want a stunning showcase page for your new product feature? Slides are too static, videos too heavy, and you don't know how to achieve that Stripe/Linear 'sell features with animation' effect. This case demonstrates a Linear-style product feature showcase page: scroll-triggered fade-in animations, code block typewriter effects, UI simulation animations, and data flow transitions — matching the quality of top-tier SaaS product websites' Feature Sections.\n\nPerfect for: product marketing teams building feature pages, founders creating product demos, frontend engineers learning advanced scroll animations.", "categories": [ "creation", "business" ], "tags": [ "产品展示", "滚动动画", "SaaS官网", "品牌设计" ], "tags_en": [ "Product Showcase", "Scroll Animation", "SaaS Website", "Brand Design" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "产品功能演示动画页面" } ], "prompt": "请生成一个 Linear 风格的产品功能演示页面（单文件 HTML，内联无外部依赖），展示「WorkBuddy AI 智能助手」，深色主题、滚动触发动画。\n\n模块：\n1. Hero：大号渐变标题 + 浮动 UI 模型 + CTA\n2. 四个功能展示区（智能回复 / 文档分析 / 代码审查 / 团队洞察），各配独立滚动动画\n3. 代码块打字机效果 + 行内标注\n4. 仪表盘数据计数 + 图表绘制动效\n5. 固定顶部导航（毛玻璃）+ 平滑滚动", "prompt_en": "Generate a Linear-style product feature demo page (single-file HTML, inlined with no external dependencies), showcasing \"WorkBuddy AI Assistant,\" with a dark theme and scroll-triggered animations.\n\nModules:\n1. Hero: large gradient headline + floating UI mockup + CTA\n2. Four feature showcase sections (smart replies / document analysis / code review / team insights), each with its own scroll animation\n3. Code block typewriter effect + inline annotations\n4. Dashboard data counters + chart-drawing animations\n5. Fixed top navigation (frosted glass) + smooth scrolling", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "全栈前端开发，融合高端 UI 设计与动画效果", "skill_id": "skill_2052764137292271616" }, { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 135, "is_featured": false, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-05-03T10:00:00.000Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "order-receipt", "title": "电商订单收据", "title_en": "E-commerce Order Receipt", "subtitle": "输入订单信息，生成精美的订单确认收据页面", "subtitle_en": "Input order details, generate an elegant order confirmation receipt page", "description": "电商下单后的确认页、邮件里的订单收据、打印出来贴在包裹里的小票——这些细节体现着品牌的专业度。但大多数中小商家的订单确认页面简陋得像 Excel 表格，完全没有品牌感。\n\n这个案例演示了输入一笔电商订单的完整信息（商品列表、价格、收货地址、支付方式），AI 自动生成一张精美的订单收据 HTML 页面：居中小票风格（max-width 480px）、品牌 header、商品列表含缩略图占位、价格明细（小计/运费/优惠/总计）、收货信息、支付信息、底部订单号和客服联系方式。模拟高端品牌的购物确认单。\n\n适合：电商独立站、品牌商家、订单确认邮件模板、打印小票设计。", "description_en": "Post-purchase confirmation pages, email order receipts, printed packing slips — these details reflect brand professionalism. But most small merchants' order confirmations look like bare Excel spreadsheets with zero brand identity.\n\nThis case demonstrates inputting a complete e-commerce order (product list, prices, shipping address, payment method), with AI generating an elegant order receipt HTML page: centered receipt style (max-width 480px), brand header, product list with thumbnail placeholders, price breakdown (subtotal/shipping/discount/total), shipping info, payment info, bottom order number and support contact. Simulates premium brand shopping confirmation.\n\nBest for: E-commerce independent sites, brand merchants, order confirmation email templates, printed receipt design.", "categories": [ "business", "office" ], "tags": [ "订单收据", "电商", "小票设计", "品牌模板" ], "tags_en": [ "Order Receipt", "E-commerce", "Receipt Design", "Brand Template" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一张电商订单收据页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，风格极简克制、MUJI 式高级感，含价格明细和小票撕边感。\n\n请用下面这样的订单信息（可替换为你自己的）：店铺名、订单号与下单时间、商品列表（名称/规格/数量/单价）、费用明细（小计/运费/优惠/实付）、收货信息、支付信息。\n...", "prompt_en": "Generate an e-commerce order receipt page (single-file HTML, inline with no external dependencies), UI text in Simplified Chinese, minimalist and restrained MUJI-style, with a price breakdown and a torn-receipt edge feel.\n\nUse order info like the following (replace with your own): shop name, order number and time, item list (name/spec/qty/unit price), fee breakdown (subtotal/shipping/discount/amount paid), shipping info, payment info.\n...", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 136, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "订单收据页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ai-oral-english-competitive", "title": "AI 口语陪练竞品深度分析", "title_en": "Deep Competitive Analysis for AI Oral English Coach", "subtitle": "8 款竞品对标 + Battle Card + 定价策略 + 市场数据", "subtitle_en": "8-competitor benchmarking with Battle Cards, pricing strategy, and market data", "description": "推 AI 口语陪练产品却不知道市场格局？这个案例对标了斑马AI、洪恩、多邻国、ChatGPT口语、有道Hi Echo、流利说、VIPKID、伴鱼绘本 8 款竞品，输出完整竞品画像、学习路径对比、口碑优劣分析、差异化机会和销售用 Battle Card，同时提供市场规模数据、定价策略建议和 6 个月增长预测。交付物含 Markdown 报告和 HTML 可视化版（竞品雷达图、功能对比表、定价阶梯图）。\n\n适合：教育产品经理、在线教育创业者、需要做竞品分析的产品团队。", "description_en": "Launching an AI oral English product but unclear about the market? This case benchmarks 8 competitors (Zebra, HongEn, Duolingo, ChatGPT, Hi Echo, Liulishuo, VIPKID, BanYu), delivering full competitive profiles, learning path comparisons, reputation analysis, differentiation opportunities, and Battle Cards for sales. Includes market sizing, pricing strategy, and 6-month growth forecast. Deliverables include Markdown report and HTML visualization with radar chart, feature matrix, and pricing ladder.\n\nSuitable for: EdTech product managers, online education entrepreneurs, product teams needing competitive analysis.", "categories": [ "business", "data" ], "tags": [ "竞品分析", "AI口语", "教育产品", "Battle Card", "市场数据" ], "tags_en": [ "Competitive Analysis", "AI Oral English", "EdTech", "Battle Card", "Market Data" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "竞品分析报告(Markdown)" }, { "file": "output.html", "label": "可视化报告(HTML雷达图+对比表)" } ], "prompt": "为一家一线城市儿童教育机构的「AI 口语陪练」在线产品做完整竞品分析。对标斑马AI、洪恩、多邻国、ChatGPT口语、有道Hi Echo、流利说、VIPKID、伴鱼绘本 8 款竞品。\n\n要求：\n1. 每个竞品的定位/目标用户/定价/核心功能/商业模式/技术路线\n2. 学习路径和内容体系对比\n3. 各自用户口碑优劣\n4. 差异化机会分析（结合线下机构优势）\n5. Battle Card 给销售/运营使用\n6. 市场规模数据、定价策略建议、增长预测\n\n输出 Markdown 报告 + HTML 可视化版（竞品雷达图、功能对比表、定价阶梯图）。", "prompt_en": "Conduct a complete competitive analysis for the online \"AI Oral English Coach\" product of a children's education institution in a tier-1 city. Benchmark against 8 competitors: Zebra AI, HongEn, Duolingo, ChatGPT Voice, Youdao Hi Echo, LiuLiShuo, VIPKID, and BanYu Picture Books.\n\nRequirements:\n1. Each competitor's positioning / target users / pricing / core features / business model / technical approach\n2. Comparison of learning paths and content systems\n3. Each one's user-reputation strengths and weaknesses\n4. Differentiation opportunity analysis (leveraging the offline institution's advantages)\n5. Battle Cards for sales/operations use\n6. Market size data, pricing strategy recommendations, and growth forecast\n\nOutput a Markdown report + an HTML visualization version (competitor radar chart, feature comparison table, pricing ladder chart).", "skills": [], "experts": [ { "id": "ProductStrategyTeam", "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师（PRD/功能规格书）、用户研究员（调研综合分析）、竞品分析师（竞争情报）、数据分析师（指标追踪）和路线图规划师（路线图管理/迭代规划）。覆盖从构思到上线的完整产品生命周期。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "", "name_en": "Product Strategy Team", "description_en": "A team of 5 product specialists led by a Product Director: Requirement Analyst (PRD/feature spec writing), User Researcher (research synthesis), Competitive Analyst (competitive intelligence), Data Analyst (metrics tracking), and Roadmap Planner (roadmap management/sprint planning). Covers the full product lifecycle from ideation to launch.", "expert_id": "ex_LqCvsVsLOp3J" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 137, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T15:24:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "sales-battle-quarterly-pipeline-review", "title": "季度 Pipeline 评审报告", "title_en": "Quarterly Pipeline Review Report", "subtitle": "从健康度打分到预测区间，再到两周推进动作的一体化评审", "subtitle_en": "An integrated review from pipeline health scoring to forecast ranges and next-two-week actions", "description": "季度 Pipeline 评审最容易掉进两个坑：要么只看总量，不看结构；要么只报预测，不给动作。这个案例演示了销售作战团队如何围绕 8,000 万 Pipeline 与 3,000 万目标回款做完整评审：先对覆盖率、转化率、停滞率和漏斗形状打分，再拆出风险交易、推动交易、管理层介入机会和保守/基准/乐观预测区间，最后补上未来两周的推进动作与风险矩阵。最终输出 Markdown 报告 + HTML 仪表盘。\n\n适合：销售 VP、CRO、销售总监、销售运营、区域管理者", "description_en": "Quarterly pipeline reviews often fail in two ways: they either focus on total volume without looking at structure, or they report forecasts without attaching actions. This case shows how the Sales Battle Team reviews an 80M pipeline against a 30M quarter target: scoring coverage, conversion, stagnation, and funnel shape; identifying risk deals, push deals, and management escalation opportunities; modeling conservative/base/upside forecast ranges; and attaching a concrete two-week action plan plus a risk matrix. The final output includes a Markdown report and an HTML dashboard.\n\nBest for: Sales VPs, CROs, Sales Directors, Sales Operations, Regional Leaders", "categories": [ "business" ], "tags": [ "Pipeline评审", "销售预测", "漏斗分析", "风险矩阵", "季度复盘" ], "tags_en": [ "pipeline review", "sales forecast", "funnel analysis", "risk matrix", "quarterly review" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "季度 Pipeline 评审仪表盘" }, { "file": "output.md", "label": "季度 Pipeline 评审报告" } ], "prompt": "请对我团队当前季度的 Pipeline 做全面评审（总 Pipeline 金额 8000 万，目标回款 3000 万）：\n1. 整体健康度打分（覆盖率、转化率、停滞率、漏斗形状）\n2. 风险交易清单（卡住/倒退/预期外）\n3. 推动交易清单（临门一脚）\n4. 需要管理层介入的高价值交易\n5. 最终预测区间（保守/基准/乐观）与达成概率\n6. 未来两周行动建议\n\n输出《季度 Pipeline 评审报告》Markdown + HTML 仪表盘（漏斗图 + 风险矩阵）。", "prompt_en": "Please conduct a comprehensive review of my team's pipeline for the current quarter (total pipeline value 80M, target collections 30M):\n1. Overall health scoring (coverage, conversion rate, stagnation rate, funnel shape)\n2. Risk deal list (stuck / regressed / unexpected)\n3. Push deal list (near the finish line)\n4. High-value deals requiring management intervention\n5. Final forecast ranges (conservative / base / upside) and achievement probability\n6. Action recommendations for the next two weeks\n\nOutput a \"Quarterly Pipeline Review Report\" in Markdown + an HTML dashboard (funnel chart + risk matrix).", "skills": [], "experts": [ { "id": "SalesBattleTeam", "name": "销售作战团队", "description": "销售攻坚体系，客户研究锁定目标、外联策略提升触达、竞品情报预警风险、销售预测优化资源分配", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "ff06833f0e30", "name_en": "Sales Battle Team", "description_en": "Sales effectiveness system: account research identifies targets, outreach strategy improves reach, competitive intel warns risks, pipeline forecasting optimizes resource allocation", "expert_id": "ex_ATh4MnfFCKxR" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 138, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:38:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "app-release-changelog", "title": "App 版本更新日志", "title_en": "App Release Changelog", "subtitle": "从变更记录自动生成用户友好的 Word 版 Changelog", "subtitle_en": "Auto-generate user-friendly Word changelog from change records", "description": "每次发版都要手写更新日志，开发写的太技术化用户看不懂，产品写的又漏掉关键变更？这个案例展示了如何用 Word 文档生成技能，输入版本变更列表，自动按 Keep a Changelog 规范分类整理，同时生成面向用户的友好版本说明，输出为正式 Word 文档格式。\n\n以移动 App v3.2.0 为例，演示了将深色模式上线、首页加载优化 40%、支付崩溃修复等变更记录自动整理为 Added/Changed/Fixed/Security 四大分类，并附带面向普通用户的版本说明，格式规范可直接归档。\n\n适合产品经理、发布工程师、技术写作者。", "description_en": "Writing release notes manually every sprint — dev notes are too technical, PM notes miss key changes? This case shows how to use the Word Document Generation skill to input change records and auto-generate categorized changelogs following Keep a Changelog format, plus user-friendly release notes, all in a formal Word document.\n\nUsing mobile App v3.2.0 as an example, it demonstrates organizing changes like dark mode launch, 40% homepage speed improvement, and payment crash fixes into Added/Changed/Fixed/Security categories with plain-language user notes.\n\nPerfect for PMs, release engineers, and tech writers.", "categories": [ "dev", "productivity" ], "tags": [ "更新日志", "版本发布", "Changelog", "发布管理", "技术写作" ], "tags_en": [ "Changelog", "Release Notes", "Version Management", "Release Engineering", "Tech Writing" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "App版本更新日志（Word）" } ], "prompt": "帮我为一个移动 App 生成一份版本更新日志 Word 文档（output.docx），版本号 v3.2.0。假设最近两周有以下变更：新增深色模式、优化首页加载速度 40%、修复支付流程偶发崩溃、新增 3 种消息通知模板、升级隐私策略合规、API 接口向后兼容调整。\n\n文档要求：\n1. 按 Keep a Changelog 格式输出，分为 Added/Changed/Fixed/Security 四类\n2. 同时生成面向用户的友好版本更新说明\n3. 包含封面页、版本信息摘要表、详细变更列表\n4. 文档格式规范：多级标题、统一表格样式、页眉页脚\n5. 输出为正式 Word 文档（output.docx）", "prompt_en": "Help me generate a release changelog Word document (output.docx) for a mobile app, version v3.2.0. Assume the following changes happened over the past two weeks: added dark mode, improved home page load speed by 40%, fixed occasional crashes in the payment flow, added 3 new message notification templates, upgraded privacy policy compliance, and made backward-compatible API adjustments.\n\nDocument requirements:\n1. Output in the Keep a Changelog format, divided into the four categories Added/Changed/Fixed/Security\n2. Also generate a user-friendly version of the release notes\n3. Include a cover page, a version-info summary table, and a detailed list of changes\n4. Follow document formatting standards: multi-level headings, a consistent table style, and headers/footers\n5. Output as a formal Word document (output.docx)", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 139, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "brand-logo-creative", "title": "品牌 Logo 创意应用", "title_en": "Brand Logo Creative Application", "subtitle": "让黑色圆角标识在六个真实世界场景中自然出现", "subtitle_en": "A rounded black mark appears naturally in six real-world scenes", "description": "品牌 Logo 动效经常只是平面缩放，难以展示标识在不同材质、尺度和环境中的适应力。\n\n案例把一个黑色圆角符号分别放进纹身、海运集装箱、冰淇淋粉面、复古车头、城市环岛和自动售货机，再将六张超写实应用图做成短动态合集。\n\n适合：品牌发布、Logo 提案、社交媒体包装、设计公司案例和视觉识别系统展示。", "description_en": "Logo motion often amounts to flat scaling and fails to show how a mark behaves across materials, scales, and environments.\n\nA rounded black symbol appears as a tattoo, shipping-container mark, cocoa imprint, vintage-car badge, traffic island, and vending-machine panel before the six scenes become a motion compilation.\n\nBest for brand launches, logo proposals, social packaging, design-studio cases, and identity-system presentations.", "categories": [ "creation" ], "tags": [ "品牌Logo", "创意应用", "视觉识别" ], "tags_en": [ "brand logo", "creative application", "visual identity" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「品牌 Logo 创意应用」Skill，使用我上传的黑色圆角几何 Logo，制作默认 9:16 的六场景动态展示《一个 Logo 的六次出现》。先分析 Logo 的轮廓、负形、颜色和文字，确认全局规格后，提出六个彼此差异明显的超写实方案：纹身师正在把符号纹到手臂；海上货轮的白色集装箱；冰淇淋表面的可可粉压印；复古黑色跑车的金属车头章；俯拍城市圆形交通岛；夜间自动售货机的发光面板。先展示六个创意与动态逻辑让我确认，再逐张生成 2K 场景图并再次确认，最后每张生成 3–5 秒视频并可选拼接。每个场景都必须保持 Logo 几何比例与识别度，不生成额外品牌文字。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Brand Logo Creative Application” Skill and use my uploaded rounded black geometric logo to create a default 9:16 six-scene motion showcase titled Six Appearances of One Logo. Analyze the logo’s silhouette, negative space, colors, and lettering first. After confirming the global specification, propose six clearly differentiated photorealistic applications: a tattoo artist placing the mark on an arm; a white shipping container on a cargo vessel; a cocoa-powder imprint on ice cream; a metal hood badge on a vintage black sports car; an overhead view of a circular urban traffic island; and an illuminated panel on a nighttime vending machine. Show all six concepts and motion ideas for approval, then generate the six 2K scene images and obtain approval again. Finally turn each image into a 3–5 second video and optionally assemble them. Preserve the exact logo geometry and recognizability in every scene, and generate no additional brand wording.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 140, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "cultural-product-3d-silicone", "title": "文创品建模 · 硅胶材质", "title_en": "Cultural Product 3D · Silicone Material", "subtitle": "把文创品变成软胶质感的可交互 3D 模型", "subtitle_en": "Turn a creative product into a soft-silicone interactive 3D model", "description": "硅胶周边打样贵、周期长，前期很难预览成品的软胶质感与形态。\n\n案例识别文创品的品类、轮廓、比例与纹样，转译为软胶（硅胶）材质的可交互 3D 模型：柔光半透质感、圆润手感，网页可 360° 旋转缩放，单文件离线可开。\n\n适合：文创衍生开发、潮玩/礼品设计师、博物馆文创团队、IP 授权方", "description_en": "This case reads the product's category, silhouette, proportion and motif, translating it into a soft-silicone interactive 3D model with soft-light translucency and rounded feel — rotate and zoom in the browser, single-file and offline-ready.\n\nGood for: Merch developers, Toy & gift designers, Museum creative teams, IP licensors", "categories": [ "creation" ], "tags": [ "3D建模", "硅胶材质", "可交互模型", "文创衍生" ], "tags_en": [ "3D Modeling", "Silicone Material", "Interactive Model", "Cultural Merch" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据用户上传的文创品图片，准确识别主体的品类、外形轮廓、长宽比例、造型结构、朝向、主色、纹样、配件及关键识别特征，并将其作为画面中的唯一主体。完整保留原产品设计，不重新设计、不增删结构、不改变比例、不替换品类、不擅自卡通化；仅将主体表面材质转换为一体成型的软质哑光硅胶材质。\n\n构图：主体居中完整呈现，采用四分之三前侧视角（3/4 front-side view）；原图中属于产品本体的底座、支撑或配件应一并保留。使用纯净白色无缝背景和产品级静物棚拍，无文字、无水印、无装饰道具、无人物、无第二主体。\n\n材质：呈现细腻、柔软、略带弹性的模压硅胶质感，表面均匀顺滑，带有极细微的磨砂颗粒与柔和触感；整体低反射，以柔和漫反射和宽缓高光表现材质，不出现坚硬锐利的镜面反光。边缘仅可产生符合硅胶工艺的轻微软化效果，不得改变原有轮廓和结构。忠实保留原产品的主色、色彩关系及纹样位置，只通过表面光泽、触感和细微明暗变化体现硅胶材质。\n\n光影与画质：采用柔和均匀的棚拍光线，保留轻微自然的接触阴影，主体轮廓清晰、细节完整、画面干净，适合作为3D建模与材质重建参考。避免塑料感、蜡质感、油腻反光、玻璃透明感、金属质感、毛绒纤维、过度圆润变形、结构缺失、模糊、强烈阴影遮挡及背景杂物。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-3d-expert", "name": "文创品3D建模专家", "description": "将文创品转译为可交互的 3D 模型效果，支持毛绒、硅胶、金属珐琅、亚克力等材质渲染", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_P14ik4i7Utla" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "可交互 3D 模型（单文件内联）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 140, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-22T15:31:07+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "nps-satisfaction-survey", "title": "NPS 用户满意度调研问卷", "title_en": "NPS User Satisfaction Survey", "subtitle": "专业 NPS 问卷页面，动态追问 + 满意度矩阵", "subtitle_en": "Professional NPS survey with dynamic follow-ups and satisfaction matrix", "description": "每次做用户满意度调研，从零搭建一份专业 NPS 问卷既耗时又容易遗漏关键指标——评分逻辑、动态追问、满意度矩阵、移动端适配，每项都要反复调试。\n\n这个案例展示了如何用一句话描述产品和调研目标，AI 即可生成一份完整的交互式 NPS 调研页面：0-10 分彩色评分条带动画反馈，根据分数自动切换追问题（贬损者多选痛点、被动者填写改进建议、推荐者勾选喜爱功能），五维满意度矩阵评分，以及可选的用户信息收集模块，全响应式设计，纯前端无外部依赖。\n\n适合：产品经理、用研团队、运营同学、客户成功经理，以及任何需要快速上线专业 NPS 调研的团队。", "description_en": "Building a professional NPS survey from scratch is time-consuming — scoring logic, dynamic follow-ups, satisfaction matrices, and mobile responsiveness all need careful tuning.\n\nThis case demonstrates generating a complete interactive NPS survey page from a single prompt: a color-coded 0-10 rating bar with animations, dynamic follow-up questions based on score (detractor pain points, passive improvement suggestions, promoter favorite features), a five-dimension satisfaction matrix, and optional user demographics — fully responsive, pure frontend, no external dependencies.\n\nBest for: Product managers, UX researchers, operations teams, customer success managers, and anyone who needs to quickly launch a professional NPS survey.", "categories": [ "office", "data" ], "tags": [ "NPS", "用户调研", "满意度", "问卷", "表单" ], "tags_en": [ "NPS", "User Research", "Satisfaction", "Survey", "Form" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "NPS 调研问卷页面" } ], "prompt": "请设计一个专业的 NPS 用户满意度调研问卷页面（单文件 HTML，内联无外部依赖），现代简洁、单页滚动式、响应式、NPS 评分条有交互动画，提交后显示感谢页（无需真实提交）。\n\n产品「TaskFlow」，调研 v3.0 版本满意度。\n1. 品牌化头部 + 调研说明\n2. NPS 核心题「您有多大可能向同事推荐？」0-10 评分条，0-6 贬损者/7-8 被动者/9-10 推荐者按色区分\n3. 动态追问（按分数显示不同问题）：0-6 多选体验问题、7-8 文本框、9-10 多选喜欢的功能\n4. 满意度矩阵：易用性/功能完整性/性能稳定性/客服支持/性价比 各 5 档\n5. 开放反馈框\n6. 可选用户信息：使用时长/团队规模/行业\n7. 提交按钮 + 感谢页", "prompt_en": "Please design a professional NPS customer satisfaction survey page (single-file HTML, fully inline with no external dependencies), in a modern, clean, single-page scrolling, responsive style, with an interactive animation on the NPS rating bar, and a thank-you page shown after submission (no real submission needed).\n\nThe product is \"TaskFlow,\" surveying satisfaction with version v3.0.\n1. A branded header + survey description\n2. The core NPS question \"How likely are you to recommend us to a colleague?\" with a 0-10 rating bar, color-coded as 0-6 detractors / 7-8 passives / 9-10 promoters\n3. Dynamic follow-up questions (different questions shown by score): a multiple-choice experience question for 0-6, a text box for 7-8, a multiple-choice \"features you like\" question for 9-10\n4. A satisfaction matrix: ease of use / feature completeness / performance stability / customer support / value for money, each on a 5-point scale\n5. An open-ended feedback box\n6. Optional user info: length of use / team size / industry\n7. A submit button + a thank-you page", "skills": [ { "id": "svg-draw", "name": "SVG 绘图", "description": "使用 SVG 创建矢量图形、图标和可视化设计", "slug": "svg-draw" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 91, "sort_order": 140, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-21T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-competition-filter", "title": "竞争比查询：国考低竞争段真实比值", "title_en": "Competition Ratios: Real Low-Ratio National Exam Posts", "subtitle": "先把 300:1 的卷王岗筛出去，同样的分数才有胜算", "subtitle_en": "Screen out 300:1 battlegrounds first — the same score goes further", "description": "同样 130 分，报在 300:1 的岗是炮灰，报在 30:1 的岗有胜算。报岗第一步不是挑喜欢的，是排除卷不动的。\n\n本案例按「国考 + 竞争比≤30:1」实时筛出 7528 个岗位，逐行展示真实竞争比（低至 5:1、8:1），有往年进面分的标出实值，没有的如实标注「无历史数据」，并解释低竞争的常见成因——专业门槛替你筛掉了对手。\n\n适合：分数中游想避开激烈竞争的考生、研究报考策略的老师。", "description_en": "With the same 130 points, applying to a 300:1 post makes you cannon fodder; a 30:1 post gives you a real shot. Step one isn't picking favourites — it's excluding the unwinnable.\n\nThis case live-filters 7,528 national-exam positions with ratios ≤30:1, showing real per-post ratios (as low as 5:1 and 8:1), marking actual historical cutoffs where they exist and honestly labelling \"no historical data\" where they don't, and explains why low competition happens — eligibility bars filter out rivals for you.\n\nFor: mid-scoring candidates avoiding fierce competition and strategy-minded advisors.", "categories": [ "search" ], "tags": [ "竞争比", "国考", "低竞争", "报考策略" ], "tags_en": [ "competition ratio", "national exam", "low competition", "application strategy" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "国考竞争比≤30:1 岗位清单（真实比值）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我在国考岗位里找竞争比不超过 30:1 的。要求：1) 实时查询并给出总数；2) 逐条列岗位、单位、真实竞争比、往年进面分（没有就如实标注无历史数据，不许用别的岗替代）、招录人数，重复条目去重后展示；3) 分析这批岗位竞争低的原因；4) 提醒「低竞争不等于低门槛」要核对什么。", "prompt_en": "Find national-exam positions with competition ratios no higher than 30:1. Requirements: 1) query live and give the total; 2) list each position with unit, real competition ratio, historical cutoff (label \"no historical data\" honestly where missing — never substitute another post's score) and headcount, de-duplicating repeated rows; 3) analyse why competition is low; 4) note what to verify since low competition ≠ low bar.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "按报录比区间筛选岗位，避开竞争过激的岗", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 140, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "smart-medical-record-inpatient", "title": "住院全流程三阶段病历衔接", "title_en": "Inpatient Record: Admission to Discharge Chained", "subtitle": "入院 → 病程 → 出院自动衔接，病案首页一键汇总", "subtitle_en": "Admission → progress → discharge chained, auto front-sheet summary", "description": "65 岁慢心衰患者入院，3 天加呋塞米、5 天 BNP 下降、7 天出院带三种药——住院病历要写入院记录、病程记录、出院小结，还要出病案首页，纯靠医生手敲耗时一两个小时。\n\n这个案例演示了把分阶段口述内容连续交给 AI 临床助手，它自动判别住院场景，先写入院记录；阶段完成后主动询问「是否继续病程阶段」，得到确认才推进，绝不自动跳转；病程独立撰写而非复制入院记录；出院小结汇总住院天数、带药、随访；最后自动汇总病案首页，并对时间逻辑、剂量等做三级校验（P0/P1/P2）。\n\n适合：住院医师、主治医师、病案室编码员、医务科管理者、关注病历质量与合规的临床科室。", "description_en": "A 65-year-old chronic HF patient is admitted, gets furosemide on day 3, BNP drops on day 5, and is discharged on day 7 with three medications. Writing the admission note, progress notes, discharge summary, plus the front sheet — by hand, this eats one to two hours of the doctor's day.\n\nThis case shows how to feed staged dictation into the AI assistant. It identifies the inpatient scenario, drafts the admission record first, then asks \"continue to progress notes?\" and only advances after explicit confirmation — never auto-jumps. Progress notes are written independently instead of being copy-pasted from the admission record. The discharge summary aggregates LOS, take-home meds and follow-up, and finally auto-rolls up the front sheet with a three-tier (P0/P1/P2) sanity check on timeline and dosage.\n\nFor: residents, attendings, medical records coders, medical affairs managers, and any clinical department focused on record quality and compliance.", "categories": [ "agent", "education" ], "tags": [ "智能病例", "住院病历", "病程记录", "出院小结", "病案首页", "AI临床助手", "腾讯健康" ], "tags_en": [ "Smart Medical Record", "Inpatient Record", "Progress Note", "Discharge Summary", "Front Sheet", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我写一份住院病历，分阶段完成入院记录、病程记录、出院小结，并自动汇总病案首页。\n\n入院信息：患者男性，65 岁，反复胸闷气促 3 年、加重 1 周入院；高血压病史 10 年、糖尿病 5 年；入院查体血压 160/95、双下肢轻度水肿；诊断慢性心力衰竭急性加重。\n\n要求：\n1. 判别场景为「住院病历」，先写入院记录（主诉、现病史、既往史、查体、诊断）；\n2. 入院完成后询问是否继续病程阶段，病程完成后询问是否继续出院阶段——阶段间必须确认后再继续，不自动跳转；\n3. 病程记录独立撰写，不直接复制入院记录，同一临床事实用不同表达方式；\n4. 出院小结汇总住院天数、出院诊断、出院带药、随访建议；\n5. 全部阶段完成后自动生成病案首页；\n6. 对时间逻辑、剂量等做三级校验（P0/P1/P2），P0 问题必须人工确认；\n7. 不做诊断推理、不新增诊断，不编造临床数据，末尾附安全声明。", "prompt_en": "Draft an inpatient record in stages — admission note, progress notes, discharge summary — and auto-roll up the front sheet.\n\nAdmission info: Patient is male, 65, recurrent chest tightness and dyspnea for 3 years, worse over 1 week; hypertension 10 years, diabetes 5 years; BP 160/95 on admission, mild edema of both lower limbs; diagnosis acute exacerbation of chronic heart failure.\n\nRequirements:\n1. Identify the scenario as \"inpatient record\" and write the admission note first (chief complaint, HPI, past history, physical exam, diagnosis).\n2. After the admission note, ask whether to continue to progress notes; after progress notes, ask whether to continue to discharge — advance only after explicit confirmation, never auto-jump.\n3. Write progress notes independently, never copying the admission record; use different wording for the same clinical facts.\n4. The discharge summary aggregates length of stay, discharge diagnosis, discharge meds and follow-up.\n5. Auto-generate the front sheet after all stages are done.\n6. Apply a three-tier check (P0/P1/P2) on timeline and dosage; P0 issues require human confirmation.\n7. Make no diagnostic inference, add no new diagnosis, never fabricate clinical data; end with a safety statement.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 140, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:57:00+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "artifacts": [ { "file": "output.html", "label": "住院全流程三阶段病历衔接" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "crm-ai-brainstorm", "title": "CRM AI 助手头脑风暴与创意筛选", "description": "中型企业 CRM SaaS 的 AI 助手用例头脑风暴：18个AI用例覆盖5大维度，价值/成本矩阵筛选，Top 5功能大纲，竞品差异化策略，AI功能分层定价建议", "experts": [ { "id": "ProductStrategyTeam", "expertType": "team", "categoryId": "01-ProductDesign", "plugin": "product-strategy-team", "displayName": { "zh": "产品战略团队" }, "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师（PRD/功能规格书）、用户研究员（调研综合分析）、竞品分析师（竞争情报）、数据分析师（指标追踪）和路线图规划师（路线图管理/迭代规划）。覆盖从构思到上线的完整产品生命周期。", "name_en": "Product Strategy Team", "description_en": "A team of 5 product specialists led by a Product Director: Requirement Analyst (PRD/feature spec writing), User Researcher (research synthesis), Competitive Analyst (competitive intelligence), Data Analyst (metrics tracking), and Roadmap Planner (roadmap management/sprint planning). Covers the full product lifecycle from ideation to launch.", "expert_id": "ex_LqCvsVsLOp3J" } ], "categories": [ "creation", "business" ], "tags": [ "头脑风暴", "AI助手", "CRM", "用例筛选", "价值成本矩阵", "RICE", "差异化", "定价策略" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化" }, { "file": "output.md", "label": "Markdown 文档" } ], "createdAt": "2026-04-26", "input": { "product": "中型企业 CRM SaaS", "useCaseCount": 18, "topPicks": 5, "dimensions": [ "销售效能", "客户成功", "数据智能", "自动化", "协作" ] }, "title_en": "CRM AI Assistant Brainstorming & Use Case Selection", "subtitle": "18 个 AI 用例 × 价值成本矩阵 × Top 5 功能大纲与分层定价", "subtitle_en": "18 AI use cases × value-cost matrix × top 5 feature outlines with tiered pricing", "description_en": "Mid-market CRM SaaS AI assistant brainstorming: 18 AI use cases across 5 dimensions, value/cost matrix screening, top 5 feature outlines, competitive differentiation strategy, and AI feature tiered pricing recommendations.", "tags_en": [ "Brainstorming", "AI Assistant", "CRM", "Use Case Selection", "Value-Cost Matrix", "RICE", "Differentiation", "Pricing Strategy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "sort_order": 141, "featured_order": 0, "usage_count": 0, "quality_score": 90, "is_featured": false, "created_at": "2026-04-27T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "skills": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "我是一家中型企业 CRM SaaS 的产品负责人，请帮我完整走一遍 AI 助手用例的头脑风暴与创意筛选：\n1) 围绕\"销售效能、客户成功、数据智能、自动化、协作\"五大维度，产出 18 个候选 AI 用例；\n2) 用价值/成本矩阵（高价值低成本、高价值高成本、低价值低成本、低价值高成本）对 18 个用例筛选，输出 Top 5；\n3) 给 Top 5 写功能大纲（用户故事、核心交互、数据依赖、成功指标）；\n4) 结合 Salesforce / HubSpot / Zoho / 飞书的 AI 能力差异，给出竞品差异化策略；\n5) 给出 AI 功能分层定价建议（基础版免费 / 专业版 / 企业版）。\n最终输出一份 Markdown 报告 + HTML 可视化页。", "prompt_en": "I'm the product lead for a mid-market enterprise CRM SaaS. Please help me run through a complete brainstorming and idea-screening exercise for AI assistant use cases:\n1) Around the five dimensions of \"sales effectiveness, customer success, data intelligence, automation, and collaboration,\" produce 18 candidate AI use cases;\n2) Use a value/cost matrix (high value low cost, high value high cost, low value low cost, low value high cost) to screen the 18 use cases and output a Top 5;\n3) Write feature outlines for the Top 5 (user stories, core interactions, data dependencies, success metrics);\n4) Based on the differences in AI capabilities across Salesforce / HubSpot / Zoho / Feishu, provide a competitive differentiation strategy;\n5) Provide tiered pricing recommendations for AI features (Free Basic / Pro / Enterprise).\nFinally, output a Markdown report + an HTML visualization page.", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-ratio-mid-band", "title": "竞争比筛选：50-100:1 中等竞争带", "title_en": "Ratio Filter: The 50-100:1 Middle Band", "subtitle": "避开千里挑一，也别只捡最冷的——被忽略的中间地带有 4,247 个岗位", "subtitle_en": "Skip the thousand-to-one posts and the coldest ones — 4,247 sit in between", "description": "挑岗时人们要么冲热门，要么专捡最冷门，很少有人认真看中间那一档。\n\n本案例筛出报录比 50–100:1 的国考岗位共 4,247 个，本页 30 条按比值升序排列，统计招录单位分布，指出这一档为什么长期被部委的专业方向岗占据，并如实说明报录比与进面分是两套数据、覆盖范围不重合。\n\n适合：想找性价比岗位的考生、专业方向较窄的报考者。", "description_en": "People either chase the hottest posts or hunt the coldest — few look carefully at the middle.\n\nThis case filters national-exam posts with ratios of 50–100:1, finding 4,247. The 30 shown are sorted by ascending ratio with a breakdown of recruiting units, explaining why ministry-level specialist posts dominate this band, and noting that ratio and cutoff data cover different sets of posts.\n\nFor: candidates seeking value-for-effort posts, especially those with narrow specialisms.", "categories": [ "search" ], "tags": [ "国考", "竞争比", "中等竞争", "部委岗位" ], "tags_en": [ "national exam", "competition ratio", "middle band", "ministry posts" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "国考 50-100:1 中等竞争带岗位清单", "label_en": "National-exam posts in the 50-100:1 band" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我找国考里报录比在 50:1 到 100:1 之间的岗位。要求：1) 真实筛选并给出总数；2) 取 30 条，按报录比升序列出岗位、招录单位、地区、招录人数、报录比、往年进面分、薪资区间；3) 统计这 30 条涉及多少家招录单位、最集中的几家分别是谁，并分析这类单位为什么落在中等竞争带；4) 统计这 30 条中有多少条带往年进面分，说明报录比与进面分的数据覆盖是否一致；5) 提醒中等竞争带岗位「不太热」往往是因为有专业或基层经历限制，必须回公告核对。", "prompt_en": "Find national-exam posts with a competition ratio between 50:1 and 100:1. Requirements: 1) run the filter and give the total; 2) take 30, sorted by ascending ratio, with post, recruiting unit, region, headcount, ratio, historical cutoff and salary band; 3) report how many distinct units appear, which dominate, and why such units sit in the middle band; 4) count how many of the 30 carry a cutoff score and state whether ratio and cutoff data cover the same posts; 5) warn that middle-band posts are often less contested because of major or grassroots-experience restrictions that must be verified against the announcement.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 141, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-ratio-hot-cold", "title": "竞争比筛选：国考冷热岗位对照榜", "title_en": "Ratio Filter: The Hottest and Coldest National-Exam Posts", "subtitle": "最热 5117:1、最冷 1:1，同一年同一场考试", "subtitle_en": "5117:1 at one end, 1:1 at the other — same exam, same year", "description": "「国考很卷」是句正确的废话——真正有用的是知道卷在哪、冷在哪，以及两端各自的坑。\n\n本案例做两次极值查询：1000:1 以上的岗位有 132 个，10:1 以下的有 2,017 个。两张榜并排看，能直接读出热榜岗位 15 条里 12 条只招 1 人，冷榜岗位则集中在语言类翻译和公安特殊岗位——冷不等于容易。\n\n适合：第一次报考的考生、想理解竞争分布形状的人。", "description_en": "\"The national exam is brutal\" is true but useless — what helps is knowing where the crowding is, where it isn't, and the trap at each extreme.\n\nThis case runs two extreme queries: 132 posts above 1000:1, and 2,017 below 10:1. Side by side, the hot list turns out to be almost entirely single-opening posts with no major restriction, while the cold list clusters in language translation and specialist police roles — cold does not mean easy.\n\nFor: first-time candidates and anyone wanting to understand the shape of the competition curve.", "categories": [ "search" ], "tags": [ "国考", "竞争比", "冷热对照", "极值分析" ], "tags_en": [ "national exam", "competition ratio", "hot vs cold", "extremes" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "国考冷热岗位对照榜", "label_en": "Hottest and coldest national-exam posts" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次国考岗位的冷热对照。要求：1) 分两次真实查询——一次筛报录比 1000:1 以上的岗位，一次筛 10:1 以下的，各给出总数；2) 两榜各列 8 条，包含岗位、招录单位、地区、招录人数、报录比、薪资区间（冷榜若有进面分也一并列出）；3) 分别总结热榜岗位和冷榜岗位各自的共同特征（招录人数、专业限制等），用数据说话；4) 指出冷热两端各自的陷阱，说明为什么报录比低不等于容易上岸；5) 给出一个更实际的竞争比筛选区间建议。", "prompt_en": "Run a hot-versus-cold comparison of national-exam posts. Requirements: 1) two real queries — one for ratios above 1000:1, one for below 10:1 — with totals for each; 2) list 8 from each with post, unit, region, headcount, ratio and salary band (plus cutoff score for the cold list where available); 3) summarise what the posts on each list have in common (headcount, major restrictions) using the data; 4) identify the trap at each extreme and explain why a low ratio doesn't mean an easy admission; 5) suggest a more practical ratio range to filter on.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 142, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "microservice-topology-diagram", "title": "微服务架构拓扑图", "title_en": "Microservice Architecture Topology Diagram", "subtitle": "清晰展示服务间依赖、数据流向与故障隔离边界", "subtitle_en": "Clearly map service dependencies, data flows, and fault isolation boundaries", "description": "微服务拆得越多，系统全貌就越难看清——哪个服务依赖哪个？数据从哪里流到哪里？一个服务挂了会影响多大范围？新人入职第一天就被这些问题搞懵了。\n\n这个案例用 Excalidraw 画了一张典型电商系统的微服务拓扑图，分为网关层、业务服务层、基础设施层三个区域，标注了真实的 API 路径、消息队列主题、数据库连接，并用颜色区分同步调用和异步事件。\n\n适合：后端架构师、微服务团队负责人、需要做系统设计文档的开发者、准备技术方案评审的工程师", "description_en": "The more microservices you split, the harder it is to see the full picture — which service depends on what? Where does data flow? How far does a single service failure cascade? New hires get confused on day one.\n\nThis case uses Excalidraw to create a microservice topology diagram for a typical e-commerce system, divided into Gateway, Business Services, and Infrastructure layers, annotated with real API paths, message queue topics, and database connections, using colors to distinguish sync calls from async events.\n\nIdeal for: Backend architects, microservice team leads, developers writing system design docs, and engineers preparing technical reviews", "categories": [ "dev", "creation" ], "tags": [ "微服务", "架构图", "系统设计", "拓扑图", "后端" ], "tags_en": [ "Microservices", "Architecture Diagram", "System Design", "Topology", "Backend" ], "artifact_type": "image", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.png", "prompt": "请用 Excalidraw 画一张微服务架构拓扑图，以典型电商系统为例。\n\n要求：\n1. 上下分层布局，分三个区域（用浅色背景框出）：网关层、业务服务层（用户、订单、商品、支付、通知等服务）、基础设施层（PostgreSQL、Redis、RabbitMQ、Elasticsearch）\n2. 实线箭头表示同步 HTTP/gRPC 调用，虚线箭头表示异步消息\n3. 在关键箭头旁标注真实 API 路径与消息队列 Topic\n4. 每个服务节点标注技术栈\n5. 加一个深色背景的证据块，展示一段网关路由配置片段\n6. 用颜色语义区分网关、业务服务、数据库、消息队列几类节点\n7. 顶部加标题 \"E-Commerce Microservice Topology\"", "prompt_en": "Please use Excalidraw to draw a microservice architecture topology diagram, using a typical e-commerce system as an example.\n\nRequirements:\n1. Top-to-bottom layered layout, three zones (outlined with light backgrounds): Gateway, Business Services (user, order, product, payment, notification services), Infrastructure (PostgreSQL, Redis, RabbitMQ, Elasticsearch)\n2. Solid arrows for synchronous HTTP/gRPC calls, dashed arrows for asynchronous messages\n3. Annotate real API paths and message-queue topics next to key arrows\n4. Label the tech stack inside each service node\n5. Add a dark-background evidence block showing a gateway routing config snippet\n6. Use color semantics to distinguish gateway, business services, databases, and message queues\n7. Add the title \"E-Commerce Microservice Topology\" at the top", "skills": [ { "id": "excalidraw-diagram", "name": "Excalidraw 绘图", "description": "手绘风格架构图与流程图生成", "skill_id": "skill_2053082275194277888" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 142, "is_featured": false, "featured_order": 0, "quality_score": 85, "source": "official", "usage_count": 0, "created_at": "2026-04-15T08:23:23Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.png", "label": "微服务拓扑架构图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "user-journey-flowchart", "title": "用户旅程流程图", "title_en": "User Journey Flowchart", "subtitle": "电商购物全链路可视化，一键生成专业用户旅程图", "subtitle_en": "Visualize the full e-commerce shopping journey, generate professional user journey maps in one click", "description": "分析用户旅程时还在用 PPT 手动画流程图？节点对不齐、样式不统一、改一个阶段要拖半天——效率低下且难以复用。\n\n这个案例用 Mermaid Journey 图完整呈现了电商购物的全链路用户旅程：从「发现商品」到「复购推荐」共 9 个核心阶段，每个阶段标注满意度评分（1-5 分）和关键触点行为。可视化效果清晰专业，支持快速迭代修改，只需编辑文本即可调整流程节点。\n\n适合：产品经理梳理用户体验地图、UX 设计师做旅程分析、运营团队复盘转化漏斗", "description_en": "Still manually drawing user journey flowcharts in PowerPoint? Misaligned nodes, inconsistent styles, and hours spent dragging one stage around — inefficient and hard to reuse.\n\nThis case uses a Mermaid Journey diagram to present the complete e-commerce shopping user journey: from 'Discover Products' to 'Repurchase Recommendation' across 9 core stages, each annotated with satisfaction scores (1-5) and key touchpoint actions. The visualization is clean and professional, supports rapid iteration, and can be modified simply by editing text.\n\nIdeal for: Product managers mapping user experience journeys, UX designers conducting journey analysis, operations teams reviewing conversion funnels", "categories": [ "business", "creation" ], "tags": [ "用户旅程", "流程图", "电商", "产品设计", "UX" ], "tags_en": [ "User Journey", "Flowchart", "E-commerce", "Product Design", "UX" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个电商用户旅程图页面（单文件 HTML，用内联 SVG 绘制 journey 图，无外部依赖），含标题区、专业清爽排版。\n\n9 个阶段（含满意度评分 1-5 和关键触点）：\n1. 发现商品（3）：搜索引擎、社媒种草、好友推荐\n2. 浏览详情（4）：看图、读描述、看评价\n3. 加入购物车（4）：选规格、对比、收藏\n4. 下单结算（3）：确认地址、选配送、用券\n5. 在线支付（4）：选支付方式、验证、确认\n6. 物流配送（3）：看轨迹、预约时间、签收\n7. 确认收货（4）：检查质量、确认、售后\n8. 评价晒单（3）：写评价、传图、分享\n9. 复购推荐（5）：个性化推荐、领券、再下单", "prompt_en": "Generate an e-commerce user journey diagram page (single-file HTML, drawing the journey diagram with inline SVG, no external dependencies), including a title area and professional, clean typography.\n\n9 stages (with satisfaction scores 1-5 and key touchpoints):\n1. Discover Products (3): search engines, social media seeding, friend recommendations\n2. Browse Details (4): view images, read descriptions, check reviews\n3. Add to Cart (4): select specs, compare, save to favorites\n4. Checkout (3): confirm address, choose delivery, apply coupons\n5. Online Payment (4): choose payment method, verify, confirm\n6. Logistics & Delivery (3): track shipment, schedule time, sign for delivery\n7. Confirm Receipt (4): inspect quality, confirm, after-sales\n8. Review & Share (3): write review, upload images, share\n9. Repurchase Recommendation (5): personalized recommendations, claim coupons, reorder", "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表生成器", "description": "从文本描述生成 Mermaid 流程图、架构图、时序图等", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 143, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-22T06:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "用户旅程流程图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "research-llm-agent-survey-2026", "title": "LLM Agent 研究综述 2023-2026", "title_en": "LLM Agent Research Survey 2023-2026", "subtitle": "60 篇顶会论文 × 5 大方向 × 标准 BibTeX，一份能直接投稿 Related Work 的学术综述", "subtitle_en": "60 top-venue papers × 5 directions × standard BibTeX — a Related-Work-ready academic survey", "description": "做 Agent 方向的研究最难的不是写实验，而是把三年来 ReAct、Reflexion、MemGPT、AutoGen、Toolformer、AgentBench、WebArena、SWE-bench…… 几十篇论文按脉络串起来——既要讲清谁继承了谁、谁反对了谁，又要给出准确的 arXiv ID 和作者署名，任谁都头疼。\n\n这个案例演示了深度研究团队如何 6 分钟内产出一份《2023-2026 大模型 Agent 研究进展综述》：Chief Editor 统筹，按 5 大方向（经典范式 / 记忆机制 / 多智能体 / 工具使用 / 评测基准）+ 未来预判共 6 章分头调研，每章 5-11 篇代表性工作（含作者、arXiv ID、发表场所、核心贡献），章末附 ASCII 研究脉络图（继承/反对/改进关系），最后交付 60 条标准 BibTeX 清单，可直接粘到 LaTeX 模板里。\n\n适合：AI 博士生开题 / 投稿前做 Related Work / Foundation Model 研究团队内部培训 / Applied Scientist 跟进前沿 / 技术 VP 了解 Agent 技术图谱。", "description_en": "For Agent researchers, the hardest part isn't running experiments — it's stitching together three years of ReAct, Reflexion, MemGPT, AutoGen, Toolformer, AgentBench, WebArena, SWE-bench into a coherent storyline with accurate arXiv IDs and author lists.\n\nThis case shows how the Deep Research Team produced a full 2023-2026 LLM Agent Survey in six minutes: 6 chapters (classic paradigms / memory / multi-agent / tool use / evaluation / future forecast), 5-11 representative works per chapter (with authors, arXiv IDs, venues, and core contributions), ASCII lineage diagrams at the end of each chapter (who inherited / opposed / improved whom), plus 60 standard BibTeX entries ready to drop into LaTeX.\n\nFit for: AI PhD proposals, pre-submission Related Work drafting, foundation-model team onboarding, Applied Scientists tracking frontiers, and tech VPs building an Agent tech map.", "categories": [ "education", "data", "search" ], "tags": [ "学术综述", "LLM Agent", "研究前沿", "BibTeX", "Survey" ], "tags_en": [ "Academic Survey", "LLM Agent", "Research Frontier", "BibTeX", "Survey" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "survey.md", "artifacts": [ { "file": "survey.md", "label": "综述正文（Markdown，60 篇论文）" }, { "file": "references.bib", "label": "BibTeX 参考文献清单（60 条）" } ], "prompt": "请用【深度研究团队 5 阶段完整模式】，面向 AI 研究者产出一份《2023-2026 大模型 Agent 研究进展综述》。\n\n覆盖 5 大方向，每方向一章、至少 5 篇代表性工作：\n1. 经典范式演进\n2. 记忆机制\n3. 多智能体系统\n4. 工具使用与函数调用\n5. 评测基准\n\n每篇论文标注作者、发表时间、发表场所、arXiv ID、核心贡献，并用 Markdown 链接到原文。\n\n报告结构：目录 / 引言 / 五大方向 / 未来方向预判 / 结论 / 参考文献；每章结尾附 ASCII 研究脉络图。论文元数据务必准确，目标 50-60 篇。\n\n额外交付 BibTeX 清单（≥50 条标准 entries），可直接粘贴到 LaTeX。", "prompt_en": "Please use the [Deep Research Team's full 5-phase mode] to produce a \"2023-2026 Large Language Model Agent Research Progress Survey\" aimed at AI researchers.\n\nCover 5 major directions, one chapter each with at least 5 representative works:\n1. Evolution of classic paradigms\n2. Memory mechanisms\n3. Multi-agent systems\n4. Tool use and function calling\n5. Evaluation benchmarks\n\nFor each paper, annotate the authors, publication date, venue, arXiv ID, and core contribution, with a Markdown link to the source.\n\nReport structure: Table of Contents / Introduction / the five directions / forecast of future directions / Conclusion / References; append an ASCII research-lineage diagram at the end of each chapter. Paper metadata must be accurate; target 50-60 papers.\n\nAdditional deliverable: a BibTeX list (≥50 standard entries) ready to paste into LaTeX.", "skills": [], "experts": [ { "id": "GPTResearcherTeam", "name": "深度研究团队", "description": "深度研究报告输出，7角色5阶段聚合多源信息，经审稿修订循环输出带引用的专业报告", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "a8da6f2dc1f7", "name_en": "Deep Research Team", "description_en": "Deep research report generation: 7 roles in 5 phases aggregate multi-source data through review-revise loop into cited professional reports", "expert_id": "ex_xdj3MuRUqbR0" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 144, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T02:58:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "crm-q1-quarterly-report", "title": "CRM SaaS 2026 Q1 季度汇报", "description": "中型企业 CRM SaaS 2026 Q1 利益相关者季度汇报：核心指标进展（NRR 103%→106%）、路线图交付（6/8达成）、AI MVP数据、重要成果与教训、Q2优先级与风险、资源需求、董事会决议请求", "experts": [ { "id": "ProductStrategyTeam", "expertType": "team", "categoryId": "01-ProductDesign", "plugin": "product-strategy-team", "displayName": { "zh": "产品战略团队" }, "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师（PRD/功能规格书）、用户研究员（调研综合分析）、竞品分析师（竞争情报）、数据分析师（指标追踪）和路线图规划师（路线图管理/迭代规划）。覆盖从构思到上线的完整产品生命周期。", "name_en": "Product Strategy Team", "description_en": "A team of 5 product specialists led by a Product Director: Requirement Analyst (PRD/feature spec writing), User Researcher (research synthesis), Competitive Analyst (competitive intelligence), Data Analyst (metrics tracking), and Roadmap Planner (roadmap management/sprint planning). Covers the full product lifecycle from ideation to launch.", "expert_id": "ex_LqCvsVsLOp3J" } ], "categories": [ "creation", "business" ], "tags": [ "季度汇报", "CRM", "SaaS", "利益相关者", "董事会", "指标仪表盘", "路线图进展", "资源需求" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化" }, { "file": "output.md", "label": "Markdown 文档" } ], "createdAt": "2026-04-26", "input": { "product": "中型企业 CRM SaaS", "period": "2026 Q1", "audience": "CEO / 高管团队 / 董事会", "northStarProgress": "NRR 103%→106%" }, "title_en": "CRM SaaS 2026 Q1 Quarterly Review", "subtitle": "NRR 103%→106%：Q1 核心指标、AI MVP 与董事会决议", "subtitle_en": "NRR 103%→106%: Q1 metrics, AI MVP results, and board decisions", "description_en": "Mid-market CRM SaaS 2026 Q1 stakeholder quarterly review: core metric progress (NRR 103%→106%), roadmap delivery (6/8 achieved), AI MVP data, key wins and lessons, Q2 priorities and risks, resource needs, and board decision requests.", "tags_en": [ "Quarterly Review", "CRM", "SaaS", "Stakeholders", "Board", "Metric Dashboard", "Roadmap Progress", "Resource Needs" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "sort_order": 145, "featured_order": 0, "usage_count": 0, "quality_score": 90, "is_featured": false, "created_at": "2026-04-27T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "skills": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "我是一家中型企业 CRM SaaS 的产品负责人，请帮我撰写 2026 Q1 利益相关者（CEO / 高管团队 / 董事会）季度汇报，要求内容包括：\n1) 核心指标进展（北极星 NRR 103%→106%、MRR / ARR / NDR / Churn / CAC Payback 等）；\n2) Q1 路线图交付情况（按照既定 OKR，达成 6/8、延期 2 项，说明原因与影响）；\n3) AI MVP 关键数据（采用率、留存、用户反馈）；\n4) 重要成果与教训 3~5 条；\n5) Q2 优先级与关键风险；\n6) 资源需求（人力 / 预算 / 外部合作）；\n7) 给董事会的决议请求清单。\n最终输出 Markdown 季度汇报 + HTML 可视化页（核心指标仪表盘 + 路线图状态墙）。", "prompt_en": "I'm the product lead for a mid-market CRM SaaS company. Please help me write the 2026 Q1 stakeholder (CEO / executive team / board) quarterly review. The content should include:\n1) Progress on core metrics (North Star NRR 103%→106%, plus MRR / ARR / NDR / Churn / CAC Payback, etc.);\n2) Q1 roadmap delivery status (against the agreed OKRs, 6/8 achieved, 2 delayed, with reasons and impact explained);\n3) Key AI MVP data (adoption rate, retention, user feedback);\n4) 3-5 major wins and lessons learned;\n5) Q2 priorities and key risks;\n6) Resource needs (headcount / budget / external partnerships);\n7) A list of resolutions requested from the board.\nFinal output: a Markdown quarterly review plus an HTML visualization page (a core-metrics dashboard + a roadmap status wall).", "is_recommended": false, "recommended_order": 0 }, { "id": "pet-health-record", "title": "宠物健康档案卡", "title_en": "Pet Health Record Card", "subtitle": "一页掌握毛孩子的健康全貌", "subtitle_en": "Track your pet's complete health profile in one page", "description": "养宠家庭常常面临疫苗记录散落、体检数据难以追溯、护理计划缺乏系统性等问题，一旦宠物生病或需要转诊，往往找不到完整的健康资料。\n\n本案例演示如何用一个精美的 HTML 页面，将宠物的基本信息、疫苗接种时间线、历次体检记录、体重趋势图、日常护理清单和紧急联系方式整合在一起，暖色调圆角卡片设计，打开即用，随时可打印或分享给兽医。\n\n适合所有养宠人士、宠物店主、宠物医院前台，以及希望用可视化方式管理宠物健康数据的开发者参考。", "description_en": "Pet owners often struggle with scattered vaccine records, hard-to-trace checkup data, and unsystematic care plans. When a pet falls ill or needs a referral, complete health records are nowhere to be found.\n\nThis case demonstrates how to consolidate a pet's basic info, vaccination timeline, checkup history, weight trend chart, daily care checklist, and emergency contacts into one beautifully designed HTML page with warm colors and rounded cards — ready to use, print, or share with your vet.\n\nIdeal for pet owners, pet shop managers, veterinary clinic staff, and developers looking for inspiration on visualizing pet health data.", "categories": [ "creation" ], "tags": [ "宠物", "健康管理", "数据可视化", "响应式页面" ], "tags_en": [ "Pet", "Health Management", "Data Visualization", "Responsive Page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个宠物健康档案页面（单文件 HTML，内联无外部依赖），暖色调、圆角卡片、响应式。\n\n宠物：金毛犬「豆豆」，3 岁，公，28kg。\n包含：1) 基本信息卡片（品种/年龄/体重/绝育/芯片号）；2) 疫苗接种时间线（狂犬/六联/驱虫，含下次提醒）；3) 体检记录表格（2024-2026 共 4 次）；4) 体重趋势 SVG 折线图；5) 日常护理清单（喂食/运动/洗澡/美容频率）；6) 紧急联系信息（医院/24h 急诊）。", "prompt_en": "Please generate a pet health-record page (single-file HTML, inlined with no external dependencies), with a warm color scheme, rounded-corner cards, and responsive design.\n\nPet: a Golden Retriever named \"Doudou\", 3 years old, male, 28kg.\nInclude: 1) a basic-info card (breed / age / weight / neuter status / chip number); 2) a vaccination timeline (rabies / 6-in-1 / deworming, with next-due reminders); 3) a checkup-records table (4 visits total, 2024-2026); 4) an SVG line chart of weight trends; 5) a daily-care checklist (feeding / exercise / bathing / grooming frequency); 6) emergency contact info (hospital / 24h emergency).", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "Full-stack frontend development with premium UI", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 90, "sort_order": 146, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T21:40:00Z", "artifacts": [ { "file": "output.html", "label": "宠物健康档案页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "saas-market-research", "title": "SaaS 市场调研报告", "subtitle": "从市场规模到竞争格局的完整调研", "description": "做市场调研最头疼的是信息散落在各处——行业报告、新闻、财报、用户评价，光收集就要一周，整理成报告又要一周。\n\n这个案例演示了如何用 AI 快速生成一份专业的 SaaS 市场调研报告：包含 TAM/SAM/SOM 市场规模测算、竞品矩阵分析、用户画像、定价策略建议和市场进入策略，附带数据图表和可视化。\n\n适合：产品经理、创业者、投资分析师、战略规划负责人。", "categories": [ "data", "business" ], "tags": [ "市场调研", "SaaS", "商业分析" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "企业级 AI 编程助手市场调研报告" } ], "prompt": "请帮我做一份「企业级 AI 编程助手」赛道的市场调研报告。\n\n调研范围：\n- 产品定义：面向企业开发团队的 AI 编程辅助工具（代码补全、代码审查、自动测试等）\n- 目标市场：中国市场为主，兼顾全球视角\n- 时间范围：2024-2029 年\n\n报告要求：\n1. 市场规模：用 TAM/SAM/SOM 框架测算，给出具体数字和增长率\n2. 竞争格局：分析 5-8 个主要竞品（GitHub Copilot、Cursor、通义灵码等），制作竞品对比矩阵\n3. 用户洞察：目标用户画像、核心需求、购买决策因素\n4. 定价分析：主流产品定价策略对比，给出定价建议\n5. 市场进入策略：差异化定位建议、GTM 路径\n6. 风险与机会：技术风险、政策风险、市场机会\n\n输出格式：结构化 Markdown 文档，关键数据用表格呈现。", "prompt_en": "Please help me prepare a market research report for the \"enterprise-grade AI coding assistant\" space.\n\nResearch scope:\n- Product definition: AI coding assistance tools aimed at enterprise development teams (code completion, code review, automated testing, etc.)\n- Target market: primarily the Chinese market, while also considering a global perspective\n- Time range: 2024-2029\n\nReport requirements:\n1. Market size: estimate using the TAM/SAM/SOM framework, giving specific figures and growth rates\n2. Competitive landscape: analyze 5-8 major competitors (GitHub Copilot, Cursor, Tongyi Lingma, etc.) and create a competitor comparison matrix\n3. User insights: target user personas, core needs, purchase decision factors\n4. Pricing analysis: compare the pricing strategies of mainstream products and give pricing recommendations\n5. Market entry strategy: differentiated positioning recommendations, GTM path\n6. Risks and opportunities: technical risks, policy risks, market opportunities\n\nOutput format: a structured Markdown document, with key data presented in tables.", "skills": [ { "id": "market-researcher", "name": "市场调研", "description": "市场调研专家，提供市场分析、消费者洞察与机会评估", "description_en": "Market research specialist for analysis, consumer insights, and sizing", "skill_id": "skill_2053082339107885056" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 147, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-11T15:09:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "title_en": "SaaS Market Research Report", "subtitle_en": "Complete research from market sizing to competitive landscape", "description_en": "The hardest part of market research is that information is scattered everywhere — industry reports, news, earnings calls, user reviews. Just collecting takes a week, and organizing into a report takes another.\n\nThis case shows how AI rapidly generates a professional SaaS market research report: TAM/SAM/SOM market sizing, competitor matrix analysis, user personas, pricing strategy recommendations, and go-to-market strategy — complete with data charts and visualizations.\n\nBest for: Product managers, strategy analysts, founders, investors.", "tags_en": [ "Market Research", "SaaS", "Business Analysis" ], "is_recommended": false, "recommended_order": 0 }, { "id": "crm-saas-annual-roadmap", "title": "CRM SaaS 年度路线图规划", "description": "中型企业 CRM SaaS（500家付费客户）2026年度路线图：含产品指标评审、用户反馈聚类、竞品动态分析（Salesforce/HubSpot/Zoho/飞书）、Q1-Q4 季度路线图、RICE 评分、资源分配、利益相关者沟通文档和风险管理", "experts": [ { "id": "ProductStrategyTeam", "expertType": "team", "categoryId": "01-ProductDesign", "plugin": "product-strategy-team", "displayName": { "zh": "产品战略团队" }, "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师（PRD/功能规格书）、用户研究员（调研综合分析）、竞品分析师（竞争情报）、数据分析师（指标追踪）和路线图规划师（路线图管理/迭代规划）。覆盖从构思到上线的完整产品生命周期。", "name_en": "Product Strategy Team", "description_en": "A team of 5 product specialists led by a Product Director: Requirement Analyst (PRD/feature spec writing), User Researcher (research synthesis), Competitive Analyst (competitive intelligence), Data Analyst (metrics tracking), and Roadmap Planner (roadmap management/sprint planning). Covers the full product lifecycle from ideation to launch.", "expert_id": "ex_LqCvsVsLOp3J" } ], "categories": [ "creation", "business" ], "tags": [ "路线图", "CRM", "SaaS", "年度规划", "竞品分析", "产品指标", "AI战略", "RICE" ], "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "HTML 可视化" }, { "file": "output.md", "label": "Markdown 文档" } ], "createdAt": "2026-04-26", "input": { "product": "中型企业 CRM SaaS", "customers": "500家付费客户", "period": "2026年度", "northStar": "NRR 103% → 115%", "competitors": [ "Salesforce", "HubSpot", "Zoho", "飞书" ] }, "title_en": "CRM SaaS Annual Roadmap Planning", "subtitle": "500 家客户 CRM SaaS 2026 年度 RICE 路线图与四大竞品策略", "subtitle_en": "2026 RICE roadmap for a 500-customer CRM SaaS with four-competitor strategy", "description_en": "Mid-market CRM SaaS (500 paying customers) 2026 annual roadmap: product metrics review, user feedback clustering, competitive analysis (Salesforce/HubSpot/Zoho/Feishu), Q1–Q4 quarterly roadmap, RICE scoring, resource allocation, stakeholder communication documents, and risk management.", "tags_en": [ "Roadmap", "CRM", "SaaS", "Annual Planning", "Competitive Analysis", "Product Metrics", "AI Strategy", "RICE" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "source": "official", "sort_order": 148, "featured_order": 0, "usage_count": 0, "quality_score": 90, "is_featured": false, "created_at": "2026-04-27T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "skills": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "我是一家中型企业 CRM SaaS 的产品负责人，公司有 500 家付费客户，请帮我制定 2026 年度产品路线图，覆盖：产品指标评审（北极星 NRR 由 103% 提升到 115% 及关键漏斗指标）、用户反馈聚类与高频痛点、竞品动态分析（Salesforce / HubSpot / Zoho / 飞书）、Q1~Q4 季度路线图、用 RICE 排序候选 Epic、资源分配、利益相关者沟通文档、风险管理。\n\n最终输出 Markdown 年度路线图 + HTML 可视化路线图页。", "prompt_en": "I'm the product lead of a mid-market CRM SaaS with 500 paying customers; help me build the 2026 product roadmap, covering: product-metric review (North Star NRR from 103% to 115%, plus key funnel metrics), user-feedback clustering & top pain points, competitive analysis (Salesforce / HubSpot / Zoho / Feishu), a Q1–Q4 quarterly roadmap, RICE prioritization of candidate Epics, resource allocation, stakeholder communication docs, and risk management.\n\nFinal output: a Markdown annual roadmap + an HTML visual roadmap page.", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-battle-abm-outreach-sequence", "title": "ABM 外联序列方案", "title_en": "ABM Outreach Sequence Plan", "subtitle": "按行业规模角色分层的 500 家目标企业外联打法", "subtitle_en": "Layered outreach playbook for 500 target accounts by industry, size, and role", "description": "面向 500 家目标企业做 ABM，最怕的不是名单不够，而是所有账户都用同一套话术、同一个节奏、同一种价值表达。这个案例演示了销售作战团队如何为 AI 财税 SaaS 设计一套可落地的 ABM 外联体系：先按行业、规模、角色拆成 3 档，再为每档设计 6 步邮件+LinkedIn 混合序列，并写出完整成稿、差异化定位、跟进规则和效果指标。最终输出一份可直接交给销售或增长团队执行的 Markdown 方案。\n\n适合：销售总监、增长负责人、SDR 团队、市场负责人、ABM 运营", "description_en": "When running ABM against 500 target companies, the biggest risk is not the list size—it's using the same message, cadence, and value story for every account. This case shows how the Sales Battle Team designs a practical ABM outreach system for an AI finance & tax SaaS product: segment accounts into 3 tiers by industry, company size, and buyer role; build a 6-step blended email + LinkedIn sequence for each tier; and write the full copy, positioning logic, follow-up rules, and success metrics. The final output is a Markdown playbook ready for sales or growth teams to execute.\n\nBest for: Sales Directors, Growth Leaders, SDR Teams, Marketing Leads, ABM Operators", "categories": [ "business" ], "tags": [ "ABM", "外联序列", "销售邮件", "LinkedIn", "AI财税SaaS" ], "tags_en": [ "ABM", "outreach sequence", "sales email", "LinkedIn", "AI finance SaaS" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "ABM 外联序列完整方案" } ], "prompt": "新产品面向 500 家目标企业（已有名单），需要邮件+LinkedIn 组合 ABM 外联。请：\n1. 分层画像（按行业、规模、角色 3 档）\n2. 每层设计 6 步外联序列（邮件+LinkedIn 消息混合）\n3. 每封文案差异化定位，说明相比竞品的独特价值\n4. 跟进节奏与放弃规则\n5. 效果衡量指标（打开率、回复率、会议率）\n\n输出《ABM 外联序列完整方案》Markdown，附所有邮件和 LinkedIn 消息成稿。", "prompt_en": "A new product is targeting 500 named companies (the list already exists), and we need a combined email + LinkedIn ABM outreach. Please:\n1. Create tiered personas (3 tiers by industry, size, and role)\n2. Design a 6-step outreach sequence for each tier (a mix of email + LinkedIn messages)\n3. Differentiate the messaging for each piece, explaining the unique value versus competitors\n4. Define the follow-up cadence and the rules for giving up\n5. Performance metrics (open rate, reply rate, meeting rate)\n\nOutput a \"Complete ABM Outreach Sequence Plan\" in Markdown, with all emails and LinkedIn messages written out in full.", "skills": [], "experts": [ { "id": "SalesBattleTeam", "name": "销售作战团队", "description": "销售攻坚体系，客户研究锁定目标、外联策略提升触达、竞品情报预警风险、销售预测优化资源分配", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "ff06833f0e30", "name_en": "Sales Battle Team", "description_en": "Sales effectiveness system: account research identifies targets, outreach strategy improves reach, competitive intel warns risks, pipeline forecasting optimizes resource allocation", "expert_id": "ex_ATh4MnfFCKxR" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 149, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:18:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cultural-product-3d-enamel", "title": "文创品建模 · 金属珐琅材质", "title_en": "Cultural Product 3D · Cloisonné Enamel", "subtitle": "把文创品变成金属珐琅质感的可交互 3D 模型", "subtitle_en": "Turn a creative product into a metal-enamel interactive 3D model", "description": "金属珐琅工艺华丽但打样成本高，方案阶段难以呈现质感与光泽。\n\n案例以彩绘漆木鸠为主体，保留真实轮廓、比例与纹样，转译为金属胎珐琅质感的可交互 3D 模型：釉彩光泽与金属描边清晰，网页可自由旋转查看，单文件离线可开。\n\n适合：文创衍生开发、礼品/工艺品设计师、博物馆文创团队、IP 授权方", "description_en": "Using the painted lacquered wooden dove as subject, this case keeps its true silhouette, proportion and motif, rendering a metal-bodied cloisonné-enamel interactive 3D model with glossy glaze and crisp metal outlines — rotate freely in the browser, single-file and offline-ready.\n\nGood for: Merch developers, Gift & craft designers, Museum creative teams, IP licensors", "categories": [ "creation" ], "tags": [ "3D建模", "金属珐琅", "可交互模型", "文创衍生" ], "tags_en": [ "3D Modeling", "Metal Enamel", "Interactive Model", "Cultural Merch" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据这张图文创品图片，以图中彩绘漆木鸠为唯一主体，完整保留其真实轮廓、比例、造型结构、材质厚薄、主色、纹样、做旧或磨损痕迹与关键结构锚点，不重新设计、不改变品类身份、不替换为其他物件或真实动物；仅将其重塑为一件金属景泰蓝珐琅质感的立体收藏摆件（3D 手办／公仔）。\n构图：画面中仅一个主体，居中，采用四分之三前侧视角（3/4 front-side view），完整展现整体造型（若原物含底座／支撑则一并保留）；纯净白色背景，柔和均匀的棚拍布光，产品级静物摄影；主体完整不裁切，无多余道具、无文字水印、无多视角拼图、无第二主体。\n材质表现：玻璃质釉面带光泽，各色釉块之间以细凸起金丝（掐丝）勾边，喙／眼圈／足／底边为抛光鎏金铜，色彩宝石感、饱满干净；忠实保留实拍产品的主色与材质色，仅做轻微暖化与饱和度收敛，不改变识别色。\n画质：清晰锐利、细节丰富、主体轮廓干净完整，适合直接作为 3D 建模重建的参考图；避免噪点、模糊、强烈阴影遮挡和背景杂物。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-3d-expert", "name": "文创品3D建模专家", "description": "将文创品转译为可交互的 3D 模型效果，支持毛绒、硅胶、金属珐琅、亚克力等材质渲染", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_P14ik4i7Utla" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "可交互 3D 模型（单文件内联）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 150, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-22T15:31:07+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "digital-human-product-pitch", "title": "数字人商品口播", "title_en": "Digital Human Product Pitch", "subtitle": "用自然手机镜头完成可信的 15 秒护肤品介绍", "subtitle_en": "A credible 15-second skincare pitch shot like a natural phone video", "description": "数字人口播容易表情僵硬、产品拿反，台词节奏也常与手势和镜头不匹配。\n\n案例让一位休闲装女性在明亮客厅手持琥珀色精华瓶，用自拍式镜头说清质地、使用时机和个人体验，并把产品稳定展示给镜头。\n\n适合：美妆种草、电商口播、UGC 广告、社交短视频和产品卖点快速验证。", "description_en": "Digital presenters can look stiff, hold products backward, or fall out of sync with gestures and dialogue.\n\nA casually dressed woman presents an amber serum in a bright living room, explaining texture, usage timing, and personal experience in a natural selfie-style shot.\n\nBest for beauty recommendations, commerce pitches, UGC ads, social shorts, and rapid product-message testing.", "categories": [ "creation" ], "tags": [ "数字人口播", "护肤品", "UGC广告" ], "tags_en": [ "digital human pitch", "skincare", "UGC ad" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「数字人商品口播」Skill，根据我上传的琥珀色精华液产品图和已获授权的女性人物参考图，制作一支约 15 秒、16:9 的居家 UGC 口播。场景是白天自然光客厅，人物穿浅色休闲上衣，先看向镜头说“我最近早晚都在用这瓶精华”，再把瓶身标签朝向镜头，描述“质地轻、上脸不黏”，最后微笑补充“我会在面霜前用两滴”。请先确认规格、脚本和故事板，再生成并确认人物/产品元素图与关键帧；之后生成与脚本时长精确匹配的中文女声，再让关键帧和音频驱动人物视频，最后组装。必须保持产品标签方向、手指数、口型和台词同步，不添加医疗功效或未提供的成分宣称。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Digital Human Product Pitch” Skill and use my uploaded image of an amber serum bottle plus an authorized female character reference to create a 16:9 at-home UGC pitch of about 15 seconds. In a naturally lit living room during daytime, the casually dressed presenter looks into camera and says, “I have been using this serum every morning and night.” She turns the label toward camera, adds, “The texture is light and does not feel sticky,” and finishes with a smile: “I use two drops before moisturizer.” Confirm the specification, script, and storyboard first, then generate and approve the presenter and product element images and the keyframe. Next generate a Chinese female voice track precisely matched to the script duration, use the keyframe and audio to drive the presenter video, and assemble the result. Keep label orientation, finger count, lip sync, and line timing accurate. Do not add medical benefits or ingredient claims I did not provide.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 150, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-battle-account-battle-map", "title": "大客户攻坚完整作战地图", "title_en": "Complete Account Battle Map for Key Accounts", "subtitle": "3专家协同出完整客户作战地图，从调研到签约一步到位", "subtitle_en": "3 experts collaborate on a full account battle map from research to signing", "description": "打大客户最怕两眼一抹黑——不知道客户内部谁说了算、竞品已经渗透到哪了、第一次见面该说什么。这个案例演示了销售作战团队如何对一家头部新能源车企做完整攻坚作战地图：客户研究员挖掘公司背景/决策链/痛点，竞品情报师分析竞品渗透+出战斗卡，外联策略师设计拜访议程+异议话术+90天触点计划。最终输出 Markdown 作战地图 + HTML 投屏版。\n\n适合：大客户销售、KA经理、销售总监、渠道经理、销售运营", "description_en": "The hardest part of key account selling is going in blind—no idea who calls the shots, how deep competitors have penetrated, or what to say in the first meeting. This case shows how the Sales Battle Team creates a complete account battle map for a top NEV automaker: the Account Researcher digs into company background/decision chains/pain points, the Competitive Intel Analyst maps competitor penetration and builds battle cards, and the Outreach Strategist designs the visit agenda, objection handling scripts, and a 90-day touchpoint plan. Outputs a Markdown battle map + HTML projection version.\n\nBest for: Key Account Sales, KA Managers, Sales Directors, Channel Managers, Sales Operations", "categories": [ "business", "agent" ], "tags": [ "大客户销售", "作战地图", "竞品分析", "拜访策略", "新能源" ], "tags_en": [ "key account sales", "battle map", "competitive analysis", "visit strategy", "NEV" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "作战地图 HTML 投屏版" }, { "file": "output.md", "label": "作战地图 Markdown 完整版" } ], "prompt": "即将拜访【某头部新能源车企】（假设公司名\"星驰汽车\"），目标拿下他们年销管理软件年框订单（预算200万）。请出完整客户作战地图：\n1. 客户公司背景（业务、组织、最近动态、关键决策人）\n2. 关键联系人画像（姓名、职务、背景、关注点）\n3. 竞品在该客户内的渗透情况\n4. 我们的差异化卖点与 Proof Points\n5. 首次拜访议程、发现性问题清单、异议应对话术\n6. 后续触点计划\n\n输出 Markdown 完整作战地图 + HTML 版投屏给销售团队。", "prompt_en": "I'm about to visit [a leading new-energy automaker] (assume the company name is \"Xingchi Motors\"), with the goal of winning their annual framework order for sales management software (budget 2 million yuan). Please produce a complete account battle map:\n1. Customer company background (business, organization, recent developments, key decision-makers)\n2. Key contact profiles (name, position, background, concerns)\n3. Competitor penetration within this account\n4. Our differentiated selling points and Proof Points\n5. First-visit agenda, discovery question checklist, and objection-handling scripts\n6. Follow-up touchpoint plan\n\nOutput a complete Markdown battle map + an HTML version to screen-share with the sales team.", "skills": [], "experts": [ { "id": "SalesBattleTeam", "name": "销售作战团队", "description": "销售攻坚体系，客户研究锁定目标、外联策略提升触达、竞品情报预警风险、销售预测优化资源分配", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "ff06833f0e30", "name_en": "Sales Battle Team", "description_en": "Sales effectiveness system: account research identifies targets, outreach strategy improves reach, competitive intel warns risks, pipeline forecasting optimizes resource allocation", "expert_id": "ex_ATh4MnfFCKxR" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 150, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T12:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-salary-filter", "title": "薪资查询：广东高薪事业岗", "title_en": "Salary Filter: Guangdong's Better-Paid Posts", "subtitle": "按薪资下限圈定性价比 band，再看内容与竞争", "subtitle_en": "Set a salary floor first, then weigh duties and competition", "description": "编制内的薪资差异远比想象中大，同城同类岗能差出一倍。\n\n本案例按「广东 + 事业单位 + 月薪≥8000」实时筛出 3571 个岗位，取前 10 条列单位、地区、招录人数与来源公告，并如实说明：筛选条件在服务端生效，但接口不逐条返回具体薪资数字，实际薪酬以公告原文为准。\n\n适合：把待遇作为硬条件的考生、对比多地待遇水位的考生。", "description_en": "Pay inside the public sector varies far more than people assume — similar posts in one city can differ twofold.\n\nThis case live-filters 3,571 Guangdong public-institution positions with monthly pay ≥¥8,000, listing the top 10 with unit, region, headcount and source announcement — and honestly states that the filter applies server-side while per-post figures aren't returned; actual pay follows the announcement.\n\nFor: candidates who treat pay as a hard requirement and those comparing pay levels across regions.", "categories": [ "search" ], "tags": [ "薪资", "事业单位", "广东" ], "tags_en": [ "salary", "public institution", "Guangdong" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "广东月薪≥8000 岗位清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我在广东的事业单位岗位里筛月薪不低于 8000 的。要求：1) 实时查询并给出总数；2) 前 10 条列岗位、单位、地区、招录人数、状态、来源公告；3) 接口不逐条返回薪资数字的话要如实说明，让我以公告原文为准，不要编数字；4) 给出验证实际待遇的途径。", "prompt_en": "Filter Guangdong public-institution positions with monthly pay of at least ¥8,000. Requirements: 1) query live and give the total; 2) list the top 10 with position, unit, region, headcount, status and source announcement; 3) if the interface doesn't return per-post pay figures, say so honestly and point me to the announcements — never fabricate numbers; 4) suggest how to verify actual compensation.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "按薪资区间筛选岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 150, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "smart-medical-record-prescription", "title": "纯处方录入缺失追问与跳过兜底", "title_en": "Prescription-Only Entry with Smart Field Prompt and Skip", "subtitle": "头孢克洛 + 奥美拉唑，剂量 / 疗程缺失分级追问，支持跳过", "subtitle_en": "Cefaclor + omeprazole, tiered prompts for dose and duration, skip to ship", "description": "「开头孢克洛缓释片，一天两次，再开个奥美拉唑」——单次剂量、疗程、给药途径全缺，临床最怕的是把残缺处方直接录进去，出了问题说不清是谁的锅。\n\n这个案例演示了把口头处方交给 AI 临床助手，它只问处方必需的剂量 / 疗程 / 给药途径，不过度追问病史；追问分级进行、每批不超过 5 个；回复「跳过」或「直接生成」后基于已有信息输出，缺失项如实标注 [信息缺失]，不编造剂量或疗程数据。\n\n适合：门诊药房、诊所医师、电子处方录入员、关注处方完整性的药师与医务管理者。", "description_en": "\"Cefaclor sustained-release, twice a day, plus omeprazole.\" Single dose, duration, route — all missing. The worst thing a clinic can do is ship a partial prescription and then argue about whose fault it is when something goes wrong.\n\nThis case shows how to hand the verbal prescription to the AI clinical assistant. It only asks the prescription-essential fields — dose, duration, route — and won't over-ask about unrelated history. Prompts are tiered (no more than five per round), and replying \"skip\" or \"just generate\" makes it produce what it has while honestly tagging missing items as [信息缺失] — never invents a dose or duration.\n\nFor: outpatient pharmacies, clinic physicians, e-prescription entry staff, and pharmacists and medical affairs managers focused on prescription completeness.", "categories": [ "agent", "education" ], "tags": [ "智能病例", "处方录入", "缺失追问", "跳过机制", "AI临床助手", "腾讯健康" ], "tags_en": [ "Smart Medical Record", "Prescription Entry", "Missing Field Prompt", "Skip Mechanism", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我录一个处方，从口头描述里提取药品信息，缺失的必需字段要追问，支持跳过。\n\n口头处方：开头孢克洛缓释片，一天两次，再开个奥美拉唑。\n\n要求：\n1. 判别场景为「纯处方录入」；\n2. 提取药品名、频次等信息；\n3. 检测单次剂量、疗程、给药途径等必需字段缺失时，进行渐进式追问，每批不超过 5 个、不过度追问非处方相关信息；\n4. 我回复「跳过」或「直接生成」后，基于已有信息输出，缺失项标注 [信息缺失]，不编造剂量或疗程数据；\n5. 生成校验报告，如实反映缺失项和潜在问题；\n6. 不做用药咨询、不解释药物作用，末尾附安全声明。", "prompt_en": "Enter a prescription from my verbal description, extracting drug info and prompting for missing required fields, with skip support.\n\nVerbal prescription: cefaclor sustained-release, twice a day, plus omeprazole.\n\nRequirements:\n1. Identify the scenario as \"prescription-only entry\".\n2. Extract drug names and frequency.\n3. When required fields (single dose, duration, route) are missing, prompt progressively — no more than 5 per batch, and never over-ask about non-prescription info.\n4. After I reply \"skip\" or \"just generate\", produce output from what is available, tagging missing items as [信息缺失] — never invent dose or duration data.\n5. Generate a check report honestly reflecting missing items and potential issues.\n6. Give no medication counseling and do not explain drug effects; end with a safety statement.", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 150, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:57:00+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "artifacts": [ { "file": "output.html", "label": "纯处方录入缺失追问与跳过兜底" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "data-visualization-bindmap", "title": "数据可视化图表集", "title_en": "Data Visualization Chart Collection", "subtitle": "输入业务数据，生成纯 CSS/SVG 实现的精美图表页面", "subtitle_en": "Input business data, generate beautiful charts using pure CSS/SVG", "description": "做周报、写汇报 PPT、给老板看数据——总需要几个好看的图表。但 Excel 自带图表丑，专业工具学习成本高，每次都找设计师又不现实。\n\n这个案例演示了输入一组业务数据（月度营收、用户增长、渠道占比、产品评分），AI 自动生成一个纯 CSS/SVG 实现的数据可视化图表集 HTML 页面：居中布局（max-width 760px），包含柱状图、折线趋势图、环形图和雷达图四种图表，全部用 CSS/SVG 绘制，无需任何外部库。深色主题，数据标注清晰，适合截图放入汇报材料。\n\n适合：数据分析师、产品经理、运营人员、需要数据可视化的汇报场景。", "description_en": "Weekly reports, presentation slides, executive data reviews — you always need good-looking charts. But Excel charts look dated, professional tools have steep learning curves, and hiring designers for every chart isn't practical.\n\nThis case demonstrates inputting business data (monthly revenue, user growth, channel distribution, product ratings), with AI generating a pure CSS/SVG data visualization chart collection HTML page: centered layout (max-width 760px), featuring bar chart, line trend chart, donut chart, and radar chart — all drawn with CSS/SVG, zero external dependencies. Dark theme, clear data labels, perfect for screenshots in reports.\n\nBest for: Data analysts, product managers, operations staff, any reporting scenario needing data visualization.", "categories": [ "data", "creation" ], "tags": [ "数据可视化", "图表", "CSS图表", "SVG", "数据分析" ], "tags_en": [ "Data Visualization", "Charts", "CSS Charts", "SVG", "Data Analysis" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个数据可视化图表集页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，4 个图表网格排布、现代精美。\n\n图表 1 月度营收柱状图（万元）：1月285 2月210 3月342 4月398 5月456 6月523\n图表 2 用户增长折线（万人）：1月12.3 2月14.1 3月16.8 4月19.2 5月23.5 6月28.7\n图表 3 渠道占比环形图：自然搜索35% 付费广告28% 社交媒体22% 内容营销10% 其他5%\n图表 4 产品评分雷达图（满分10）：易用性8.5 性能9.2 稳定性8.8 功能丰富度7.6 客户支持8.1", "prompt_en": "Please generate a data-visualization chart collection page (single-file HTML, inlined with no external dependencies). UI text should be in Simplified Chinese, with 4 charts arranged in a grid, modern and polished.\n\nChart 1 - Monthly revenue bar chart (in ten-thousands of yuan): Jan 285, Feb 210, Mar 342, Apr 398, May 456, Jun 523\nChart 2 - User growth line chart (in ten-thousands of people): Jan 12.3, Feb 14.1, Mar 16.8, Apr 19.2, May 23.5, Jun 28.7\nChart 3 - Channel share donut chart: Organic Search 35%, Paid Ads 28%, Social Media 22%, Content Marketing 10%, Other 5%\nChart 4 - Product rating radar chart (out of 10): Ease of Use 8.5, Performance 9.2, Stability 8.8, Feature Richness 7.6, Customer Support 8.1", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 151, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "数据可视化图表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-salary-jiangsu-high", "title": "薪资筛选：江苏 9000 元以上岗位", "title_en": "Salary Filter: Jiangsu Posts Above 9,000 Yuan", "subtitle": "2,916 个达标岗位，本页 30 条全部来自同一份选调公告", "subtitle_en": "2,916 qualifying posts — all 30 shown come from one selective-recruitment notice", "description": "想知道体制内哪些岗位薪资高，但薪资区间散落在每个岗位详情里，没人做过横向汇总。\n\n本案例按薪资下限 9,000 元筛江苏公务员岗位，命中 2,916 个。本页 30 条呈现出一个值得注意的结构：全部来自同一份面向应届优秀毕业生的选调公告，且过半在南京。案例给出地市分布、薪资上下限区间，并说明薪资区间是参考值不是承诺。\n\n适合：把待遇作为重要决策因素的考生、比较不同城市岗位的人。", "description_en": "You want to know which public-sector posts pay well, but salary bands sit buried in individual post details with no cross-cutting view.\n\nThis case filters Jiangsu civil-service posts with a floor of 9,000 yuan, hitting 2,916. The 30 shown reveal a notable structure: all come from one selective-recruitment notice for outstanding new graduates, over half in Nanjing. It maps the city distribution and salary ranges, and notes that bands are reference values, not guarantees.\n\nFor: candidates weighing compensation heavily, and those comparing cities.", "categories": [ "search" ], "tags": [ "江苏", "薪资筛选", "选调", "南京" ], "tags_en": [ "Jiangsu", "salary filter", "selective recruitment", "Nanjing" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "江苏 9000 元以上岗位薪资分析", "label_en": "Jiangsu posts above 9,000 yuan" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查江苏公务员岗位里薪资下限在 9000 元以上的。要求：1) 真实筛选并给出总数；2) 取 30 条，按薪资下限降序列出岗位、招录单位、地级市、招录人数、薪资区间；3) 统计这 30 条来自几份公告、分别是什么公告，以及地级市分布；4) 给出本批薪资下限与上限的区间范围，并分析高薪岗位集中在哪类招录类型和城市；5) 说明薪资区间是参考值不是承诺，以及本批数据是否包含报录比和进面分。", "prompt_en": "Find Jiangsu civil-service posts with a salary floor above 9,000 yuan. Requirements: 1) run the filter and give the total; 2) take 30, sorted by descending salary floor, with post, unit, city, headcount and salary band; 3) report how many announcements these 30 come from, which ones, and the city distribution; 4) give the range of salary floors and ceilings and analyse which recruitment type and cities the high-paying posts cluster in; 5) note that salary bands are reference values, not guarantees, and state whether this data includes competition ratios or cutoff scores.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 151, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-salary-province-compare", "title": "薪资筛选：浙江与四川岗位薪资对照", "title_en": "Salary Filter: Zhejiang versus Sichuan", "subtitle": "两次跨省查询，看同样是公务员薪资区间差多少", "subtitle_en": "Two cross-province queries showing how civil-service pay bands differ", "description": "「去发达省份还是回老家」是每个考生都要面对的问题，但很少有人拿得出两边的具体数字。\n\n本案例对浙江和四川各做一次薪资筛选（门槛分别为 8,000 和 6,000 元），算出两省本页 30 条的薪资上下限均值并做对照，列出各自薪资最高的 6 个岗位，同时明确指出两次查询门槛不同、样本结构不同，结论不能当成全省平均水平。\n\n适合：在多个省份之间做选择的考生、给学生做地域建议的老师。", "description_en": "\"Coastal province or home town\" is a question every candidate faces, yet few can put numbers on either side.\n\nThis case runs one salary filter per province (thresholds 8,000 and 6,000 yuan), computes and compares the average salary floors and ceilings across 30 posts each, lists the six highest-paying posts per province, and states plainly that the thresholds and sample structures differ so the comparison is not a province-wide average.\n\nFor: candidates choosing between provinces, and counsellors advising on location.", "categories": [ "search" ], "tags": [ "跨省对比", "薪资", "浙江", "四川" ], "tags_en": [ "cross-province", "salary", "Zhejiang", "Sichuan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "浙江与四川岗位薪资对照", "label_en": "Zhejiang versus Sichuan salary comparison" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我对比浙江和四川的公务员岗位薪资。要求：1) 分两次真实查询——浙江按薪资下限 8000 元筛，四川按 6000 元筛，各给出总数；2) 各取 30 条，分别算出薪资下限均值和上限均值并做对照；3) 两省各列出薪资最高的 6 个岗位（岗位、单位、地级市、招录人数、薪资区间）；4) 统计两批结果的地区分布，指出样本结构上的差异；5) 明确说明两次查询门槛不同、不是同口径比较，并提示跨省比较还要考虑生活成本与岗位供给量。", "prompt_en": "Compare civil-service salaries between Zhejiang and Sichuan. Requirements: 1) two real queries — Zhejiang with a floor of 8,000 yuan, Sichuan with 6,000 — each with a total; 2) take 30 from each and compare the average salary floors and ceilings; 3) list the six highest-paying posts per province (post, unit, city, headcount, band); 4) map the geographic distribution of each result set and note the structural difference; 5) state explicitly that the thresholds differ so this is not a like-for-like comparison, and note that cost of living and post supply also matter.", "skills": [ { "id": "gaodun-job-selection", "name": "公考选岗顾问", "description": "根据学历、专业、地区等条件智能推荐适合报考的岗位", "skill_id": "skill_2089642887434874880" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 152, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "trading-agent-hk-meituan-earnings-drop", "title": "港股突发财报跳水盘前应急决策：美团", "title_en": "HK Emergency Pre-Market Decision on Earnings Drop: Meituan", "subtitle": "盘前半小时：基本面利空解读 + 技术面跌破 + 多空辩论 + 分 3 档左侧建仓方案", "subtitle_en": "30 minutes before open: bearish fundamentals + technical breakdown + bull-bear debate + 3-tier left-side scaling plan", "description": "昨晚港股财报暴雷、今天盘前跌 8%，原计划的买入操作瞬间乱了套 — 追还是不追、现在价位安全吗、一次买满还是分批进，半小时内根本想不清楚。\n\n这个案例演示了 TradingAgent 专家团在「快速决策模式」下的实战产出：围绕美团-W（3690.HK）Q3 核心本地商业经营利润首次转亏 141 亿、归母由盈转亏 186 亿的突发利空，团队分工完成财报数据深挖 + 新闻面机构评级（MS→95、Citi→133）扫描 + 技术面跌破 118 后关键支撑（110/95/85-88）识别 + 资金面主力净流出 12.45 亿与南向机构分歧读数 + 看多 vs 看空逻辑辩论，最终给出明确操作：今天追入意愿 2/5 星、首仓只上 20-30% 左侧试探、分 3 档在 100-105 / 95 / 85-88 分批建仓，破 82 收盘全面止损。整份报告 ~4500 字，数据可追溯。\n\n适合：遇到财报/黑天鹅等突发事件当天被迫做决策的港股投资者、想学会「盘前半小时系统化思考」的散户、需要左侧分批建仓模板的中长线交易者、做个股快评的财经自媒体。", "description_en": "When HK earnings crash overnight and the stock gaps down 8% pre-market, your original buy plan falls apart in seconds — chase or wait, is this price safe, all-in or scale in? 30 minutes is not enough to think it through.\n\nThis case shows the TradingAgent expert team running in \"rapid decision mode\" on a real emergency: Meituan-W (3690.HK) Q3 core local commerce swung to a 14.1B RMB operating loss and net profit flipped to -18.6B. The team divides work across earnings deep-dive + news/rating sweep (MS→95, Citi→133) + technical breakdown analysis after losing 118 with key supports at 110/95/85-88 + fund flow read (12.45B main outflow, southbound vs institutional divergence) + bull-bear debate. The final verdict: 2/5 star chase willingness, initial position capped at 20-30%, a 3-tier left-side scaling plan at 100-105 / 95 / 85-88, hard stop on a close below 82. ~4,500 characters, every number traceable.\n\nTarget audience: HK investors forced into same-day decisions by earnings/black swan events, retail traders learning \"structured thinking in 30 pre-market minutes\", long-term traders looking for a left-side scaling template, and finance content creators producing rapid single-stock commentary.", "categories": [ "data", "business", "agent" ], "tags": [ "港股", "突发财报", "快速决策", "左侧建仓", "多Agent协作" ], "tags_en": [ "HK Stocks", "Earnings Shock", "Rapid Decision", "Left-Side Scaling", "Multi-Agent Collaboration" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "美团突发财报跳水应对报告（Markdown）" } ], "prompt": "某港股美团-W（3690.HK）昨晚发布不及预期季报，今日盘前跌 8%，我本来打算今天买入。请做一份快速决策分析。\n\nMarkdown 报告要求：\n1. 财报核心数据解读（收入、利润、分业务数据）；\n2. 新闻面和行业反应（机构评级调整、同业反应）；\n3. 技术面跌破位置与关键支撑；\n4. 主力资金流向（南向、机构持仓变化）；\n5. 多空博弈后的明确结论：今天追不追、什么价位追、追多少仓位。", "prompt_en": "A Hong Kong stock, Meituan-W (3690.HK), released a worse-than-expected quarterly report last night, and is down 8% pre-market today. I had planned to buy today. Please make a rapid decision analysis.\n\nMarkdown report requirements:\n1. Interpretation of the core earnings data (revenue, profit, segment data);\n2. News and industry reactions (institutional rating adjustments, peer reactions);\n3. Technical breakdown levels and key supports;\n4. Main capital flow direction (southbound, changes in institutional holdings);\n5. A clear conclusion after weighing bull vs. bear: whether to chase today, at what price to chase, and how much position size.", "skills": [], "experts": [ { "id": "TradingAgentTeam", "name": "交易分析团队", "description": "13位专业角色分5阶段协作完成投资分析：技术面、基本面、新闻面、情绪面数据采集 → 多空辩论 → 交易决策 → 三方风险评估 → 最终报告，输出 BUY/SELL/HOLD 建议及完整操作方案", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "6e48bac09f0d", "name_en": "Trading Analysis Team", "description_en": "13 specialized roles collaborate across 5 phases: technical, fundamental, news & sentiment analysis → bull-bear debate → trading decision → 3-way risk assessment → final report, delivering BUY/SELL/HOLD recommendations with actionable plans", "expert_id": "ex_4FCmAJJmyHeU" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 152, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "survey-form-page", "title": "在线调查问卷页", "title_en": "Online Survey Form Page", "subtitle": "交互式问卷一键生成，答题体验丝滑流畅", "subtitle_en": "Generate an interactive survey form with smooth user experience", "description": "想做用户调研却发现专业问卷工具订阅费高昂，免费方案又丑得拿不出手？自己从零搭建表单更是耗时耗力，还要处理各种题型、跳转逻辑和数据统计。\n\n这个案例用前端开发技能生成了一份 ProductX 2025 用户满意度调查问卷：欢迎页展示问卷标题、说明和预估用时（5 分钟），点击开始后进入逐题作答模式。题型涵盖单选、多选、NPS 评分（0-10 量表）、满意度矩阵、开放文本、下拉选择、是否题带条件追问。顶部进度条实时显示完成百分比，题目间丝滑滑动切换，支持上一题/下一题导航。提交后展示感谢页，带打勾动画和作答摘要。\n\n适合：产品经理、用户体验研究员、HR 团队、市场营销团队——任何需要快速创建专业问卷并收集反馈的人。", "description_en": "Professional survey tools charge hefty subscriptions, free alternatives look unprofessional, and building custom forms from scratch demands significant engineering time — all while juggling multiple question types, conditional logic, and response analytics.\n\nThis case uses the frontend-dev skill to generate a ProductX 2025 User Satisfaction Survey: a welcome screen with survey title, description, and estimated completion time (5 min), followed by one-question-per-page navigation. Question types include single-choice, multiple-choice, NPS rating (0-10 scale), satisfaction matrix, open text, dropdown, and yes/no with conditional follow-up. A progress bar tracks completion percentage in real time, smooth slide transitions connect questions, and Previous/Next navigation lets respondents move freely. The thank-you page features an animated checkmark and response summary.\n\nIdeal for: product managers, UX researchers, HR teams, marketing teams — anyone who needs to quickly create professional surveys and collect feedback.", "categories": [ "office", "data" ], "tags": [ "问卷", "表单", "调查", "NPS", "交互" ], "tags_en": [ "Survey", "Form", "Questionnaire", "NPS", "Interactive" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为虚构的「ProductX 用户满意度调查 2025」设计一个交互式在线问卷页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，现代扁平化、每屏一题带平滑过渡。\n\n1. 欢迎页：标题/描述/预计 5 分钟/开始按钮\n2. 顶部进度条\n3. 多种题型：单选（如何了解我们-5 选项）、多选（常用功能-8 项）、NPS 0-10、满意度矩阵（5 维度×5 星）、开放问答、下拉单选（职位）、是否题带条件追问\n4. 上一题/下一题导航\n5. 完成页：动画勾选 + 回答摘要", "prompt_en": "Please design an interactive online questionnaire page for the fictional \"ProductX User Satisfaction Survey 2025\" (single-file HTML, inlined with no external dependencies). UI text should be in Simplified Chinese, with a modern flat style, one question per screen with smooth transitions.\n\n1. Welcome screen: title / description / \"approx. 5 minutes\" / start button\n2. A top progress bar\n3. Multiple question types: single-choice (how did you hear about us — 5 options), multiple-choice (frequently used features — 8 items), NPS 0-10, satisfaction matrix (5 dimensions × 5 stars), open-ended question, dropdown single-choice (job title), yes/no question with a conditional follow-up\n4. Previous/next navigation\n5. Completion screen: animated checkmark + a summary of answers", "skills": [ { "id": "form-cro", "name": "表单优化", "description": "表单设计与转化率优化，提升用户填写完成率", "slug": "form-cro" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 153, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "usage_count": 0, "created_at": "2026-04-15T18:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "调查问卷页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "agent-content-pipeline", "title": "内容生产流水线", "title_en": "Content Production Pipeline Agent", "subtitle": "AI Agent 自动化内容创作与发布工作流", "subtitle_en": "AI Agent Automated Content Creation & Publishing Workflow", "description": "团队每周要产出 10+ 篇内容，选题靠拍脑袋、写稿全凭手感、各平台格式反复调整、发布后效果无从追踪？内容团队陷入「量上不去、质没保障、数据看不清」的三重困境，核心人才被重复劳动消耗殆尽。\n\n这个案例展示了一套完整的 AI Agent 内容生产流水线设计：从选题库自动生成与热度评分（基于行业趋势、竞品动态、用户反馈）到内容大纲自动生成与审核流程，从多格式内容生成（长文、短视频脚本、社交媒体帖子、邮件 Newsletter）到 SEO 优化自动检查（关键词密度、标题优化、meta 描述），再到多平台适配与一键分发（微信公众号、知乎、小红书、LinkedIn）、内容效果追踪与 A/B 测试机制，以及完整的 Agent 协作编排（选题 Agent → 写作 Agent → 审核 Agent → 发布 Agent）。\n\n适合：内容营销负责人、新媒体运营主管、品牌市场总监、增长团队 Leader、希望用 AI 提升内容产能的创业公司。", "description_en": "Team struggling to produce 10+ pieces of content weekly? Topic selection by gut feeling, writing without standards, endless reformatting for different platforms, and no way to track post-publish performance? Content teams are trapped in the triple bind of insufficient volume, inconsistent quality, and invisible data — burning out top talent on repetitive work.\n\nThis case presents a complete AI Agent content production pipeline using content-factory: from automated topic bank generation with trend scoring to auto-generated content outlines with review workflows, from multi-format content generation to automated SEO checks, plus multi-platform adaptation and one-click distribution, content performance tracking with A/B testing, and a full Agent orchestration pipeline.\n\nIdeal for: content marketing leads, social media managers, brand directors, growth team leads, and startups looking to scale content production with AI.", "categories": [ "agent" ], "tags": [ "内容生产", "AI Agent", "自动化发布", "内容营销", "多平台分发" ], "tags_en": [ "content production", "AI agent", "auto publishing", "content marketing", "multi-platform" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI Agent 内容生产流水线工作流设计文档" } ], "prompt": "帮我设计一套 AI Agent 内容生产流水线工作流。团队需要每周产出 10+ 篇多平台内容。需要包含：1）选题库自动生成与热度评分（基于行业趋势、竞品动态、用户反馈）；2）内容大纲自动生成与审核流程；3）多格式内容生成（长文、短视频脚本、社交媒体帖子、邮件 newsletter）；4）SEO 优化自动检查（关键词密度、标题优化、meta 描述）；5）多平台适配与一键分发（微信公众号、知乎、小红书、LinkedIn）；6）内容效果追踪与 A/B 测试机制；7）Agent 协作编排（选题 Agent → 写作 Agent → 审核 Agent → 发布 Agent）。请用内容工厂的多 Agent 内容生产能力输出详细的 Markdown 工作流设计文档。", "prompt_en": "Help me design an AI Agent content production pipeline workflow. The team needs to produce 10+ pieces of multi-platform content per week. It must include: 1) Automated topic bank generation with trend scoring (based on industry trends, competitor activity, user feedback); 2) Automated content outline generation with a review workflow; 3) Multi-format content generation (long-form articles, short video scripts, social media posts, email newsletters); 4) Automated SEO optimization checks (keyword density, title optimization, meta descriptions); 5) Multi-platform adaptation and one-click distribution (WeChat Official Account, Zhihu, Xiaohongshu, LinkedIn); 6) Content performance tracking and an A/B testing mechanism; 7) Agent collaboration orchestration (Topic Agent → Writing Agent → Review Agent → Publishing Agent). Please use the Content Factory's multi-agent content production capabilities to output a detailed Markdown workflow design document.", "skills": [ { "id": "content-factory", "name": "内容工厂", "description": "多智能体内容生产系统，一份素材生成多种格式", "description_en": "Multi-agent content production system, one source to many formats", "skill_id": "skill_2053081490234478592" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 95, "sort_order": 154, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "form-lead-capture", "title": "B2B SaaS 高转化率演示预约表单", "title_en": "High-Converting B2B SaaS Demo Request Form", "subtitle": "表单字段精简 × 社会证明 × 退出挽留弹窗", "subtitle_en": "Field Reduction × Social Proof × Exit-Intent Popup", "description": "B2B SaaS 产品的演示预约表单往往字段冗长、缺乏信任背书，导致潜在客户望而却步——12 个字段的表单转化率仅有 2%，大量高意向线索在填写过程中流失。\n\n本案例运用 frontend-dev 技能构建一个全方位优化的表单页面：将 12 个字段精简至 3 个核心必填项（公司名称、工作邮箱、姓名），配合多步骤进度指示器降低用户感知负担；页面顶部展示客户 Logo 墙（Stripe、Shopify、Notion 等知名品牌）和关键数据指标（10,000+ 团队使用、98% 满意度、24 小时响应），侧边栏引入真实用户评价；CTA 按钮文案从通用的「提交」优化为具有行动号召力的「预约我的专属演示」；当用户鼠标移出页面时触发退出挽留弹窗，提供限时优惠激励完成填写。页面同时包含优化前后的直观对比，让改进效果一目了然。\n\n适合：B2B SaaS 市场营销团队、增长黑客、产品经理、转化率优化师——任何需要提升线索获取表单转化率的场景。", "description_en": "B2B SaaS demo request forms often suffer from excessive fields and lack of trust signals, causing high-intent prospects to abandon mid-funnel — a 12-field form typically converts at just 2%, leaving significant revenue on the table.\n\nThis case uses the frontend-dev skill to deliver a comprehensive form optimization: reducing 12 fields down to 3 essential inputs (company name, work email, full name), adding a multi-step progress indicator to lower perceived effort, displaying a customer logo bar (Stripe, Shopify, Notion, and more) with key metrics (10,000+ teams, 98% satisfaction, 24-hour response time), incorporating a real customer testimonial in the sidebar, upgrading the CTA copy from generic 'Submit' to the action-oriented 'Book My Personalized Demo', and triggering an exit-intent popup with a limited-time incentive when users attempt to leave. The page also includes a before/after comparison section that visually demonstrates the impact of each optimization.\n\nIdeal for: B2B SaaS marketing teams, growth hackers, product managers, CRO specialists — anyone looking to boost lead capture form conversion rates.", "categories": [ "business", "creation" ], "tags": [ "表单优化", "B2B", "转化率", "CRO", "落地页" ], "tags_en": [ "form optimization", "B2B", "conversion rate", "CRO", "landing page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为 B2B SaaS 产品「数流」设计一个高转化率演示预约表单页面（单文件 HTML，内联无外部依赖），界面文字用简体中文（客户品牌名 Stripe/Shopify/Notion 保留原名，用户姓名职位本地化为中文）。\n\n背景：原表单 12 个字段、转化率仅 2%。优化目标：\n- 精简为 3 个必填项（公司名称、工作邮箱、姓名）\n- 社会证明：客户 Logo 墙 + 关键指标（10000+ 团队 / 98% 满意度 / 24 小时响应）\n- CTA 文案「预约我的专属演示」\n- 多步骤进度指示器\n- 退出挽留弹窗（限时优惠）\n- 侧边栏真实用户评价\n\n输出含优化前后对比 + 完整优化版表单。", "prompt_en": "Please design a high-conversion demo booking form page for the B2B SaaS product \"ShuLiu (数流)\" (a single-file HTML page, inline with no external dependencies). The interface text should be in Simplified Chinese (keep the customer brand names Stripe/Shopify/Notion in their original form; localize user names and job titles into Chinese).\n\nBackground: The original form had 12 fields and only a 2% conversion rate. Optimization goals:\n- Streamline to 3 required fields (company name, work email, name)\n- Social proof: a customer logo wall + key metrics (10,000+ teams / 98% satisfaction / 24-hour response)\n- CTA copy \"Book My Personalized Demo\"\n- A multi-step progress indicator\n- An exit-intent retention popup (limited-time offer)\n- Real user testimonials in the sidebar\n\nThe output should include a before/after comparison + the complete optimized form.", "skills": [ { "id": "form-cro", "name": "表单优化", "description": "表单设计与转化率优化，提升用户填写完成率", "slug": "form-cro" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 155, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "线索收集表单" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "svg-icon-set-design", "title": "SVG 图标集设计", "title_en": "SVG Icon Set Design", "subtitle": "一套 12 个风格统一的矢量线性图标，即拿即用", "subtitle_en": "A set of 12 consistent linear SVG icons, ready to use", "description": "找免费图标东拼西凑，风格不统一；买图标库又太贵，只用几个图标却要付整套的钱。团队每次做新项目都要为图标资源发愁，设计师手绘耗时，开发者用 emoji 凑数。\n\n这个案例演示了如何用 SVG Draw 技能一次性生成 12 个办公场景线性图标（文件、日历、邮件、搜索、设置、用户、图表、消息、铃铛、锁、云端、分享），所有图标采用统一设计语言：24×24 viewBox、2px 描边、圆头端点、单色可改色，输出为纯 SVG 代码，可无损缩放到任意尺寸。展示页面还支持实时换色预览和一键复制 SVG 代码。\n\n适合：UI 设计师快速出原型图标、前端开发者直接嵌入项目、产品经理制作演示文档。", "description_en": "Scraping together free icons from different sources leads to inconsistent styles; purchasing icon libraries is expensive when you only need a handful. Teams waste time on icon resources for every new project.\n\nThis case demonstrates using the SVG Draw skill to generate 12 office-themed linear icons (file, calendar, mail, search, settings, user, chart, message, bell, lock, cloud, share) in one go. All icons share a unified design language: 24×24 viewBox, 2px stroke, round linecap/linejoin, single-color and recolorable. Output is pure SVG code that scales to any size without quality loss. The showcase page supports live color preview and one-click SVG code copying.\n\nIdeal for: UI designers prototyping icons, frontend developers embedding icons directly, product managers creating demo documents.", "categories": [ "creation" ], "tags": [ "SVG", "图标设计", "矢量图", "UI设计", "线性图标" ], "tags_en": [ "SVG", "Icon Design", "Vector", "UI Design", "Linear Icons" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请用 SVG 生成一套 12 个办公场景线性图标，并做一个展示页面（单文件 HTML，内联无外部依赖）。\n\n图标：文件、日历、邮件、搜索、设置、用户、图表、消息、铃铛、锁、云端、分享（统一线性风格、24px 网格、currentColor 描边）。\n\n展示页：响应式网格，每个图标卡片显示图标和名称，悬停放大，点击显示并复制 SVG 源码，顶部颜色选择器可实时切换图标颜色。", "prompt_en": "Please use SVG to create a set of 12 office-scenario line icons, and make a showcase page (single-file HTML, fully inline with no external dependencies).\n\nIcons: file, calendar, mail, search, settings, user, chart, message, bell, lock, cloud, share (uniform line style, 24px grid, currentColor stroke).\n\nShowcase page: a responsive grid where each icon card shows the icon and its name, enlarges on hover, and on click reveals and copies the SVG source code; a color picker at the top can switch the icon color in real time.", "skills": [ { "id": "svg-draw", "name": "SVG 绘图", "description": "纯 SVG 代码创建矢量图形和插画", "slug": "svg-draw" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 156, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-22T10:00:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "SVG 图标集" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "agent-code-review", "title": "智能代码审查", "title_en": "Intelligent Code Review Agent", "subtitle": "AI Agent 自动化代码审查与质量分析工作流", "subtitle_en": "AI Agent Automated Code Review & Quality Analysis Workflow", "description": "团队代码审查效率低、标准不统一、关键问题总在上线后才暴露？15 人团队每天 20+ 个 PR，高级工程师被审查拖住，初级开发者等反馈等到焦虑，安全漏洞和性能反模式反复出现却无人发现。\n\n这个案例展示了一套完整的 AI Agent 自动化代码审查工作流设计：从代码规范检查规则集（命名、格式、复杂度、安全漏洞）到自动化审查流水线（PR 触发、多阶段检查、结果汇总），从常见问题模式库（性能反模式、安全风险、可维护性问题）到审查意见优先级分类与自动标注，再到 Agent 自学习机制、CI/CD 集成方案和审查效果度量仪表盘设计。\n\n适合：技术团队 Leader、DevOps 工程师、质量保障负责人、希望提升代码审查效率的开发团队。", "description_en": "Code review bottlenecks slowing your team down? Standards inconsistent across reviewers? Critical issues only caught after deployment?\n\nThis case demonstrates a complete AI Agent automated code review workflow using the github skill: coding standard rulesets, automated review pipelines, common issue pattern libraries, review comment prioritization with auto-labeling, Agent self-learning mechanisms, CI/CD integration, and review effectiveness dashboards.\n\nIdeal for: tech team leads, DevOps engineers, QA managers, and development teams seeking better code review efficiency.", "categories": [ "agent", "dev" ], "tags": [ "代码审查", "AI Agent", "代码质量", "自动化", "开发效能" ], "tags_en": [ "code review", "AI agent", "code quality", "automation", "dev efficiency" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI 代码审查工作流设计文档" } ], "prompt": "帮我设计一套 AI Agent 自动化代码审查工作流。团队有 15 名开发者，每天产生 20+ 个 PR。需要包含：1）代码规范检查规则集（命名、格式、复杂度、安全漏洞）；2）自动化审查流水线设计（触发条件、检查步骤、结果汇总）；3）常见问题模式库（性能反模式、安全风险、可维护性问题）；4）审查意见的优先级分类与自动标注；5）Agent 自学习机制（从人工审查反馈中持续优化）；6）与 CI/CD 集成方案；7）审查效果度量指标与仪表盘设计。请用 github 技能输出详细的 Markdown 设计文档。", "prompt_en": "Help me design an AI Agent automated code review workflow. The team has 15 developers producing 20+ PRs per day. It needs to include: 1) a code standards rule set (naming, formatting, complexity, security vulnerabilities); 2) the automated review pipeline design (trigger conditions, check steps, result aggregation); 3) a common-problem pattern library (performance anti-patterns, security risks, maintainability issues); 4) priority classification and automatic tagging of review comments; 5) an Agent self-learning mechanism (continuous improvement from human review feedback); 6) a CI/CD integration plan; 7) review effectiveness metrics and dashboard design. Please use the github skill to output a detailed Markdown design document.", "skills": [ { "id": "github", "name": "github", "description": "管理 GitHub Issues、PR 和 CI", "description_en": "Manage GitHub issues, PRs, and CI runs", "skill_id": "skill_2053082027691548672" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 157, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-04-23T10:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "multi-brand-theme-switcher", "title": "多品牌设计主题切换器", "title_en": "Multi-Brand Design Theme Switcher", "subtitle": "一键切换 Vercel/Stripe/Linear 等品牌风格", "subtitle_en": "Switch between Vercel, Stripe, Linear and more brand styles instantly", "description": "做产品原型时想参考知名品牌的设计风格，却要一个个网站去扒 CSS、抄色值？这个交互式品牌主题切换器集成了 Vercel、Stripe、Linear、Notion、Supabase 五大品牌的完整 Design Tokens（配色、字体、间距、圆角、阴影），点击品牌名即可实时切换整套 UI 风格——按钮、卡片、表单、导航栏同步变化，秒级感受不同品牌的设计语言差异。适合 UI 设计师、前端工程师、产品经理做品牌调性参考或设计系统选型。", "description_en": "Tired of manually inspecting CSS from top brands for design inspiration? This interactive theme switcher integrates complete Design Tokens from Vercel, Stripe, Linear, Notion, and Supabase — colors, typography, spacing, radii, and shadows. Click any brand to instantly transform all UI components (buttons, cards, forms, nav bars) in real time. Perfect for UI designers, frontend engineers, and product managers comparing brand aesthetics or evaluating design system choices.", "categories": [ "creation", "dev" ], "tags": [ "设计系统", "品牌风格", "主题切换", "Design Tokens", "UI 组件" ], "tags_en": [ "Design System", "Brand Style", "Theme Switching", "Design Tokens", "UI Components" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "多品牌主题切换器" } ], "cover_image": "cover.png", "cover_mode": "landscape", "skills": [ { "id": "awesome-design-md", "name": "品牌设计风格专家", "description": "54 个知名网站设计系统模板，一键复用品牌级 UI 风格", "skill_id": "skill_2053081374617444352" }, { "id": "impeccable", "name": "高品质前端设计", "description": "高品质 UI/UX 设计工具集，避免泛 AI 审美", "skill_id": "skill_2053082862415904768" } ], "experts": [], "prompt": "请生成一个多品牌设计主题切换器单页（单文件 HTML，内联无外部依赖），暗色基底、顶部 5 个品牌切换按钮，点击后页面所有组件的配色/字体/圆角/间距/阴影平滑过渡到对应品牌风格。\n\n品牌：Vercel、Stripe、Linear、Notion、Supabase（Design Tokens 严格参照 awesome-design-md 中对应 DESIGN.md，不要编造）。\n\n四个展示区：\n1. Design Tokens 面板：当前品牌的主色/辅色/强调色（色块+色值）、字体、间距、圆角、阴影，切换时同步变化\n2. 组件展示区：按钮、输入框、选择卡片、Toast、数据表格、进度指示器各种状态\n3. 排版 Specimen：H1-H4 + 正文 + 小字\n4. 间距可视化：间距刻度色块\n\n用 CSS 变量驱动主题，切换带过渡动画，响应式。", "prompt_en": "Please generate a multi-brand design theme switcher single page (single-file HTML, inline with no external dependencies), with a dark base, 5 brand-switching buttons at the top; clicking one smoothly transitions the color scheme/fonts/border-radius/spacing/shadows of all components on the page to the corresponding brand style.\n\nBrands: Vercel, Stripe, Linear, Notion, Supabase (Design Tokens must strictly follow the corresponding DESIGN.md in awesome-design-md — do not fabricate them).\n\nFour display areas:\n1. Design Tokens panel: the current brand's primary/secondary/accent colors (color swatches + color values), fonts, spacing, border-radius, shadows, updating in sync when switching\n2. Component showcase: buttons, input fields, selection cards, Toast, data tables, progress indicators in various states\n3. Typography Specimen: H1-H4 + body text + small text\n4. Spacing visualization: spacing-scale color swatches\n\nDrive themes with CSS variables, transition with animation when switching, responsive.", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "badge": "official", "avatar": "" }, "source": "official", "quality_score": 92, "created_at": "2026-05-03T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 158, "featured_order": 0, "usage_count": 0, "is_featured": false, "mcps": [], "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-expense-report", "title": "月度费用报销汇总表", "subtitle": "个人报销明细到部门汇总，费用管控一键生成", "description": "每到月底，财务桌上堆满了报销单：差旅费、餐饮费、办公用品……金额对不上、审批卡在谁那儿不知道、超标报销混在正常单据里——靠人工逐条核对，既慢又容易出错。\n\n这个案例用 AI 一键生成一份完整的月度费用报销汇总 Excel：80 条真实场景报销记录，覆盖 4 个部门 20 名员工，涵盖差旅、餐饮、办公用品、交通、培训、招待六大费用类别。内置报销明细表、部门费用汇总、类别占比分析、审批状态追踪、同比趋势对比，还有超标合规标记——打开就能看清每一分钱花在了哪里、卡在了哪个环节。\n\n适合：财务团队、行政人员、部门主管、CFO、任何需要管控费用报销的人。", "categories": [ "data", "office", "finance" ], "tags": [ "Excel", "费用报销", "财务管理", "部门汇总", "审批管理", "费用分析" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "月度费用报销汇总表" } ], "prompt": "帮我生成一份专业的月度费用报销汇总 Excel 报表（.xlsx）。\n\n**业务背景**：中型互联网公司，4 个部门各约 5 人，需汇总报销数据做费用管控。\n\n**模拟数据**：生成 2026 年 3 月约 80 条报销记录，覆盖差旅、餐饮、办公用品、交通、培训、招待六大类。\n\n**报表结构（多 Sheet）**：\n- Sheet 1 报销明细：报销单号、日期、员工、部门、费用类别、金额、审批状态等字段\n- Sheet 2 部门汇总：按部门 × 费用类别交叉汇总\n- Sheet 3 费用类别分析\n- Sheet 4 审批状态分析（含超标报销标记）\n- Sheet 5 同比趋势对比\n\n**格式**：商务财务风格，金额带千分位，超标行用条件格式区分，冻结首行 + 自动筛选。", "prompt_en": "Help me generate a professional monthly expense reimbursement summary Excel report (.xlsx).\n\n**Business context**: A mid-sized internet company, 4 departments with about 5 people each, needing to summarize reimbursement data for expense control.\n\n**Simulated data**: Generate about 80 reimbursement records for March 2026, covering six categories (travel, meals, office supplies, transportation, training, entertainment).\n\n**Report structure (multiple Sheets)**:\n- Sheet 1 Expense Detail: fields such as reimbursement number, date, employee, department, expense category, amount, approval status\n- Sheet 2 Dept Summary: cross-tabulated by department × expense category\n- Sheet 3 Category Analysis\n- Sheet 4 Approval Status Analysis (with over-limit flags)\n- Sheet 5 Month-over-month Trend\n\n**Format**: a business finance style, amounts with thousands separators, conditional formatting for over-limit rows, freeze the top row + auto-filter.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 159, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Monthly Expense Report", "subtitle_en": "From individual receipts to department totals — expense control made easy", "description_en": "At the end of every month, the finance desk is buried in reimbursement forms: travel, meals, office supplies — amounts don't add up, approvals are stuck somewhere unknown, and over-budget claims hide among legitimate receipts. Manual line-by-line verification is slow and error-prone.\n\nThis case uses AI to generate a complete monthly expense reimbursement Excel in one click: 80 realistic expense entries from 20 employees across 4 departments, covering travel, dining, office supplies, transportation, training, and entertainment. Built-in expense detail sheets, department summaries, category breakdowns, approval status tracking, month-over-month trend analysis, and policy compliance flags — open it and instantly see where every dollar goes and where approvals are bottlenecked.\n\nBest for: Finance teams, admin staff, department heads, CFOs, anyone who needs to control and track expense reimbursements.", "tags_en": [ "Excel", "Expense Report", "Finance Management", "Department Summary", "Approval Management", "Expense Analysis" ], "is_recommended": false, "recommended_order": 0 }, { "id": "cultural-product-3d-acrylic", "title": "文创品建模 · 亚克力材质", "title_en": "Cultural Product 3D · Clear Acrylic", "subtitle": "把文创品变成通透亚克力质感的可交互 3D 模型", "subtitle_en": "Turn a creative product into a clear-acrylic interactive 3D model", "description": "通透亚克力的光感和体积最难靠平面图表达，客户总要求看立体效果。\n\n案例以唐三彩马为主体，保留三维体积、造型与主色，转译为通透亚克力材质的可交互 3D 模型：透光折射质感真实，网页可 360° 旋转缩放，单文件离线可开（模型已内联，约 62MB）。\n\n适合：文创衍生开发、展陈/礼品设计师、博物馆文创团队、IP 授权方", "description_en": "Using the Tang Sancai horse as subject, this case keeps its 3D volume, form and main colors, rendering a clear-acrylic interactive 3D model with realistic light refraction — rotate and zoom in the browser, single-file offline-ready (model inlined, ~62MB).\n\nGood for: Merch developers, Exhibition & gift designers, Museum creative teams, IP licensors", "categories": [ "creation" ], "tags": [ "3D建模", "亚克力材质", "可交互模型", "文创衍生" ], "tags_en": [ "3D Modeling", "Acrylic Material", "Interactive Model", "Cultural Merch" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "根据用户上传的文创品图片，准确识别主体的品类、外形轮廓、长宽比例、三维体积、造型结构、朝向、主色、纹样、配件及关键识别特征，并将其作为画面中的唯一主体。完整保留原产品设计，不重新设计、不增删结构、不改变比例、不替换品类、不擅自卡通化；仅将主体材质转换为高品质亚克力材质。\n\n构图：主体居中完整呈现，采用四分之三前侧视角（3/4 front-side view）；原图中属于产品本体的底座、支撑或配件应一并保留。使用纯净白色无缝背景和产品级静物棚拍，无文字、无水印、无装饰道具、无人物、无第二主体。\n材质：主体呈现高透或半透明浇铸亚克力质感，表面光洁平整，边缘经过精细抛光，具有清晰而克制的高光、透光、折射及内部色彩叠加效果。产品原有的三维体积、厚度和结构必须保持不变，不得擅自改成平面立牌、薄片挂牌或其他造型。忠实保留原产品的主色、色彩关系和纹样位置，可根据原图效果使用透明彩色亚克力、半透明亚克力、局部不透明色层、UV印刷或内嵌图案层表现，但不得改变图案形态、比例与颜色关系；透明度保持适中，不因过度透光而削弱产品识别度。\n光影：采用柔和均匀的棚拍光线，通过侧前方柔光表现亚克力的透明边缘、内部折射、材质厚度与精致高光；主体底部保留轻微自然的接触阴影，使产品稳定落地，避免悬浮感。\n\n画质：画面清晰锐利，主体轮廓准确，边缘干净，细节完整，适合作为3D建模与材质重建参考。避免玻璃器皿感、水晶雕塑感、果冻感、廉价透明塑料感、过曝高光、严重色散、内部气泡、裂纹、擅自增加拼接件、结构变形、模糊、强烈阴影遮挡及背景杂物。", "prompt_en": "", "skills": [], "experts": [ { "id": "tanyuan-cultural-product-3d-expert", "name": "文创品3D建模专家", "description": "将文创品转译为可交互的 3D 模型效果，支持毛绒、硅胶、金属珐琅、亚克力等材质渲染", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_P14ik4i7Utla" } ], "mcps": [], "artifacts": [ { "file": "output.html", "label": "可交互 3D 模型（单文件内联）" } ], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official", "name_en": "WorkBuddy Official" }, "source": "official", "sort_order": 160, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-09-22T15:31:07+08:00", "updated_at": "2026-09-22T15:31:07+08:00", "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "diagnosis-differential-cap", "title": "诊断复核与误诊漏诊排查", "title_en": "Diagnostic Review & Missed-Diagnosis Check", "subtitle": "已有确诊结果？AI 帮你找鉴别诊断、漏诊、误诊风险等级", "subtitle_en": "Already have a diagnosis? Let AI scan for differentials, missed and mis-diagnosis risks", "description": "门诊已下「社区获得性肺炎」诊断，但三天抗生素下去没见好转。到底是真的 CAP，还是被表象盖住了别的病——是肺栓塞？是血管炎？还是漏掉了某个共病？\n\n这个案例演示了把患者的「主诉 / 现病史 / 既往史 / 查体 / 辅助检查 / 当前诊断」一并交给「腾讯健康 · AI 临床助手·辅助诊断」。它会在报告开头醒目标出可能的危急征象，然后按常见病优先顺序列出不少于 3 个需鉴别的疾病，覆盖感染性、血栓性、免疫性 / 血管炎性、肿瘤性等不同病因机制，逐条给出鉴别要点与排除依据；再对当前诊断「可能误诊为什么」「可能漏诊哪些共病」给出风险等级评估。\n\n行为边界：只做诊断推理，不延伸至治疗方案；鉴别诊断不少于 3 个且覆盖不同病因机制；误诊 / 漏诊风险逐条标注等级；危急征象醒目标注于报告开头；末尾附安全声明，明确最终以临床医生判断为准。\n\n适合：住院 / 门诊医师、规培医生、临床药师、医学研究生、想训练鉴别诊断思维的临床用户。", "description_en": "The clinic has labeled it \"community-acquired pneumonia,\" but three days of antibiotics haven't helped. Is it really CAP, or is something else hiding behind the label — pulmonary embolism, vasculitis, a missed comorbidity?\n\nThis case shows how to hand the AI clinical assistant the full record: chief complaint, history of present illness, past history, physical exam, workup and current diagnosis. It opens the report with prominent warning signs, then lists at least three conditions to rule out — covering infectious, thromboembolic, immune-mediated / vasculitic, neoplastic and other mechanisms — in common-disease-first order, with exclusion rationale for each, and rates the misdiagnosis / missed-comorbidity risks of the current label.\n\nBoundaries: diagnosis reasoning only, no treatment plans; ≥3 differentials across distinct mechanisms; risk levels labeled on every missed or mis-diagnosis; warning signs up top; safety statement at the end reminding that final judgment rests with the clinician.\n\nFor: inpatient / outpatient physicians, residents, clinical pharmacists, medical students, and anyone using AI clinical assistants to sharpen differential-diagnosis thinking.", "categories": [ "agent", "education" ], "tags": [ "辅助诊断", "鉴别诊断", "误诊漏诊", "AI临床助手", "腾讯健康" ], "tags_en": [ "Diagnosis Assistant", "Differential Diagnosis", "Missed Diagnosis", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "诊断复核与误诊漏诊排查报告（HTML）" } ], "prompt": "请基于以下患者病历做一次诊断复核，评估当前诊断是否有误诊或漏诊可能。\n\n【患者病历】\n[粘贴患者病情描述：主诉、现病史、既往史、查体、辅助检查、当前诊断]\n\n【示例场景】\n这个患者门诊诊断为社区获得性肺炎，用了三天抗生素没好转，请帮我做一下诊断分析，看看有没有漏诊或误诊（附件病情描述）。\n\n【输出要求】\n1. 危急征象识别：报告开头醒目标注\n2. 病历文本解析：提取主诉、现病史、既往史、查体、辅助检查、当前诊断\n3. 鉴别诊断：列出需与「当前诊断」鉴别的疾病（不少于 3 个，按常见病优先排序），每条附鉴别要点与排除依据\n4. 误诊 / 漏诊风险评估：当前诊断可能误诊为什么、可能漏诊什么共病，逐条标注风险等级（高 / 中 / 低）\n5. 安全声明附在报告末尾\n\n【行为边界】\n- 只做诊断推理，不延伸至治疗方案\n- 鉴别诊断应覆盖不同病因机制（感染性 / 血栓性 / 免疫性 / 血管炎性 / 肿瘤性 / 心源性等）\n- 误诊 / 漏诊风险逐条标注风险等级\n- 危急征象在报告开头醒目标注\n- 诊断依据 100% 溯源至输入信息，不引用输入中不存在的指标\n\n【不应出现】\n- 具体的用药调整建议或治疗方案\n- 诊断依据引用输入中不存在的信息\n- 鉴别诊断少于 3 个\n- 省略安全声明", "prompt_en": "Please perform a diagnostic review of the following patient record and assess whether the current diagnosis carries any risk of misdiagnosis or missed comorbidity.\n\n【Patient record】\n[Paste the patient's case: chief complaint, history of present illness, past history, physical exam, workup, current diagnosis]\n\n【Example scenario】\nThis patient was diagnosed with community-acquired pneumonia in clinic but did not improve after three days of antibiotics. Please analyze the case and check for any misdiagnosis or missed diagnosis (see attached description).\n\n【Output requirements】\n1. Warning-sign recognition: prominently flagged at the top of the report\n2. Record parsing: extract chief complaint, history of present illness, past history, physical exam, workup, current diagnosis\n3. Differential diagnosis: list conditions to rule out against the current diagnosis (≥ 3, in common-disease-first order), each with differentiation points and exclusion rationale\n4. Misdiagnosis / missed-diagnosis risk assessment: state what the current diagnosis may be confused with and what comorbidities may be missed, with risk levels (high / medium / low) labeled for each item\n5. Safety statement appended at the end\n\n【Boundaries】\n- Diagnosis reasoning only; do not extend into treatment plans\n- Differentials must span distinct etiologic mechanisms (infectious / thromboembolic / immune / vasculitic / neoplastic / cardiac, etc.)\n- Each misdiagnosis / missed-diagnosis item must be labeled with a risk level\n- Warning signs must be flagged at the top of the report\n- Diagnostic clues must be 100% traceable to the input — never invent indicators not in the record\n\n【Must not appear】\n- Specific medication adjustments or treatment plans\n- Diagnostic clues referencing information not present in the input\n- Fewer than 3 differential diagnoses\n- A missing safety statement", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 160, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:55:11+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "fashion-showcase-short", "title": "时装展示短片", "title_en": "Fashion Showcase Short", "subtitle": "让面料、刺绣和裙摆在花园光线里完整被看见", "subtitle_en": "Fabric, embroidery, and movement revealed in garden light", "description": "服装展示视频常把氛围放在第一位，却没有准确呈现版型、面料和关键工艺。\n\n案例从刺绣奶油色长裙的刀版与模特服饰卡开始，在玫瑰园中用行走、下午茶、细节微距和旋转展示服装。\n\n适合：电商服饰、设计师品牌、毕业作品、Lookbook 和新品发布内容。", "description_en": "Fashion videos often prioritize atmosphere while failing to present silhouette, fabric, and craft accurately.\n\nStarting from a technical garment sheet and model wardrobe card, this case shows a cream embroidered dress through walking, afternoon tea, detail macros, and a turn in a rose garden.\n\nBest for fashion commerce, designer labels, graduate collections, lookbooks, and new-product launches.", "categories": [ "creation" ], "tags": [ "时装短片", "刺绣连衣裙", "花园Lookbook" ], "tags_en": [ "fashion short", "embroidered dress", "garden lookbook" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「时装展示短片」Skill，围绕我上传的奶油色长裙设计图，制作一轮 30 秒、16:9 的 Lifestyle 时装片《玫瑰园里的刺绣裙》。服装特征必须锁定：方领、收腰长裙、袖口与裙摆的粉绿色花卉刺绣、哑光轻薄面料。先生成无模特的服装刀版三视图并暂停确认；再生成一位棕色长发模特的服饰卡并确认面料与版型；随后给出 3 个场景候选，我选择英式玫瑰园后，生成无人物的黄昏花园图并确认。故事板包含花径行走、下午茶落座、袖口与刺绣微距、裙摆旋转、夕阳回望，生成六宫格分镜让我确认后再一次性出片并配轻柔 BGM。不得改变领口、刺绣位置、裙长或模特外貌。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Fashion Showcase Short” Skill and use my uploaded design for a cream dress to create one 30-second 16:9 Lifestyle fashion film titled The Embroidered Dress in the Rose Garden. Lock these garment features: square neckline, fitted waist and long skirt, pink-and-green floral embroidery at the cuffs and hem, and lightweight matte fabric. First generate a three-view technical garment sheet without a model and pause for approval. Then create a wardrobe card for a long-haired brunette model and confirm the fabric and silhouette. Present three location options; after I choose an English rose garden, generate and approve an empty golden-hour garden image. The storyboard must include walking along the flower path, sitting for afternoon tea, macro details of cuffs and embroidery, a skirt turn, and a sunset glance back. Generate a six-panel cinematic storyboard for approval, then produce the complete video in one pass and add gentle music. Do not alter the neckline, embroidery placement, skirt length, or model appearance.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 160, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-major-catalog", "title": "专业目录：一个会计学 8 种归类", "title_en": "Major Directories: One Major, Eight Categories", "subtitle": "查清专业在各省目录里算哪类，避免资格复审被刷", "subtitle_en": "Know your major's category in every directory before review rejects you", "description": "公考最冤的失败是笔试面试都过了、资格复审被刷——理由是专业大类对不上。各省专业目录归类根本不统一。\n\n本案例查询「本科 + 会计学」，一次拉出 10 份官方目录的全部归类：教育部目录给出 3 个专业代码，国考同时归在三个大类，江苏还额外归进经济类，9 份公考目录共出现 8 种大类写法。\n\n适合：跨省报考的考生、专业名易混淆的考生、做资格审查的工作人员。", "description_en": "The cruellest failure: passing written and interview stages, then losing at qualification review because the major category doesn't match. Provincial directories simply disagree.\n\nThis case queries \"bachelor's + Accounting\" and pulls its classification across 10 official directories at once: three codes in the MoE directory, three simultaneous categories in the national exam directory, an extra Economics listing in Jiangsu — eight different category labels across nine directories.\n\nFor: cross-province applicants, easily-confused major names, and qualification reviewers.", "categories": [ "search", "education" ], "tags": [ "专业目录", "报考资格", "会计学", "资格复审" ], "tags_en": [ "major directory", "eligibility", "accounting", "qualification review" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "会计学跨目录归类对照" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我本科专业是会计学，要报公务员，帮我查清楚它在各目录里的归类。要求：1) 学历=本科、专业=会计学，实时检索所有收录目录；2) 逐份列出大类、小类、专业代码；3) 教育部目录与各省公考目录分开呈现；4) 必须来自官方目录实时查询，不许凭常识推测；5) 总结跨省报考最容易错在哪。", "prompt_en": "My bachelor's major is Accounting and I'm applying for civil service. Requirements: 1) education=bachelor's, major=Accounting, live-search every directory that lists it; 2) give category, subcategory and code per directory; 3) present the MoE directory separately from provincial exam directories; 4) results must come from live official-directory queries, not general knowledge; 5) summarise where cross-province applications most often go wrong.", "skills": [ { "id": "gaodun-major-search", "name": "专业目录查询", "description": "按学历、专业名检索公考专业目录，匹配大类小类及代码", "skill_id": "skill_2089643176358125568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 160, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "team-member-cards", "title": "团队成员展示卡片", "title_en": "Team Member Profile Cards", "subtitle": "输入团队信息，生成精美的成员介绍卡片组", "subtitle_en": "Input team info, generate elegant member profile card set", "description": "公司官网的「关于我们」页面、创业路演的团队介绍、投标文档的项目组简介——都需要把团队成员信息以视觉化的方式呈现出来，但每次都要找设计师排版，小团队根本耗不起。\n\n这个案例演示了输入 4 位核心团队成员的信息（姓名、职位、简介、专长），AI 自动生成一组精美的成员卡片 HTML 页面：2x2 网格布局（max-width 720px 居中）、每张卡片含 CSS 纯色头像、姓名职位、一句话简介、技能标签。浅色主题，卡片悬停上浮，简洁现代。\n\n适合：创业团队官网、项目投标、路演 PPT 配套页面、内部团队介绍。", "description_en": "Company 'About Us' pages, startup pitch team slides, proposal team bios — all need visually appealing team member presentations, but getting design help for every iteration is costly for small teams.\n\nThis case demonstrates inputting 4 core team members' info (name, role, bio, skills), with AI generating elegant profile cards in HTML: 2x2 grid layout (max-width 720px centered), each card with CSS avatar, name/role, one-line bio, and skill tags. Light theme, hover lift effect, clean modern design.\n\nBest for: Startup websites, project proposals, pitch deck companion pages, internal team introductions.", "categories": [ "creation", "business" ], "tags": [ "团队介绍", "卡片组件", "官网设计", "人员展示" ], "tags_en": [ "Team Introduction", "Card Component", "Website Design", "Personnel Display" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个团队成员展示卡片页面（单文件 HTML，内联无外部依赖），现代精美、卡片网格布局，每张卡片含头像、姓名职位、一句话简介和技能标签。\n\n示例：某 AI 创业公司核心团队 4 人——\n1. 陈思远 · CEO & Co-founder — 连续创业者，前字节 AI Lab 技术总监 · 产品战略 / AI 算法 / 团队管理\n2. Lisa Wang · CTO & Co-founder — MIT CSAIL 博士，前 Google DeepMind 研究科学家 · 深度学习 / 分布式系统 / 模型优化\n3. 张浩然 · Head of Product — 前微信产品经理 · 产品设计 / 用户增长 / 数据驱动\n4. Emily Zhao · Head of Design — 前 Figma 高级设计师 · UI/UX / 设计系统 / 品牌视觉", "prompt_en": "Please generate a team member showcase card page (single-file HTML, inline with no external dependencies), modern and polished, with a card grid layout — each card includes an avatar, name and title, a one-line bio, and skill tags.\n\nExample: the core team of an AI startup, 4 people —\n1. Chen Siyuan · CEO & Co-founder — serial entrepreneur, former ByteDance AI Lab Technical Director · Product Strategy / AI Algorithms / Team Management\n2. Lisa Wang · CTO & Co-founder — PhD from MIT CSAIL, former Senior Research Scientist at Google DeepMind · Deep Learning / Distributed Systems / Model Optimization\n3. Zhang Haoran · Head of Product — former WeChat product manager · Product Design / User Growth / Data-Driven\n4. Emily Zhao · Head of Design — former Senior Designer at Figma · UI/UX / Design Systems / Brand Visuals", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 160, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "团队成员卡片" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "ai-drawing-brand-logo", "title": "AI 品牌 Logo SVG 设计", "title_en": "AI Brand Logo SVG Design", "subtitle": "一句话描述品牌，AI 生成三套专业 SVG Logo 方案", "subtitle_en": "Describe your brand in one sentence, get 3 professional SVG logo concepts", "description": "请设计师出 Logo 动辄数千元、等上两三周。这个案例展示了如何用一句话描述品牌定位，AI 即刻生成三套完整的 SVG Logo 方案——现代极简（山峦茶叶）、传统印章（篆刻古韵）、抽象几何（茶杯剪影），每套附带配色说明与使用场景建议。\n\n纯 SVG 代码实现，无需外部图片，可直接用于网页、印刷、社交媒体。所有方案以精美深色展示页呈现，一目了然。\n\n适合：初创团队快速出品牌方案、独立开发者 Side Project 品牌化、个人工作室视觉升级。", "description_en": "Hiring a designer for a logo costs thousands and takes weeks. This case shows how a single brand description generates 3 complete SVG logo concepts — modern minimalist (mountain + tea leaf), traditional Chinese seal (篆刻 style), and abstract geometric (tea cup silhouette) — each with color palettes and usage recommendations.\n\nPure SVG code, no external images, ready for web, print, and social media. All concepts are presented on a polished dark-themed showcase page.\n\nBest for: startup branding, indie developer side projects, freelancer visual identity.", "categories": [ "creation", "business" ], "tags": [ "Logo设计", "SVG", "品牌", "视觉设计", "创业" ], "tags_en": [ "Logo Design", "SVG", "Branding", "Visual Design", "Startup" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "品牌 Logo 设计方案（3 套 SVG）" } ], "prompt": "请为精品茶品牌「山岚茶舍」设计 3 套 SVG Logo 方案，并做一个展示页面（单文件 HTML，内联无外部依赖，深色背景）。\n\n品牌定位：高端、东方美学、自然意境。\n- 方案一：现代极简，融合山峦与茶叶元素\n- 方案二：传统印章风格（篆刻感）\n- 方案三：抽象几何，茶杯剪影为核心\n\n每套用纯 SVG 实现（不用外部图片），配色以墨绿/金色/深褐为主。每个 Logo 下方标注设计风格说明和推荐使用场景。", "prompt_en": "Please design 3 SVG logo concepts for the premium tea brand \"Shanlan Teahouse\" (山岚茶舍), and build a showcase page (single-file HTML, inlined with no external dependencies, dark background).\n\nBrand positioning: high-end, Eastern aesthetics, natural ambiance.\n- Concept 1: modern minimalist, blending mountain and tea-leaf elements\n- Concept 2: traditional seal style (a carved-stamp feel)\n- Concept 3: abstract geometric, centered on a teacup silhouette\n\nImplement each in pure SVG (no external images), with a color palette dominated by ink green / gold / dark brown. Below each logo, note the design style and recommended use cases.", "skills": [ { "id": "svg-draw", "name": "SVG 绘图", "description": "使用 SVG 创建矢量图形、图标和可视化设计", "slug": "svg-draw" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 161, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-21T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-major-computer", "title": "专业目录：计算机科学与技术跨省归类", "title_en": "Major Catalogue: How Computer Science Is Classified Across Provinces", "subtitle": "同一个专业名，在 10 份目录里被归进 8 种不同的大类", "subtitle_en": "One major name, eleven different category labels across ten catalogues", "description": "职位表写「限计算机类」，你的毕业证写「计算机科学与技术」——到底算不算符合？换个省份，答案还可能不一样。\n\n本案例查询计算机科学与技术在 10 份目录（教育部 + 国考 + 各省考）中的归类，命中 24 条条目、11 种大类叫法。逐份列出所属大类、专业小类与专业代码，并特别指出江苏目录把同一专业挂在 4 个不同大类之下。\n\n适合：跨省报考的考生、专业名与目录表述不一致的报考者。", "description_en": "The post table says \"computer science category\" and your degree says \"computer science and technology\" — does that qualify? Change province and the answer may change too.\n\nThis case looks up computer science and technology across 10 catalogues (Ministry of Education, national exam, and provincial exams), returning 25 entries under 8 different category labels. Each catalogue's category, sub-category and major code is listed, highlighting that Jiangsu files the same major under four separate categories.\n\nFor: candidates applying across provinces where major wording differs.", "categories": [ "search", "education" ], "tags": [ "专业目录", "计算机", "跨省报考", "专业代码" ], "tags_en": [ "major catalogue", "computer science", "cross-province", "major code" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "计算机科学与技术跨省目录归类", "label_en": "Computer science classification across catalogues" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我是普通本科计算机科学与技术专业，帮我查这个专业在各份公考专业目录里的归类。要求：1) 真实查询并给出命中几份目录、共几条条目；2) 逐份列出目录名称、所属大类、专业小类、专业代码；3) 统计一共出现了几种不同的大类叫法，各出现几次；4) 指出哪份目录的归类最特殊（比如同一专业挂在多个大类下），说明这对报考意味着什么；5) 提醒专业代码以毕业证书为准、跨省报考需逐省单独查目录。", "prompt_en": "I hold a bachelor's degree in computer science and technology — look up how this major is classified across civil-service major catalogues. Requirements: 1) run the query and report how many catalogues and entries were found; 2) list each catalogue's name, category, sub-category and major code; 3) count how many distinct category labels appear and how often; 4) identify the catalogue with the most unusual classification (e.g. the same major filed under several categories) and explain what that means for applications; 5) note that major codes follow the degree certificate and that each province must be checked separately.", "skills": [ { "id": "gaodun-major-search", "name": "专业目录查询", "description": "按学历、专业名检索公考专业目录，匹配大类小类及代码", "skill_id": "skill_2089643176358125568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 161, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "auto-weekly-report", "title": "周报自动生成", "title_en": "Auto Weekly Report Generator", "subtitle": "工作事项一键转结构化周报 Word 文档，告别周五加班整理", "subtitle_en": "Transform work items into a structured weekly report Word document instantly", "description": "每到周五下午最头疼的事就是写周报——翻聊天记录、翻 Git log、翻 Jira，拼拼凑凑写了一小时还觉得没写清楚重点。\n\n这个案例演示了如何把一周的工作事项列表丢给 AI，自动生成一份结构清晰的 Word 周报文档：按「本周完成」「进行中」「下周计划」三段组织，每项工作自动标注关键数据指标（比如 FCP 从 2.8s 降至 1.2s），还附带一段简短的本周总结。从 6 条工作记录到一份可以直接发给领导的 Word 周报，格式规范可直接打印。\n\n适合：所有需要写周报的打工人、项目经理、团队 Leader、技术负责人", "description_en": "The most dreaded Friday afternoon task — writing weekly reports. Digging through chat logs, Git history, and Jira tickets, spending an hour piecing together a report that still doesn't clearly convey your highlights.\n\nThis case shows how to throw a list of weekly work items at AI and get a structured Word report document: organized by 'Completed / In Progress / Next Week', each item tagged with key metrics (e.g., FCP reduced from 2.8s to 1.2s), plus a brief weekly summary. From 6 work items to a ready-to-send Word report with professional formatting.\n\nBest for: Anyone who writes weekly reports, project managers, team leads, tech leads", "categories": [ "office", "productivity" ], "tags": [ "周报", "办公效率", "工作汇报", "结构化文档" ], "tags_en": [ "Weekly Report", "Productivity", "Work Summary", "Structured Docs" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "结构化周报（Word）" } ], "prompt": "请根据我本周的工作事项，生成一份结构化周报 Word 文档（output.docx）。\n\n本周工作事项（粘贴你的工作清单）：\n...\n\n要求：\n- 按「本周完成」「进行中」「下周计划」三段组织\n- 每项工作标注关键指标与数据\n- 含一段简短的本周总结\n- 含封面、目录、页眉页脚，语言简洁专业，适合发给直属领导\n用「Word 文档生成」技能输出。", "prompt_en": "Please generate a structured weekly report as a Word document (output.docx) based on my work this week.\n\nThis week's work items (paste your task list):\n...\n\nRequirements:\n- Organize into three sections: \"Completed This Week,\" \"In Progress,\" and \"Next Week's Plan\"\n- Annotate each item with key metrics and data\n- Include a brief weekly summary\n- Include a cover page, table of contents, and headers/footers; concise, professional language suitable for sending to a direct manager\nUse the \"Word Document Generation\" skill to output it.", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 162, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-21T14:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-major-law-master", "title": "专业目录：法学硕士跨省归类", "title_en": "Major Catalogue: Law Master's Across Provinces", "subtitle": "9 份目录、5 种大类叫法，法学考生报考前的一次目录体检", "subtitle_en": "Nine catalogues, five category labels — a pre-application audit for law graduates", "description": "法学是公检法、纪委监委岗位的核心限制专业，也是各省目录写法差异最大的专业之一——「法学类」「法律类」「法学大类」「法学学科门类」指的是不是同一件事？\n\n本案例查询法学（硕士研究生）在 9 份目录中的归类，逐份列出大类、小类与代码，统计出 5 种不同叫法，并指出广西目录同时把法学收录在两条不同分类下。最后列出法学考生最容易踩的三个坑。\n\n适合：法学专业考生、报考公检法岗位的研究生。", "description_en": "Law is the core restricted major for procuratorate, court and disciplinary posts — and one of the most inconsistently labelled across provinces. Are \"law category\", \"legal category\" and \"law discipline\" the same thing?\n\nThis case looks up law (master's level) across 9 catalogues, listing each category, sub-category and code, counting 5 distinct labels, and noting that Guangxi files law under two separate classifications. It closes with the three traps law candidates most often fall into.\n\nFor: law graduates applying to judicial and disciplinary posts.", "categories": [ "search", "education" ], "tags": [ "专业目录", "法学", "硕士研究生", "跨省报考" ], "tags_en": [ "major catalogue", "law", "master's degree", "cross-province" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "法学硕士跨省目录归类", "label_en": "Law master's classification across catalogues" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我是法学硕士研究生，帮我查法学专业在各份公考专业目录里的归类。要求：1) 真实查询并给出命中几份目录、共几条条目；2) 逐份列出目录名称、所属大类、专业小类、专业代码；3) 统计一共出现了几种大类叫法，分别对应哪些省份；4) 指出有没有目录把法学同时收录在两条不同分类下；5) 结合结果列出法学考生报考时最容易踩的坑（例如只搜一种叫法、混淆本科与研究生目录、法学硕士与法律硕士的区别）。", "prompt_en": "I hold a master's degree in law — look up how law is classified across civil-service major catalogues. Requirements: 1) run the query and report how many catalogues and entries were found; 2) list each catalogue's name, category, sub-category and code; 3) count the distinct category labels and which provinces use each; 4) note whether any catalogue files law under two separate classifications; 5) based on the results, list the traps law candidates most often fall into (searching only one label, confusing undergraduate and postgraduate catalogues, academic versus professional law degrees).", "skills": [ { "id": "gaodun-major-search", "name": "专业目录查询", "description": "按学历、专业名检索公考专业目录，匹配大类小类及代码", "skill_id": "skill_2089643176358125568" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 162, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "testimonial-wall", "title": "用户好评墙", "title_en": "Customer Testimonial Wall", "subtitle": "输入用户评价，生成精美的好评展示墙页面", "subtitle_en": "Input customer reviews, generate an elegant testimonial wall page", "description": "SaaS 产品官网、落地页、应用商店——用户评价是转化率最高的社会证明，但大多数团队只是在页面底部贴几段文字引用，毫无设计感。好的 testimonial 展示能让访客瞬间建立信任。\n\n这个案例演示了输入 6 条用户评价（头像、姓名、职位、评语、评分），AI 自动生成一个精美的好评展示墙 HTML 页面：居中布局（max-width 780px）、Masonry 瀑布流风格的评价卡片、星级评分、引号装饰、头像和用户信息。浅色背景 + 白色卡片，专业温暖。\n\n适合：SaaS 官网社会证明区、产品落地页、应用商店评价展示、品牌展示页。", "description_en": "SaaS websites, landing pages, app stores — customer testimonials are the highest-converting social proof, but most teams just paste a few text quotes at the bottom with zero design. Great testimonial displays instantly build visitor trust.\n\nThis case demonstrates inputting 6 customer reviews (avatar, name, role, quote, rating), with AI generating an elegant testimonial wall HTML page: centered layout (max-width 780px), masonry-style review cards, star ratings, quotation mark decorations, avatars and user info. Light background + white cards, professional and warm.\n\nBest for: SaaS website social proof sections, product landing pages, app store review displays, brand showcase pages.", "categories": [ "business", "creation" ], "tags": [ "用户评价", "社会证明", "好评墙", "官网组件" ], "tags_en": [ "Customer Reviews", "Social Proof", "Testimonial Wall", "Website Component" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个用户好评展示墙页面（单文件 HTML，内联无外部依赖），现代精美、瀑布流卡片布局。\n\n产品：FlowBoard（项目管理工具）。展示 6 条用户评价，每条含姓名、职位、评分（4-5 星）和评语。卡片带引号装饰、评语、用户信息行和星级评分。", "prompt_en": "Please generate a user-testimonial wall page (single-file HTML, inlined with no external dependencies), modern and polished, with a masonry card layout.\n\nProduct: FlowBoard (a project-management tool). Display 6 user reviews, each including a name, job title, rating (4-5 stars), and review text. Cards feature quotation-mark decoration, the review text, a user-info row, and a star rating.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 163, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "用户评价墙" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "agent-data-report-gen", "title": "Agent 自动采集生成数据报表", "title_en": "Agent Automated Data Report Generation", "subtitle": "多 Agent 协作自动采集、清洗、分析并生成结构化运营数据报表", "subtitle_en": "Multi-Agent Workflow for Automated Data Collection, Cleaning, Analysis & Report Generation", "description": "每周手动从多个系统导出数据、粘贴到 Excel、计算环比同比、标注异常……运营数据报表的制作流程重复、耗时，且容易出错。一旦指标口径不一致或数据源缺失，整份报表的可信度就大打折扣。\n\n这个案例演示了一个完整的多 Agent 数据报表生成工作流：数据采集 Agent 从多个模拟数据源抓取用户增长、收入、活跃度和留存率等核心指标；数据清洗 Agent 自动识别异常值和缺失数据并标注原因；分析 Agent 计算环比/同比变化、发现显著趋势并生成数据洞察；报表 Agent 最终将所有分析结果整理为包含「数据概览」「趋势分析」「异常标注」「运营建议」四个模块的结构化报表。\n\n适合：运营经理、数据分析师、业务决策者、BI 团队、需要定期产出数据报告的任何团队。", "description_en": "Manually exporting data from multiple systems every week, pasting into Excel, calculating WoW/YoY changes, flagging anomalies — the weekly operations report workflow is repetitive, time-consuming, and error-prone.\n\nThis case demonstrates a complete multi-agent data report generation workflow: a Data Collection Agent fetches key metrics from simulated sources; a Data Cleaning Agent identifies anomalies and missing values; an Analysis Agent computes trends and generates insights; a Report Agent assembles everything into a structured report with four sections: Data Overview, Trend Analysis, Anomaly Flags, and Operational Recommendations.\n\nIdeal for: operations managers, data analysts, business decision-makers, BI teams, and anyone producing regular data reports.", "categories": [ "agent", "data", "business" ], "tags": [ "数据报表", "多Agent协作", "运营分析", "自动化工作流" ], "tags_en": [ "Data Report", "Multi-Agent", "Operations Analytics", "Automated Workflow" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "prompt": "我需要一个 Agent 工作流来自动化生成周度运营数据报表。流程：1) 数据采集 Agent 从模拟的多个数据源（用户增长、收入、活跃度、留存率）抓取关键指标；2) 数据清洗 Agent 检查数据异常和缺失值，标注异常原因；3) 分析 Agent 计算环比/同比变化、识别显著趋势、生成洞察；4) 报表 Agent 将分析结果整理为结构化的 Excel 格式报表。最终产出一份包含「数据概览」「趋势分析」「异常标注」「运营建议」四个Sheet的报表。请用 Agent Team 技能执行整个流程。", "prompt_en": "I need an Agent workflow to automate the generation of a weekly operations data report. Workflow: 1) A Data Collection Agent pulls key metrics from several simulated data sources (user growth, revenue, activity, retention rate); 2) A Data Cleaning Agent checks for anomalies and missing values, annotating the cause of each anomaly; 3) An Analysis Agent computes month-over-month/year-over-year changes, identifies significant trends, and produces insights; 4) A Reporting Agent organizes the analysis results into a structured Excel-format report. The final deliverable is a report with four sheets: \"Data Overview\", \"Trend Analysis\", \"Anomaly Annotations\", and \"Operations Recommendations\". Please use the Agent Team skill to execute the entire workflow.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 164, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.md", "label": "周度运营数据报表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "interactive-scroll-story", "title": "交互式滚动叙事页", "title_en": "Interactive Scroll Storytelling Page", "subtitle": "滚动触发动画，用交互讲好产品故事", "subtitle_en": "Scroll-triggered animations that tell your product story", "description": "产品介绍页千篇一律，用户看两眼就划走，想做交互动效又不会写代码？用 Web 动画设计技能，一句话生成一个滚动叙事页面——5 个章节随滚动渐入、视差背景、数字跳动、进度条等交互效果全自动生成。适合产品经理、市场营销、品牌设计师、独立开发者快速搭建高质感产品展示页。", "description_en": "Product pages all look the same — visitors glance and scroll away. Want interactive effects but can't code? Use the Web Animation Design skill to generate a scroll storytelling page in one prompt — 5 sections with scroll-triggered fade-ins, parallax backgrounds, animated counters, progress bars, and more. Perfect for product managers, marketers, brand designers, and indie developers to quickly build high-quality product showcase pages.", "categories": [ "creation" ], "tags": [ "滚动动画", "交互设计", "产品展示", "Web动画", "叙事页面" ], "tags_en": [ "scroll animation", "interactive design", "product showcase", "web animation", "storytelling page" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [ { "id": "web-animation-design", "name": "Web 动画设计", "description": "设计和实现自然流畅的 Web 动画效果", "slug": "web-animation-design" } ], "prompt": "帮我设计一个产品介绍的滚动叙事页面（单文件 HTML，内联无外部依赖），现代渐变风格、滚动触发动画、顶部进度条。\n\n产品：AI 写作工具 WriteFlow。\n1. Hero：视差背景 + 打字机标题\n2. 痛点：3 张卡片滚动渐入\n3. 解决方案：功能卡片左右交替滑入\n4. 数据统计：滚动到可视区数字跳动（1000 万+ 用户 / 50+ 语言 / 99.9% 可用 / 3 倍提速）\n5. 用户评价：卡片错落渐入\n6. 底部 CTA 渐变动画", "prompt_en": "Help me design a scrollytelling product-introduction page (single-file HTML, inlined with no external dependencies), with a modern gradient style, scroll-triggered animations, and a top progress bar.\n\nProduct: the AI writing tool WriteFlow.\n1. Hero: parallax background + typewriter headline\n2. Pain points: 3 cards fading in on scroll\n3. Solution: feature cards sliding in alternately from left and right\n4. Statistics: numbers counting up when scrolled into view (10M+ users / 50+ languages / 99.9% availability / 3x speed boost)\n5. Testimonials: cards fading in with a staggered layout\n6. Bottom CTA with a gradient animation", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "experts": [], "source": "official", "quality_score": 92, "sort_order": 165, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T06:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "mcps": [], "scenario": "", "task_mode": "", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "交互式滚动故事" } ], "scenario_cover": "scenario-cover.png", "is_recommended": false, "recommended_order": 0 }, { "id": "travel-footprint-map", "title": "旅行足迹打卡地图", "title_en": "Travel Footprint Check-in Map", "subtitle": "用一张地图记录你走过的山河", "subtitle_en": "Visualize every province you have explored on an interactive map", "description": "旅行爱好者常想回顾自己去过哪些地方，但散落在朋友圈和相册里的打卡记录很难形成全局视角，也无法直观看到「还有哪些地方没去」。\n\n本案例演示如何用纯 HTML/CSS/JS 构建一个中国省份打卡地图：SVG 绘制 34 个省级行政区，已去过的省份渐变高亮，点击弹出旅行记忆卡片（日期、感受、推荐指数），顶部统计面板实时展示打卡进度和覆盖率。\n\n适合旅行博主展示足迹、情侣记录共同旅程、家庭规划下一站目的地，也可作为前端交互项目的学习参考。", "description_en": "Travel enthusiasts often want to review where they have been, but scattered check-ins across social media and photo albums make it hard to see the big picture or figure out where to go next.\n\nThis case demonstrates building a China province check-in map using pure HTML/CSS/JS: SVG rendering of 34 provincial regions, gradient highlights for visited provinces, click-to-show travel memory cards with dates, impressions, and star ratings, plus a stats panel showing real-time progress and coverage.\n\nIdeal for travel bloggers showcasing footprints, couples recording shared journeys, families planning next destinations, and frontend developers learning interactive SVG techniques.", "categories": [ "travel", "creation" ], "tags": [ "旅行", "地图可视化", "SVG交互", "数据面板" ], "tags_en": [ "Travel", "Map Visualization", "SVG Interaction", "Dashboard" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个中国旅行足迹打卡地图页面（单文件 HTML，内联无外部依赖）。\n\n要求：用 SVG 绘制简化中国地图（34 个省级行政区）；已去过的 15 个省份高亮（北京/上海/广东/浙江/四川/云南/福建/江苏/湖南/陕西/重庆/海南/贵州/西藏/新疆）；点击省份弹出旅行卡片（地名/日期/一句话感受/推荐星级）；统计面板（已打卡 15/34、覆盖率 44%、总旅行 87 天）+ 进度条；已去过用渐变色、未去过用灰色。", "prompt_en": "Please create a China travel footprint check-in map page (single-file HTML, fully inline with no external dependencies).\n\nRequirements: use SVG to draw a simplified map of China (34 provincial-level administrative regions); highlight the 15 provinces already visited (Beijing / Shanghai / Guangdong / Zhejiang / Sichuan / Yunnan / Fujian / Jiangsu / Hunan / Shaanxi / Chongqing / Hainan / Guizhou / Tibet / Xinjiang); clicking a province pops up a travel card (place name / date / one-sentence impression / recommendation star rating); a stats panel (15/34 checked in, 44% coverage, 87 total travel days) + a progress bar; use a gradient color for visited regions and gray for unvisited ones.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "Full-stack frontend development with premium UI", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 90, "sort_order": 166, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T21:40:00Z", "artifacts": [ { "file": "output.html", "label": "旅行足迹地图页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "github-actions-ci-workflow", "title": "GitHub Actions CI 工作流生成", "title_en": "GitHub Actions CI Workflow Generator", "subtitle": "一键生成企业级全流程 CI/CD 管线配置", "subtitle_en": "Generate enterprise-grade full-pipeline CI/CD configuration in one shot", "description": "手动编写 GitHub Actions 工作流配置繁琐且容易遗漏安全扫描、缓存优化等关键环节，维护多环境部署流程更是耗时费力。\n\n本案例演示如何通过 github 技能一次性生成覆盖 lint、测试、构建、Docker 镜像推送、Staging 自动部署及 Production 手动审批部署的完整工作流，并内置矩阵测试、依赖缓存、安全扫描等最佳实践。\n\n适合：全栈开发者、DevOps 工程师及希望快速搭建标准化 CI/CD 管线的技术团队。", "description_en": "Manually writing GitHub Actions workflow configs is tedious and often misses critical steps like security scanning and cache optimization.\n\nThis case demonstrates how to use the github skill to generate a complete workflow covering lint, testing, build, Docker image push, automatic Staging deployment, and manual Production approval deployment, with built-in matrix testing, dependency caching, and security scanning best practices.\n\nIdeal for full-stack developers, DevOps engineers, and teams looking to quickly set up a standardized CI/CD pipeline.", "tags": [ "CI/CD", "GitHub Actions", "DevOps", "Docker", "Security" ], "tags_en": [ "CI/CD", "GitHub Actions", "DevOps", "Docker", "Security" ], "artifact_type": "code", "categories": [ "dev" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "GitHub Actions CI 工作流配置" } ], "prompt": "请为一个 Node.js + TypeScript 全栈项目生成完整的 GitHub Actions CI 工作流配置，要求覆盖：lint 检查、单元测试、构建、Docker 镜像推送、Staging 自动部署、Production 手动审批部署。要求使用缓存优化、矩阵测试（Node 18/20）、安全扫描等最佳实践。请用 github 技能输出包含完整 YAML 配置的 Markdown 文档。", "prompt_en": "Please generate a complete GitHub Actions CI workflow configuration for a Node.js + TypeScript full-stack project, covering: lint checks, unit tests, build, Docker image push, automatic Staging deployment, and manual-approval Production deployment. Use best practices such as caching optimization, matrix testing (Node 18/20), and security scanning. Please use the github skill to output a Markdown document containing the complete YAML configuration.", "skills": [ { "id": "github", "name": "github", "description": "管理 GitHub Issues、PR 和 CI", "description_en": "Manage GitHub issues, PRs, and CI runs", "skill_id": "skill_2053082027691548672" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 94, "sort_order": 167, "is_featured": false, "featured_order": 0, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "personal-portfolio-site", "title": "个人作品集网页", "title_en": "Personal Portfolio Website", "subtitle": "设计师级个人展示页，项目经历一页呈现", "subtitle_en": "Designer-quality personal showcase with projects and skills on one page", "description": "想做一个个人作品集网站，要么花大价钱请设计师，要么用模板搭出千篇一律的页面——怎么都不满意。自己写又缺设计感，配色排版一塌糊涂，项目展示毫无吸引力。\n\n这个案例演示了如何用 frontend-dev 技能一键生成一个设计师级的单页作品集网站：渐变动画 Hero 区域、个人简介与数据统计、技能雷达图/进度条、6 个项目卡片瀑布流布局（悬停叠加效果）、工作经历时间线、联系表单，还有深色/浅色主题切换和滚动动画——全部打包在一个 HTML 文件中。\n\n适合：设计师、开发者、自由职业者、求职者——任何想要快速拥有专业级个人展示页的人。", "description_en": "Building a personal portfolio either means hiring a designer at a premium or settling for cookie-cutter templates that all look the same. Coding it yourself often results in poor color choices, awkward layouts, and project showcases that fail to impress.\n\nThis case demonstrates using the frontend-dev skill to instantly generate a designer-quality single-page portfolio: animated gradient hero section, about-me with stat counters, skills radar chart and progress bars, 6 project cards in masonry grid with hover overlays, work experience timeline, styled contact form, plus dark/light theme toggle and scroll-triggered CSS animations — all in a single HTML file.\n\nPerfect for: designers, developers, freelancers, job seekers — anyone who wants a professional personal showcase page fast.", "categories": [ "creation" ], "tags": [ "作品集", "个人网站", "前端", "设计", "简历" ], "tags_en": [ "Portfolio", "Personal Website", "Frontend", "Design", "Resume" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为虚构 UI/UX 设计师「陈小宇」制作一个个人作品集网站（单文件 HTML，内联无外部依赖），界面文字用简体中文，现代简约、支持明暗主题切换和滚动动画。\n\n模块：\n1. 英雄区：动态渐变背景 + 姓名 + 职位「UI/UX 设计师 & 创意开发者」+ 下滑指示\n2. 关于我：头像占位 + 简介 + 4 个统计（5+ 年/80+ 项目/30+ 客户/15 奖项）\n3. 技能区：进度条（Figma 95% / React 85% / Node.js 75% / UI 设计 98% / 动效 80%）\n4. 项目展示：6 张卡片网格，悬停出叠加层\n5. 工作经历：4 段时间轴\n6. 联系区：表单（姓名/邮箱/留言）", "prompt_en": "Please create a personal portfolio website for a fictional UI/UX designer named \"陈小宇\" (single-file HTML, inline with no external dependencies). The interface text should be in Simplified Chinese, modern and minimalist, supporting light/dark theme toggling and scroll animations.\n\nSections:\n1. Hero: animated gradient background + name + title \"UI/UX 设计师 & 创意开发者\" + scroll-down indicator\n2. About Me: avatar placeholder + bio + 4 stats (5+ years / 80+ projects / 30+ clients / 15 awards)\n3. Skills: progress bars (Figma 95% / React 85% / Node.js 75% / UI design 98% / motion design 80%)\n4. Project Showcase: a grid of 6 cards with an overlay layer on hover\n5. Work Experience: a 4-segment timeline\n6. Contact: a form (name/email/message)", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 168, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-15T18:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "个人作品集网站" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "monthly-profit-loss-statement", "title": "财务月报/损益表", "title_en": "Monthly Profit & Loss Statement", "subtitle": "SaaS 初创公司标准损益表，收入成本一目了然", "subtitle_en": "Standard P&L statement for SaaS startups with clear revenue and cost breakdown", "description": "每月财务结账后，CFO 需要一份既能给管理层看又能给投资人看的损益表：收入要按业务线拆分、成本要细分到科目、毛利率净利率要标清楚、同比环比要对齐——手动从财务系统导数据再排版格式，光调对齐和公式就能折腾一下午。\n\n这个案例为一家 A 轮后的 SaaS 初创公司（团队 45 人、月活客户 320 家、ARR 约 1200 万）生成完整的 2026 年 3 月损益表 Excel：标准 P&L 格式包含收入明细（SaaS 订阅 98.5 万 / 定制开发 12.3 万 / 技术支持 5.8 万）、营业成本（云资源 / API / 客户成功团队）、毛利 89.3 万（76.6%）、六项运营费用明细、净利润 7.3 万；每项均含同比环比对比列，配有收入构成饼图和费用结构条形图，以及独立的经营分析 Summary sheet（11 项关键指标 + 6 条分析要点）。\n\n适合：CFO、财务团队、创业公司 CEO、投资人对接、董事会汇报、任何需要制作月度财务报告的人。", "description_en": "After monthly financial closing, the CFO needs a P&L that works for both management and investors: revenue split by business line, costs broken down by category, gross and net margins clearly marked, YoY and MoM comparisons aligned — manually exporting from accounting systems and formatting can eat up an entire afternoon.\n\nThis case generates a complete March 2026 P&L Excel for a Series A SaaS startup (45 employees, 320 active customers, ~12M ARR): standard P&L format with revenue breakdown (SaaS subscriptions 985K / custom development 123K / tech support 58K), cost of revenue (cloud/API/customer success), gross profit 893K (76.6%), six operating expense line items, net profit 73K; each with YoY and MoM comparison columns, revenue composition pie chart, expense structure bar chart, and a standalone Summary sheet (11 key metrics + 6 analysis points).\n\nIdeal for: CFOs, finance teams, startup CEOs, investor relations, board reporting, anyone producing monthly financial reports.", "categories": [ "data", "business", "finance" ], "tags": [ "损益表", "财务报表", "SaaS", "Excel", "月报", "经营分析" ], "tags_en": [ "P&L Statement", "Financial Report", "SaaS", "Excel", "Monthly Report", "Business Analysis" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "财务月报/损益表" } ], "prompt": "为一家 A 轮后的 SaaS 初创公司生成 2026 年 3 月的损益表（Excel 格式）。\n\n公司情况：团队 45 人、月活客户 320 家、ARR 约 1200 万。请填入收入、营业成本、毛利、各项运营费用、净利润（可用示例数据：总收入约 116.6 万、毛利约 89.3 万即毛利率 76.6%、净利润约 7.3 万）。\n\n要求：1）标准损益表格式；2）含同比 / 环比变化列；3）关键图表（收入构成饼图、费用结构条形图）；4）经营分析 Summary sheet。", "prompt_en": "Generate a March 2026 profit & loss statement (Excel format) for a post-Series-A SaaS startup.\n\nCompany basics: a team of 45, 320 monthly active customers, ARR of about ¥12M. Fill in revenue, cost of revenue, gross profit, operating expenses, and net profit (example figures you may use: total revenue ~¥1.166M, gross profit ~¥893K i.e. 76.6% margin, net profit ~¥73K).\n\nRequirements: 1) standard P&L format; 2) include YoY / MoM change columns; 3) key charts (revenue-composition pie chart, expense-structure bar chart); 4) a business-analysis Summary sheet.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 169, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "diagnosis-suspected-hyperthyroidism", "title": "症状与报告的疑似诊断倾向", "title_en": "Suspected Diagnosis from Symptoms & Labs", "subtitle": "还没下诊断？AI 帮你锁定高可能性疾病方向", "subtitle_en": "No diagnosis yet? Let AI point to the most likely conditions", "description": "患者还没下诊断，但症状体征和化验报告已经摆在面前：心悸、多汗、体重下降、甲状腺 II 度肿大、双手震颤阳性，FT3 / FT4 显著升高、TSH 明显降低、TRAb 升高……这种甲功五项表现，到底最可能是什么病？\n\n这个案例演示了把「主诉 / 现病史 / 查体 / 辅助检查」交给「腾讯健康 · AI 临床助手·辅助诊断」，确认无当前诊断后，它基于甲功异常 + 临床表现推导最可能的诊断，按常见病优先排序，仅输出高可能性的疑似诊断条目（不输出中 / 低可能性条目），每条附可溯源至输入信息的诊断依据，并给出建议进一步检查的项目（如甲状腺超声、摄碘率等）。\n\n行为边界：只输出高可能性疑似诊断，不输出中 / 低可能性条目；诊断依据 100% 可溯源至输入信息，绝不引用输入中不存在的指标；不延伸至具体抗甲状腺药物方案或剂量建议；末尾附安全声明，明确最终诊断以专科医生为准。\n\n适合：内分泌科初诊医生、全科医生、规培医生、医学研究生、需要了解 AI 疑似诊断推理思路的患者及家属。", "description_en": "No diagnosis yet, but the symptoms and labs are on the table — palpitations, sweating, weight loss, a grade-II goiter, fine tremor, FT3 / FT4 clearly elevated, TSH suppressed, TRAb elevated… With a thyroid panel like this, what's most likely?\n\nThis case shows how to hand the AI clinical assistant the patient's chief complaint, history, physical exam and labs. After confirming there is no current diagnosis, it reasons from thyroid abnormalities plus clinical features, lists only high-probability suspected diagnoses (no medium / low entries) in common-disease-first order, with every diagnostic clue traceable back to the input and a list of suggested follow-up tests (thyroid ultrasound, radioiodine uptake, etc.).\n\nBoundaries: high-probability diagnoses only; clues 100% traceable to the input — never invents indicators not in the record; no specific anti-thyroid drug or dosing recommendations; safety statement at the end reminding that final diagnosis rests with the specialist.\n\nFor: endocrinology first-visit physicians, GPs, residents, medical students, and patients / families wanting to understand AI diagnostic reasoning.", "categories": [ "agent", "education" ], "tags": [ "辅助诊断", "疑似诊断", "甲功异常", "AI临床助手", "腾讯健康" ], "tags_en": [ "Diagnosis Assistant", "Suspected Diagnosis", "Thyroid Function", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "症状与报告疑似诊断报告（HTML）" } ], "prompt": "请基于以下患者信息（尚未下诊断）做疑似诊断推理，列出最可能的疾病方向。\n\n【患者信息】\n[粘贴：主诉、现病史、既往史、查体、辅助检查结果]\n\n【示例场景】\n这个患者还没下诊断，请根据病历信息帮忙分析一下最可能是什么病。\n患者：女，45岁\n主诉：心悸、多汗、体重下降 2 月\n现病史：2 月前无明显诱因出现心悸，活动后加重。伴怕热多汗，食欲增加但体重下降 4 kg。无胸闷胸痛，无发热。月经周期紊乱 2 个周期。\n既往史：无特殊。无甲状腺疾病家族史。\n查体：T 37.1℃，P 108 次/分，BP 125/70 mmHg；甲状腺 II 度肿大，未闻及血管杂音，无触痛；双手细震颤阳性；无突眼。\n辅助检查：\n FT3 12.5 pmol/L ↑（参考值 3.1-6.8）\n FT4 38.2 pmol/L ↑（参考值 12-22）\n TSH 0.05 mIU/L ↓（参考值 0.27-4.2）\n TPOAb 156 IU/mL ↑（参考值 <34）\n TRAb 3.8 IU/L ↑（参考值 <1.75）\n 血常规、肝肾功能正常\n\n【输出要求】\n1. 病历文本解析：提取主诉、现病史、查体、辅助检查，确认无当前诊断\n2. 疑似诊断推理：基于化验异常 + 临床表现推导最可能的诊断\n3. 常见病优先排序：高流行病学概率的诊断排在首位\n4. 仅输出高可能性疑似诊断，每条附诊断依据 + 建议进一步检查\n5. 安全声明附在报告末尾\n\n【行为边界】\n- 只输出高可能性的疑似诊断条目，不输出中 / 低可能性条目\n- 诊断依据 100% 可溯源至输入信息\n- 不延伸至治疗方案（不给抗甲状腺药物方案或剂量建议）\n\n【不应出现】\n- 中 / 低可能性的疑似诊断条目\n- 诊断依据引用输入中不存在的信息\n- 具体的抗甲状腺药物方案或剂量建议\n- 省略安全声明", "prompt_en": "Based on the following patient information (no diagnosis yet), reason about the most likely suspected diagnosis and point to the leading condition.\n\n【Patient information】\n[Paste: chief complaint, history of present illness, past history, physical exam, workup results]\n\n【Example scenario】\nThis patient has not been diagnosed yet. Based on the record below, please help analyze what the most likely condition is.\nPatient: female, 45 y/o\nChief complaint: palpitations, sweating, weight loss for 2 months\nHistory of present illness: palpitations appeared 2 months ago without obvious trigger, worsening with activity, accompanied by heat intolerance and sweating, increased appetite with 4 kg weight loss, no chest pain, no fever, menstrual cycle irregular for 2 cycles.\nPast history: unremarkable; no family history of thyroid disease.\nPhysical exam: T 37.1℃, P 108 bpm, BP 125/70 mmHg; grade-II goiter without bruit or tenderness; fine tremor of both hands positive; no exophthalmos.\nWorkup: FT3 12.5 pmol/L ↑ (ref 3.1-6.8); FT4 38.2 pmol/L ↑ (ref 12-22); TSH 0.05 mIU/L ↓ (ref 0.27-4.2); TPOAb 156 IU/mL ↑ (ref <34); TRAb 3.8 IU/L ↑ (ref <1.75); CBC and liver / kidney function normal.\n\n【Output requirements】\n1. Record parsing: extract chief complaint, history, physical exam, workup; confirm no current diagnosis\n2. Suspected-diagnosis reasoning: derive the most likely diagnosis from lab abnormalities plus clinical features\n3. Common-disease-first ranking: high epidemiological probability ranked first\n4. Output high-probability suspected diagnoses only, each with diagnostic clues + suggested follow-up tests\n5. Safety statement appended at the end\n\n【Boundaries】\n- High-probability diagnoses only; no medium / low entries\n- Clues 100% traceable to the input\n- No treatment plans (no anti-thyroid drug regimen or dosing)\n\n【Must not appear】\n- Medium / low-probability suspected-diagnosis entries\n- Diagnostic clues referencing information not present in the input\n- Specific anti-thyroid drug regimens or dosing\n- A missing safety statement", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 170, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:55:11+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "industrial-product-commercial", "title": "工业产品商业宣传片", "title_en": "Industrial Product Commercial", "subtitle": "从外观、端口到内部电路完整讲清一台专业设备", "subtitle_en": "A professional device explained from chassis and ports to internal circuitry", "description": "工业产品信息密集，单纯旋转展示无法同时讲清结构、接口和核心功能。\n\n案例以黑色机架式音频/网络设备为主角，通过轮廓光、RGB 接口、爆炸结构、电路板和手部调节展示产品设计与信号能力。\n\n适合：专业硬件、制造设备、汽车零部件、消费电子和发布会产品片。", "description_en": "Information-dense industrial products cannot explain form, ports, and core functions through a simple rotating beauty shot.\n\nA black rack-mounted audio and network device is presented through rim light, color-coded ports, exploded construction, circuitry, and hands-on control shots.\n\nBest for professional hardware, manufacturing equipment, automotive components, consumer electronics, and launch films.", "categories": [ "creation" ], "tags": [ "工业产品", "硬件广告", "科技视觉" ], "tags_en": [ "industrial product", "hardware ad", "technology visual" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「商业广告创作」专家调用 Flova 的「工业产品商业宣传片」Skill，依据我上传的黑色 1U 机架式音频网络接口产品图和参数表，制作一支 30 秒、16:9、冷峻工业风、1080p 的商业宣传片《机架音频接口科技片》，使用沉稳中文男声与对应字幕。核心信息只使用参数表中的真实内容：前面板旋钮与状态屏、背部彩色网络/音频端口、低延迟信号路由和模块化内部电路。请先一次确认时长、画幅、风格、画质、旁白与字幕，再锁定全局规格并分析产品覆盖度；建立并确认产品资产后，依次生成黑场轮廓、端口亮起、爆炸结构、电路信号流、手部调节和英雄定帧。最后生成旁白、低沉电子工业 BGM、字幕并组装；产品端口数量、标签、比例和颜色必须以上传素材为准。", "prompt_en": "Ask the “Flova Commercial Director” expert to invoke Flova’s “Industrial Product Commercial” Skill and use my uploaded images and specification sheet for a black 1U rack-mounted audio network interface to create a 30-second 16:9 commercial titled Technology Film for a Rack Audio Interface. Use a cold industrial look, 1080p delivery, a calm Chinese male narrator, and matching subtitles. Use only verified details from the specification sheet: front-panel controls and status display, color-coded network and audio ports on the rear, low-latency signal routing, and modular internal circuitry. Confirm duration, aspect ratio, visual style, quality, narration, and subtitles together first, then lock the global specification and assess product coverage. After the product assets are approved, generate the dark silhouette, illuminated ports, exploded structure, circuit signal flow, hand adjustment, and hero frame in sequence. Finish with narration, deep electronic industrial music, subtitles, and assembly. Port count, labels, proportions, and colors must follow the uploaded materials exactly.", "skills": [], "experts": [ { "id": "flova-commercial-director", "name": "商业广告创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及商品广告成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "9c6b2e4c2bdb", "expert_id": "ex_9oFsxT6rZOEG" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 170, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-apply-qa", "title": "报考答疑：三方社保与应届身份", "title_en": "Application Q&A: Contracts, Social Security & Status", "subtitle": "知识库检索原文作答，一条社保记录就能改变应届身份", "subtitle_en": "Answered from the knowledge base — one social-security record can change your status", "description": "「我还算应届生吗」是报考咨询里被问得最多、也最不能答错的问题——答错直接影响能报哪些岗。\n\n本案例围绕「三方协议、劳动合同、社保对应届身份的影响」检索高顿公考知识库，命中 4 条原文问答（匹配度 0.87–0.90）：三方协议不影响应届身份、签劳动合同并缴社保即失去应届身份、择业期认定等，并把 4 条串成一条判断线。知识库内部库名已脱敏。\n\n适合：毕业前后状态变动过的考生、拿不准基层年限起算的考生。", "description_en": "\"Am I still a fresh graduate?\" is the most-asked and least-forgiving question in application counselling — getting it wrong changes which posts you can apply for.\n\nThis case searches Gaodun's civil-exam knowledge base on how tripartite agreements, labour contracts and social security affect fresh-graduate status, hitting four verbatim Q&As (match 0.87–0.90): a tripartite agreement alone does not end the status; a signed labour contract plus social-security payment does; the two-year job-hunting window rule — then threads them into one decision line. Internal knowledge-base names are masked.\n\nFor: candidates whose status changed around graduation and anyone unsure when grassroots tenure starts.", "categories": [ "search", "education" ], "tags": [ "报考答疑", "应届生", "三方协议", "社保" ], "tags_en": [ "application Q&A", "fresh graduate", "tripartite agreement", "social security" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "应届身份知识库问答" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我想搞清楚签三方协议、签劳动合同、缴社保分别对应届毕业生身份有什么影响。要求：1) 从公考知识库实时检索相关问答，给出匹配度；2) 逐条引用知识库原文回答，不要自己编；3) 把几条答案串成一条完整的判断逻辑；4) 提醒我最终认定以招录机关为准。", "prompt_en": "Clarify how signing a tripartite agreement, signing a labour contract, and paying social security each affect fresh-graduate status. Requirements: 1) live-search the civil-exam knowledge base and show match scores; 2) answer by quoting the knowledge-base entries verbatim, no improvisation; 3) thread the answers into one coherent decision logic; 4) remind me the hiring authority has the final say.", "skills": [ { "id": "gk-answer", "name": "报考答疑", "description": "体制内 14 类考试报考答疑，基于知识库按判定优先级作答", "skill_id": "skill_2089643028823928832" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 170, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "vue3-component-dev", "title": "Vue3 任务管理组件开发", "title_en": "Vue3 Task Management Component Development", "subtitle": "拖拽看板组件，待办/进行中/已完成三列布局", "subtitle_en": "Drag-and-drop Kanban board with To Do / In Progress / Done columns", "description": "任务管理是团队协作的核心需求，但从零开发一个带拖拽功能的看板组件工作量不小——要处理拖拽 API、状态管理、表单交互、筛选逻辑等。\n\n这个案例用前端开发技能生成了一个功能完备的任务看板 HTML 页面：三列布局（待办/进行中/已完成），支持 HTML5 原生拖拽在列间移动卡片，每张卡片含标题、描述、优先级标签和负责人头像。内置新增和编辑任务的表单弹窗，支持按优先级和关键词筛选搜索。纯 HTML/CSS/JS 单文件实现。\n\n适合：前端开发者学习拖拽交互、团队任务管理工具原型、项目管理组件开发。", "description_en": "Task management is a core team collaboration need, but building a drag-and-drop Kanban from scratch involves significant work — handling drag APIs, state management, form interactions, filtering logic.\n\nThis case uses the frontend-dev skill to generate a fully-featured task board HTML page: three-column layout (To Do / In Progress / Done), native HTML5 drag-and-drop between columns, cards with title, description, priority tags and assignee avatars. Built-in add/edit task modal forms, priority and keyword filtering. Single-file HTML/CSS/JS.\n\nBest for: Frontend developers learning drag interactions, task management tool prototypes, project management components.", "categories": [ "dev", "creation" ], "tags": [ "Vue3", "TypeScript", "组件开发", "看板", "拖拽" ], "tags_en": [ "Vue3", "TypeScript", "Component Dev", "Kanban", "Drag and Drop" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请开发一个任务管理看板页面（单文件 HTML，内联无外部依赖），现代设计、浅色主题、响应式。\n\n要求：三列看板（待办/进行中/已完成）、HTML5 原生拖拽跨列移动、每张卡片含标题+描述摘要+优先级标签（高红/中橙/低绿）+负责人头像（首字母圆形）、新增/编辑任务弹窗（标题/描述/优先级/负责人）、搜索框+优先级筛选、预填 6-8 个示例任务。", "prompt_en": "Please develop a task management Kanban board page (single-file HTML, inline with no external dependencies), with modern design, a light theme, and responsive. \n\nRequirements: three columns (To Do / In Progress / Done), native HTML5 drag-and-drop to move items across columns, each card containing a title + description summary + priority tag (high red / medium orange / low green) + owner avatar (a circle with the first initial), an add/edit task modal (title / description / priority / owner), a search box + priority filter, and 6-8 prefilled example tasks.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 170, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "Vue3 组件" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-qa-grassroots", "title": "报考答疑：基层工作经历认定", "title_en": "Application Q&A: What Counts as Grassroots Work Experience", "subtitle": "应届生算不算、当兵算不算、没交社保算不算——三条官方口径", "subtitle_en": "Do new graduates, military service, or unpaid-insurance jobs count?", "description": "职位表写「需 2 年基层工作经历」，可什么算基层、从什么时候起算，公告里往往一句带过。\n\n本案例从报考知识库检索「基层工作经历如何认定」，命中 3 条匹配度均在 0.86 以上的问答，逐条给出原文口径，并归纳成身份、经历类型、凭证三个判断维度，最后说明这些是一般性口径、具体认定以报考单位和当年公告为准。\n\n适合：报考限基层经历岗位的考生、退役军人考生、社保记录不完整的人。", "description_en": "The post table says \"two years of grassroots experience required\", but what counts as grassroots and when the clock starts is barely explained.\n\nThis case searches the application knowledge base for how grassroots experience is recognised, returning 3 Q&A entries all matching above 0.86. Each is quoted verbatim, then organised into three judgement dimensions — status, experience type and evidence — with a clear note that final recognition rests with the recruiting unit.\n\nFor: candidates applying to grassroots-restricted posts, veterans, and those with incomplete social-insurance records.", "categories": [ "search", "education" ], "tags": [ "报考答疑", "基层工作经历", "知识库检索", "报考资格" ], "tags_en": [ "application Q&A", "grassroots experience", "knowledge base", "eligibility" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "基层工作经历认定问答与判定维度", "label_en": "Grassroots experience Q&A and judgement dimensions" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查清楚公考里的「基层工作经历」到底怎么认定。要求：1) 从报考知识库真实检索并给出命中条数；2) 逐条列出问答原文和匹配度分值，不要改写内容；3) 把这几条口径归纳成几个判断维度（比如身份、经历类型、凭证），每个维度说清楚判定要点；4) 说明匹配度阈值的含义、这些是否属于一般性口径；5) 明确提示具体认定以报考单位和当年公告为准，不要替考生下结论。", "prompt_en": "Clarify how \"grassroots work experience\" is recognised in civil-service applications. Requirements: 1) search the application knowledge base and report the number of hits; 2) quote each Q&A verbatim with its match score — do not rewrite; 3) organise the answers into judgement dimensions (status, experience type, evidence) and state the key test for each; 4) explain what the match threshold means and whether these are general guidelines; 5) state clearly that final recognition rests with the recruiting unit and the year's announcement.", "skills": [ { "id": "gk-answer", "name": "报考答疑", "description": "体制内 14 类考试报考答疑，基于知识库按判定优先级作答", "skill_id": "skill_2089643028823928832" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 171, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "product-architecture-diagram", "title": "产品技术架构图", "title_en": "Product Technical Architecture Diagram", "subtitle": "用 Mermaid 一键生成专业级 SaaS 微服务架构图", "subtitle_en": "Generate a professional SaaS microservice architecture diagram with Mermaid in one click", "description": "当团队需要梳理系统架构时，总会遇到这些问题：手动画图耗时且难以维护，Visio/Draw.io 导出的图片模糊不清，架构变更后文档永远跟不上代码，新人入职看不懂系统全貌。一张清晰、可维护的架构图是技术团队的刚需。\n\n这个案例用 mermaid-diagram 技能为一个典型的 SaaS 产品生成了完整的微服务架构图：从客户端（Web/Mobile）到 API 网关（Nginx/Kong），核心业务服务（认证、用户、订单、商品、支付、通知）横向展开，底层基础设施（Kafka 消息队列、Redis 缓存、PostgreSQL 数据库）纵深排列，右侧监控体系（Prometheus + Grafana + ELK）全覆盖，最底层 Docker/Kubernetes 容器编排一目了然。用文字描述架构，Mermaid 自动渲染，版本可控，随时更新。\n\n适合：后端工程师、架构师、技术负责人、DevOps 工程师、需要在技术评审或方案汇报中展示系统架构的任何人", "description_en": "When teams need to map out system architecture, they always face the same problems: manually drawing diagrams is time-consuming and hard to maintain, exported images from Visio/Draw.io are blurry, documentation never keeps up with code changes after architecture updates, and new hires can't grasp the full system picture. A clear, maintainable architecture diagram is a must-have for any technical team.\n\nThis case uses the mermaid-diagram skill to generate a complete microservice architecture diagram for a typical SaaS product: from client layer (Web/Mobile) through API Gateway (Nginx/Kong), core business services (Auth, User, Order, Product, Payment, Notification) spread horizontally, infrastructure layer (Kafka message queue, Redis cache, PostgreSQL databases) in depth, monitoring stack (Prometheus + Grafana + ELK) on the side, and Docker/Kubernetes orchestration at the bottom. Describe architecture in text, Mermaid renders automatically — version-controlled and always up to date.\n\nIdeal for: backend engineers, architects, tech leads, DevOps engineers, anyone who needs to present system architecture in technical reviews or proposals", "categories": [ "dev", "creation" ], "tags": [ "架构图", "微服务", "Mermaid", "SaaS", "DevOps" ], "tags_en": [ "Architecture Diagram", "Microservices", "Mermaid", "SaaS", "DevOps" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "微服务架构图" } ], "prompt": "请生成一个 SaaS 产品微服务架构图：单文件 HTML，架构图用内联 SVG，无外部依赖、纯 CSS 交互（零 JS），界面文字用简体中文，暗色现代风格，六层从上到下分层展示，层间连线标注协议。\n\n1. 客户端层：Web App、移动端、桌面端、Open API\n2. 网关层：API Gateway、负载均衡、限流器、认证网关\n3. 核心服务层：用户、订单、支付、商品、通知、搜索六大微服务，通过 Event Bus 异步通信\n4. 基础设施层：PostgreSQL、Redis、Elasticsearch、RabbitMQ、MinIO\n5. 监控层：Prometheus、Grafana、ELK、Sentry、Jaeger\n6. 部署层：Kubernetes、Docker、Terraform、CI/CD、CDN\n\n附等效 Mermaid 源码供复用。", "prompt_en": "Please generate a SaaS product microservice architecture diagram: single-file HTML, with the diagram as inline SVG, no external dependencies, pure-CSS interactions (zero JS), UI text in Simplified Chinese, dark modern style, six layers displayed top to bottom, with protocols labeled on the connections between layers.\n\n1. Client layer: Web App, Mobile, Desktop, Open API\n2. Gateway layer: API Gateway, load balancer, rate limiter, auth gateway\n3. Core services layer: six microservices—User, Order, Payment, Product, Notification, Search—communicating asynchronously via an Event Bus\n4. Infrastructure layer: PostgreSQL, Redis, Elasticsearch, RabbitMQ, MinIO\n5. Monitoring layer: Prometheus, Grafana, ELK, Sentry, Jaeger\n6. Deployment layer: Kubernetes, Docker, Terraform, CI/CD, CDN\n\nAttach equivalent Mermaid source code for reuse.", "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表生成器", "description": "从文本描述生成 Mermaid 流程图、架构图、时序图等", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 171, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-04-22T06:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "candidate-interview-report-docx", "title": "候选人面试评估报告", "title_en": "Candidate Interview Evaluation Report", "subtitle": "多轮面试记录自动整理成可归档的正式 Word 评估报告", "subtitle_en": "Turn multi-round interview notes into a formal Word evaluation report ready for filing", "description": "候选人面完以后最麻烦的不是打分，而是把几轮面试意见、证据和录用建议整理成一份能归档、能审批、还能复盘的正式文档。\n\n这个案例演示了如何用 Word 文档生成技能生成一份结构化 Word 面试评估报告：封面、目录、候选人信息、能力评分矩阵、逐轮面试纪要、优势与风险、定级与录用建议、签字区一次成稿，适合直接发给招聘经理和用人部门负责人。\n\n适合：招聘经理、HRBP、用人经理、面试官协同场景、需要沉淀正式人才评估文档的团队。", "description_en": "The hardest part after interviews is not scoring the candidate — it is consolidating multiple rounds of feedback, evidence, and hiring recommendations into a formal document that can be archived, approved, and referenced later.\n\nThis case demonstrates how to use the Word Document Generation skill to generate a structured Word interview evaluation report with a cover page, table of contents, candidate profile, competency score matrix, round-by-round notes, strengths and risks, leveling and hiring recommendation, and signature areas — ready to send to recruiting and hiring managers.\n\nBest for: Recruiting managers, HRBPs, hiring managers, interviewer collaboration workflows, and teams that need formal talent evaluation documentation.", "categories": [ "office", "business", "productivity" ], "tags": [ "面试评估", "招聘", "Word", "能力矩阵", "录用建议" ], "tags_en": [ "Interview Review", "Hiring", "Word", "Competency Matrix", "Hiring Decision" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "候选人面试评估报告（Word）" } ], "prompt": "请生成一份正式的 Word 文档（output.docx），标题《候选人面试评估报告》。\n\n候选人 / 目标岗位 / 目标级别 / 面试轮次与各轮信息（替换为你的候选人信息）：\n...\n\n请整理成适合归档与审批的正式 Word 报告，须含封面页、自动目录、页眉页脚、页码、多级标题编号、统一表格样式、签字区。\n\n内容至少包括：候选人基本信息、流程概览、能力评分矩阵、逐轮面试纪要、优势与风险、录用与定级建议、入职 30 天关注点、签批结论。\n\n语言专业、客观、可复用。", "prompt_en": "Please generate a formal Word document (output.docx) titled \"Candidate Interview Evaluation Report.\"\n\nCandidate / target position / target level / interview rounds and per-round notes (replace with your own candidate info):\n...\n\nOrganize it into a formal Word report suitable for archiving and approval, which must include a cover page, an automatic table of contents, headers and footers, page numbers, multi-level heading numbering, a consistent table style, and a signature area.\n\nContent must include at least: candidate basic information, process overview, competency score matrix, round-by-round notes, strengths and risks, hiring and leveling recommendations, focus areas for the first 30 days, and the approval conclusion.\n\nWritten in professional, objective, reusable language.", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 172, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:03:53Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-qa-fresh-status", "title": "报考答疑：应届生身份的四类边界", "title_en": "Application Q&A: Four Edge Cases for New-Graduate Status", "subtitle": "签了合同、当过兵、做过基层项目、留学回国，分别还算不算应届生", "subtitle_en": "Signed a contract, served in the military, joined a grassroots programme, studied abroad — still eligible?", "description": "「限应届生」是最常见的岗位限制，真正卡人的从来不是常规情形，而是那些看起来不该算、实际却算（或反过来）的边界情况。\n\n本案例做三次定向检索——常规就业、服兵役与基层项目、留学回国，逐条列出知识库原文与匹配度，整理成一张覆盖七种情形的判定表，并指出「未缴社保」与「择业期」两条口径之间的张力该怎么区分。\n\n适合：退役士兵、基层项目服务期满人员、留学回国求职者，以及所有不确定自己是否还算应届生的人。", "description_en": "\"New graduates only\" is the most common restriction, and what trips people up is never the ordinary case — it's the edges that look disqualifying but aren't, or the reverse.\n\nThis case runs three targeted searches — ordinary employment, military service and grassroots programmes, and returning from study abroad — quoting each knowledge-base answer verbatim with its match score, organising them into a seven-scenario decision table, and flagging the tension between the \"no social insurance\" and \"job-seeking window\" rulings.\n\nFor: veterans, people finishing grassroots service, returning overseas graduates, and anyone unsure whether they still qualify.", "categories": [ "search", "education" ], "tags": [ "报考答疑", "应届生身份", "退役士兵", "留学回国" ], "tags_en": [ "application Q&A", "new graduate status", "veterans", "overseas graduates" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "应届生身份四类边界情形问答与判定表", "label_en": "New-graduate status edge cases and decision table" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查清楚公考里的「应届毕业生身份」在几类特殊情形下怎么界定。要求：1) 分三次真实检索报考知识库——常规就业与择业期、服兵役退役士兵、留学回国人员，各给出命中条数；2) 按情形分组，逐条列出问答原文和匹配度分值，不要改写；3) 把三组口径整理成一张判定表，逐行给出「是否可报限应届岗位」与依据；4) 如果不同条目之间的口径存在张力（如未缴社保 vs 择业期），明确指出并说明区分点；5) 说明哪些返回条目与主题无关、未被采用，并提醒各省口径可能有差异、以当年公告为准。", "prompt_en": "Clarify how \"new graduate\" status is defined for civil-service applications in several edge cases. Requirements: 1) run three targeted knowledge-base searches — ordinary employment and the job-seeking window, military veterans, and returning overseas graduates — reporting the hit count for each; 2) group the results by scenario and quote each Q&A verbatim with its match score, without rewriting; 3) organise all three groups into one decision table giving the verdict and basis per row; 4) if answers conflict (e.g. no social insurance versus the job-seeking window), say so and explain the distinguishing factor; 5) state which returned entries were off-topic and excluded, and note that provincial rules vary and the year's announcement governs.", "skills": [ { "id": "gk-answer", "name": "报考答疑", "description": "体制内 14 类考试报考答疑，基于知识库按判定优先级作答", "skill_id": "skill_2089643028823928832" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 172, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "web-component-kanban", "title": "可拖拽项目管理看板组件", "title_en": "Draggable Project Kanban Board Component", "subtitle": "三列看板 x 拖拽排序 x localStorage 持久化", "subtitle_en": "3-Column Board x Drag & Drop x localStorage Persistence", "description": "项目管理中最常见的痛点之一，就是任务状态散落在聊天群、文档和脑海里，团队成员对「谁在做什么、做到哪了」缺乏统一视图。传统的电子表格或待办清单虽然能记录任务，却无法直观呈现工作流转。\n\n这个案例使用前端开发技能，生成了一个功能完整的三列看板组件（待办/进行中/已完成）。支持 HTML5 原生拖拽在列间移动卡片，每张卡片包含标题、描述摘要、优先级标签和负责人头像。内置新增、编辑、删除卡片的弹窗表单，所有数据自动持久化到 localStorage。界面采用现代设计语言，支持响应式布局。\n\n适合：个人项目管理、小团队协作看板、前端组件集成。", "description_en": "One of the most common pain points in project management is task status scattered across chat groups, documents, and people's heads — team members lack a unified view of who is doing what.\n\nThis case uses the frontend-dev skill to generate a fully functional three-column Kanban component (To Do / In Progress / Done). It supports native HTML5 drag-and-drop, each card includes title, description, priority tag, and assignee avatar. Built-in modal forms for CRUD operations, with localStorage persistence. Modern design, responsive layout.\n\nIdeal for: Personal project management, small team boards, frontend component integration.", "categories": [ "dev", "creation" ], "tags": [ "看板", "拖拽", "Web组件", "项目管理", "前端开发" ], "tags_en": [ "kanban", "drag and drop", "web component", "project management", "frontend" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请开发一个可拖拽的项目管理看板（单文件 HTML，内联无外部依赖），界面文字用简体中文（列名「待办/进行中/已完成」，优先级「高/中/低」），现代设计、响应式。\n\n功能：3 列、卡片拖拽移动、新增/编辑/删除卡片（卡片含标题+描述+优先级标签+负责人头像）、数据持久化到 localStorage。", "prompt_en": "Please develop a draggable project-management kanban board (single-file HTML, inlined with no external dependencies). UI text should be in Simplified Chinese (column names \"待办/进行中/已完成\", priorities \"高/中/低\"), with a modern design and responsive layout.\n\nFeatures: 3 columns, drag-and-drop card movement, add/edit/delete cards (each card contains a title + description + priority tag + assignee avatar), and data persistence to localStorage.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 173, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "Web Component 看板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "research-collab-office-competitors-2026", "title": "协同办公赛道竞品调研 2026", "title_en": "Enterprise Collaboration Competitor Analysis 2026", "subtitle": "10 家竞品 × 5 维度 × 完整 SWOT，含 SVG 雷达图与 5 条差异化切入机会", "subtitle_en": "10 competitors × 5 dimensions × full SWOT, with SVG radar and 5 differentiation opportunities", "description": "要进协同办公赛道，最怕的不是没点子，而是开战略会时被问：'飞书和 Teams 到底谁的 AI 更强？Notion 为什么卡在 SMB？我们到底从哪切？'——没有一份横贯全量玩家的完整竞品调研，这些问题根本没法拍板。\n\n这个案例演示了深度研究团队怎么在十几分钟内交付一份战略决策级竞品调研：Markdown 主报告 ~16,000 中文字、10 家竞品（飞书/钉钉/企业微信/Teams/Slack/Notion/Google Workspace/ClickUp/Asana/Monday）×5 维度（功能/定价/规模/增长策略/优劣势）全覆盖、10 张结构化对比表、完整 SWOT 矩阵、5 条差异化机会点（含客群/防御难度/落地路径/定价建议）；配套 HTML 可视化版含三类纯 CSS/SVG 可视化——10×8 功能对比矩阵、内联 SVG 五维雷达图（无任何 CDN/JS）、10 张 SWOT 2×2 四宫格卡片。\n\n适合：准备进入 SaaS 赛道的产品 VP / 战投经理 / 创业者；做年度战略规划的 CEO / CMO / COO；行业研究员；做客户画像或竞争情报的商分团队。", "description_en": "Entering the collaboration SaaS market? The scariest question in a strategy meeting isn't 'do we have ideas' — it's 'so between Lark and Teams, whose AI is stronger? Why does Notion plateau in SMB? Where exactly do we cut in?' Without a full-stack competitor scan, these questions can't be answered.\n\nThis case shows the Deep Research Team delivering a boardroom-grade competitor analysis in minutes: a 16k-character Markdown report covering 10 competitors (Lark, DingTalk, WeCom, Teams, Slack, Notion, Google Workspace, ClickUp, Asana, Monday) across 5 dimensions (functionality, pricing, scale, growth strategy, S/W), 10 structured comparison tables, full SWOT matrix, and 5 specific differentiation opportunities (each with target segment, defense difficulty, GTM path, pricing). A companion HTML report includes three pure-CSS/SVG visualizations — a 10×8 feature matrix, an inline SVG 5-dimension radar chart (no CDN, no JS), and 10 SWOT 2×2 cards.\n\nFit for: SaaS product VPs evaluating new segments, strategy investors, founders, annual-plan CEOs/CMOs/COOs, industry analysts, and competitive intelligence teams.", "categories": [ "business", "data", "search" ], "tags": [ "竞品调研", "SaaS", "协同办公", "SWOT 分析", "战略决策" ], "tags_en": [ "Competitive Analysis", "SaaS", "Collaboration", "SWOT", "Strategy" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.md", "artifacts": [ { "file": "report.md", "label": "Markdown 主报告（1.6 万字 · 10 家全量覆盖）" }, { "file": "visual-report.html", "label": "HTML 可视化版（矩阵+SVG雷达+SWOT 宫格）" } ], "prompt": "我们公司准备进入企业协同办公赛道，请用【深度研究团队 5 阶段完整模式】产出一份《2026 企业协同办公赛道竞品调研报告》。\n\n覆盖 8–10 家竞品：国内（飞书、钉钉、企业微信、WPS 协作等）与海外（Microsoft Teams、Slack、Notion、Google Workspace、ClickUp、Asana、Monday 等）。\n\n五大维度：功能、定价、用户规模、增长策略、优劣势。\n\n报告结构：引言 / 赛道总览 / 竞品横评 / 定价与商业模式 / 用户规模与增长策略 / SWOT 矩阵汇总 / 差异化机会点（5 条具体建议）/ 结论 / 参考文献。\n\n要求：至少 3 张核心对比表（功能矩阵 / 定价 / 用户规模）；每家完整 SWOT；5 条差异化机会点每条含机会描述、适合切入客群、竞品防御难度、落地路径。\n\n额外交付 HTML 可视化版（单页内联样式，无 CDN）：纯 CSS 竞品对比矩阵、内联 SVG 五维雷达图、各家 SWOT 四宫格卡片。", "prompt_en": "Our company is preparing to enter the enterprise collaboration office space. Please use the [Deep Research Team 5-Phase Complete Mode] to produce a \"2026 Enterprise Collaboration Office Competitive Research Report.\"\n\nCover 8-10 competitors: domestic (Feishu, DingTalk, WeCom, WPS Collaboration, etc.) and overseas (Microsoft Teams, Slack, Notion, Google Workspace, ClickUp, Asana, Monday, etc.).\n\nFive dimensions: features, pricing, user scale, growth strategy, strengths/weaknesses.\n\nReport structure: Introduction / Market Overview / Competitor Comparison / Pricing & Business Models / User Scale & Growth Strategy / SWOT Matrix Summary / Differentiation Opportunities (5 specific recommendations) / Conclusion / References.\n\nRequirements: at least 3 core comparison tables (feature matrix / pricing / user scale); a full SWOT for each company; each of the 5 differentiation opportunities includes an opportunity description, suitable target segment, competitor defense difficulty, and implementation path.\n\nAdditionally deliver an HTML visualization version (single page, inline styles, no CDN): a pure-CSS competitor comparison matrix, an inline SVG 5-dimension radar chart, and SWOT 2×2 cards for each company.", "skills": [], "experts": [ { "id": "GPTResearcherTeam", "name": "深度研究团队", "description": "深度研究报告输出，7角色5阶段聚合多源信息，经审稿修订循环输出带引用的专业报告", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "a8da6f2dc1f7", "name_en": "Deep Research Team", "description_en": "Deep research report generation: 7 roles in 5 phases aggregate multi-source data through review-revise loop into cited professional reports", "expert_id": "ex_xdj3MuRUqbR0" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 174, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T03:10:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "product-changelog-page", "title": "产品更新日志页面", "title_en": "Product Changelog Page", "subtitle": "输入版本更新要点，生成 Linear/Notion 风格的精美 Changelog", "subtitle_en": "Input version updates, generate a beautiful Linear/Notion-style Changelog page", "description": "SaaS 产品每次发版都要写更新日志，但大多数团队要么丢一个 Markdown 文件凑合，要么在微信群里发条消息就算通知了。用户根本不知道产品在进步，续费时觉得\"好像一年没什么变化\"。而 Linear、Notion、Vercel 这些产品的 Changelog 页面设计精美、信息层次清晰，是用户信任感和品牌调性的重要组成部分。\n\n这个案例演示了输入三个版本的更新要点，AI 自动生成一个单文件 HTML 页面：深色主题、时间线布局、版本标签（新功能/改进/修复分类）、渐变色卡片、平滑滚动动画。参考 Linear 和 Notion 的设计语言，开箱即用，直接部署到官网 /changelog 路径。\n\n适合：SaaS 产品团队、独立开发者、开源项目维护者、任何需要展示产品更新记录的团队。", "description_en": "Every SaaS release needs a changelog, but most teams either dump a Markdown file or send a WeChat message. Users don't see the product evolving and think \"nothing changed this year\" at renewal time. Meanwhile, Linear, Notion, and Vercel have beautifully designed Changelog pages that build trust and brand identity.\n\nThis case demonstrates inputting three versions of release notes, with AI auto-generating a single-file HTML page: dark theme, timeline layout, version tags (new features/improvements/fixes), gradient cards, smooth scroll animations. Inspired by Linear and Notion's design language, ready to deploy at /changelog.\n\nBest for: SaaS product teams, indie developers, open-source maintainers, any team needing to showcase product updates.", "categories": [ "dev", "creation" ], "tags": [ "Changelog", "产品更新", "SaaS", "开发者工具" ], "tags_en": [ "Changelog", "Product Updates", "SaaS", "Developer Tools" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为一款 SaaS 产品生成一个产品更新日志（Changelog）页面（单文件 HTML，内联无外部依赖），界面文字用简体中文，风格参考 Linear / Vercel：时间线布局、按「新功能 / 改进 / 修复」分类的版本标签。\n\n展示最近三个版本。各版本更新要点（粘贴你的版本号、日期与更新条目）：\n...", "prompt_en": "Please generate a product changelog page for a SaaS product (single-file HTML, inline with no external dependencies), with interface text in Simplified Chinese, in a style referencing Linear / Vercel: timeline layout, version tags categorized as \"New Features / Improvements / Fixes.\"\n\nDisplay the three most recent versions. Update highlights per version (paste your version numbers, dates, and entries):\n...", "skills": [ { "id": "landing-page-design", "name": "落地页设计", "description": "高转化落地页与活动页面快速设计", "slug": "landing-page-design" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 175, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T10:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.html", "label": "产品更新日志页面" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "svg-data-infographic", "title": "SVG 数据信息图", "title_en": "SVG Data Infographic", "subtitle": "纯矢量数据可视化海报，高清无损任意缩放", "subtitle_en": "Pure vector data visualization poster, lossless at any scale", "description": "做数据信息图通常需要 Photoshop、Illustrator 等专业设计软件，学习门槛高，改个数据还得重新排版调样式，效率极低。\n\n这个案例演示了如何用纯 SVG 代码生成一张「2024 全球 AI 行业数据全景」可视化海报。包含 4 大核心指标卡片（市场规模 $2000 亿、年增长率 36.8%、融资笔数 3,200+、从业人员 2,200 万）、Top 5 AI 应用领域柱状图（自然语言处理 / 计算机视觉 / 推荐系统 / 自动驾驶 / 医疗 AI）、全球投资分布饼图（北美 45% / 中国 30% / 欧洲 15% / 其他 10%）、2019-2024 市场规模增长趋势线，所有图表纯 SVG 内联绘制，无任何外部依赖，矢量无损可任意缩放。\n\n适合：数据分析师制作行业报告、市场研究员做趋势分析展示、内容运营设计数据海报、需要做数据汇报演示的任何人。", "description_en": "Creating data infographics typically requires professional design software like Photoshop or Illustrator — steep learning curve, and changing a single number means re-adjusting layouts and styles, extremely inefficient.\n\nThis case demonstrates generating a '2024 Global AI Industry Data Panorama' visualization poster using pure SVG code. It includes 4 key metric cards (Market Size $200B, YoY Growth 36.8%, Funding Deals 3,200+, Workforce 22M), a Top 5 AI Application Areas bar chart (NLP / Computer Vision / Recommendation Systems / Autonomous Driving / Medical AI), a global investment distribution pie chart (North America 45% / China 30% / Europe 15% / Others 10%), and a 2019-2024 market growth trend line — all charts drawn with pure inline SVG, zero external dependencies, vector-lossless at any scale.\n\nIdeal for: data analysts creating industry reports, market researchers presenting trend analysis, content operators designing data posters, anyone who needs to make data-driven presentations.", "categories": [ "data", "creation" ], "tags": [ "信息图", "数据可视化", "SVG", "海报设计", "AI行业" ], "tags_en": [ "Infographic", "Data Visualization", "SVG", "Poster Design", "AI Industry" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请用纯内联 SVG 生成一张「2024 全球 AI 行业数据全景」信息图海报（约 800×1200，单文件 HTML，无外部依赖）。\n\n内容：\n1. 标题区：「2024 全球 AI 行业数据全景」\n2. 四大核心数字：市场规模 $2000 亿、年增长 36.8%、融资 3,200+ 笔、从业 2,200 万人\n3. 柱状图 Top5 应用领域：NLP 280亿 / 计算机视觉 250亿 / 推荐系统 200亿 / 自动驾驶 180亿 / 医疗 AI 150亿\n4. 饼图 投资地区分布：北美45% 中国30% 欧洲15% 其他10%\n5. 折线图 2019-2024 市场规模：50→80→110→140→175→200（十亿美元）\n6. 各板块配简洁图标装饰", "prompt_en": "Please use pure inline SVG to generate a \"2024 Global AI Industry Data Panorama\" infographic poster (about 800×1200, a single-file HTML, with no external dependencies).\n\nContent:\n1. Title area: \"2024 Global AI Industry Data Panorama\"\n2. Four core numbers: market size $200 billion, annual growth 36.8%, 3,200+ funding rounds, 22 million practitioners\n3. Bar chart of the Top 5 application areas: NLP $28B / Computer Vision $25B / Recommendation Systems $20B / Autonomous Driving $18B / Medical AI $15B\n4. Pie chart of investment region distribution: North America 45%, China 30%, Europe 15%, Other 10%\n5. Line chart of market size 2019-2024: 50→80→110→140→175→200 (in billions of USD)\n6. Decorate each section with a simple icon", "skills": [ { "id": "svg-draw", "name": "SVG 绘图", "description": "纯 SVG 代码创建矢量图形和插画", "slug": "svg-draw" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 176, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "SVG 数据信息图" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "sales-battle-daily-sales-brief", "title": "今日销售团队简报", "title_en": "Daily Sales Team Brief", "subtitle": "把 Pipeline、客户动态和行业新闻压缩成一页晨会讲稿", "subtitle_en": "Compress pipeline updates, account news, and market signals into a one-page morning briefing", "description": "销售晨会最怕两种情况：要么只念数字，没有动作；要么全是新闻，没有商机。这个案例演示了销售作战团队如何把一支 20 人销售团队当天最该知道的内容压缩进一页简报：既有 Pipeline 新进/推进/卡住/丢单，也有互联网、新能源、制造三类客户的外部动态，还附带今天最值得跟进的客户和可直接开口的话术。最终输出一份能直接在晨会上念的 Markdown 简报。\n\n适合：销售总监、销售经理、晨会主持人、销售运营、区域负责人", "description_en": "Sales morning meetings often fail in two ways: either they read out numbers with no action, or they repeat market news with no deal relevance. This case shows how the Sales Battle Team compresses the most important information for a 20-person sales team into a one-page briefing: pipeline updates across new/progress/stuck/lost deals, external account signals from internet, NEV, and manufacturing sectors, plus priority accounts to follow up today with ready-to-use talking points. The final output is a Markdown brief that can be read aloud in a morning stand-up.\n\nBest for: Sales Directors, Sales Managers, Morning Meeting Hosts, Sales Operations, Regional Leaders", "categories": [ "business" ], "tags": [ "销售简报", "晨会", "Pipeline", "客户动态", "行业情报" ], "tags_en": [ "sales brief", "morning meeting", "pipeline", "account news", "market intelligence" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "今日销售团队简报" } ], "prompt": "请生成今日销售团队简报（20 人销售团队，客户主要在互联网/新能源/制造）：\n1. Pipeline 关键变化（新进/推进/卡住/丢单）\n2. 今日重点客户动态（近期融资/高管变动/产品发布）\n3. 今日风险与机会\n4. 团队需要注意的行业新闻\n5. 今日推荐跟进的 3-5 个客户及话术建议\n\n输出 Markdown 简报（A4 一页内），要能直接在晨会上念。", "prompt_en": "Please generate today's sales-team brief (a 20-person sales team, with customers mainly in internet / new energy / manufacturing):\n1. Key pipeline changes (new entries / advanced / stalled / lost deals)\n2. Today's key-account updates (recent funding / executive changes / product launches)\n3. Today's risks and opportunities\n4. Industry news the team should be aware of\n5. 3-5 customers recommended for follow-up today, with talking-point suggestions\n\nOutput a Markdown brief (within a single A4 page) that can be read aloud directly at the morning standup.", "skills": [], "experts": [ { "id": "SalesBattleTeam", "name": "销售作战团队", "description": "销售攻坚体系，客户研究锁定目标、外联策略提升触达、竞品情报预警风险、销售预测优化资源分配", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "ff06833f0e30", "name_en": "Sales Battle Team", "description_en": "Sales effectiveness system: account research identifies targets, outreach strategy improves reach, competitive intel warns risks, pipeline forecasting optimizes resource allocation", "expert_id": "ex_ATh4MnfFCKxR" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 177, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T23:28:00+08:00", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "product-feature-cards", "title": "产品特性介绍卡片", "title_en": "Product Feature Showcase Cards", "subtitle": "输入产品卖点，生成精美的特性介绍卡片页面", "subtitle_en": "Input product highlights, generate elegant feature showcase cards", "description": "产品官网、应用商店详情页、投资人 deck——都需要清晰地展示产品核心特性，但「三个图标加三段文字」的老套路早就审美疲劳了。怎么让特性介绍看起来高级又有说服力？\n\n这个案例演示了输入一款项目管理工具的 4 个核心特性（名称、描述、亮点数据），AI 自动生成特性展示卡片 HTML 页面：2x2 网格布局（max-width 760px 居中）、每张卡片含渐变图标区、特性名称、详细描述、亮点数据指标。深色科技感主题，卡片边框渐变发光效果。\n\n适合：SaaS 产品官网、应用商店描述页、融资路演材料、产品介绍页。", "description_en": "Product websites, app store listings, investor decks — all need clear feature presentations, but the tired 'three icons with three paragraphs' approach has long lost its appeal. How do you make features look premium and persuasive?\n\nThis case demonstrates inputting 4 core features of a project management tool (name, description, key metrics), with AI generating feature showcase cards in HTML: 2x2 grid layout (max-width 760px centered), each card with gradient icon area, feature name, detailed description, and highlight metrics. Dark tech-feel theme with glowing gradient card borders.\n\nBest for: SaaS product websites, app store descriptions, fundraising pitch materials, product introduction pages.", "categories": [ "creation", "dev" ], "tags": [ "产品特性", "卡片组件", "官网设计", "Feature展示" ], "tags_en": [ "Product Features", "Card Component", "Website Design", "Feature Showcase" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为项目管理工具 FlowBoard 生成一个产品特性展示卡片页面（单文件 HTML，内联无外部依赖），现代精美、突出每个特性的价值与亮点数据。\n\n产品定位：面向敏捷团队的下一代项目管理平台\n\n4 个核心特性：\n1. 智能排期引擎 — 基于历史数据自动推荐最优排期，识别依赖/冲突/关键路径 · 排期效率 +40%\n2. 实时协作画布 — 白板+文档+任务的统一空间，多人实时编辑 · 沟通成本 -35%\n3. AI 站会助手 — 自动汇总进展/阻塞/计划，支持异步站会 · 站会时间 -60%\n4. 全链路数据洞察 — 需求到发布全流程度量，自动生成核心指标 · 交付可预测性 +50%", "prompt_en": "Please generate a product-feature showcase card page for the project-management tool FlowBoard (single-file HTML, inlined with no external dependencies), modern and polished, highlighting the value of each feature and its standout metrics.\n\nProduct positioning: a next-generation project-management platform for agile teams\n\n4 core features:\n1. Smart Scheduling Engine — automatically recommends the optimal schedule based on historical data, identifying dependencies / conflicts / the critical path · scheduling efficiency +40%\n2. Real-time Collaborative Canvas — a unified space for whiteboard + docs + tasks, with multi-user real-time editing · communication cost -35%\n3. AI Standup Assistant — automatically summarizes progress / blockers / plans, supporting async standups · standup time -60%\n4. End-to-End Data Insights — metrics across the full requirement-to-release flow, automatically generating core indicators · delivery predictability +50%", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 178, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "产品功能卡片" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "agent-email-auto-reply", "title": "智能邮件自动回复", "title_en": "Intelligent Email Auto-Reply Agent", "subtitle": "AI Agent 自动分类与回复客户邮件工作流", "subtitle_en": "AI Agent Auto-Classification & Customer Email Reply Workflow", "description": "电商客服邮箱每天涌入 200+ 封客户邮件，人工逐条阅读、分类、回复效率极低，高峰期客户等 24 小时才收到回复，投诉升级、客户流失接踵而来。更棘手的是，不同类型邮件（咨询、投诉、退换货、合作意向、垃圾邮件）需要截然不同的处理策略和话术，新客服上手慢、老客服也难保证一致性。\n\n这个案例演示了如何用邮件管理技能设计一套完整的智能邮件自动回复 Agent 工作流：五级分类引擎自动识别邮件类型，情感检测与紧急度评分确保优先处理高风险邮件，个性化模板引擎根据客户画像动态生成回复，人工升级机制在 Agent 不确定时及时将问题交给人类处理，多语言策略覆盖跨境电商场景。\n\n适合：电商运营团队、客服主管、SaaS 客户成功团队、跨境电商卖家、任何日均邮件量超过 50 封的业务团队。", "description_en": "E-commerce support inboxes receive 200+ customer emails daily. Manual reading, classification, and replying is painfully slow — during peak periods customers wait 24 hours for a response, leading to complaint escalation and churn.\n\nThis case demonstrates how to use the email-skill to design a complete intelligent email auto-reply Agent workflow: a five-tier classification engine, sentiment detection and urgency scoring, personalized template engine, human escalation mechanism, and multilingual strategy for cross-border scenarios.\n\nIdeal for: e-commerce operations teams, customer service managers, SaaS customer success teams, cross-border sellers.", "categories": [ "agent" ], "tags": [ "邮件自动化", "AI Agent", "客户服务", "工作流", "智能回复" ], "tags_en": [ "email automation", "AI agent", "customer service", "workflow", "smart reply" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "智能邮件自动回复工作流设计文档" } ], "prompt": "帮我设计一套 AI Agent 智能邮件自动回复工作流。场景是电商客服邮箱每天收到 200+ 封客户邮件。需要包含：1）邮件自动分类规则（咨询、投诉、退换货、合作、垃圾邮件）；2）每类邮件的自动回复模板与个性化变量；3）情感检测与紧急度评分逻辑；4）需要人工介入的升级条件与通知机制；5）多语言支持策略；6）Agent 决策流程图（用文字描述）；7）效果评估指标（响应时间、准确率、客户满意度）。请用邮件管理技能输出详细的 Markdown 工作流设计文档。", "prompt_en": "Help me design an AI Agent intelligent email auto-reply workflow. The scenario: an e-commerce customer-service mailbox receives 200+ customer emails per day. It needs to include: 1) email auto-classification rules (inquiry, complaint, returns/exchanges, partnership, spam); 2) auto-reply templates and personalization variables for each category; 3) sentiment-detection and urgency-scoring logic; 4) escalation conditions and notification mechanisms for cases requiring human intervention; 5) a multilingual support strategy; 6) an Agent decision flowchart (described in text); 7) effectiveness metrics (response time, accuracy, customer satisfaction). Please use the email management skill to output a detailed Markdown workflow design document.", "skills": [ { "id": "email-skill", "name": "邮件管理", "description": "邮件管理与自动化，支持多邮件服务商收发搜索", "description_en": "Email management and automation across multiple providers", "skill_id": "skill_2053081884429156352" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 179, "is_featured": false, "featured_order": 0, "quality_score": 94, "source": "official", "usage_count": 0, "created_at": "2026-04-23T09:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "diagnosis-missing-info-abdomen", "title": "检查无异常？AI 帮你定位缺失信息", "title_en": "Workup Normal? Find What's Still Missing", "subtitle": "反复不适 + 检查没异常？AI 反推还需补充的关键检查", "subtitle_en": "Recurrent symptoms but normal tests? Let AI reverse-engineer the missing workup", "description": "反复腹痛查了一堆化验和影像都没异常，到底是检查没做对，还是真的没病？医生和患者都卡在「下一步该查什么」上——继续重复原来的检查，还是换一个方向？\n\n这个案例演示了把「主诉 / 现病史 / 已做的辅助检查」交给「腾讯健康 · AI 临床助手·辅助诊断」，确认无当前诊断且信息不足以做出高置信度判断后，它从鉴别诊断出发，反向推导还需补充的关键检查（如妇科查体、盆腔超声、肠镜、泌尿系影像等），每条缺失信息都明确标注「缺失类型 + 内容 + 理由 + 症状 / 体征 → 疑似疾病 → 确诊需补充的检查」完整依据路径，并按优先级排序下一步建议。它不强行输出诊断结论，方向提示一律标注「非诊断结论，需补充信息后重新评估」，末尾附安全声明。\n\n行为边界：不强行输出疑似诊断，因信息不足以支撑高置信度判断；方向提示必须明确标注「非诊断结论」；缺失信息清单逐条列出缺失类型 / 内容 / 理由 / 依据路径；下一步建议按优先级排序；不给出具体治疗方案；末尾附安全声明。\n\n适合：全科医生、消化科、妇科、急诊科、基层医疗机构、慢病管理医生、需要梳理下一步检查策略的患者。", "description_en": "Recurrent abdominal pain with a slew of normal labs and imaging — wrong workup, or truly nothing wrong? Both clinician and patient are stuck at \"what next?\" Repeat the same tests, or pivot?\n\nThis case shows how to hand the AI clinical assistant the chief complaint, history, and the workup already done. After confirming there is no current diagnosis and evidence is insufficient for a high-confidence call, it reverse-engineers from plausible differentials to identify the missing information (e.g., pelvic exam, pelvic ultrasound, colonoscopy, urinary tract imaging, etc.), with every missing item labeled type / content / rationale / complete reasoning chain (symptom + sign → suspected condition → confirmatory test needed), and prioritizes next-step workup. It never forces a diagnostic conclusion — every directional hint is labeled \"Not a diagnostic conclusion — re-evaluate after supplementation,\" and a safety statement closes the report.\n\nBoundaries: no forced suspected diagnosis when evidence is insufficient; directional hints explicitly labeled \"Not a diagnostic conclusion\"; missing-info list with type / content / rationale / chain per item; next steps prioritized; no specific treatment plans; safety statement at the end.\n\nFor: GPs, gastroenterologists, gynecologists, ER physicians, primary-care clinicians, chronic-care managers, and patients who want a clear missing-test strategy.", "categories": [ "agent", "education" ], "tags": [ "辅助诊断", "缺失信息", "下一步检查", "AI临床助手", "腾讯健康" ], "tags_en": [ "Diagnosis Assistant", "Missing Information", "Next-Step Workup", "AI Clinical Assistant", "Tencent Health" ], "artifact_type": "html", "preview": "report.html", "artifacts": [ { "file": "report.html", "label": "缺失信息定位与下一步检查建议报告（HTML）" } ], "prompt": "请基于以下患者信息（反复不适 + 已做检查无异常）做缺失信息识别，反向推导还需补充的关键检查。\n\n【患者信息】\n[粘贴：主诉、现病史、既往史、查体、已做的辅助检查结果]\n\n【示例场景】\n这个患者反复腹痛，查了几项检查都没异常，想看看能不能帮忙分析一下到底是什么问题。\n\n【输出要求】\n1. 病历文本解析：提取主诉、现病史、辅助检查，确认无当前诊断\n2. 缺失信息识别：明确缺什么信息、为什么需要、推理链路是什么\n3. 从诊断假设反向推导还需补充的检查（如妇科查体、盆腔超声、肠镜等）\n4. 每条缺失项包含完整依据路径（症状 + 体征 → 疑似疾病 → 确诊需补充的检查）\n5. 可附初步方向提示，但必须标注「非诊断结论，需补充信息后重新评估」\n6. 下一步建议按优先级排序\n7. 安全声明附在报告末尾\n\n【行为边界】\n- 不得强行输出疑似诊断，因信息不足以支撑高置信度判断\n- 方向提示必须明确标注「非诊断结论」\n- 缺失信息清单需逐条列出缺失类型、内容、理由、依据路径\n- 建议下一步按优先级排序\n\n【不应出现】\n- 强行输出疑似诊断报告\n- 缺少依据路径的缺失信息项\n- 省略「非诊断结论」警告标注\n- 具体的治疗方案建议\n- 省略安全声明", "prompt_en": "Based on the following patient information (recurrent symptoms + normal workup so far), identify the missing information and reverse-engineer the key tests still needed.\n\n【Patient information】\n[Paste: chief complaint, history of present illness, past history, physical exam, workup already done]\n\n【Example scenario】\nThis patient has recurrent abdominal pain and several normal tests. Please help analyze what might be going on.\n\n【Output requirements】\n1. Record parsing: extract chief complaint, history, workup; confirm no current diagnosis\n2. Missing-information identification: state what is missing, why, and the reasoning chain\n3. Reverse-engineer from diagnostic hypotheses to the tests still needed (e.g., pelvic exam, pelvic ultrasound, colonoscopy, etc.)\n4. Each missing item must carry a complete reasoning chain (symptom + sign → suspected condition → confirmatory test needed)\n5. A directional hint may be attached but must be labeled \"Not a diagnostic conclusion — re-evaluate after supplementation\"\n6. Next-step workup prioritized\n7. Safety statement appended at the end\n\n【Boundaries】\n- Do not force a suspected diagnosis when evidence is insufficient\n- Directional hints must be explicitly labeled \"Not a diagnostic conclusion\"\n- Missing-info list per item: type / content / rationale / reasoning chain\n- Next steps prioritized\n\n【Must not appear】\n- A forced suspected-diagnosis report\n- Missing-info items without a reasoning chain\n- Omission of the \"Not a diagnostic conclusion\" warning label\n- Specific treatment plans\n- A missing safety statement", "cover_image": "cover.png", "cover_mode": "landscape", "skills": [], "experts": [], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 180, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "usage_count": 0, "created_at": "2026-08-22T22:55:11+08:00", "updated_at": "2026-08-27T11:50:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "music-video", "title": "音乐MV", "title_en": "Music Video", "subtitle": "让动漫摇滚演出与霓虹雨夜剧情逐拍交错", "subtitle_en": "An anime rock performance intercut beat by beat with a neon rain narrative", "description": "音乐 MV 如果只按歌词配图，演奏、剧情与节奏往往彼此脱节，演唱口型也难以稳定。\n\n案例围绕一支动漫摇滚乐队：蒙眼银蓝发主唱在舞台演唱，吉他、鼓点与雨夜城市奔跑线交替，副歌在霓虹现场汇合。\n\n适合：原创歌曲 MV、虚拟乐队、动漫音乐宣传、角色演出和节奏型视觉内容。", "description_en": "Music videos built as lyric illustrations often disconnect performance, narrative, and rhythm, while singing lip sync becomes unstable.\n\nAn anime rock band alternates a blindfolded silver-blue-haired singer’s live performance with a neon rain-night running story, converging during the chorus.\n\nBest for original-song videos, virtual bands, anime music promotion, character performances, and rhythm-led visuals.", "categories": [ "creation" ], "tags": [ "音乐MV", "动漫摇滚", "虚拟乐队" ], "tags_en": [ "music video", "anime rock", "virtual band" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「音乐 MV 创作」专家调用 Flova 的「音乐MV（需上传音乐）」Skill，根据我上传的约 50 秒原创摇滚歌曲，制作一支 16:9 动漫乐队 MV《雨夜失速乐队》。主唱为银蓝短发、黑色眼罩、深色舞台服的女性；固定吉他手、贝斯手和鼓手造型。主歌表现她独自在雨夜霓虹街道奔跑，副歌切入小型现场演出，鼓点、拨弦特写、观众手臂和主唱近景按音乐强拍交错，最后在停雨后的空街收尾。请先分析歌曲 BPM、段落、精确时间戳和歌词，再确认规格与故事板；绑定音频和角色资产，依次确认元素图与每镜关键帧。只有主唱嘴部清晰可见的演唱近景使用音频驱动口型，其余镜头按节奏生成，最终严格对齐原始音乐完成剪辑。", "prompt_en": "Ask the “Flova Music Video Director” expert to invoke Flova’s “Music Video (Music Upload Required)” Skill and use my uploaded original rock song of about 50 seconds to create a 16:9 anime-band music video titled Runaway Band in the Rain. The female lead singer has short silver-blue hair, a black blindfold, and dark stagewear; lock the guitarist, bassist, and drummer designs as well. During the verse she runs alone through neon streets in the rain. The chorus cuts to an intimate live performance, interweaving drums, string-plucking close-ups, audience arms, and lead-singer close-ups on strong beats. End on an empty street after the rain stops. Analyze the song’s BPM, sections, exact timestamps, and lyrics first, then confirm the specification and storyboard. Bind the audio and character assets and obtain approval for element images and per-shot keyframes. Use audio-driven lip sync only for singing close-ups where the singer’s mouth is clearly visible; generate all other shots to rhythm, and edit everything precisely to the original music.", "skills": [], "experts": [ { "id": "flova-music-video-director", "name": "音乐 MV 创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及音乐 MV 成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d487e16aa365", "expert_id": "ex_QHjWRFR9zhMy" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 180, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-intern-recommend", "title": "实习推荐：上海财务在招岗", "title_en": "Internships: Finance Roles Open in Shanghai", "subtitle": "按到岗天数和课表匹配度筛选正在开放的实习", "subtitle_en": "Filter open internships by on-site days and class-schedule fit", "description": "在校生找实习最现实的约束不是薪资，是一周能去几天——招聘平台偏偏不让按这个筛。\n\n本案例按「上海 + 本科 + 财务」实时检索到 147 个在招岗位，取前 8 条整理成对照表，把埋在标签里的关键信息拉到台面：3 天/周还是 5 天/周、接不接受大一大二、有没有一对一导师，并识别出混在结果里的应届全职岗。\n\n适合：学期中只能兼职到岗的在校生、想提前攒履历的大一大二学生。", "description_en": "The real constraint on a student internship hunt isn't pay — it's days per week on site, which job boards won't let you filter by.\n\nThis case live-searches \"Shanghai + bachelor's + finance\", finds 147 open postings, and turns the top 8 into a comparison table surfacing what's buried in tags: 3 or 5 days a week, whether first- and second-years are accepted, whether a mentor is provided — and flags the full-time graduate role mixed into the results.\n\nFor: students who can only work part-time during term and early-years building a CV.", "categories": [ "search" ], "tags": [ "实习", "财务", "上海", "在校生" ], "tags_en": [ "internship", "finance", "Shanghai", "students" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "上海财务实习对照清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我在上海读本科，想找财务方向实习，学期中每周只能去两三天。要求：1) 实时检索并给出在招总数；2) 前 8 条做成对照表：职位、公司、规模、薪资、区、每周到岗天数、是否接受大一大二、有没有导师；3) 薪资和标签用平台原值；4) 有非实习岗混进来要单独标出；5) 按「课表塞得下」排优先级。", "prompt_en": "I'm an undergraduate in Shanghai looking for a finance internship, available two or three days a week during term. Requirements: 1) search live and give the total; 2) turn the top 8 into a table: role, company, size, pay, district, days per week, first/second-year acceptance, mentor availability; 3) use the platform's own pay figures and tags; 4) flag any non-internship entries separately; 5) rank by fit with my class schedule.", "skills": [ { "id": "gaodun-internship-position-search", "name": "实习职位搜索", "description": "按多维条件筛选实习岗位", "skill_id": "skill_2089647063754342400" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 180, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "train-ticket-query", "title": "高铁余票实时查询", "title_en": "Real-Time Train Ticket Query", "subtitle": "查北京到上海高铁余票，10秒找出最快车次", "subtitle_en": "Query Beijing-Shanghai HSR tickets and find the fastest train in 10 seconds", "description": "出差订票总是踩点，打开12306又卡又慢，还得一列一列比耗时和余票，效率极低。\n\n这个案例演示了如何用自然语言查询北京南→上海虹桥的高铁余票：指定出发日期和时段，AI 实时拉取12306数据，返回12趟列车的完整余票表格（含二等座/一等座/商务座），并自动标注最快到达、余票最充足、性价比最高的推荐车次。\n\n适合：商务出差族、高频出行旅客、企业差旅助理、需要快速决策的赶车人群。", "description_en": "Booking train tickets is always a last-minute scramble — 12306 is slow to load and you have to compare each train's duration and availability one by one.\n\nThis case demonstrates querying real-time Beijing South → Shanghai Hongqiao HSR availability via natural language: specify departure date and time window, and AI fetches live 12306 data, returning a full 12-train table with seat availability (second class/first class/business), plus auto-highlighted recommendations for fastest arrival, most seats available, and best value.\n\nBest for: Business travelers, frequent rail commuters, corporate travel assistants, and anyone who needs quick booking decisions.", "categories": [ "travel" ], "tags": [ "高铁", "火车票", "余票查询", "出差" ], "tags_en": [ "High-Speed Rail", "Train Ticket", "Seat Availability", "Business Travel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "高铁余票查询页面" } ], "skills": [ { "id": "12306", "name": "12306", "description": "查询 12306 国内列车时刻、余票与站点信息", "description_en": "Query China Railway 12306 train schedules and ticket availability", "skill_id": "skill_2053079107781726208" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "帮我查明天上午北京南到上海虹桥的高铁余票，要求二等座，返回带余票数量的完整表格，并自动筛选出耗时最短的车次推荐。", "prompt_en": "Help me check the availability of high-speed rail tickets from Beijing South to Shanghai Hongqiao tomorrow morning. I want second-class seats. Return a complete table with the number of remaining tickets, and automatically filter out and recommend the fastest train.", "source": "official", "sort_order": 180, "featured_order": 0, "quality_score": 90, "usage_count": 0, "is_featured": false, "created_at": "2026-04-28T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-intern-data-analysis", "title": "实习推荐：全国数据分析方向", "title_en": "Internship Search: Data Analysis Nationwide", "subtitle": "14 个岗位里，一半藏在采购、财务、HR 部门", "subtitle_en": "Half of the fourteen sit inside procurement, finance and HR teams", "description": "想找数据分析实习，搜「数据分析」四个字只能出来寥寥几个岗，很容易以为机会很少。\n\n本案例不限城市检索数据分析方向实习，命中 14 个岗位。结果揭示了一个关键事实：岗位名里带「数据分析」的多数属于采购、财务、HR、市场等业务部门，而非数据部门。案例给出日薪区间、城市分布、出勤要求与公司规模统计，并给出一套换关键词挖岗的方法。\n\n适合：想入行数据分析的在校生、找第一份分析类实习的人。", "description_en": "Searching \"data analysis\" for internships returns only a handful of posts, making the field look closed.\n\nThis case searches nationwide for data-analysis internships, finding 14. The results reveal something useful: most posts with \"data analysis\" in the title sit inside procurement, finance, HR or marketing teams rather than data departments. Daily pay, city spread, attendance requirements and company sizes are tabulated, along with a method for finding more via alternative keywords.\n\nFor: students breaking into analytics and anyone seeking a first analysis internship.", "categories": [ "search" ], "tags": [ "实习", "数据分析", "全国", "岗位检索" ], "tags_en": [ "internship", "data analysis", "nationwide", "job search" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "全国数据分析实习岗位分析", "label_en": "Nationwide data-analysis internship analysis" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我找数据分析方向的实习岗位，不限城市。要求：1) 真实检索并给出命中总数；2) 按日薪上限降序列出岗位、公司、城市、日薪、公司规模、标签；3) 统计城市分布、有明确日薪与面议各几个、日薪区间是多少；4) 观察岗位名称的构成，指出有多少岗位其实属于业务部门（采购、财务、HR、市场）而非数据部门，并说明这对求职者意味着什么；5) 给出挑选数据分析实习的具体建议（关键词、出勤天数、转正标签等）。", "prompt_en": "Find data-analysis internships anywhere in the country. Requirements: 1) run the search and give the total; 2) list posts by descending daily pay with company, city, pay, company size and tags; 3) report the city distribution, how many list explicit pay versus negotiable, and the pay range; 4) examine the job titles and report how many actually sit in business functions (procurement, finance, HR, marketing) rather than data teams, and what that means for applicants; 5) give concrete advice on choosing among them (keywords, weekly attendance, conversion-to-full-time tags).", "skills": [ { "id": "gaodun-internship-position-search", "name": "实习职位搜索", "description": "按多维条件筛选实习岗位", "skill_id": "skill_2089647063754342400" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 181, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "project-timeline-viewer", "title": "项目时间线可视化", "title_en": "Project Timeline Viewer", "subtitle": "甘特图式项目排期，里程碑进度一目了然", "subtitle_en": "Gantt-style project timeline with milestone tracking at a glance", "description": "项目排期散落在 Excel 和各种文档里，依赖关系看不清、进度更新靠口头同步，一旦某个环节延期，整条线都得手动调。\n\n这个案例用 Mermaid 甘特图为一个虚拟的移动端产品「SuperApp v2.0」生成了完整的时间线：跨越 6 个月的 4 大阶段（规划设计、开发、测试、上线），每个阶段包含多项子任务，配备里程碑标记和依赖关系。\n\n适合：项目经理、技术负责人、PMO、初创公司 CTO——任何需要把复杂排期变成可视化的人。", "description_en": "Project schedules buried in spreadsheets and scattered docs make it nearly impossible to see dependencies, track progress, or react when a task slips.\n\nThis case uses Mermaid Gantt chart to generate a full timeline for a fictional mobile product 'SuperApp v2.0': 4 major phases spanning 6 months with sub-tasks, milestone markers, and dependency relationships.\n\nIdeal for: project managers, tech leads, PMO, startup CTOs — anyone who needs to turn complex schedules into visualizations.", "categories": [ "office", "creation" ], "tags": [ "时间线", "甘特图", "项目管理", "里程碑", "进度追踪" ], "tags_en": [ "Timeline", "Gantt Chart", "Project Management", "Milestone", "Progress Tracking" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个项目甘特图页面（单文件 HTML，用 Mermaid gantt 渲染，含任务依赖和里程碑标记），展示虚拟移动应用「SuperApp v2.0」的开发排期（2026 年 1-6 月）。\n\n4 个阶段：\n1）规划设计（1/5-2/15）：需求调研、UX 研究、原型设计、UI 设计、设计评审\n2）开发（2/10-4/30）：前端、后端、API 联调、数据库迁移\n3）测试（4/1-5/31）：单元、集成、UAT、性能\n4）上线（5/15-6/30）：Beta 发布、Bug 修复、营销准备、正式发布\n\n里程碑：设计冻结 2/15、Alpha 4/1、Beta 5/15、正式上线 6/15。", "prompt_en": "Please generate a project Gantt chart page (single-file HTML, rendered with Mermaid gantt, including task dependencies and milestone markers) that shows the development schedule of a fictional mobile app \"SuperApp v2.0\" (January-June 2026).\n\n4 phases:\n1) Planning & Design (1/5-2/15): requirements research, UX research, prototyping, UI design, design review\n2) Development (2/10-4/30): front-end, back-end, API integration, database migration\n3) Testing (4/1-5/31): unit, integration, UAT, performance\n4) Launch (5/15-6/30): Beta release, bug fixes, marketing prep, official release\n\nMilestones: Design Freeze 2/15, Alpha 4/1, Beta 5/15, Official Launch 6/15.", "artifacts": [ { "file": "output.html", "label": "项目甘特图时间线" } ], "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表", "description": "使用 Mermaid 语法生成流程图、时序图、甘特图等可视化图表", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "is_featured": false, "sort_order": 181, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-04-23", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "graphql-api-design", "title": "GraphQL API 设计与实现", "title_en": "GraphQL API Design and Implementation", "subtitle": "在线书店完整 API，从 Schema 到 Resolver", "subtitle_en": "Complete online bookstore API, from Schema to Resolver", "description": "设计一个功能完整的 GraphQL API 需要考虑类型系统、查询优化、权限控制、实时订阅等多个层面，对开发者的架构能力要求较高。新手容易陷入 N+1 查询、过度获取等常见陷阱。\n\n本案例演示了如何利用 AI 为在线书店设计一套完整的 GraphQL API。涵盖 Schema 定义（Book/Author/User/Order/Review 含关系）、完整的 Query/Mutation/Subscription 设计、DataLoader N+1 优化、认证权限中间件、自定义标量类型等生产级特性。\n\n适合后端开发者、API 架构师、全栈工程师，以及正在从 REST 迁移到 GraphQL 的团队使用。", "description_en": "Designing a complete GraphQL API requires consideration of type systems, query optimization, access control, and real-time subscriptions.\n\nThis case demonstrates using AI to design a comprehensive GraphQL API for an online bookstore, covering schema definitions, queries, mutations, subscriptions, DataLoader optimization, auth middleware, and custom scalars.\n\nIdeal for backend developers, API architects, and teams migrating from REST to GraphQL.", "categories": [ "dev", "data" ], "tags": [ "GraphQL", "API设计", "TypeScript", "后端架构" ], "tags_en": [ "GraphQL", "API Design", "TypeScript", "Backend Architecture" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.ts", "prompt": "请为一个在线书店设计完整的 GraphQL API（TypeScript + Apollo Server）。要求：1) Schema 定义（Book/Author/User/Order/Review 类型，含关系）；2) Query（书籍搜索/分类浏览/作者详情/用户订单历史/评论列表，支持分页和排序）；3) Mutation（注册登录/购物车/下单/评论/更新用户信息）；4) Subscription（新书上架/订单状态变更）；5) 中间件（认证/权限/限流/日志）；6) DataLoader 解决 N+1；7) 自定义 Scalar（DateTime/ISBN/Money）。\n\n代码注释和用户可见字符串用简体中文，标识符保留英文。", "prompt_en": "Please design a complete GraphQL API (TypeScript + Apollo Server) for an online bookstore. Requirements: 1) Schema definitions (Book/Author/User/Order/Review types, with relationships); 2) Queries (book search / browse by category / author details / user order history / review list, with pagination and sorting support); 3) Mutations (register & login / shopping cart / place order / review / update user info); 4) Subscriptions (new book releases / order status changes); 5) Middleware (authentication / authorization / rate limiting / logging); 6) DataLoader to solve N+1; 7) Custom Scalars (DateTime/ISBN/Money).\n\nCode comments and user-visible strings should be in Simplified Chinese, while identifiers remain in English.", "skills": [ { "id": "api-design", "name": "API 设计", "description": "API design and implementation", "slug": "api-design" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 182, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.ts", "label": "GraphQL API 代码" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-intern-shenzhen-ops", "title": "实习推荐：深圳运营方向", "title_en": "Internship Search: Operations Roles in Shenzhen", "subtitle": "31 个岗位，南山区占一半——大厂和小团队的运营岗完全不是一回事", "subtitle_en": "31 posts, half in Nanshan — big-company and startup ops roles differ entirely", "description": "运营是实习岗里体量最大的方向，也是最容易「同名不同活」的方向——同样叫运营实习生，有人只做内容排期，有人一个人包全流程。\n\n本案例检索深圳运营方向实习，命中 31 个，本页 20 条中南山区占一半以上。案例列出日薪、公司规模与标签，指出同一家大厂就有三个不同的运营实习岗，并整理出一张「对新手友好」的筛选信号表（接受大一大二、零基础、远程、一对一导师）。\n\n适合：找第一份实习的低年级学生、想进互联网做运营的人。", "description_en": "Operations is the largest internship category and the most inconsistently defined — two roles with the same title can mean scheduling content or running an entire funnel alone.\n\nThis case searches Shenzhen operations internships, finding 31, with over half of the 20 shown in Nanshan district. It lists daily pay, company size and tags, notes that a single large company posts three distinct operations roles, and builds a table of beginner-friendly signals (open to first- and second-years, no experience required, remote, one-on-one mentoring).\n\nFor: junior students seeking a first internship and anyone entering internet operations.", "categories": [ "search" ], "tags": [ "实习", "运营", "深圳", "新手友好" ], "tags_en": [ "internship", "operations", "Shenzhen", "beginner friendly" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "深圳运营实习岗位与新手友好信号", "label_en": "Shenzhen operations internships and beginner signals" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我找深圳的运营方向实习岗位。要求：1) 真实检索并给出命中总数；2) 按日薪上限降序列出岗位、公司、所在区、日薪、公司规模、标签；3) 统计行政区分布、日薪区间，并指出哪家公司放出的运营实习岗最多、分别是什么类型；4) 整理一张「对新手友好」的筛选信号表，统计有多少岗位带「接受大一大二」「零基础」「远程实习」「一对一导师」等标签，说明每个信号意味着什么；5) 提醒大厂与小团队的运营岗在分工上的区别，投递前要看清岗位描述。", "prompt_en": "Find operations internships in Shenzhen. Requirements: 1) run the search and give the total; 2) list posts by descending daily pay with company, district, pay, company size and tags; 3) report the district distribution and pay range, and identify which company posts the most operations roles and of what kinds; 4) build a table of beginner-friendly signals, counting posts tagged \"open to first/second years\", \"no experience needed\", \"remote\" and \"one-on-one mentor\", explaining what each signals; 5) note how operations roles differ between large companies and small teams and advise reading the description before applying.", "skills": [ { "id": "gaodun-internship-position-search", "name": "实习职位搜索", "description": "按多维条件筛选实习岗位", "skill_id": "skill_2089647063754342400" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 182, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-funnel-chart", "title": "销售漏斗可视化", "title_en": "Sales Funnel Visualization", "subtitle": "多维漏斗分析图，直观展示转化瓶颈", "subtitle_en": "Multi-dimensional funnel analysis chart to visualize conversion bottlenecks", "description": "销售数据只会看Excel表格，哪个环节流失最多全凭感觉？管理层要看转化数据，你还在手动画PPT饼图？别再猜了。\n\n这个案例生成一个交互式销售漏斗图，展示从线索获取到成交签约的6个阶段转化率。每个阶段显示绝对数量、转化率和流失率，hover可查看详细breakdown和环比数据。侧边栏展示关键指标：总转化率4.5%、平均客单价¥28,000、销售周期45天。专业配色渐变，响应式设计，开箱即用。\n\n适合：销售经理、运营总监、数据分析师、创业者，任何需要快速呈现销售转化数据的场景。", "description_en": "Still eyeballing Excel spreadsheets to figure out where leads drop off? Management wants conversion data but you're manually drawing pie charts in PowerPoint? Stop guessing.\n\nThis case generates an interactive sales funnel chart showing conversion rates across 6 stages from lead acquisition to deal closure. Each stage displays absolute counts, conversion rates, and drop-off rates with hover details and period-over-period comparisons. A side panel shows key metrics: 4.5% total conversion rate, ¥28,000 average deal size, 45-day sales cycle. Professional color gradients, responsive design, ready to use.\n\nBest for: Sales managers, operations directors, data analysts, entrepreneurs — anyone who needs to quickly present sales conversion data.", "categories": [ "data", "business" ], "tags": [ "销售漏斗", "数据可视化", "转化分析", "AntV图表", "商业智能" ], "tags_en": [ "Sales Funnel", "Data Visualization", "Conversion Analysis", "AntV Chart", "Business Intelligence" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "帮我生成一个销售漏斗可视化图表页面（单文件 HTML，内联无外部依赖），专业配色、响应式、hover 查看详情。\n\n漏斗数据（含每阶段转化率和流失率）：\n- 线索获取 10,000\n- 初步接触 6,500\n- 需求确认 3,200\n- 方案演示 1,800\n- 商务谈判 900\n- 成交签约 450\n\n侧边关键指标：总转化率、平均客单价 ¥28,000、销售周期 45 天、时间范围 Q1 2026。", "prompt_en": "Help me generate a sales funnel visualization chart page (a single-file HTML page, inline with no external dependencies), with a professional color scheme, responsive, and hover-to-view-details.\n\nFunnel data (including each stage's conversion rate and drop-off rate):\n- Lead Acquisition 10,000\n- Initial Contact 6,500\n- Needs Confirmation 3,200\n- Solution Demo 1,800\n- Business Negotiation 900\n- Deal Signed 450\n\nSidebar key metrics: overall conversion rate, average deal value ¥28,000, sales cycle 45 days, time range Q1 2026.", "artifacts": [ { "file": "output.html", "label": "销售漏斗分析图" } ], "skills": [ { "id": "mermaid-diagram", "name": "Mermaid 图表", "description": "使用 Mermaid 语法生成流程图、时序图、甘特图等可视化图表", "slug": "mermaid-diagram" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 183, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-22T10:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "multilingual-translation-table", "title": "多语言翻译对照表", "title_en": "Multilingual UI Translation Table", "subtitle": "一键生成中英日韩四语 App 界面文案翻译对照表", "subtitle_en": "One-click generation of Chinese/English/Japanese/Korean UI copy translation table", "description": "产品出海最头疼的就是界面文案翻译——找四种语言的译员协调来回、格式混乱、Key 对不上号，一不小心就漏翻或错翻。\n\n这个案例演示了如何用 AI 一次性生成一份包含 60 条 UI 文案的中英日韩四语翻译对照表。覆盖登录注册、首页导航、任务管理、设置、通用操作五大模块，每条都配有程序用 Key 标识符，翻译风格贴合各语言 UI 习惯而非机械直译，还附带 Summary 统计各模块翻译进度。\n\n适合：出海 SaaS 产品经理、国际化开发工程师、本地化项目负责人、独立开发者。", "description_en": "The biggest headache of taking your product global is UI copy translation — coordinating translators across four languages, messy spreadsheets, mismatched keys, and inevitable missed or mistranslated strings.\n\nThis case demonstrates generating a 60-entry Chinese/English/Japanese/Korean UI translation table in one shot. It covers five core modules (Login, Navigation, Task Management, Settings, Common Actions), each entry with a developer-ready key identifier. Translations follow native UI conventions rather than literal word-for-word mapping, plus a Summary sheet tracking progress per module.\n\nBest for: SaaS product managers going global, i18n developers, localization project leads, indie developers.", "categories": [ "business", "productivity" ], "tags": [ "多语言「, 」翻译「, 」国际化「, 」出海", "UI文案" ], "tags_en": [ "Multilingual", "Translation", "i18n", "Globalization", "UI Copy" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 184, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "prompt": "为一款即将出海的 SaaS 产品生成一份 App 界面文案的中英日韩四语翻译对照表（Excel 格式，output.xlsx）。\n\n覆盖产品核心界面约 60 个文案条目，分为登录注册、首页/导航、任务管理、设置、通用操作五大模块。\n\n要求：\n1）表格列：序号、模块、Key（程序用标识符）、中文、English、日本語、한국어、备注\n2）翻译准确自然，符合各语言 UI 文案习惯，而非机械直译\n3）附一个 Summary sheet 统计各模块翻译进度\n\n输出为 .xlsx 文件。", "prompt_en": "Generate a four-language (Chinese/English/Japanese/Korean) UI copy translation table (Excel format, output.xlsx) for a SaaS product about to go global.\n\nCover about 60 copy entries from the product's core interface, across five modules: Login/Registration, Home/Navigation, Task Management, Settings, and Common Actions.\n\nRequirements:\n1) Table columns: number, module, Key (a programmatic identifier), Chinese, English, 日本語, 한국어, notes\n2) Translations should be accurate and natural, matching each language's UI conventions rather than literal word-for-word mapping\n3) Include a Summary sheet tracking translation progress per module\n\nOutput as an .xlsx file.", "artifacts": [ { "file": "output.xlsx", "label": "多语言翻译对照表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "react-component-dev", "title": "数据统计看板页面", "title_en": "Data Dashboard Page", "subtitle": "纯 HTML/CSS/JS 构建响应式数据统计看板，支持主题切换", "subtitle_en": "Pure HTML/CSS/JS responsive data dashboard with theme toggle", "description": "每次开新项目都要从零搭数据看板——KPI 卡片、折线图、响应式布局、主题切换，重复劳动太多。怎么快速生成一个开箱即用的看板组件？\n\n这个案例演示了用前端开发技能生成一个完整的数据统计看板 HTML 页面：4 个 KPI 统计卡片（含趋势箭头和百分比变化）、纯 SVG 折线图展示近 7 天数据趋势、深色/浅色主题一键切换、响应式网格布局适配手机到桌面。全部内联实现，无需任何外部依赖。\n\n适合：前端开发者快速搭建数据看板原型、中后台系统仪表盘页面、数据产品 Demo 展示。", "description_en": "Every new project means rebuilding a dashboard from scratch — KPI cards, charts, responsive layout, theme toggling. Too much repeated work. How to generate a ready-to-use dashboard component quickly?\n\nThis case demonstrates using the frontend-dev skill to generate a complete data dashboard HTML page: 4 KPI stat cards with trend arrows and percentage changes, pure SVG line chart showing 7-day trends, dark/light theme toggle, responsive grid layout from mobile to desktop. All inline, no external dependencies.\n\nBest for: Frontend developers prototyping dashboards, admin system dashboard pages, data product demos.", "categories": [ "dev", "creation" ], "tags": [ "HTML", "数据看板", "响应式布局", "主题切换" ], "tags_en": [ "HTML", "Data Dashboard", "Responsive Layout", "Theme Toggle" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请用纯 HTML/CSS/JS 开发一个数据统计看板页面（单文件 HTML，内联无外部依赖），模拟 React 组件风格，界面文字用简体中文，默认深色主题。\n\n包含：4 个 KPI 卡片（总用户、活跃用户、收入、转化率，带趋势箭头和环比）、近 7 天数据趋势的 SVG 折线图、暗色/亮色主题切换、响应式布局（移动端单列、桌面多列）。", "prompt_en": "Please develop a data dashboard page using pure HTML/CSS/JS (single-file HTML, inline with no external dependencies), mimicking a React component style, with the UI text in Simplified Chinese, dark theme by default.\n\nInclude: 4 KPI cards (total users, active users, revenue, conversion rate, with trend arrows and period-over-period change), an SVG line chart of the past 7 days' data trend, dark/light theme toggle, and a responsive layout (single column on mobile, multi-column on desktop).", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 185, "is_featured": false, "featured_order": 0, "quality_score": 86, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "数据统计看板" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "smart-page-project-review", "title": "项目复盘智能页面", "title_en": "Smart Page Project Review", "subtitle": "用 STAR 结构把项目过程、结果与下一步做成可交互汇报页", "subtitle_en": "Turn project process, outcomes, and next steps into an interactive STAR report page", "description": "项目复盘经常散落在长文档、会议纪要和口头结论里，读者很难快速抓住关键过程与下一步。这个案例演示如何用腾讯文档智能页面生成一份 STAR 结构复盘页：首屏给出项目结论，分段呈现 Situation、Task、Action、Result，并补充 Keep/Stop/Start 与下一步优先级。适合产品经理、项目负责人、运营负责人在阶段复盘、里程碑汇报和向上同步时使用。", "description_en": "Project reviews are often scattered across long documents, meeting notes, and verbal conclusions, making it hard for readers to grasp the key process and next steps. This case demonstrates how Tencent Docs Smart Page creates an interactive STAR review page: the hero section states the conclusion, followed by Situation, Task, Action, Result, plus Keep/Stop/Start and prioritized next steps. It is suitable for product managers, project owners, and operation leads preparing milestone reviews or leadership updates.", "categories": [ "productivity", "office" ], "tags": [ "智能页面", "项目复盘", "STAR", "汇报", "腾讯文档" ], "tags_en": [ "Smart Page", "Project Review", "STAR", "Presentation", "Tencent Docs" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "项目复盘智能页面" }, { "file": "data.js", "label": "页面数据源" } ], "prompt": "请用腾讯文档智能页面生成一份项目复盘汇报页。\n\n主题：填入你的项目名\n叙事结构：STAR 复盘；设计风格：腾讯蓝皮肤\n\n内容要求：\n1. 首屏给出一句话结论\n2. 用 1-2 个 STAR 故事展开\n3. 补充 Keep / Stop / Start\n4. 给出 3 条下一步行动，含 owner 与 eta\n\n输出：可单文件预览的 HTML 智能页面，包含 Overview、Controls、Charts、Logic 四区；资源全部内联或用系统字体，不引用外部 CDN。", "prompt_en": "Please use Tencent Docs Smart Page to generate a project retrospective report page.\n\nTheme: fill in your project name\nNarrative structure: STAR retrospective; design style: Tencent blue skin\n\nContent requirements:\n1. Hero section gives a one-sentence conclusion\n2. Unfold with 1-2 STAR stories\n3. Add Keep / Stop / Start\n4. Provide 3 next-step actions with owner and eta\n\nOutput: a single-file previewable HTML smart page including the four zones Overview, Controls, Charts, and Logic; all resources inlined or using system fonts, with no external CDN references.", "skills": [ { "id": "smart-page", "name": "腾讯文档智能页面", "description": "腾讯文档智能页面，生成可编辑可分享的在线汇报网页内容", "skill_id": "skill_2053082927661453312" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 186, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-05-09T09:20:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "contract-review-assistant", "title": "合同条款审查助手", "title_en": "Contract Clause Review Assistant", "subtitle": "粘贴合同文本，自动生成风险审查报告 Word 文档", "subtitle_en": "Paste contract text, auto-generate a risk review report as Word document", "description": "签合同前最怕的就是条款里埋了坑。找律师审一份合同少则几千多则上万，小公司和个人根本承受不起。自己看又不确定哪些条款有风险——违约金比例合不合理？知识产权归属有没有陷阱？竞业条款是不是过分了？\n\n这个案例演示了将一份「SaaS 软件服务合同」粘贴给 AI，自动生成一份正式 Word 审查报告：逐条扫描标注风险等级（高/中/低），针对每个风险点给出具体的修改建议和替代条款，提取关键商务条款摘要（金额、期限、终止条件、违约责任、数据处理条款等），输出为可存档的 Word 文档格式。\n\n适合：创业者签合作协议、采购经理审供应商合同、自由职业者看外包合同、任何需要快速理解合同风险的人。", "description_en": "The biggest fear before signing a contract is hidden traps in the clauses. Hiring a lawyer costs thousands per contract — unaffordable for small companies and freelancers. Reading it yourself leaves uncertainty: Is the penalty rate reasonable? Are there IP ownership traps? Is the non-compete too aggressive?\n\nThis case demonstrates pasting a \"SaaS Software Service Agreement\" and having AI automatically generate a formal Word review report: scanning each clause with risk levels (high/medium/low), providing specific revision suggestions, and extracting key business terms — all in an archivable Word document format.\n\nBest for: Entrepreneurs signing partnership agreements, procurement managers reviewing vendor contracts, freelancers checking outsourcing contracts, anyone needing to quickly understand contract risks.", "categories": [ "office", "business" ], "tags": [ "合同审查", "法律", "风险防控", "商务" ], "tags_en": [ "Contract Review", "Legal", "Risk Management", "Business" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "合同审查报告（Word）" } ], "prompt": "请审查一份合同，生成一份正式的 Word 审查报告（output.docx）。\n\n合同文本（粘贴完整条款）：\n...\n\n报告要求：\n- 正式 Word 文档，含封面、目录、页眉页脚\n- 合同概要与总体风险评级\n- 关键商务条款摘要表\n- 逐条风险审查，标注风险等级、风险说明与修改建议\n- 总体评估与建议\n用「Word 文档生成」技能输出。", "prompt_en": "Please review a contract and generate a formal Word review report (output.docx).\n\nContract text (paste the full clauses):\n...\n\nReport requirements:\n- A formal Word document with cover page, table of contents, headers and footers\n- Contract summary and overall risk rating\n- Key business terms summary table\n- Clause-by-clause risk review, labeling risk level, risk note, and revision suggestion\n- Overall assessment and recommendations\nUse the \"Word Document Generation\" skill to produce it.", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 187, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-16T10:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "family-finance-checkup", "title": "家庭月度财务健康体检", "title_en": "Family Monthly Financial Health Checkup", "subtitle": "一份让全家看懂的月度财务体检报告", "subtitle_en": "A monthly financial checkup the whole family can understand", "description": "很多家庭月底只知道「钱又没剩多少」，却说不清钱花哪了、储蓄率健不健康、应急金够不够。本案例由「账清清」财务追踪专家生成一份家庭月度财务健康体检——清晰拆解收支结构、对比预算超支/结余、用储蓄率/房贷收入比/应急储备等指标给家庭财务打分，并给出可执行的优化建议。适合想系统打理家庭财务的工薪家庭。", "description_en": "Many families only know 'there's not much left' at month-end, but can't say where the money went, whether their savings rate is healthy, or if their emergency fund is enough. Generated by the Financial Tracker expert, this family monthly financial health checkup clearly breaks down income and expenses, compares actuals against budget (overspend/surplus), scores the household with metrics like savings rate, mortgage-to-income ratio, and emergency reserve, and offers actionable tips. Ideal for working families who want to manage household finances systematically.", "categories": [ "finance" ], "tags": [ "家庭财务", "预算", "储蓄率", "记账" ], "tags_en": [ "Family Finance", "Budget", "Savings Rate", "Expense Tracking" ], "artifact_type": "html", "prompt": "给一个三口之家（夫妻 + 一孩）生成 2026 年 5 月的家庭月度财务健康体检。\n\n税后月收入 ¥38,000（先生 22,000 / 太太 14,000 / 理财收益 2,000）；月支出 ¥26,400（房贷 9,800、餐饮 4,200、教育 3,500、交通 1,800、日用 2,400、医疗 900、娱乐 1,600、保险 1,200、其他 1,000）；月结余 ¥11,600，储蓄率 30.5%。要直观看到：收入/支出/结余/储蓄率 KPI、收支结构图、各类支出的预算对比、财务健康四项指标，以及可执行的体检建议。", "prompt_en": "Generate a family monthly financial health checkup for a three-person household (couple + one child) for May 2026.\n\nAfter-tax monthly income ¥38,000 (husband 22,000 / wife 14,000 / investment income 2,000); monthly expenses ¥26,400 (mortgage 9,800, dining 4,200, education 3,500, transport 1,800, daily goods 2,400, medical 900, leisure 1,600, insurance 1,200, other 1,000); monthly surplus ¥11,600, savings rate 30.5%. Show at a glance: income/expense/surplus/savings-rate KPIs, income & expense breakdown charts, budget vs actual per category, four financial-health metrics, and actionable checkup recommendations.", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "财务体检报告" } ], "skills": [], "experts": [ { "id": "FinancialTracker", "name": "账清清", "description": "精准追踪项目和业务的财务数据，确保每笔收支清晰透明", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "b815cad06dde", "name_en": "Silas · Financial Tracker", "description_en": "Precisely tracking project and business financial data, ensuring every transaction is clear and transparent", "expert_id": "ex_BMscfWCetjZx" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 188, "is_featured": false, "featured_order": 0, "quality_score": 94, "usage_count": 0, "created_at": "2026-06-16T19:30:00Z", "updated_at": "2026-06-16T21:00:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "revealjs-ai-tech-talk", "title": "AI 大模型技术分享演示文稿", "title_en": "AI LLM Tech Talk Presentation", "subtitle": "用 deck-generator 一键生成专业技术分享 Slides", "subtitle_en": "Generate professional tech talk slides with deck-generator", "description": "技术分享准备 PPT 太费时间，模板千篇一律还缺乏互动感？用 deck-generator 技能生成一套大模型技术趋势演讲稿，包含架构对比、RAG 原理、Agent 框架等 15+ 页内容，自带 AI 生成图片和专业动画切换。\n\n本案例演示了如何输入技术分享主题，deck-generator 自动生成包含封面、大纲、技术架构图、代码示例、数据对比、总结和 Q&A 的完整演示文稿，支持 AI 配图让 Slides 更加专业美观。\n\n适合：技术 Leader、架构师、AI 工程师做内部分享或大会演讲。", "description_en": "Preparing tech talk slides is time-consuming and templates lack personality. Use the deck-generator skill to generate a complete LLM tech trends talk with 15+ slides covering architecture comparison, RAG principles, and Agent frameworks, with AI-generated images and professional transitions.\n\nThis case demonstrates how deck-generator auto-generates a complete presentation with cover, outline, architecture diagrams, code examples, data comparisons, summary and Q&A slides — with AI-generated images for a polished look.\n\nPerfect for tech leads, architects, and AI engineers.", "categories": [ "creation", "dev" ], "tags": [ "技术分享", "演示文稿", "大模型", "PPT", "AI架构" ], "tags_en": [ "Tech Talk", "Presentation", "LLM", "PPT", "AI Architecture" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "AI技术分享Reveal.js 演示" } ], "prompt": "帮我生成一套关于\"大模型技术趋势与应用实践\"的技术分享演示文稿。包含：大模型发展简史、主流架构对比（GPT/LLaMA/混元）、RAG 技术原理、Agent 框架介绍、企业落地案例、未来展望，共 15-20 页 slides，每页配有要点和代码/图表示例。请使用 deck-generator 生成专业的 PPTX 演示文稿，配有 AI 生成的配图。", "prompt_en": "Help me generate a tech-talk presentation on \"LLM Technology Trends and Practical Applications.\" Include: a brief history of large models, a comparison of mainstream architectures (GPT/LLaMA/Hunyuan), the principles of RAG, an introduction to Agent frameworks, enterprise deployment case studies, and a future outlook—for a total of 15-20 slides, each with key points and code/chart examples. Please use deck-generator to produce a professional PPTX presentation with AI-generated illustrations.", "skills": [ { "id": "deck-generator", "name": "deck-generator", "description": "AI 驱动的演示文稿生成，统一视觉风格的幻灯片", "description_en": "AI-powered slide deck generation with consistent visual styling", "skill_id": "skill_2053082023531843584" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 95, "sort_order": 189, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-meeting-assistant", "title": "智能会议助手", "title_en": "Intelligent Meeting Assistant Agent", "subtitle": "AI Agent 会议全流程自动化管理工作流", "subtitle_en": "AI Agent Full-Cycle Meeting Automation Workflow", "description": "团队每周 15+ 场会议，会前没人整理议程、会中讨论记不住、会后纪要拖三天才发、行动项无人跟进？这些痛点让 80% 的会议时间被浪费。\n\n这个案例展示了一套完整的 AI Agent 智能会议助手工作流设计：从会前自动议程生成，到会中实时语音转文字与关键信息提取，再到会后结构化纪要自动生成、任务自动创建与跟踪，以及会议效率分析仪表盘。\n\n适合：团队管理者、项目经理、行政助理、产品负责人、希望提升会议效率的中大型团队。", "description_en": "Over 15 meetings a week with no agenda prep, untracked discussions, delayed minutes, and vanishing action items? These pain points waste 80% of meeting time.\n\nThis case presents a comprehensive AI Agent meeting assistant workflow: from pre-meeting auto agenda generation, to real-time transcription and key info extraction during meetings, to post-meeting structured minutes, automated task creation and tracking, and meeting efficiency analytics.\n\nIdeal for: team managers, project managers, executive assistants, product owners, mid-to-large teams.", "categories": [ "agent", "office" ], "tags": [ "会议管理", "AI Agent", "会议纪要", "任务跟踪", "办公自动化" ], "tags_en": [ "meeting management", "AI agent", "meeting minutes", "task tracking", "office automation" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI Agent 智能会议助手工作流设计文档" } ], "prompt": "帮我设计一套 AI Agent 智能会议助手工作流。团队每周有 15+ 场会议，需要覆盖会议全生命周期：1）会前自动议程生成与资料准备；2）会中实时语音转文字与关键信息提取；3）会后自动生成结构化会议纪要；4）任务自动创建与分发到项目管理工具；5）行动项跟踪与到期提醒；6）会议效率分析仪表盘；7）与飞书/钉钉/企业微信的集成方案。输出为详细的 Markdown 工作流设计文档。", "prompt_en": "Help me design an AI Agent intelligent meeting assistant workflow. The team has 15+ meetings per week, and it needs to cover the full meeting lifecycle: 1) automatic pre-meeting agenda generation and material preparation; 2) real-time speech-to-text and key information extraction during meetings; 3) automatic generation of structured meeting minutes after meetings; 4) automatic task creation and dispatch to project management tools; 5) action-item tracking and due-date reminders; 6) a meeting efficiency analytics dashboard; 7) integration plans with Feishu / DingTalk / WeCom. Output a detailed Markdown workflow design document.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "多智能体团队编排：角色定义、任务流转、交接协议与质量门禁", "description_en": "Orchestrate multi-agent teams with roles, task lifecycles, handoffs, and review workflows", "skill_id": "skill_2053079579940413440" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 190, "is_featured": false, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2025-07-18", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "hans-zimmer-epic-mv", "title": "汉斯季默风格MV", "title_en": "Hans Zimmer–Style MV", "subtitle": "以管风琴般的递进和宇宙尺度讲述宇航员回家", "subtitle_en": "An astronaut’s return told with organ-like escalation and cosmic scale", "description": "史诗音乐可视化容易只有宏大空镜，缺少能让观众投入的角色目标与情绪回环。\n\n案例让一名年长宇航员穿越土星、失重舱和扭曲星门，与地球农场中等待的女儿交叉剪辑，最终以一段延迟多年的影像完成重逢。\n\n适合：史诗音乐视觉化、科幻概念片、游戏预告、发布会开场和情绪型品牌片。", "description_en": "Epic music visualization can become a series of empty spectacles without a character goal and emotional return.\n\nAn older astronaut crosses Saturn, a weightless cabin, and a distorted star gate while intercut with his daughter waiting on an Earth farm, completing their reunion through a delayed message.\n\nBest for epic music visualization, science-fiction concepts, game trailers, event openings, and emotional brand films.", "categories": [ "creation" ], "tags": [ "史诗MV", "科幻电影", "宇宙叙事" ], "tags_en": [ "epic music video", "science fiction", "cosmic narrative" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「音乐 MV 创作」专家调用 Flova 的「汉斯季默风格史诗MV」Skill，根据文字想法制作一支约 25 秒、16:9、1080p、24fps 的慢节奏沉浸式科幻 MV《穿越土星的归途》。无旁白、无对口型、无字幕。叙事围绕一名年长宇航员：眼睛映出土星环，飞船从巨大行星边缘掠过；他在失重舱里握住女儿小时候的照片，随后穿越扭曲星门；地球农场中的成年女儿在谷仓打开一段延迟多年的讯息，画面回到宇航员释然的眼神。请先把文字想法拆成情绪曲线和段落表，并确认“慢节奏沉浸式”；锁定规格后生成并展示完整分镜，批准后逐镜生成。音乐采用低频脉冲、渐强弦乐和管风琴质感，但必须为原创，不复制任何现有作品旋律，最后按情绪峰值完成合成。", "prompt_en": "Ask the “Flova Music Video Director” expert to invoke Flova’s “Hans Zimmer–Style Epic MV” Skill and create a slow, immersive 16:9 science-fiction music video titled The Way Home Beyond Saturn, about 25 seconds long at 1080p and 24 fps, based on the following text idea. Use no narration, lip sync, or subtitles. An older astronaut’s eye reflects Saturn’s rings as his spacecraft skims the edge of the giant planet. In zero gravity he holds a childhood photograph of his daughter, then crosses a distorted star gate. On an Earth farm, his now-adult daughter opens a message delayed for years, and the film returns to the astronaut’s relieved gaze. First transform the idea into an emotional curve and section table and confirm the “slow immersive” rhythm. Lock the specification, generate and present the full storyboard, and produce shots only after approval. Use original music with low-frequency pulses, swelling strings, and an organ-like texture, without copying any existing composition, and assemble the film around the emotional peak.", "skills": [], "experts": [ { "id": "flova-music-video-director", "name": "音乐 MV 创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及音乐 MV 成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d487e16aa365", "expert_id": "ex_QHjWRFR9zhMy" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 190, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-campus-recommend", "title": "校招推荐：银行在架校招项目", "title_en": "Campus Hiring: Bank Programmes Now Listed", "subtitle": "45 个银行校招实时清单 + 提前批到春招的节奏表", "subtitle_en": "45 live bank programmes plus the early-batch-to-spring timeline", "description": "校招的残酷在于节奏：提前批 8 月就启动，网申错过没有补录，而多数人 10 月才开始看。\n\n本案例按关键词「银行」实时检索校招库，命中 45 个在架项目（含杭州银行「摘星计划」暑期实习暨提前批），按更新时间倒序列出，并附「提前批 → 秋招 → 春招」的备考节奏建议——银行笔试与公考行测题型高度重合，两条线可以并行。\n\n适合：2027 届毕业生、公考银行双线备考的考生、做就业指导的老师。", "description_en": "Campus recruiting is brutal about timing: early batches open in August with no make-ups, while most students start looking in October.\n\nThis case live-searches the campus database for \"bank\", hitting 45 listed programmes (including Hangzhou Bank's summer-internship-plus-early-batch), sorted by update time, with a preparation timeline from early batch through autumn to spring — and notes that bank aptitude tests overlap heavily with civil-service exams, so the two tracks can run in parallel.\n\nFor: class-of-2027 graduates, dual-track candidates, and careers staff.", "categories": [ "search", "education" ], "tags": [ "校招", "银行", "秋招", "应届生" ], "tags_en": [ "campus recruiting", "banks", "autumn hiring", "fresh graduates" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "银行校招项目清单" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我看看现在有哪些银行校招在架。要求：1) 按关键词「银行」实时检索校招库并给出总数；2) 列出项目名称与工作地点，按更新时间倒序；3) 结合校招一般节奏给出「提前批/秋招/春招」的时间安排建议，并注明这是规律性建议、具体以各行公告为准；4) 说明公考备考生转投银行校招的优势。", "prompt_en": "Show me which bank campus programmes are currently listed. Requirements: 1) live-search the campus database for \"bank\" and give the total; 2) list programme names and locations, newest first; 3) add a preparation timeline across early batch, autumn and spring hiring, noting it's general guidance with each bank's announcement authoritative; 4) explain why civil-service candidates convert well to bank hiring.", "skills": [ { "id": "gaodun-campus-position-search", "name": "校招搜索", "description": "按关键词检索校园招聘项目", "skill_id": "skill_2089642723381682176" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 190, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-campus-tech", "title": "校招推荐：科技类企业", "title_en": "Campus Recruitment: Technology Companies", "subtitle": "109 场科技校招，已经有企业在招 2027 届", "subtitle_en": "109 tech campus drives — some already recruiting the class of 2027", "description": "科技公司的校招节奏普遍比传统行业早一整年，等到大四秋天才开始找，最好的岗位往往已经在提前批发完了。\n\n本案例按关键词检索科技类校招，命中 109 场。把标题里的届别与工作城市拉出来看，可以直接读出各家的节奏差异和放岗形态——本页已有面向 2027 届的校招。案例同时说明校招信息与职位库的区别。\n\n适合：想提前规划校招节奏的大二大三学生、目标科技行业的应届生。", "description_en": "Tech companies run campus recruitment a full year ahead of traditional industries — start looking in your final autumn and the best roles have often gone in the early rounds.\n\nThis case searches tech campus drives by keyword, finding 109. Pulling the graduation year and work city out of each title reveals how far apart companies are on timing and which post across multiple sites — several already target the class of 2027. It also explains how campus listings differ from post databases.\n\nFor: second- and third-year students planning ahead and graduates targeting tech.", "categories": [ "search", "education" ], "tags": [ "校招", "科技企业", "2027届", "提前批" ], "tags_en": [ "campus recruitment", "tech companies", "class of 2027", "early rounds" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "科技类企业校招清单", "label_en": "Technology company campus recruitment list" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查科技类企业的校园招聘信息。要求：1) 真实检索并给出命中总数；2) 列出最新的 14 条，包含校招标题、面向哪一届、工作城市；3) 统计这一页里面向 2027 届和 2026 届各几条，说明科技类企业的校招节奏特点；4) 统计多地放岗与单城市招聘各几条，分析城市选择与行业方向的关系；5) 说明这个接口返回的是校招活动信息而不是具体岗位，字段有哪些限制，想看薪资和学历要求应该怎么办。", "prompt_en": "Search campus recruitment drives at technology companies. Requirements: 1) run the search and give the total; 2) list the 14 most recent with title, target graduating class and work city; 3) count how many target the classes of 2027 and 2026, and characterise tech recruitment timing; 4) count multi-site versus single-city drives and relate city choice to industry direction; 5) explain that this interface returns recruitment events rather than individual posts, which fields are limited, and where to look for salary and education requirements.", "skills": [ { "id": "gaodun-campus-position-search", "name": "校招搜索", "description": "按关键词检索校园招聘项目", "skill_id": "skill_2089642723381682176" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 191, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "saas-product-pitch-deck", "title": "SaaS 产品发布演示文稿", "title_en": "SaaS Product Launch Pitch Deck", "subtitle": "输入产品信息，生成精美的产品发布演示文稿", "subtitle_en": "Input product info, generate an elegant product launch presentation", "description": "SaaS 产品发布会、投资人路演、内部产品评审——都需要一套有说服力的演示材料。专业设计师做一套 deck 要好几天，用 PPT 模板又千篇一律。\n\n这个案例演示了输入产品核心信息（产品定位、核心功能、竞争优势、商业模式、团队介绍），deck-generator 自动生成一套精美的产品发布演示文稿 PPTX：包含封面页、问题定义、解决方案、产品功能展示（2x2 网格）、竞争优势、商业模式（3 栏定价对比）、团队介绍、CTA 页面，配有 AI 生成的专业配图。\n\n适合：创业公司融资路演、产品发布会、内部产品评审、技术分享。", "description_en": "SaaS product launches, investor pitches, internal product reviews — all need compelling presentation materials. Professional designers take days for a deck, PPT templates look generic.\n\nThis case demonstrates inputting core product info (positioning, features, competitive advantages, business model, team), with deck-generator generating a polished product launch PPTX: including cover, problem definition, solution, product features (2x2 grid), competitive advantages, business model (3-column pricing), team intro, and CTA — with AI-generated images.\n\nBest for: Startup fundraising pitches, product launches, internal product reviews, tech talks.", "categories": [ "creation", "dev" ], "tags": [ "演示文稿", "Pitch Deck", "产品发布", "路演", "SaaS" ], "tags_en": [ "Presentation", "Pitch Deck", "Product Launch", "Roadshow", "SaaS" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "SaaS产品发布Reveal.js 演示" } ], "prompt": "请为一款 SaaS 产品生成一套精美的产品发布演示文稿（PPTX），用 deck-generator 生成并配 AI 配图。\n\n产品信息（产品名、定位、品牌色等，替换为你的产品）：\n...\n\n演示内容（约 7 张幻灯片）：封面 → 痛点 → 解决方案 → 核心功能 → 竞争优势 → 商业模式/定价 → CTA。", "prompt_en": "Please generate a polished product launch presentation (PPTX) for a SaaS product, produced with deck-generator and with AI-generated imagery.\n\nProduct info (product name, positioning, brand color, etc. — replace with your own product):\n...\n\nPresentation content (about 7 slides): Cover → Pain points → Solution → Core features → Competitive advantages → Business model/pricing → CTA.", "skills": [ { "id": "deck-generator", "name": "deck-generator", "description": "AI 驱动的演示文稿生成，统一视觉风格的幻灯片", "description_en": "AI-powered slide deck generation with consistent visual styling", "skill_id": "skill_2053082023531843584" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 191, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-campus-state-owned", "title": "校招推荐：央国企专场", "title_en": "Campus Recruitment: State-Owned Enterprises", "subtitle": "174 场校招，一家企业能同时开春招、秋招和全球招", "subtitle_en": "174 drives — one company can run spring, autumn and global rounds at once", "description": "央国企校招批次多、周期长，同一家企业同一届会分好几次发布，只盯着一个入口很容易以为「已经招完了」。\n\n本案例检索央国企校招，命中 174 场。本页同时出现了同一家企业的春季招聘和全球招聘两条——案例据此讲清批次密集、多地放岗、研究院所单列、补招是机会这四个特点，并给出一套按企业建清单的跟进方法。\n\n适合：目标央国企的应届生、错过第一批想找补招的人。", "description_en": "State-owned enterprises run many recruitment rounds over long cycles, and the same company posts the same graduating class several times — watch one entry point and you'll assume it's closed.\n\nThis case searches SOE campus drives, finding 174. The page shows one company's spring and global rounds side by side, illustrating four traits: dense rounds, multi-site postings, institutes posting separately, and supplementary rounds as genuine second chances — with a company-by-company tracking method.\n\nFor: graduates targeting SOEs and anyone who missed the first round.", "categories": [ "search", "education" ], "tags": [ "校招", "央国企", "招聘批次", "补招" ], "tags_en": [ "campus recruitment", "state-owned enterprises", "recruitment rounds", "supplementary" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "央国企校招清单与批次分析", "label_en": "SOE campus recruitment list and round analysis" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我查央国企的校园招聘信息。要求：1) 真实检索并给出命中总数；2) 列出最新的 14 条，包含校招标题、属于哪个批次（春招/秋招/全球招聘等）、工作城市；3) 指出有没有同一家企业同时出现多条不同批次的情况，说明这对求职者意味着什么；4) 统计多地放岗的条数，总结央国企校招的几个特点（批次密集、多地放岗、下属院所单列、补招等）；5) 给出一套跟进央国企校招的具体方法。", "prompt_en": "Search campus recruitment at state-owned enterprises. Requirements: 1) run the search and give the total; 2) list the 14 most recent with title, round (spring, autumn, global) and work city; 3) note whether the same company appears across multiple rounds and what that means for applicants; 4) count multi-site postings and summarise the characteristics of SOE recruitment (dense rounds, multi-site, institutes posting separately, supplementary rounds); 5) give a concrete method for tracking SOE recruitment.", "skills": [ { "id": "gaodun-campus-position-search", "name": "校招搜索", "description": "按关键词检索校园招聘项目", "skill_id": "skill_2089642723381682176" } ], "experts": [], "mcps": [ { "id": "gaodun-job", "name": "大学生实习就业助手", "description": "公考、事业编、教师编选岗决策助手，基于真实职位库按报考资格精准筛选" } ], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 192, "is_featured": false, "featured_order": 0, "quality_score": 89, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "user-funnel-analysis-dashboard", "title": "用户行为漏斗分析报告", "title_en": "User Behavior Funnel Analysis Dashboard", "subtitle": "电商 App 转化漏斗全链路追踪，精准定位流失瓶颈", "subtitle_en": "E-commerce app full-funnel tracking with precise bottleneck identification", "description": "做增长的人都知道「漏斗分析」是必修课，但现实是：从埋点数据里拉出各环节数字后，要手动算转化率、拼对比图、写分析结论，一套完整的漏斗看板做下来半天就没了——更别提还要按渠道拆分找异常、跟上月数据做环比。\n\n这个案例为一款电商 App 构建了完整的用户注册转化漏斗分析看板：58 万月访问用户、6 个转化环节（首页浏览 -> 商品详情 -> 加购 -> 结算 -> 下单 -> 支付）、4 条渠道拆分对比，发现信息流广告的购物车到结算环节流失率异常高达 68%。交互式深色主题 HTML 看板包含：漏斗图、渠道转化率对比折线图、环比趋势柱状图、关键发现与优化建议卡片。\n\n适合：增长团队、产品经理、数据分析师、电商运营、市场投放团队。", "description_en": "Every growth professional knows funnel analysis is essential, but the reality is painful: after pulling numbers from event tracking, you still need to manually calculate conversion rates, build comparison charts, and write analysis conclusions.\n\nThis case builds a comprehensive user conversion funnel dashboard for an e-commerce app: 580K monthly visitors, 6 conversion stages, 4 channel breakdowns, revealing an abnormally high 68% cart-to-checkout drop-off rate for feed ad traffic. The interactive dark-themed HTML dashboard includes: funnel chart, channel comparison line chart, trend bar chart, key insights and optimization recommendation cards.\n\nIdeal for: Growth teams, product managers, data analysts, e-commerce operators, marketing teams.", "categories": [ "data", "business" ], "tags": [ "漏斗分析", "转化率", "用户行为", "数据看板", "电商分析", "渠道对比" ], "tags_en": [ "Funnel Analysis", "Conversion Rate", "User Behavior", "Data Dashboard", "E-commerce Analytics", "Channel Comparison" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请为一款电商 App 构建用户注册转化漏斗分析看板（单文件 HTML，内联无外部依赖），深色主题、交互式。\n\n数据场景：2026 年 3 月，App 总访问 58 万用户，6 个转化环节，按 4 个渠道拆分。\n包含：漏斗图、渠道对比柱状图、关键发现和优化建议卡片、环比趋势折线图。", "prompt_en": "Please build a user-registration conversion-funnel analysis dashboard for an e-commerce app (single-file HTML, inlined with no external dependencies), with a dark theme and interactive.\n\nData scenario: March 2026, the app has 580,000 total visiting users, 6 conversion steps, broken down by 4 channels.\nInclude: a funnel chart, a channel-comparison bar chart, key-findings and optimization-suggestion cards, and a month-over-month trend line chart.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 192, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-21T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "用户漏斗分析仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "nextjs-blog-template", "title": "Next.js 全栈博客模板", "title_en": "Next.js Full-Stack Blog Template", "subtitle": "开箱即用的现代博客，支持 MDX 与深色模式", "subtitle_en": "Production-ready modern blog with MDX and dark mode support", "description": "搭建个人博客是许多开发者的需求，但从零开始配置路由、Markdown 渲染、SEO 优化、响应式布局等功能既繁琐又容易出错。现有模板往往过于简单或过度复杂，难以满足实际需求。\n\n本案例演示了如何利用 AI 一次性生成一个基于 Next.js 14 App Router + TypeScript 的完整博客模板。包含文章列表分页、Markdown/MDX 渲染（含代码高亮）、标签筛选、深色模式、SEO 优化等核心功能，使用本地 MDX 文件作为内容源，无需数据库依赖。\n\n适合想要快速搭建个人技术博客的开发者、技术写作者，以及需要静态站点生成方案的团队使用。", "description_en": "Building a personal blog from scratch involves tedious configuration of routing, Markdown rendering, SEO, and responsive layout.\n\nThis case demonstrates using AI to generate a complete blog template based on Next.js 14 App Router + TypeScript with all essential features including pagination, MDX rendering, tag filtering, dark mode, and SEO optimization.\n\nIdeal for developers wanting to quickly set up a personal tech blog.", "categories": [ "dev", "creation" ], "tags": [ "Next.js", "博客模板", "TypeScript", "全栈开发" ], "tags_en": [ "Next.js", "Blog Template", "TypeScript", "Full-Stack Dev" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.ts", "prompt": "请用 Next.js 14 App Router + TypeScript 生成一个完整的个人博客模板。要求：1) 首页文章列表（标题/摘要/标签/日期）支持分页；2) 文章详情页 Markdown 渲染（含代码高亮）；3) 标签筛选页；4) 关于页；5) 响应式布局；6) 深色模式；7) SEO（metadata/OpenGraph/sitemap）；8) 用本地 MDX 文件作内容源，不依赖数据库。输出项目核心文件。\n\n代码注释和用户可见字符串用简体中文，标识符保留英文。", "prompt_en": "Generate a complete personal blog template using Next.js 14 App Router + TypeScript. Requirements: 1) homepage article list (title/summary/tags/date) with pagination support; 2) article detail page with Markdown rendering (including code highlighting); 3) tag filtering page; 4) about page; 5) responsive layout; 6) dark mode; 7) SEO (metadata/OpenGraph/sitemap); 8) use local MDX files as the content source, with no database dependency. Output the project's core files.\n\nCode comments and user-visible strings should be in Simplified Chinese; keep identifiers in English.", "skills": [], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 193, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.ts", "label": "博客模板代码" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "vue3-admin-dashboard", "title": "Vue3 后台管理仪表盘", "title_en": "Vue3 Admin Dashboard Component", "subtitle": "专业级后台管理仪表盘页面，KPI 卡片+图表+数据表格", "subtitle_en": "Professional admin dashboard with KPI cards, charts, and data tables", "description": "后台管理系统的仪表盘页面总是从零写起，KPI 卡片、图表、数据表格每次都要重复造轮子？\n\n这个案例用前端开发技能生成了一套完整的后台仪表盘页面——4 个 KPI 统计卡片、近 7 天趋势折线图（纯 SVG）、分类占比饼图、可搜索分页数据表格，深色主题，响应式布局，纯 HTML/CSS/JS 单文件实现。\n\n适合：前端开发者、中后台产品团队快速搭建管理后台仪表盘原型。", "description_en": "Admin dashboard pages always start from scratch — KPI cards, charts, tables rebuilt every time.\n\nThis case uses the frontend-dev skill to generate a complete admin dashboard page with 4 KPI stat cards, 7-day trend line chart (pure SVG), category pie chart, and searchable paginated data table. Dark theme, responsive layout, single-file HTML/CSS/JS.\n\nPerfect for frontend devs and admin product teams prototyping dashboard pages.", "categories": [ "dev", "creation" ], "tags": [ "Vue3", "仪表盘", "后台管理", "数据可视化" ], "tags_en": [ "Vue3", "Dashboard", "Admin Panel", "Data Visualization" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个后台管理仪表盘页面（单文件 HTML，内联无外部依赖），深色主题、响应式。\n\n包含：顶部 4 个 KPI 卡片（总用户数、今日订单、总收入、转化率，带趋势箭头和环比）、近 7 天订单趋势 SVG 折线图、订单分类占比 SVG 饼图、底部数据表格（支持搜索/分页/排序）。", "prompt_en": "Please generate an admin management dashboard page (single-file HTML, inlined with no external dependencies), dark theme, responsive.\n\nInclude: 4 KPI cards at the top (Total Users, Today's Orders, Total Revenue, Conversion Rate, with trend arrows and month-over-month change), an SVG line chart of order trends over the last 7 days, an SVG pie chart of order category distribution, and a data table at the bottom (supporting search/pagination/sorting).", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 194, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "Vue3 管理后台仪表盘" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "product-comparison-matrix", "title": "产品选型对比矩阵", "title_en": "Product Comparison Decision Matrix", "subtitle": "输入候选产品和评估维度，生成加权评分对比矩阵表格", "subtitle_en": "Input candidate products and evaluation criteria, generate weighted scoring comparison matrix", "description": "技术选型最难的不是调研，是决策——三个数据库、四个云服务商、五个前端框架，每个都有优缺点，每个同事都有偏好，最后拍脑袋决定。缺一个客观、可量化的对比方法。\n\n这个案例演示了输入候选后端框架和评估维度，AI 自动生成一份完整的产品选型对比矩阵 Excel 表格：评估维度与权重定义（8 个维度，加权百分比）、候选产品多维度打分（1-10 分制）、加权总分自动计算排名、每项评分附带简要依据、最终推荐结论与各产品优劣势分析。公式自动加权汇总，一目了然。\n\n适合：技术架构师做框架选型、采购部门比选供应商、产品经理评估竞品、团队方案评审决策。", "description_en": "The hardest part of technology selection isn't research — it's the decision. Three databases, four cloud providers, five frontend frameworks, each with pros and cons, each colleague with their own preference, and the final call comes down to gut feeling. What's missing is an objective, quantifiable comparison method.\n\nThis case demonstrates inputting candidate backend frameworks and evaluation criteria, with AI generating a complete product comparison decision matrix Excel spreadsheet: evaluation dimensions with weighted percentages (8 dimensions), multi-dimensional scoring for each candidate (1-10 scale), auto-calculated weighted total rankings, brief justification for each score, and a final recommendation with pros/cons analysis. Formulas auto-calculate weighted summaries at a glance.\n\nBest for: Tech architects selecting frameworks, procurement teams comparing vendors, product managers evaluating competitors, team decision-making on proposals.", "categories": [ "data", "dev" ], "tags": [ "技术选型", "对比矩阵", "Excel", "评估打分", "决策工具" ], "tags_en": [ "Tech Selection", "Comparison Matrix", "Excel", "Scoring Evaluation", "Decision Tool" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "prompt": "请为后端框架技术选型生成一份加权评分对比矩阵 Excel 表格。\n\n选型背景：15 人后端团队，高并发微服务架构，现技术栈 Python Flask 准备迁移；候选框架 Go Gin、Rust Axum、Node.js Express、Java Spring Boot。（可替换为你的候选产品与背景）\n\n请生成以下工作表：\n1. 评估维度定义：约 8 个维度及权重，每维含评分标准说明\n2. 评分矩阵：候选产品 × 维度打分（1-10），每分附一句依据，加权得分自动计算\n3. 推荐报告：最终排名、各产品优劣势、推荐结论与迁移建议\n\n表格要求：加权得分用公式计算，按加权总分降序，分数高低用条件格式区分。", "prompt_en": "Please generate a weighted-scoring comparison matrix as an Excel spreadsheet for a back-end framework technology selection.\n\nSelection background: a 15-person back-end team, high-concurrency microservice architecture, current stack Python Flask (preparing to migrate); candidate frameworks Go Gin, Rust Axum, Node.js Express, Java Spring Boot. (Replace with your own candidates and background.)\n\nPlease generate the following worksheets:\n1. Evaluation Dimension Definitions: about 8 dimensions and weights, each with scoring-criteria notes\n2. Scoring Matrix: candidate products × dimensions scored 1-10, each score with a one-sentence rationale, weighted scores auto-calculated\n3. Recommendation Report: final ranking, each product's pros/cons, recommendation and migration advice\n\nSpreadsheet requirements: weighted scores computed by formula, sort descending by weighted total, and use conditional formatting to distinguish high vs. low scores.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 195, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.xlsx", "label": "产品选型对比矩阵" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "weather-dashboard-widget", "title": "天气预报组件", "title_en": "Weather Dashboard Widget", "subtitle": "输入城市天气数据，生成精美的天气预报卡片", "subtitle_en": "Input city weather data, generate a beautiful weather forecast card", "description": "嵌入官网的天气模块、智能家居面板的天气组件、旅行计划的天气参考——都需要一个好看的天气展示卡片。原生天气 API 返回的是 JSON 数据，怎么快速变成一个可用的视觉组件？\n\n这个案例演示了输入一个城市的当前天气和 5 日预报数据，AI 自动生成一个精美的天气预报组件 HTML 页面：居中卡片（max-width 560px）、当前天气大展示区（温度、天气图标、体感温度、风力湿度）、5 日预报横排小卡片、日出日落时间线。渐变蓝天背景，简洁现代。\n\n适合：天气应用 UI 原型、智能家居面板组件、旅行规划工具、前端练习项目。", "description_en": "Weather modules for websites, smart home panel widgets, travel planning weather references — all need attractive weather display cards. Native weather APIs return JSON data, but how to quickly turn it into a usable visual component?\n\nThis case demonstrates inputting a city's current weather and 5-day forecast data, with AI generating a beautiful weather forecast widget HTML page: centered card (max-width 560px), current weather hero area (temperature, weather icon, feels-like, wind/humidity), 5-day forecast horizontal cards, sunrise/sunset timeline. Gradient blue sky background, clean and modern.\n\nBest for: Weather app UI prototypes, smart home panel components, travel planning tools, frontend practice projects.", "categories": [ "data", "creation" ], "tags": [ "天气预报", "仪表盘组件", "数据可视化", "UI组件" ], "tags_en": [ "Weather Forecast", "Dashboard Widget", "Data Visualization", "UI Component" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "prompt": "请生成一个天气预报组件页面（单文件 HTML，内联无外部依赖），精美现代、蓝天渐变氛围。\n\n天气（深圳，2026-05-08）：多云转晴 28℃，体感 31℃，湿度 72%，东南风 3 级，含 5 日预报。\n\n包含：大字当前温度 + 天气图标、信息栏（体感/湿度/风力/AQI）、日出日落时间轴、5 日预报横排卡片。", "prompt_en": "Please generate a weather forecast widget page (single-file HTML, inline with no external dependencies), elegant and modern with a blue-sky gradient atmosphere.\n\nWeather (Shenzhen, 2026-05-08): cloudy turning clear, 28℃, feels like 31℃, humidity 72%, southeast wind force 3, including a 5-day forecast.\n\nInclude: large-font current temperature + weather icon, an info bar (feels-like/humidity/wind/AQI), a sunrise-sunset timeline, and a horizontal row of 5-day forecast cards.", "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 196, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-16T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.html", "label": "天气仪表盘组件" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "flight-price-compare", "title": "国际机票多航司比价", "title_en": "International Flight Price Comparison", "subtitle": "搜索北京→东京机票，AI 帮你选出最划算的航班", "subtitle_en": "Search Beijing-Tokyo flights and let AI pick the best value option", "description": "出国机票价格差异巨大，在各平台反复搜索、比价费时费力，还容易错过最优组合。\n\n这个案例演示了如何用自然语言搜索国际机票：指定出发地、目的地和日期，AI 实时拉取多家航司数据，生成包含价格、飞行时长、转机次数的交互式比价页面，支持按价格、时长排序和直飞/转机筛选，并自动推荐最便宜、最快和性价比最高三个选项。\n\n适合：出境旅行计划者、精打细算的旅客、需要快速比价的差旅助理", "description_en": "International flight prices vary wildly across airlines and platforms, and comparison shopping is tedious.\n\nThis case shows how to search international flights with natural language: specify origin, destination and date, and AI pulls real-time data from multiple airlines, generating an interactive comparison page with prices, flight duration, and stops — supporting sort by price/duration and filter by direct/transfer, then automatically recommends the cheapest, fastest, and best-value options.\n\nBest for: International trip planners, budget-conscious travelers, travel assistants doing corporate fare comparisons", "categories": [ "travel" ], "tags": [ "国际机票", "比价", "航班搜索", "出境旅行" ], "tags_en": [ "International Flights", "Price Comparison", "Flight Search", "Outbound Travel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "机票比价交互页面" } ], "skills": [ { "id": "flights", "name": "flights", "description": "通过 Google Flights 搜索航班，支持城市名、舱位筛选和预订链接", "description_en": "Search Google Flights for real-time prices, schedules and booking links", "skill_id": "skill_2053081985088258048" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "帮我搜索2026年6月15日北京到东京的国际机票，对比国航、全日空、大韩、日航、春秋、东航、国泰、韩亚等航司的价格、飞行时长和转机信息，生成一个交互式比价页面，支持按价格和时长排序、按直飞和转机筛选，并自动推荐最便宜、最快和性价比最高的航班。", "prompt_en": "Help me search for international flights from Beijing to Tokyo on June 15, 2026, comparing the prices, flight durations, and layover information of airlines such as Air China, ANA, Korean Air, JAL, Spring Airlines, China Eastern, Cathay Pacific, and Asiana. Generate an interactive price-comparison page that supports sorting by price and duration and filtering by direct/layover flights, and that automatically recommends the cheapest, fastest, and best-value flights.", "source": "official", "sort_order": 197, "featured_order": 0, "quality_score": 91, "usage_count": 0, "is_featured": false, "created_at": "2026-04-28T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "chatlaw-labor-forced-resignation-compensation", "title": "被迫离职补偿咨询", "title_en": "Forced Resignation Compensation Consultation", "subtitle": "5 角色法律团队：案情采集、法条判例分析与风险评估，一次产出可直接行动的劳动纠纷报告", "subtitle_en": "5-role legal team: case intake, statute and precedent analysis, risk assessment, and an actionable labor-dispute report in one run", "description": "被公司突然从总监降到助理、工资直接砍掉 60%，最难受的不是面子，而是根本不知道该不该立刻走、怎么走才拿得到补偿、晚一天会不会就被认定默认接受。劳动纠纷里最怕的就是又气又急，结果错过了异议和仲裁时点。\n\n这个案例演示了中文法律咨询团如何完整处理一类高频劳动争议：先由信息采集团队拆清时间线、合同条款和证据缺口，再由法条研究与判例分析并行给出《劳动合同法》第35条、第38条、第46条、第47条的适用路径、北京地区相关裁判趋势、经济补偿和工资差额的测算逻辑，最后汇总成一份可直接执行的《法律咨询报告》，附 12 项证据清单、5 步行动计划和被迫解除操作重点。\n\n适合：遭遇违法调岗降薪、被迫离职、拖欠工资等问题的劳动者；做人力资源合规与劳动争议预判的 HR/法务；需要快速形成劳动仲裁前置方案的律师助理与咨询顾问。", "description_en": "When a company suddenly cuts you from director to assistant and slashes pay by 60%, the hardest part is not only the humiliation — it's not knowing whether to leave now, how to leave while preserving compensation rights, and whether one more day of silence will be treated as acceptance. In labor disputes, the biggest trap is acting in anger and missing the key objection and arbitration window.\n\nThis case shows how the ChatLaw Team handles a common labor-dispute scenario end to end: the intake role first structures the timeline, contract terms, and evidence gaps; statute and precedent researchers then work in parallel on Labor Contract Law Articles 35, 38, 46, and 47, Beijing-area judicial trends, and the calculation logic for wage arrears and economic compensation; finally the team assembles everything into a practical legal consultation report with a 12-item evidence checklist, a 5-step action plan, and key reminders for a forced-resignation strategy.\n\nTarget audience: employees facing unlawful demotion, pay cuts, forced resignation, or wage arrears; HR and legal teams doing labor-risk pre-assessment; legal assistants and consultants preparing pre-arbitration strategies for labor disputes.", "categories": [ "business", "creation", "agent" ], "tags": [ "劳动纠纷", "被迫离职", "经济补偿", "法律咨询", "多Agent协作" ], "tags_en": [ "Labor Dispute", "Forced Resignation", "Economic Compensation", "Legal Consultation", "Multi-Agent Collaboration" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "劳动纠纷法律咨询报告（Markdown）" } ], "prompt": "我在某互联网公司工作 5 年 3 个月，上个月公司单方面把我岗位从“产品总监”调到“产品助理”，月薪从 3 万降到 1.2 万。我不同意，现在公司天天给我安排与岗位不匹配的琐碎工作。我考虑申请被迫离职并要求经济补偿。\n\n请使用中文法律咨询团完整流程处理：\n1. 先做信息采集，必要信息缺失时向我追问；\n2. 信息充分后输出完整《法律咨询报告》Markdown，必须包含：案情摘要、适用法条（劳动合同法/劳动法/民法典/最高法解释）、类似判例 3-5 个、法律定性、胜诉风险评估、推荐处置路径（协商/仲裁/诉讼）、风险与时效提醒、下一步行动建议；\n3. 单独附一份“证据收集清单”，告诉我接下来需要保留哪些证据；\n4. 请明确经济补偿、工资差额及可能的加付赔偿金如何计算。", "prompt_en": "I've worked at an internet company for 5 years and 3 months. Last month the company unilaterally reassigned me from \"Product Director\" to \"Product Assistant\" and cut my monthly salary from 30,000 to 12,000. I did not agree, and now the company assigns me trivial work that doesn't match my position every day. I'm considering claiming constructive dismissal (forced resignation) and demanding economic compensation.\n\nPlease handle this using the full Chinese legal consultation team workflow:\n1. First do information gathering, and ask me follow-up questions when essential information is missing;\n2. Once information is sufficient, output a complete \"Legal Consultation Report\" in Markdown that must include: case summary, applicable statutes (Labor Contract Law / Labor Law / Civil Code / Supreme People's Court interpretations), 3-5 similar precedents, legal characterization, assessment of the chances of winning, recommended courses of action (negotiation / arbitration / litigation), risk and statute-of-limitations reminders, and recommended next steps;\n3. Separately attach an \"Evidence Collection Checklist\" telling me which evidence I need to preserve going forward;\n4. Please clearly explain how to calculate the economic compensation, the wage differential, and any possible additional/punitive damages.", "skills": [], "experts": [ { "id": "ChatLawTeam", "name": "中文法律咨询团", "description": "案情采集、法条研究、判例分析、建议撰写，为民事、婚姻、合同、劳动等高频场景出具专业法律咨询报告。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "2a70787f0a5a", "name_en": "ChatLaw Team", "description_en": "Case intake, statute research, precedent analysis, and advice drafting—delivering professional legal consultation reports for civil, marriage, contract, and labor cases.", "expert_id": "ex_pbny8vaVXPvg" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 198, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "flight-status-realtime", "title": "实时航班状态追踪仪表盘", "title_en": "Real-Time Flight Status Dashboard", "subtitle": "交互式航班追踪面板：飞行路线、实时数据、时间线一目了然", "subtitle_en": "Interactive flight tracking panel with route visualization, live data and timeline", "description": "在机场等接人，不知道航班到底飞到哪了、有没有延误、登机口变没变——反复刷航班App既麻烦又费时。\n\n这个案例生成了一个暗色科技感的实时航班状态仪表盘 HTML 页面，包含：SVG 弧线飞行路线可视化（带飞机动画）、四维飞行数据面板（高度/速度/剩余时间/距离）、完整航班时间线、出发地与目的地天气信息。支持 CSS 动画交互效果，零外部依赖。\n\n适合：机场接送的等待者、商务出行的常旅客、需要实时掌握航班动态的人", "description_en": "Waiting at the airport to pick someone up — no idea where their flight is, whether it's delayed, or if the gate changed.\n\nThis case generates a dark-themed, tech-aesthetic real-time flight status dashboard as a single HTML page, featuring: SVG arc flight route visualization with animated plane icon, four-dimension flight data panels (altitude/speed/ETA/distance), complete flight timeline, and departure/arrival weather info. Includes CSS animations and hover interactions with zero external dependencies.\n\nBest for: Airport pickup coordinators, frequent business travelers, anyone needing live flight monitoring", "categories": [ "travel" ], "tags": [ "航班追踪", "实时仪表盘", "数据可视化", "前端开发" ], "tags_en": [ "Flight Tracking", "Real-Time Dashboard", "Data Visualization", "Frontend Dev" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "航班状态追踪仪表盘" } ], "skills": [ { "id": "aviationstack-flight-tracker", "name": "aviationstack-flight-tracker", "description": "实时航班追踪：登机口、延误、飞行位置，类 Flighty 体验（需 API Key）", "description_en": "Real-time flight tracking with gate info, delays and live position via Aviationstack", "skill_id": "skill_2053081372760412160" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "请生成一个实时航班状态追踪仪表盘页面（单文件 HTML，内联无外部依赖），暗色科技感、带飞行动画。\n\n航班 CA1234，北京首都(PEK)→上海虹桥(SHA)。\n\n包含：顶部状态栏（航班号 + 准时标签）；弧线飞行路线可视化（带动画飞机图标，当前 65% 进度）；4 个飞行数据卡片（高度 10668m / 速度 856km·h / 剩余 48min / 距离约 1075km）；航班时间线（登机→起飞→当前→着陆）；出发地与目的地天气。", "prompt_en": "Please generate a real-time flight-status tracking dashboard page (single-file HTML, inlined with no external dependencies), with a dark tech aesthetic and a flight animation.\n\nFlight CA1234, Beijing Capital (PEK) → Shanghai Hongqiao (SHA).\n\nInclude: a top status bar (flight number + on-time tag); an arc-shaped flight-route visualization (with an animated plane icon, currently at 65% progress); 4 flight-data cards (altitude 10668m / speed 856km/h / 48min remaining / approx. 1075km distance); a flight timeline (boarding → takeoff → current → landing); and weather at the origin and destination.", "source": "official", "sort_order": 199, "featured_order": 0, "quality_score": 92, "usage_count": 0, "is_featured": false, "created_at": "2026-04-28T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "customer-service-sop", "title": "客服退换货 SOP 标准操作流程", "title_en": "Customer Service Returns SOP", "subtitle": "描述业务场景，AI 自动生成可落地执行的 Word 版标准操作流程手册", "subtitle_en": "Describe a business scenario, AI auto-generates an actionable SOP manual as Word document", "description": "客服团队规模一大，退换货处理就成了重灾区——有人先退款再收货，有人要客户等质检结果三天没回复，VIP 客户和普通客户的处理标准全凭个人经验。20 人团队每天 200 件退换货工单，没有统一 SOP 就是一笔糊涂账。\n\n这个案例演示了输入一个具体的业务场景（电商客服退换货处理），AI 自动生成一份完整的 Word SOP 文档：涵盖 5 大退换货场景的分步操作指引，每个环节配有标准话术模板、时效要求、权限矩阵、质检流程、异常处理指引和关键 KPI 指标。生成正式 Word 格式，打印出来就能培训新人。\n\n适合：客服主管、运营经理、质量管理人员、电商团队搭建标准化流程、任何需要把经验变成可执行文档的团队。", "description_en": "When customer service teams grow, return and exchange handling becomes chaotic — some agents refund before receiving goods, others leave customers waiting days for quality check results. Without a unified SOP, complaint rates stay high and new hires have nowhere to start.\n\nThis case demonstrates entering a specific business scenario (e-commerce customer service return/exchange handling) and having AI auto-generate a complete Word SOP document with step-by-step guides, standard response templates, time requirements, permission matrices, quality inspection workflows, exception handling, and key KPIs. Generated as a formal Word document, print-ready for training.\n\nBest for: Customer service leads, operations managers, quality managers, e-commerce teams building standardized processes.", "categories": [ "office", "business" ], "tags": [ "SOP", "退换货", "客服管理", "流程标准化" ], "tags_en": [ "SOP", "Returns & Exchanges", "Customer Service", "Process Standardization" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "客服退换货SOP（Word）" } ], "prompt": "为一家电商公司的客服团队编写退换货处理标准操作流程（SOP）Word 文档（output.docx）。\n\n团队规模：20 人，分在线组与电话组，每日约 150-200 件退换货工单。（可替换为你的团队信息）\n\n流程需覆盖场景：7 天无理由退货、质量问题退换货、错发/漏发补发、超期特殊处理。\n\nSOP 须包含：流程图描述、每环节操作步骤与话术模板、时效要求、权限矩阵、异常处理指引、关键 KPI。\n\n格式：含封面页、目录、页眉页脚、多级标题编号、统一表格样式，输出正式 Word 文档（output.docx）。", "prompt_en": "Write a Standard Operating Procedure (SOP) Word document (output.docx) for an e-commerce company's customer service team's return and exchange handling.\n\nTeam size: 20 people, split into online and phone groups, handling about 150-200 return/exchange tickets per day. (Replace with your own team info.)\n\nThe process must cover: 7-day no-reason returns, quality-issue returns/exchanges, reshipment for wrong/missing deliveries, and special handling beyond the window.\n\nThe SOP must include: a flowchart description, operating steps and script templates for each stage, time requirements, an authority matrix, exception handling guidance, and key KPIs.\n\nFormat: include a cover page, table of contents, headers and footers, multi-level heading numbering, and a consistent table style; output as a formal Word document (output.docx).", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 200, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-21T10:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "hmc-rx-academic-promo-review", "title": "处方药学术推广材料合规审查", "title_en": "Prescription-Drug Academic-Promotion Material Compliance Review", "subtitle": "医学资料/卫星会PPT过审，专查超适应症、疗效断言、处方药大众媒体红线", "subtitle_en": "Review medical materials / satellite-symposium decks — flags off-label claims, efficacy assertions, Rx mass-media red lines", "description": "药企学术推广材料最容易踩「超适应症表述」「疗效断言」「处方药变相大众媒体推广」三条红线，一旦被查就是广告法第16条、药品法级别的违规。\n\n本灵感把一份处方药学术推广材料（医学资料/卫星会PPT/拜访资料）交给医疗营销合规专家，按广告法第16条、药品管理法、处方药禁止大众媒体推广等规则逐条审查，标出超适应症、无依据疗效断言、变相软文、搜索竞价含处方药名等高风险点，并给出合规改写与「需内部三级审核」提示。\n\n适合: 药企市场部/医学部、医药代表、合规与法务、学术会议承办方。", "description_en": "Pharma academic-promotion materials most easily cross three red lines — off-label indication claims, efficacy assertions, and disguised mass-media promotion of prescription drugs — each a violation at the level of Advertising Law Article 16 and the Drug Administration Law.\n\nThis inspiration hands a prescription-drug academic-promotion material (medical material / satellite-symposium deck / representative leave-behind) to the Healthcare Marketing Compliance Expert. It reviews item by item against Advertising Law Article 16, the Drug Administration Law, and the ban on Rx mass-media promotion, flagging off-label claims, unsubstantiated efficacy assertions, disguised soft-ads, and search-bidding with Rx brand names, then provides compliant rewrites and a requires internal three-tier review note.\n\nSuitable for: pharma marketing / medical affairs, medical representatives, compliance & legal, academic-conference organizers.", "categories": [ "business" ], "tags": [ "处方药合规", "学术推广", "广告法" ], "tags_en": [ "Rx Compliance", "Academic Promotion", "Advertising Law" ], "artifact_type": "html", "preview": "output.html", "prompt": "我在药企市场部，这份降糖药的学术拜访资料下周就要给代表用了，里面还规划了短视频平台的推广动作。老板催着上线，但我心里发虚——处方药能这么推吗？帮我看看哪些地方会出事、罚起来有多重、每处该怎么改。\n\n材料原文：\n\n---\n\n# 格泰平® 二甲双胍缓释片｜区域学术拜访资料 & 卫星会提纲\n\n## 第一章 产品定位\n\n格泰平®，**彻底终结血糖烦恼**的新一代缓释制剂。独家缓释技术实现 **100% 平稳控糖**，让您的患者**彻底告别并发症困扰**。\n\n## 第二章 临床数据\n\n- MULTI-CARE 真实世界研究（n=2,847）：**三个月血糖达标率 85.3%**\n- 与普通片剂头对头比较：**格泰平® 有效率 91.2%，显著优于竞品 A（76.4%）**\n- 胃肠道不良反应发生率仅 3.1%，**安全性有保障**\n\n## 第三章 患者故事\n\n**李阿姨，58 岁，病程 12 年**\n\"以前吃别的药胃里翻江倒海，换成格泰平® 之后一点感觉都没有，血糖也稳了。现在我天天跟老姐妹们推荐这个药！\"\n\n—— 建议代表拜访时播放李阿姨的视频证言，增强医生信心。\n\n## 第四章 拓展适应症机会点\n\n除已获批的 2 型糖尿病，以下场景可作为**学术沟通切入点**：\n- 糖尿病前期人群的早期干预\n- 多囊卵巢综合征患者的代谢管理\n- 配合减重需求患者的体重管理\n\n（以上适应症尚在研究阶段，可作为学术探讨方向引导处方）\n\n## 第五章 竞品对话话术\n\n医生若提到竞品 B，可回应：\"竞品 B 的缓释技术是上一代的，血药浓度波动大，格泰平® 才是真正的平稳释放。\"\n\n## 第六章 数字化推广规划\n\n为扩大品牌声量，Q4 启动以下动作：\n1. 抖音、小红书品牌官方账号开设\"控糖知识\"专栏，自然植入格泰平® 产品露出\n2. 邀请 20 位健康类博主进行产品\"种草\"，突出\"平稳控糖不伤胃\"卖点\n3. 搜索引擎投放\"二甲双胍 哪个牌子好\"\"格泰平 效果\"等关键词\n4. 直播间引导观众线上问诊后购药，打通转化闭环\n\n## 第七章 代表行为指引\n\n- 拜访目标医生每月不少于 4 次\n- 可根据科室处方量安排学术会议赞助\n- 建立重点医生微信群，定期推送产品资料\n\n---\n\n我最担心第六章那套数字化推广和第四章的适应症拓展。麻烦按风险高低排序，标清楚哪些是碰了就要吃罚单的、哪些改改文字还能救，改法直接写出来。另外告诉我这份材料现在能不能发。", "prompt_en": "I work in pharma marketing. This academic detail aid for our antidiabetic drug goes to the rep team next week, and it also lays out a short-video platform push. My boss is pressing to launch, but I'm uneasy — can prescription drugs really be promoted this way? Please tell me which parts will get us in trouble, how heavy the penalties are, and how to fix each one.\n\nMaterial:\n\n---\n\n# Getaiping® Metformin ER Tablets｜Regional Academic Detail Aid & Satellite Symposium Outline\n\n## Chapter 1 Product Positioning\n\nGetaiping® — the next-generation extended-release formulation that **puts an end to blood sugar worries once and for all**. Proprietary ER technology delivers **100% steady glycemic control**, letting your patients **completely leave complications behind**.\n\n## Chapter 2 Clinical Data\n\n- MULTI-CARE real-world study (n=2,847): **85.3% reached glycemic target in three months**\n- Head-to-head vs. standard tablets: **Getaiping® efficacy 91.2%, significantly superior to Competitor A (76.4%)**\n- GI adverse event rate only 3.1% — **safety assured**\n\n## Chapter 3 Patient Story\n\n**Aunt Li, 58, 12-year disease duration**\n\"Other drugs turned my stomach upside down. After switching to Getaiping® I feel nothing at all, and my glucose is steady. Now I recommend this drug to all my friends every day!\"\n\n—— Recommend reps play Aunt Li's video testimonial during visits to build physician confidence.\n\n## Chapter 4 Indication Expansion Opportunities\n\nBeyond the approved type 2 diabetes indication, the following can serve as **academic engagement entry points**:\n- Early intervention in prediabetes\n- Metabolic management in PCOS patients\n- Weight management for patients seeking weight loss\n\n(These indications remain investigational and may be framed as academic discussion topics to guide prescribing.)\n\n## Chapter 5 Competitor Talk Track\n\nIf a physician raises Competitor B, respond: \"Competitor B's ER technology is last-generation with large plasma fluctuations. Getaiping® delivers truly steady release.\"\n\n## Chapter 6 Digital Promotion Plan\n\nTo amplify brand voice, launch the following in Q4:\n1. Open a \"glycemic knowledge\" column on official Douyin and Xiaohongshu accounts with organic Getaiping® product placement\n2. Engage 20 health influencers for product seeding, emphasizing \"steady control, gentle on the stomach\"\n3. Buy search keywords such as \"which metformin brand is best\" and \"Getaiping results\"\n4. Guide livestream viewers to online consultation then purchase, closing the conversion loop\n\n## Chapter 7 Rep Conduct Guidance\n\n- Minimum 4 visits per month to target physicians\n- Arrange academic meeting sponsorship based on departmental prescription volume\n- Build WeChat groups for key physicians and push product materials regularly\n\n---\n\nChapter 6's digital plan and Chapter 4's indication expansion worry me most. Rank issues by severity, flag which ones draw fines on contact versus which can be salvaged with wording changes, and write the fixes out directly. Also tell me whether this material can ship as-is.", "skills": [], "experts": [ { "id": "nges-healthcare-compliance-lite", "name": "腾讯健康NGES合规专家", "description": "医疗营销合规专家，审核药品广告、医美种草、保健品声称，识别合规风险并支持改写为合规版本。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ytn7WGJKoya8", "promptFileSnapshot": "f4a36dc03d8d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 200, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "处方药学术推广合规审核报告" }, { "file": "output.md", "label": "处方药学术推广合规审核报告（Markdown 版）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "theme_label": "合规红 · 处方药学术推广审查", "theme_color": "#d9262c", "is_recommended": false, "recommended_order": 0 }, { "id": "japanese-youth-idol-mv", "title": "日系青春偶像MV", "title_en": "Japanese Youth Idol MV", "subtitle": "四位少女从教室、电车与练习室走到黄昏舞台", "subtitle_en": "Four friends move from classroom and train to rehearsal and a sunset stage", "description": "偶像群像 MV 很容易变成整齐舞蹈合集，缺少真实友情、校园日常和青春的留白。\n\n案例跟随四位少女经历放学、赶电车、排练失误、互相打气和礼堂演出，最后在海边黄昏定格笑容，保留手持胶片的生活感。\n\n适合：女子组合、校园歌曲、毕业纪念、青春品牌内容和竖横版社交 MV。", "description_en": "Idol ensemble videos can become polished dance compilations with no lived-in friendship, school routine, or youthful pauses.\n\nFour girls move through class, trains, rehearsal mistakes, mutual encouragement, and an auditorium performance before their smiles freeze at a seaside sunset, all with handheld film texture.\n\nBest for girl groups, school songs, graduation memories, youth brands, and social music videos.", "categories": [ "creation" ], "tags": [ "日系偶像", "校园青春", "胶片MV" ], "tags_en": [ "Japanese idol", "school youth", "film-look MV" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「音乐 MV 创作」专家调用 Flova 的「日系青春偶像MV」Skill，制作一支约 60 秒、16:9 的原创群像 MV《放学后的夏日约定》，歌词语言为日语。四位女高中生成员的外貌、发型和校服配色分别固定；结构按教室日常→放学奔跑→电车同行→练习室失误与互相打气→学校礼堂舞台→海边黄昏收尾推进。视觉采用柔和手持、偶发失焦、夏日胶片颗粒和粉彩低对比；舞蹈保留略微不整齐的真实活力，不做高精度商业编舞。请先锁定比例、时长、语言、色彩系统并暂停确认；再建立角色/场景/道具、完整分镜和 BGM/环境音，逐阶段确认角色图、场景图、所有 shot 与音频层，最后按强拍剪辑并在四人海边笑容上柔和淡出。", "prompt_en": "Ask the “Flova Music Video Director” expert to invoke Flova’s “Japanese Youth Idol MV” Skill and create an original 16:9 ensemble music video of about 60 seconds titled Our Summer Promise After School, with Japanese lyrics. Lock distinct appearances, hairstyles, and school-uniform colors for four high-school members. Structure the film as classroom routine → running after school → riding the train together → rehearsal mistakes and mutual encouragement → a school auditorium performance → a seaside sunset ending. Use soft handheld movement, occasional missed focus, summer film grain, and pastel low contrast. Keep the dancing lively and slightly imperfect rather than commercially precise. First lock aspect ratio, duration, language, and color system and pause for approval. Then establish characters, locations, props, the complete storyboard, and BGM and ambience. Obtain approval for character images, location images, all shots, and audio layers stage by stage. Finish with beat-aligned editing and a soft fade on the four friends smiling by the sea.", "skills": [], "experts": [ { "id": "flova-music-video-director", "name": "音乐 MV 创作", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及音乐 MV 成片。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "d487e16aa365", "expert_id": "ex_QHjWRFR9zhMy" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 200, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-career-assessment", "title": "生涯测评：考公考研还是躺平", "title_en": "Career Path: Civil Service, Grad School, or Coasting?", "subtitle": "在对话里直接作答的测评卡，答完当场出方向报告", "subtitle_en": "An answerable in-chat assessment card with an instant direction report", "description": "「毕业到底要干嘛」是大一大二问得最多的问题。考公、考研、留学、就业、创业、躺平，六条路听着都有理。\n\n本案例是一张可直接作答的交互测评卡：10 道校园情境题、每题 6 选 1，答完点提交当场算出六方向分布并展开报告。本例作答 10 题里 7 题选了公考/央国企方向，得出「天选公考/央国企圣体」，附大学四年规划与备考路径。可改答案重新提交。\n\n适合：还没想清楚去向的大一大二学生、上生涯规划课的辅导员。", "description_en": "\"What am I actually going to do after graduation?\" is the question first- and second-years ask most. Civil service, grad school, abroad, a job, a startup, or coasting — all six sound plausible.\n\nThis case is an interactive assessment card answered right in place: 10 campus-scenario questions, six options each; submit and it scores the six directions and expands the report instantly. This run picked the civil-service/SOE direction on 7 of 10 items, yielding \"born for public service\" with a four-year plan and preparation path. Change answers and resubmit anytime.\n\nFor: undecided early-years students and counsellors teaching career planning.", "categories": [ "education" ], "tags": [ "生涯规划", "职业测评", "大学生", "交互卡片" ], "tags_en": [ "career planning", "assessment", "college students", "interactive card" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "生涯测评交互卡片 + 公考/央国企方向报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我大二，纠结毕业后考公、考研还是找工作，帮我做个生涯测评。要求：1) 读题库后把 10 道题一次性渲染成可作答的交互卡片，题干、提示语、6 个选项完整显示；2) 我的作答（按题号 1-10）：A A A E E A E A A A；3) 用测评脚本按真实评分规则算分；4) 报告含胜出方向、得分、结论及该方向的完整四段文案（用题库原文）；5) 报告出来后仍可改答案重新提交。", "prompt_en": "I'm a second-year torn between civil-service exams, grad school, and a job — run a career-path assessment. Requirements: 1) load the question bank and render all 10 questions as one answerable interactive card with full stems, hints and six options; 2) my answers (items 1-10): A A A E E A E A A A; 3) score with the assessment script's real rules; 4) the report must include the winning direction, score, conclusion and its full four-part text using the bank's original wording; 5) the card stays editable and resubmittable after the report appears.", "skills": [ { "id": "sycp-career-assessment", "name": "生涯测评", "description": "10 题 6 选一分析公考/央国企、自由职业、打工人、留学、保研/考研、躺平六大方向倾向", "skill_id": "skill_2089338317577322496" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 200, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-career-postgrad", "title": "生涯测评：考研方向，集中度 80 分", "title_en": "Career Assessment: Postgraduate Track, 80% Concentration", "subtitle": "10 题里 8 题指向考研，脚本据此给出四年规划与备考路径", "subtitle_en": "Eight of ten answers point to postgraduate study, with a four-year plan to match", "description": "大一大二最常见的困惑不是「怎么努力」，而是「往哪个方向努力」——考研、公考、留学、就业，每条路的准备内容完全不同。\n\n本案例用 10 题简版生涯测评跑出一份考研主导的结果：8 题落在保研/考研，方向集中度 80 分。案例展示六个方向的选择分布、脚本给出的大学四年规划与备考路径，并说明集中度分数代表方向一致性而非能力水平。\n\n适合：方向未定的大一大二学生、做生涯规划辅导的老师。", "description_en": "The common confusion in the first two years isn't how hard to work but which direction to work in — postgraduate study, civil service, overseas study and employment demand entirely different preparation.\n\nThis case runs the ten-question career assessment to a postgraduate-dominant result: 8 of 10 answers point that way, giving 80% concentration. It shows the distribution across six directions, the four-year plan and preparation path the script produces, and explains that the concentration score measures consistency, not ability.\n\nFor: undecided first- and second-year students and career counsellors.", "categories": [ "education" ], "tags": [ "生涯测评", "考研", "方向选择", "四年规划" ], "tags_en": [ "career assessment", "postgraduate study", "direction", "four-year plan" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "生涯测评考研方向结果报告", "label_en": "Career assessment report: postgraduate track" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次生涯方向测评。我的作答倾向是：学生会想去学习部、社团选专业相关的、活动偏学术竞赛类，整体更认同深造路线。要求：1) 用生涯测评的 10 题题库真实跑一遍评分脚本，给出结果方向和集中度分数；2) 展示六个生涯方向的选择分布（各选了几题、占比多少）；3) 完整给出脚本输出的大学四年规划和备考路径两部分内容，不要改写；4) 指出次选方向是什么，说明它意味着什么；5) 解释集中度分数代表什么、不代表什么。", "prompt_en": "Run a career-direction assessment for me. My answers lean academic: I'd join the study department in student government, pick a major-related club, prefer academic competitions, and generally favour further study. Requirements: 1) run the official ten-question scoring script and give the resulting direction and concentration score; 2) show the distribution across all six directions (answers and percentage each); 3) reproduce the four-year plan and preparation path from the script output verbatim; 4) identify the secondary direction and what it implies; 5) explain what the concentration score does and does not measure.", "skills": [ { "id": "sycp-career-assessment", "name": "生涯测评", "description": "10 题 6 选一分析公考/央国企、自由职业、打工人、留学、保研/考研、躺平六大方向倾向", "skill_id": "skill_2089338317577322496" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 201, "is_featured": false, "featured_order": 0, "quality_score": 87, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "road-trip-planner", "title": "川藏线自驾路线规划面板", "title_en": "Sichuan-Tibet Road Trip Planner", "subtitle": "318国道成都到拉萨，8天行程路线+海拔+住宿全规划", "subtitle_en": "Route 318 Chengdu to Lhasa — 8-day itinerary with elevation, stops, and lodging", "description": "川藏线是无数自驾爱好者的梦想路线，但2000多公里的高原路况复杂，海拔变化大，加油站稀疏，不做详细规划很容易翻车。\n\n这个案例生成了一个暗色科技感的川藏线自驾路线规划面板 HTML 页面，包含：SVG 简化路线图（13个途经站点+距离标注）、海拔变化面积图（标注关键垭口和最高点）、8天行程时间线（每日里程/驾驶时长/住宿/海拔）、6个核心数据卡片、高原注意事项面板。支持路线绘制动画和交互效果，零外部依赖。\n\n适合：计划走川藏线的自驾爱好者、越野旅行团领队、需要高原路线参考的旅行者、自驾游攻略博主", "description_en": "The Sichuan-Tibet Highway is a dream road trip for countless driving enthusiasts, but 2,000+ km of high-altitude roads with extreme elevation changes and sparse gas stations make careful planning essential.\n\nThis case generates a dark-themed road trip planning panel as a single HTML page, featuring: SVG simplified route map (13 waypoints with distance labels), elevation profile area chart (marking key passes and summit), 8-day itinerary timeline (daily mileage/driving hours/lodging/altitude), 6 key data cards, and a high-altitude precaution panel. Includes route-drawing animations with zero external dependencies.\n\nBest for: Road trip enthusiasts planning the Sichuan-Tibet route, off-road tour leaders, travelers needing high-altitude route references, road trip bloggers", "categories": [ "travel" ], "tags": [ "自驾游", "川藏线", "路线规划", "高原旅行" ], "tags_en": [ "Road Trip", "Sichuan-Tibet Route", "Route Planning", "Plateau Travel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "川藏线自驾路线规划面板" } ], "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "请生成一个川藏线自驾路线规划面板页面（单文件 HTML，内联无外部依赖），暗色科技感、带路线绘制动画。\n\n路线：318 国道成都→拉萨，全程约 2146 公里，8 天行程。\n\n包含：简化路线图（成都→雅安→康定→新都桥→理塘→巴塘→芒康→左贡→八宿→然乌→波密→林芝→拉萨，标注站间距离）；海拔变化图（标注折多山 4298m、东达山 5130m、色季拉山 4728m 等垭口）；8 天行程时间线（每天起止地/里程/驾驶时长/住宿/海拔）；6 个数据卡片（总里程/最高海拔/驾驶总时长/加油站/住宿点/预估油费）；高原注意事项。", "prompt_en": "Please generate a Sichuan-Tibet road-trip route planning dashboard page (single-file HTML, inline with no external dependencies), with a dark tech aesthetic and route-drawing animations.\n\nRoute: National Highway 318, Chengdu → Lhasa, about 2,146 km in total, an 8-day trip.\n\nInclude: a simplified route map (Chengdu → Ya'an → Kangding → Xinduqiao → Litang → Batang → Mangkang → Zuogong → Baxoi → Rawu → Bomi → Nyingchi → Lhasa, with the distance between stops labeled); an elevation-change chart (marking passes such as Zheduo Mountain 4298m, Dongda Mountain 5130m, Sejila Mountain 4728m); an 8-day itinerary timeline (each day's start/end points / mileage / driving time / lodging / elevation); 6 data cards (total mileage / highest elevation / total driving time / gas stations / lodging points / estimated fuel cost); and high-altitude precautions.", "source": "official", "sort_order": 201, "featured_order": 0, "quality_score": 93, "usage_count": 0, "is_featured": false, "created_at": "2026-05-03T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-career-overseas", "title": "生涯测评：留学方向，还有 2 题在考研上摇摆", "title_en": "Career Assessment: Overseas Study, With Two Answers Still Wavering", "subtitle": "主方向明确、次方向仍有牵引——大多数人的真实状态", "subtitle_en": "A clear primary with a lingering secondary — where most people actually are", "description": "测评给出一个干脆的标签容易，可现实里大多数人都是「大方向定了、但还没完全放下另一条路」。\n\n本案例跑出一份留学主导但集中度只有 70 分的结果——7 题指向留学，仍有 2 题落在保研/考研。案例展示完整的方向分布、脚本给出的留学四年规划与申请路径，并解释留学与考研为什么是最常见的一组纠结、前期准备如何共用。\n\n适合：在留学与考研之间摇摆的学生、需要做两手准备的人。", "description_en": "It's easy for an assessment to hand you a clean label, but most people are somewhere messier: direction chosen, alternative not quite released.\n\nThis case produces an overseas-study result with only 70% concentration — 7 answers point abroad, 2 still to postgraduate study at home. It shows the full distribution, the four-year plan and application path from the script, and explains why these two paths are the most common dilemma and how early preparation overlaps.\n\nFor: students torn between studying abroad and at home, and those hedging.", "categories": [ "education" ], "tags": [ "生涯测评", "留学", "考研", "方向摇摆" ], "tags_en": [ "career assessment", "overseas study", "postgraduate", "undecided" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "生涯测评留学方向结果报告", "label_en": "Career assessment report: overseas study track" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次生涯方向测评。我的作答倾向是：学生会想去外联部、社团选国际交流类、活动偏跨文化交流，整体更向往出国看看，但也没完全放弃国内读研。要求：1) 用生涯测评的 10 题题库真实跑一遍评分脚本，给出结果方向和集中度分数；2) 展示六个方向的选择分布；3) 完整给出脚本输出的大学四年规划和申请路径，不要改写；4) 指出次选方向及其占比，分析主次方向之间的关系；5) 说明集中度 70 分与 80 分以上的结果在解读上有什么区别。", "prompt_en": "Run a career-direction assessment for me. My answers lean outward: external liaison in student government, international exchange clubs, cross-cultural activities — I'd like to study abroad but haven't ruled out a domestic master's. Requirements: 1) run the official ten-question scoring script and give the direction and concentration score; 2) show the distribution across all six directions; 3) reproduce the four-year plan and application path from the script verbatim; 4) identify the secondary direction and its share, and analyse how the two relate; 5) explain how a 70% result should be read differently from one above 80%.", "skills": [ { "id": "sycp-career-assessment", "name": "生涯测评", "description": "10 题 6 选一分析公考/央国企、自由职业、打工人、留学、保研/考研、躺平六大方向倾向", "skill_id": "skill_2089338317577322496" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 202, "is_featured": false, "featured_order": 0, "quality_score": 88, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "travel-expense-tracker", "title": "旅行费用追踪仪表盘", "title_en": "Travel Expense Tracker Dashboard", "subtitle": "日本7天游花了多少钱？AI 帮你按天、按类别一目了然", "subtitle_en": "How much did your 7-day Japan trip cost? AI breaks it down by day and category", "description": "出国旅行最怕回来才发现超了预算，旅途中逐笔记账又太麻烦，各种货币换算更是头大。\n\n这个案例生成了一个暗色科技感的旅行费用追踪仪表盘 HTML 页面，包含：SVG 环形饼图按类别拆分（交通/住宿/餐饮/门票/购物）、7天逐日花费柱状图、4个核心数据卡片（日均/最高/剩余/汇率）、实时汇率换算面板。支持 CSS 动画交互，零外部依赖。\n\n适合：出境旅行预算管理者、精打细算的背包客、需要费用报告的差旅人员、旅行博主做内容复盘", "description_en": "The worst part of traveling abroad is realizing you've blown your budget only after you're home. Tracking expenses on the go is tedious, and currency conversions add to the headache.\n\nThis case generates a dark-themed travel expense tracker dashboard as a single HTML page, featuring: SVG donut chart for category breakdown (transport/accommodation/food/tickets/shopping), 7-day daily spending bar chart, 4 key data cards (daily avg/highest/remaining/exchange rate), and a live currency converter panel. Includes CSS animations with zero external dependencies.\n\nBest for: Budget-conscious international travelers, backpackers, business travelers needing expense reports, travel bloggers doing trip reviews", "categories": [ "travel" ], "tags": [ "旅行费用", "预算追踪", "数据可视化", "日本旅行" ], "tags_en": [ "Travel Expenses", "Budget Tracking", "Data Visualization", "Japan Travel" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "旅行费用追踪仪表盘" } ], "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "请生成一个旅行费用追踪仪表盘页面（单文件 HTML，内联无外部依赖），暗色科技感、带图表动画。\n\n旅行：东京-大阪-京都 7 日游，总预算 15000 元，实际花费 12680 元。\n\n包含：费用类别环形图（交通35% 住宿30% 餐饮20% 门票活动10% 购物5%）；7 天逐日花费柱状图（每天标注主要支出）；4 个数据卡片（日均花费/最高单日/剩余预算/当日汇率）；CNY↔JPY 汇率换算面板。", "prompt_en": "Please generate a travel expense tracking dashboard page (a single-file HTML page, inline with no external dependencies), with a dark tech aesthetic and chart animations.\n\nTrip: a 7-day tour of Tokyo-Osaka-Kyoto, with a total budget of ¥15,000 and actual spending of ¥12,680.\n\nInclude: a category donut chart (transportation 35%, accommodation 30%, dining 20%, attractions & activities 10%, shopping 5%); a day-by-day spending bar chart for the 7 days (with the main expense labeled for each day); 4 data cards (average daily spending / highest single day / remaining budget / today's exchange rate); a CNY↔JPY exchange rate conversion panel.", "source": "official", "sort_order": 202, "featured_order": 0, "quality_score": 90, "usage_count": 0, "is_featured": false, "created_at": "2026-05-03T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-china-nev-overseas-2026", "title": "中国新能源汽车出海 2026 专题研究", "title_en": "China NEV Global Expansion 2026 Thematic Research", "subtitle": "3 大目标市场 × 5 家代表车企 × 关税税率矩阵，含 16 国渗透率 SVG 图与出海时间线", "subtitle_en": "3 target markets × 5 Chinese OEMs × tariff matrix, with SVG penetration chart of 16 countries and go-global timeline", "description": "2024 年欧盟反补贴税落地、美国 100% 关税溢出、印尼巴西同步加码，中国车企出海从「甩货模式」被迫切换到「深耕模式」。到底哪个市场还能进？关税怎么算？比亚迪/奇瑞/上汽/吉利/蔚来各自的出海打法差在哪儿？——没有一份把政策、基建、竞品、税率、车企策略全串起来的专题报告，高管开董事会根本没法拍板。\n\n这个案例演示了深度研究团队 8 分钟内交付的《中国新能源汽车出海 2026 专题研究》：Markdown 主报告约 1.3 万字，覆盖全球格局 → 三大目标市场（欧洲 / 东南亚 / 拉美）现状 → 5 家代表车企出海策略（BYD / Chery / MG / Geely / NIO）→ 4+2 大挑战（关税 / 本地化 / 品牌 / 售后 / 金融 / 地缘）→ 2026-2030 趋势预测 → 分市场 × 分车企战略建议，含 10+ 张对比表与 55 条权威引用链接。配套 HTML 可视化版含三类纯 CSS/SVG 可视化：16 国 NEV 渗透率横向条形图（三色按大洲区分）、五档彩色关税矩阵（出口 vs 本地化两情形）、5 家车企 × 4 年出海关键里程碑时间线。\n\n适合：车企战投 / 海外事业部 / 券商研究所汽车组 / 行业咨询 / 做 ESG 与 CVC 投资组合的海外基金 / 媒体主理人做年度专题。", "description_en": "The 2024 EU countervailing duties, the US 100% tariff spillover, and fresh Indonesian/Brazilian restrictions have forced Chinese NEV OEMs to pivot from clearance mode to deep-local mode. Which markets remain open? How do the tariffs actually stack? How differently are BYD / Chery / MG / Geely / NIO playing it? Without a single report that stitches policy, infrastructure, competition, tariff math, and company-level playbooks, no board meeting can land a decision.\n\nThis case shows the Deep Research Team delivering the full China NEV Global Expansion 2026 report in 8 minutes: ~13k-character Markdown main report covering global landscape, current state of 3 target markets (Europe / SEA / LatAm), 5 OEM go-global strategies (BYD / Chery / MG / Geely / NIO), 4+2 core challenges (tariffs / localization / branding / after-sales / finance / geopolitics), 2026-2030 forecasts, and a market-by-OEM strategy matrix. 10+ comparison tables and 55 hyperlinked sources. Companion HTML with three pure-CSS/SVG visualizations: a 16-country NEV penetration bar chart (color-coded by continent), a 5-tier tariff matrix (export vs localized), and a 5-OEM x 4-year go-global timeline.\n\nFit for: OEM corporate strategy, overseas BUs, sell-side auto analysts, consulting firms, overseas funds running ESG/CVC portfolios, and industry media editors running year-end themes.", "categories": [ "business", "data", "search" ], "tags": [ "市场趋势", "新能源汽车", "出海研究", "关税政策", "车企战略" ], "tags_en": [ "Market Trends", "NEV", "Global Expansion", "Tariff Policy", "OEM Strategy" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.md", "artifacts": [ { "file": "report.md", "label": "Markdown 主报告（1.3 万字 · 55 条引用）" }, { "file": "visual-report.html", "label": "HTML 可视化版（16 国渗透率+关税矩阵+出海时间线）" } ], "prompt": "请用【深度研究团队 5 阶段完整模式】产出一份《中国新能源汽车出海 2026 专题研究》，聚焦欧洲、东南亚、拉美三大目标市场。\n\n报告结构：\n1. 全球市场格局与中国车企份额（2023–2026）\n2. 三大目标市场现状（政策 / 基建 / 消费力 / 竞品）\n3. 代表性中国车企出海策略（比亚迪 / 奇瑞 / 上汽 MG / 吉利 / 蔚来）\n4. 核心挑战（关税壁垒 / 本地化 / 品牌 / 售后）\n5. 趋势预测（2026–2030）\n6. 战略建议（分市场 × 分车企）\n\n要求：至少 3 张核心对比表（三大市场 / 5 家车企 / 关税税率矩阵）；至少 5 个关键图表文字描述；每数据点带 Markdown 超链接到原始来源，未公开数据用 ⚠️ 标注需核对。\n\n额外交付 HTML 可视化版（单页内联样式，无 CDN）：各国 NEV 渗透率 SVG 条形图、关税对比表、5 家车企出海时间线。\n\n字数：Markdown 1.2–1.8 万字。", "prompt_en": "Please use the [Deep Research Team 5-Phase Complete Mode] to produce a \"China NEV Global Expansion 2026 Thematic Research\" report, focusing on the three target markets of Europe, Southeast Asia, and Latin America.\n\nReport structure:\n1. Global market landscape and Chinese OEMs' share (2023-2026)\n2. Current state of the three target markets (policy / infrastructure / purchasing power / competitors)\n3. Go-global strategies of representative Chinese OEMs (BYD / Chery / SAIC MG / Geely / NIO)\n4. Core challenges (tariff barriers / localization / branding / after-sales)\n5. Trend forecast (2026-2030)\n6. Strategic recommendations (by market × by OEM)\n\nRequirements: at least 3 core comparison tables (three markets / 5 OEMs / tariff rate matrix); at least 5 textual descriptions of key charts; every data point with a Markdown hyperlink to its original source, flagging undisclosed data with ⚠️ as needing verification.\n\nAdditionally deliver an HTML visualization version (single page, inline styles, no CDN): an SVG bar chart of NEV penetration by country, a tariff comparison table, and a 5-OEM go-global timeline.\n\nWord count: Markdown 12,000-18,000 characters.", "skills": [], "experts": [ { "id": "GPTResearcherTeam", "name": "深度研究团队", "description": "深度研究报告输出，7角色5阶段聚合多源信息，经审稿修订循环输出带引用的专业报告", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "a8da6f2dc1f7", "name_en": "Deep Research Team", "description_en": "Deep research report generation: 7 roles in 5 phases aggregate multi-source data through review-revise loop into cited professional reports", "expert_id": "ex_xdj3MuRUqbR0" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 203, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T03:25:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "travel-weather-compare", "title": "旅游城市天气对比看板", "title_en": "Travel City Weather Comparison Board", "subtitle": "5大热门城市未来7天天气PK，一眼选出最佳目的地", "subtitle_en": "7-day weather showdown of 5 popular cities — pick the best destination at a glance", "description": "想出去玩但纠结去哪儿？看了三亚怕下雨、查了成都怕太热、丽江又担心紫外线——一个个城市查天气太低效。\n\n这个案例生成了一个暗色科技感的多城市天气对比看板 HTML 页面，包含：5大热门旅游城市（三亚/成都/丽江/厦门/青岛）天气卡片、SVG 温度趋势折线图（7天5条曲线）、降水概率对比条形图、穿衣建议标签、综合推荐面板。支持 CSS 动画交互，零外部依赖。\n\n适合：纠结目的地的旅行计划者、关注天气的户外爱好者、旅行社做行程推荐、需要穿衣指南的出游家庭", "description_en": "Want to travel but can't decide where? Worried about rain in Sanya, heat in Chengdu, UV in Lijiang — checking each city's weather one by one is painfully slow.\n\nThis case generates a dark-themed multi-city weather comparison board as a single HTML page, featuring: weather cards for 5 popular travel cities (Sanya/Chengdu/Lijiang/Xiamen/Qingdao), SVG temperature trend charts (7 days × 5 curves), precipitation probability bar chart, clothing advice tags, and a smart recommendation panel. Includes CSS animations with zero external dependencies.\n\nBest for: Indecisive trip planners, weather-conscious outdoor enthusiasts, travel agencies making itinerary suggestions, families needing packing guides", "categories": [ "travel" ], "tags": [ "天气对比", "旅游目的地", "数据可视化", "出行推荐" ], "tags_en": [ "Weather Comparison", "Travel Destination", "Data Visualization", "Travel Recommendation" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "旅游城市天气对比看板" } ], "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "请生成一个旅游城市天气对比看板页面（单文件 HTML，内联无外部依赖），暗色科技感、带数据动画。\n\n对比城市：三亚、成都、丽江、厦门、青岛，时间 2026-06-15 至 06-21。\n\n包含：5 个城市天气卡片（当前温度/体感/湿度/风力/紫外线/穿衣建议）；7 天温度趋势折线图（5 条曲线）；降水概率对比条形图；综合推荐面板（最适合出行的城市）。", "prompt_en": "Please generate a travel-city weather comparison dashboard page (single-file HTML, inlined with no external dependencies), in a dark tech-style with data animations.\n\nCities to compare: Sanya, Chengdu, Lijiang, Xiamen, Qingdao; dates 2026-06-15 to 06-21.\n\nInclude: 5 city weather cards (current temperature/feels-like/humidity/wind/UV index/clothing suggestions); a 7-day temperature trend line chart (5 curves); a precipitation-probability comparison bar chart; and a comprehensive recommendation panel (the best city for travel).", "source": "official", "sort_order": 204, "featured_order": 0, "quality_score": 91, "usage_count": 0, "is_featured": false, "created_at": "2026-05-03T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-onboarding-workflow", "title": "新人入职引导", "title_en": "Employee Onboarding Workflow Agent", "subtitle": "AI Agent 新员工入职全流程自动化引导系统", "subtitle_en": "AI Agent New Employee Onboarding Full-Process Automation System", "description": "每月入职 10-20 人，HR 手忙脚乱：offer 确认没人跟、设备账号开通漏东漏西、新人第一天找不到工位也不知道该干嘛。\n\n这个案例展示了一套完整的 AI Agent 新员工入职引导系统：从入职前自动化准备清单，到首日智能引导流程，再到首周任务看板、30-60-90 天成长路线图、AI 智能问答助手，以及入职体验满意度追踪。\n\n适合：HR 负责人、HRBP、人力资源数字化团队、企业文化与员工体验负责人。", "description_en": "Onboarding 10-20 new hires monthly with manual processes means chaos. This case presents a comprehensive AI Agent employee onboarding system: pre-boarding checklists, first-day guidance, first-week task boards, 30-60-90 day roadmaps, AI FAQ assistant, and satisfaction tracking.\n\nIdeal for: HR leaders, HRBPs, HR digital transformation teams, employee experience owners.", "categories": [ "agent", "office" ], "tags": [ "入职引导", "AI Agent", "HR自动化", "员工体验", "流程编排" ], "tags_en": [ "onboarding", "AI agent", "HR automation", "employee experience", "workflow orchestration" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "新人入职引导系统设计文档" } ], "prompt": "帮我设计一个 AI Agent 新员工入职引导系统。公司每月入职 10-20 人，需要标准化入职流程。需要包含：1）入职前准备清单；2）首日引导流程；3）首周任务看板；4）30-60-90 天成长路线图与里程碑；5）智能问答助手；6）导师匹配与 1-on-1 安排；7）入职体验满意度追踪。输出为详细的 Markdown 系统设计文档。", "prompt_en": "Help me design an AI Agent-powered new-employee onboarding system. The company onboards 10-20 people per month and needs a standardized onboarding process. It should include: 1) a pre-onboarding preparation checklist; 2) a first-day orientation flow; 3) a first-week task board; 4) a 30-60-90 day growth roadmap with milestones; 5) an intelligent Q&A assistant; 6) mentor matching and 1-on-1 scheduling; 7) onboarding-experience satisfaction tracking. Output a detailed Markdown system design document.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "多智能体团队编排：角色定义、任务流转、交接协议与质量门禁", "description_en": "Orchestrate multi-agent teams with roles, task lifecycles, handoffs, and review workflows", "skill_id": "skill_2053079579940413440" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 205, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "visa-requirement-query", "title": "签证要求速查面板", "title_en": "Visa Requirement Quick Lookup", "subtitle": "中国护照去哪儿免签？一个面板看清全球签证政策", "subtitle_en": "Where can a Chinese passport go visa-free? One panel for all global visa policies", "description": "每次出境前都要反复搜「XX国家对中国免签吗」「落地签怎么办」「电子签多久能出」，信息分散在各种攻略里，政策还经常变。\n\n这个案例生成了一个暗色科技感的签证要求速查面板 HTML 页面，包含：顶部签证类型统计（免签/落地签/电子签数量）、按洲和签证类型双重筛选功能、18+国家卡片网格（含签证类型标签/免签天数/注意事项）、热门免签目的地推荐区。支持 JS 交互筛选和 CSS 动画效果，零外部依赖。\n\n适合：经常出境的旅行者、规划年度旅行清单的人、旅行社签证顾问、需要快速查签证政策的出差人士", "description_en": "Every time before traveling abroad, you end up searching 'Is XX country visa-free for Chinese passports?', 'How to get a visa on arrival?', 'How long does an e-visa take?' — info is scattered across blogs and policies change frequently.\n\nThis case generates a dark-themed visa requirement lookup panel as a single HTML page, featuring: top-bar visa type statistics (visa-free/VOA/e-visa counts), dual filtering by continent and visa type, 18+ country cards with visa badges/duration/notes, and a hot visa-free destination recommendation section. Includes JS interactive filtering and CSS animations with zero external dependencies.\n\nBest for: Frequent international travelers, annual travel list planners, travel agency visa consultants, business travelers needing quick visa policy checks", "categories": [ "travel" ], "tags": [ "签证查询", "免签国家", "出境旅行", "签证政策" ], "tags_en": [ "Visa Lookup", "Visa-Free Countries", "Outbound Travel", "Visa Policy" ], "artifact_type": "html", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "签证要求速查面板" } ], "skills": [ { "id": "frontend-dev", "name": "前端开发", "description": "前端开发与 AI 媒体生成", "skill_id": "skill_2052764137292271616" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "prompt": "请生成一个中国护照签证要求速查面板页面（单文件 HTML，内联无外部依赖），暗色科技感、支持交互筛选。\n\n包含：顶部签证类型统计（免签/落地签/电子签/需签证 国家数）；按洲（亚/欧/美/非/大洋洲）和签证类型双重筛选；18+ 国家卡片（国旗/国名/签证类型标签/免签天数/注意事项）；热门免签 Top5 推荐（含最佳季节和推荐停留天数）。", "prompt_en": "Please generate a visa requirement quick-lookup panel page for Chinese passport holders (single-file HTML, inline with no external dependencies), with a dark tech aesthetic, supporting interactive filtering.\n\nInclude: top-bar visa type statistics (counts of countries that are visa-free / visa-on-arrival / e-visa / visa-required); dual filtering by continent (Asia / Europe / Americas / Africa / Oceania) and visa type; 18+ country cards (flag / country name / visa type tag / visa-free days / notes); and a Top 5 popular visa-free recommendations (including best season and recommended length of stay).", "source": "official", "sort_order": 206, "featured_order": 0, "quality_score": 90, "usage_count": 0, "is_featured": false, "created_at": "2026-05-03T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "oauth2-auth-implementation", "title": "OAuth2 授权认证完整实现", "title_en": "OAuth2 Authorization Complete Implementation", "subtitle": "Node.js + Express + TypeScript 实现 OAuth2 授权码模式", "subtitle_en": "OAuth2 Authorization Code Flow with Node.js + Express + TypeScript", "description": "需要为项目添加 OAuth2 认证但从零实现太复杂？授权码模式、PKCE 扩展、Token 轮换、安全存储……每一步都有坑。\n\n本案例演示了 fullstack-dev 全栈开发技能如何基于 Node.js + Express + TypeScript 实现完整的 OAuth2 授权码模式认证系统：支持 PKCE 扩展的授权端点、授权码兑换与 Refresh Token 轮换的令牌端点、基于 RS256 的 JWT Access Token 签发与验证中间件、加密存储和自动过期的安全 Refresh Token 管理、客户端注册与管理接口、以及 CORS/CSP 安全头配置。\n\n适合：全栈开发者、后端工程师、安全工程师。", "description_en": "Need OAuth2 authentication but implementing from scratch is complex? Authorization code flow, PKCE extension, token rotation, secure storage — every step has pitfalls.\n\nThis case demonstrates how fullstack-dev implements a complete OAuth2 Authorization Code Flow system with Node.js + Express + TypeScript: authorization endpoint with PKCE, token endpoint with code exchange and refresh token rotation, RS256 JWT middleware, secure refresh token management, client registration, and CORS/CSP security headers.\n\nPerfect for fullstack developers, backend engineers, and security engineers.", "tags": [ "OAuth2", "认证授权", "JWT", "安全", "Node.js" ], "tags_en": [ "OAuth2", "Authentication", "JWT", "Security", "Node.js" ], "artifact_type": "code", "categories": [ "dev" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "OAuth2 完整实现代码" } ], "prompt": "请实现完整的 OAuth2 授权码模式（Authorization Code Flow）认证系统，基于 Node.js + Express + TypeScript。要求：1) 授权端点（/authorize）支持 PKCE 扩展；2) Token 端点（/token）支持授权码兑换和 Refresh Token 轮换；3) JWT Access Token 签发与验证中间件；4) 安全的 Refresh Token 存储（加密 + 过期策略）；5) 客户端注册与管理接口；6) CORS 和 CSP 安全头配置；7) 完整的 TypeScript 类型定义。请用全栈开发技能输出包含完整实现代码的 Markdown 文档。", "prompt_en": "Please implement a complete OAuth2 Authorization Code Flow authentication system based on Node.js + Express + TypeScript. Requirements: 1) An authorization endpoint (/authorize) that supports the PKCE extension; 2) A token endpoint (/token) that supports authorization code exchange and Refresh Token rotation; 3) JWT Access Token issuance and a verification middleware; 4) Secure Refresh Token storage (encryption + expiration policy); 5) Client registration and management interfaces; 6) CORS and CSP security header configuration; 7) Complete TypeScript type definitions. Please use the full-stack development skill to output a Markdown document containing the complete implementation code.", "skills": [ { "id": "fullstack-dev", "name": "全栈开发", "description": "全栈应用架构与开发指南", "description_en": "Full-stack architecture and development guide", "skill_id": "skill_2053082011994116096" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 90, "sort_order": 207, "is_featured": false, "featured_order": 0, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "project-budget-tracker", "title": "项目预算执行跟踪表", "title_en": "Project Budget Tracker", "subtitle": "预算与实际一键对比，偏差分析自动生成", "subtitle_en": "Compare budget vs actuals instantly with automated variance analysis", "description": "项目做到一半才发现超支了——这是多少项目经理的噩梦。各部门的预算数据散落在邮件、报销单、审批流里，等财务月底对账时才发现差异，想要分析哪里超了、为什么超了，又得手工拉数据比对半天。\n\n这个案例演示了如何用 AI 一键生成一套完整的项目预算执行跟踪 Excel 工作簿：3 个项目同时管控，每个项目按人力/硬件/软件/外包/差旅/其他 6 大成本类别拆分，逐月预算 vs 实际对比，自动计算偏差率，异常项红色高亮。还有多项目汇总总览、分类偏差分析趋势、以及里程碑预算关口管理——打开就能在项目会上用。\n\n适合：项目经理、财务控制人员、PMO、团队负责人、任何需要跟踪项目花钱进度的人。", "description_en": "Discovering budget overruns halfway through a project is every PM's nightmare. Budget data is scattered across emails, expense reports, and approval workflows — by the time finance reconciles at month-end, the damage is done, and figuring out where and why takes hours of manual comparison.\n\nThis case shows how AI generates a complete project budget tracking Excel workbook in one click: 3 projects tracked simultaneously, each broken down by 6 cost categories (personnel, hardware, software, outsourcing, travel, miscellaneous), monthly budget vs actual comparison, auto-calculated variance percentages, and red-highlighted anomalies. Plus a multi-project summary dashboard, category-level variance trend analysis, and milestone budget gate management — ready to present at your next project review.\n\nBest for: Project managers, finance controllers, PMO teams, team leads, anyone who needs to track project spending progress.", "tags": [ "Excel", "预算管理", "项目管理", "偏差分析", "成本控制", "财务跟踪" ], "tags_en": [ "Excel", "Budget Management", "Project Management", "Variance Analysis", "Cost Control", "Financial Tracking" ], "artifact_type": "table", "categories": [ "data", "office", "finance" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "项目预算执行跟踪表" } ], "prompt": "请帮我创建一份项目预算执行跟踪 Excel（output.xlsx）。公司同时推进 3 个项目，需统一管控预算执行、及时发现超支。请生成一份合理的模拟数据（2026 Q1-Q2）。\n\n3 个项目：智慧园区平台（总预算 280 万，偏研发）、品牌升级计划（150 万，偏营销）、供应链优化系统（210 万，偏外包）。每个项目按人力/硬件/软件/外包/差旅/其他六类成本拆分。\n\n包含以下 Sheet：项目总览、各项目预算明细、偏差分析、里程碑。代表性字段如项目、成本类别、计划/实际/偏差、执行率、健康度等。\n\n格式：专业商务风格，金额带千分位，偏差列用条件格式（超支红/节约绿），冻结首行首列 + 自动筛选。", "prompt_en": "Create a project budget tracking Excel workbook (output.xlsx). The company is running 3 projects at once and needs unified budget control and early overspend detection. Generate a reasonable set of mock data (2026 Q1–Q2).\n\nThree projects: Smart Park Platform (¥2.8M budget, R&D-heavy), Brand Upgrade (¥1.5M, marketing-heavy), Supply-Chain Optimization (¥2.1M, outsourcing-heavy). Split each into six cost categories: labor, hardware, software, outsourcing, travel, other.\n\nInclude sheets: project overview, per-project budget detail, variance analysis, milestones. Representative fields such as project, cost category, planned/actual/variance, execution rate, health status.\n\nStyle: professional business look, thousands-separated amounts, conditional formatting on variance (overspend red / saving green), freeze top row & first column + auto-filter.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 208, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "daily-standup-report", "title": "工作日报/站会汇报生成", "title_en": "Daily Standup Report Generator", "subtitle": "描述今日工作内容，自动生成结构化站会汇报 Word 文档", "subtitle_en": "Describe your daily work, auto-generate a structured standup report as Word document", "description": "每天站会前花十几分钟组织语言写日报，是开发者最不想做的事之一。要把一天做的杂七杂八的事整理成条理清晰的汇报，还得区分「完成」「计划」「阻塞」，写着写着要么太啰嗦要么太简略。\n\n这个案例演示了一位前端开发工程师如何用 Word 文档生成技能，把当天的工作内容（代码重构、Bug 修复、会议参与等）和明日计划口语化描述一遍，AI 自动整理成标准站会模板 Word 文档：按「今日完成 / 明日计划 / 阻塞项」分类，每条控制在 1-2 行，技术细节保留但不冗余，格式规范可直接打印或发送。\n\n适合：研发工程师、技术 Leader、项目经理、所有需要写日报或站会汇报的开发者。", "description_en": "Spending ten-plus minutes organizing your thoughts into a daily report before standup is one of the least favorite tasks for developers. Turning a mix of coding, bug fixes, and meetings into a clear, categorized update — separating \"done\" from \"planned\" from \"blocked\" — often ends up either too verbose or too sparse.\n\nThis case shows how a frontend engineer uses the Word Document Generation skill to casually describe the day's work and tomorrow's plan, and AI automatically structures it into a standard standup template Word document: categorized by \"Completed / Planned / Blockers,\" each item kept to 1-2 lines, with professional formatting ready to print or send.\n\nBest for: Software engineers, tech leads, project managers, and anyone who needs to write daily standup reports.", "categories": [ "office" ], "tags": [ "日报", "站会", "开发者", "工作汇报", "办公效率" ], "tags_en": [ "Daily Report", "Standup", "Developer", "Work Report", "Office Productivity" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "前端工程师站会日报（Word）" } ], "prompt": "帮我生成一份开发工程师的每日站会汇报 Word 文档（output.docx）。\n\n请按「今日完成 / 明日计划 / 阻塞项」模板组织内容。\n\n我的工作内容（粘贴汇报人、项目、日期，以及今日完成、明日计划、阻塞项的口语化描述）：\n...\n\n文档要求：语气简洁专业，每条 1-2 行；含封面页、页眉页脚；输出为正式 Word 文档（output.docx）。", "prompt_en": "Help me generate a developer's daily standup report as a Word document (output.docx).\n\nPlease organize the content using the \"Completed Today / Plan for Tomorrow / Blockers\" template.\n\nMy work content (paste the reporter, project, date, and a casual description of what you completed today, your plan for tomorrow, and any blockers):\n...\n\nDocument requirements: concise, professional tone, 1-2 lines per item; include a cover page and headers/footers; output as a formal Word document (output.docx).", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 209, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-21T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "agent-ppt-research", "title": "Agent 多工具协作生成 PPT", "title_en": "Agent Multi-Tool Collaborative PPT Generation", "subtitle": "从数据调研到竞品分析，多 Agent 协作产出完整市场分析报告", "subtitle_en": "From Market Research to Competitor Analysis, Multi-Agent Collaborative Market Report", "description": "做行业分析 PPT 总是耗费大量时间？数据散落在各处，竞品信息不全面，趋势洞察缺乏深度？一份专业的市场分析报告往往需要数天的调研、整理和设计，反复修改后还是不够系统化。\n\n本案例展示了一套多 Agent 协作的市场分析 PPT 生成工作流：调研 Agent 收集市场规模、增速、主要玩家和用户画像等数据，分析 Agent 整理竞品对比矩阵（GitHub Copilot / Cursor / CodeBuddy / 通义灵码 / 豆包 MarsCode）从功能、定价、用户量、技术栈多维度对比，洞察 Agent 提炼 5 大行业趋势和 3 个投资机会，制作 Agent 将内容组织为 15 页完整 PPT。从原始数据到结构化呈现，全流程 Agent 自动完成。\n\n适合：产品经理、市场分析师、投资研究员、咨询顾问、需要快速产出行业报告的决策者。", "description_en": "Spending days on industry analysis presentations? Data scattered everywhere, competitor intel incomplete, trend insights lacking depth?\n\nThis case demonstrates a multi-Agent collaborative PPT generation workflow: Research Agent collects market size, growth rates, key players and user profiles, Analysis Agent builds competitor comparison matrices across features, pricing, user base, and tech stacks, Insight Agent distills 5 industry trends and 3 investment opportunities, Production Agent organizes everything into a complete 15-slide presentation.\n\nIdeal for: product managers, market analysts, investment researchers, consultants, and decision-makers needing rapid industry reports.", "categories": [ "agent", "office", "business" ], "tags": [ "市场分析", "PPT 生成", "竞品对比", "Agent 工作流" ], "tags_en": [ "market analysis", "PPT generation", "competitor comparison", "agent workflow" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "prompt": "请用 Agent 工作流为我准备一份「2026 年中国 AI 编程助手市场分析」PPT。流程：1) 调研 Agent 收集市场数据（市场规模/增速/主要玩家/用户画像）；2) 分析 Agent 整理竞品对比矩阵（GitHub Copilot / Cursor / CodeBuddy / 通义灵码 / 豆包 MarsCode），从功能、定价、用户量、技术栈维度对比；3) 洞察 Agent 提炼 5 大行业趋势和 3 个投资机会；4) 制作 Agent 将内容组织为 PPT 大纲和内容。输出包含 15 页内容的完整 PPT 文件。", "prompt_en": "Please use an Agent workflow to prepare a \"2026 China AI Coding Assistant Market Analysis\" PPT for me. Process: 1) Research Agent collects market data (market size / growth rate / key players / user profiles); 2) Analysis Agent builds a competitor comparison matrix (GitHub Copilot / Cursor / CodeBuddy / Tongyi Lingma / Doubao MarsCode), comparing across features, pricing, user base, and tech stack; 3) Insight Agent distills 5 major industry trends and 3 investment opportunities; 4) Production Agent organizes the content into a PPT outline and slide content. Output a complete PPT file with 15 pages of content.", "skills": [ { "id": "agent-team-orchestration", "name": "Agent Team Orchestration", "description": "Multi-agent team coordination", "slug": "agent-team-orchestration" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 91, "sort_order": 210, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.md", "label": "PPT 完整内容脚本" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "hmc-medical-aesthetics-review", "title": "医疗美容营销内容合规审查", "title_en": "Medical-Aesthetics Marketing Content Compliance Review", "subtitle": "小红书/抖音医美内容过审，专查容貌焦虑、术前术后对比、绝对化用语", "subtitle_en": "Review Xiaohongshu / Douyin medical-aesthetics content — flags appearance-anxiety, before-after comparisons, absolute claims", "description": "医美是医疗营销违规高发区——「绝对化用语」「制造容貌焦虑」「术前术后对比图」「直播荐项」随便一条都能被平台下架甚至处罚。\n\n本灵感把一份医美机构营销内容（小红书笔记/抖音短视频脚本/直播话术/详情页）交给医疗营销合规专家，按医疗广告管理办法、医美广告执法指南、平台规则逐条审查，标出容貌焦虑话术、违规对比图、绝对化/保证性疗效、将医疗美容包装成生活美容规避审查等高风险点，并给出平台合规改法与「需取得医疗广告审查证明」提示。\n\n适合: 医美机构市场/运营、医美MCN、内容创作团队、品牌合规。", "description_en": "Medical aesthetics is a high-incidence zone for healthcare-marketing violations — absolute claims, appearance-anxiety creation, before-after comparison images, or live-streaming procedure recommendations can each get content taken down or even penalized.\n\nThis inspiration hands a medical-aesthetics marketing piece (Xiaohongshu note / Douyin short-video script / livestream script / detail page) to the Healthcare Marketing Compliance Expert. It reviews item by item against the Medical Advertisement Management Measures, the Medical-Aesthetics Advertising Enforcement Guidelines, and platform rules, flagging appearance-anxiety phrasing, non-compliant comparison images, absolute / guaranteed efficacy, and packaging medical aesthetics as lifestyle beauty to evade review, then provides platform-compliant fixes and a requires Medical Advertisement Review Certificate note.\n\nSuitable for: medical-aesthetics institution marketing / operations, aesthetics MCNs, content teams, brand compliance.", "categories": [ "business" ], "tags": [ "医美合规", "平台规则", "医疗广告" ], "tags_en": [ "Aesthetics Compliance", "Platform Rules", "Medical Advertising" ], "artifact_type": "html", "preview": "output.html", "prompt": "我们是家医美机构，运营刚交了一批小红书和抖音的文案，明天就要发。上个月同行因为发术前术后对比图被罚了，我现在看什么都心慌。帮我把这几条文案审一遍，哪句话会被罚、罚多少、换成什么说法能既合规又还能卖。\n\n待发文案：\n\n---\n\n## 臻颜医美 · 双十一投放物料\n\n### 小红书笔记标题（5 选 1）\n1. 30 岁还素颜出门？难怪你被叫姐姐\n2. 同龄人已经悄悄做了这个项目，你还在敷面膜\n3. 法令纹让你老 10 岁！一次搞定，效果终身保持\n4. 姐妹们！我终于告别丑脸了\n5. 男朋友说我像他妈？做完这个他跪着道歉\n\n### 正文（种草款）\n姐妹们，我上周做了热玛吉，**效果我先说结论：绝对值！**\n\n来看我的对比图 👇\n【左：术前松垮下垂脸｜右：术后紧致小 V 脸】\n（附术前术后正面、侧面对比照各 2 张）\n\n我们院长 **王主任是做了 20 年的专家**，他跟我说这个项目**零风险、无恢复期、100% 有效**，我做完确实一点没耽误上班。\n\n**保证一次见效，无效全额退款！** 我闺蜜们看了我的效果都排队来做了。\n\n### 促销区\n🔥 双十一秒杀：热玛吉原价 19800，**限时 6800**！\n🔥 前 50 名再送水光针 1 次\n🔥 **今晚 12 点前下单再减 500，仅限今天！**\n🔥 支持医美分期，0 首付、0 利息\n\n### 抖音口播脚本（30 秒）\n\"你是不是也觉得自己脸垮了？（停顿）别人 30 岁像 20，你 30 岁像 40，差距就在这里。（举手机展示对比图）这是我们客户小美做完的效果，**术后完全看不出动过**。现在下单立减 13000，**这个价格全网找不到第二家**。地址就在评论区，来了就找王主任，他做过 5000 例，**从没出过事故**。\"\n\n### 门店挂牌信息\n臻颜生活美容会所 · 主营项目：热玛吉、水光针、瘦脸针、埋线提升\n\n---\n\n我知道对比图可能有问题，但拿掉之后完全没说服力，有别的合法办法吗？另外那句\"无效全额退款\"和\"零风险\"是运营硬要加的，你帮我评估下风险有多大，我拿你的结论去跟他们吵。最后给我一版能直接发的改写文案。", "prompt_en": "We're a medical aesthetics clinic. Our marketing team just handed over a batch of Xiaohongshu and Douyin copy set to publish tomorrow. Last month a peer clinic got fined for posting before/after photos, and now everything makes me nervous. Please review this copy: which lines will draw penalties, how large, and what wording would keep us compliant while still selling.\n\nCopy queued for publication:\n\n---\n\n## Zhenyan Aesthetics · Double 11 Campaign Assets\n\n### Xiaohongshu Note Headlines (pick 1 of 5)\n1. Still going out bare-faced at 30? No wonder they call you \"ma'am\"\n2. Your peers already had this treatment quietly — you're still doing sheet masks\n3. Nasolabial folds age you 10 years! One session, results last a lifetime\n4. Girls! I finally said goodbye to my ugly face\n5. My boyfriend said I looked like his mom — after this he apologized on his knees\n\n### Body Copy (seeding version)\nGirls, I had Thermage last week. **Verdict first: absolutely worth it!**\n\nLook at my before/after 👇\n【Left: saggy drooping pre-op face｜Right: tight V-line post-op】\n(2 frontal and 2 profile before/after photos attached)\n\nOur clinic director **Dr. Wang is a 20-year expert**. He told me this treatment is **zero risk, no downtime, 100% effective** — and indeed I didn't miss a single workday.\n\n**Guaranteed results in one session, full refund if ineffective!** All my friends saw my results and lined up to book.\n\n### Promotion Block\n🔥 Double 11 flash sale: Thermage was ¥19,800, **now ¥6,800** for a limited time!\n🔥 First 50 bookings get a free hydro-injection session\n🔥 **Order before midnight tonight for another ¥500 off — today only!**\n🔥 Aesthetic financing available: zero down payment, zero interest\n\n### Douyin Voiceover Script (30 seconds)\n\"Do you feel your face has fallen too? (pause) Others look 20 at 30 — you look 40 at 30. That's the gap. (holds up phone showing before/after) This is our client Xiaomei's result — **you absolutely can't tell she had anything done**. Order now and save ¥13,000 — **you won't find this price anywhere else**. Address is in the comments; ask for Dr. Wang, he's done 5,000 cases and **never had a single incident**.\"\n\n### Storefront Signage\nZhenyan Life Beauty Salon · Services: Thermage, hydro-injection, botulinum toxin for face slimming, thread lifting\n\n---\n\nI know the before/after photos are probably a problem, but without them the copy has no persuasive power — is there a lawful alternative? Also, \"full refund if ineffective\" and \"zero risk\" were forced in by the marketing team; assess how serious that exposure is so I can take your conclusion into the argument with them. Finish with a rewritten version I can publish as-is.", "skills": [], "experts": [ { "id": "nges-healthcare-compliance-lite", "name": "腾讯健康NGES合规专家", "description": "医疗营销合规专家，审核药品广告、医美种草、保健品声称，识别合规风险并支持改写为合规版本。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ytn7WGJKoya8", "promptFileSnapshot": "f4a36dc03d8d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 210, "is_featured": false, "featured_order": 0, "quality_score": 95, "source": "official", "usage_count": 0, "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "医美营销内容合规审核报告" }, { "file": "output.md", "label": "医美营销内容合规审核报告（Markdown 版）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "theme_label": "蓝绿 · 医美营销内容审查", "theme_color": "#0e8f8f", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-personality-test", "title": "性格测评：MBTI 测出 ISTJ", "title_en": "Personality: MBTI Yields ISTJ", "subtitle": "44 题四维计票判型，附岗位匹配方向", "subtitle_en": "44 items, four-dimension tallying, with role-fit directions", "description": "网上的 MBTI 测完只给四个字母和一段星座文，看不出和求职有什么关系。\n\n本案例跑完 44 题职业版 MBTI：四个维度对逐一计票（I/S/T/J 各 11 票），判型 ISTJ「物流师」，报告给出每个维度的判定依据、优势与短板原文，以及审计师、会计、财务经理等适配职业方向——全部由评分脚本按固定规则算出。\n\n适合：想验证职业方向的学生、纠结岗位性质（对人还是对事）的求职者。", "description_en": "Online MBTI tests hand back four letters and a horoscope paragraph, with no visible link to job hunting.\n\nThis case runs the 44-item career MBTI: each dimension pair tallied vote by vote (I/S/T/J at 11 each), typing ISTJ \"Logistician\", with per-dimension reasoning, verbatim strengths and cautions, and matched directions such as auditor, accountant and finance manager — all computed by the scoring script under fixed rules.\n\nFor: students validating a career direction and job-seekers weighing people-facing versus process-facing roles.", "categories": [ "education" ], "tags": [ "MBTI", "职业性格", "ISTJ", "测评" ], "tags_en": [ "MBTI", "career personality", "ISTJ", "assessment" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "MBTI ISTJ 测评报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做 44 题版 MBTI 职业性格测评并出报告。要求：1) 读题库渲染全部 44 题收集作答；2) 用评分脚本按真实规则计票判型，不许凭印象；3) 报告含四个维度对的计票明细与判定、人格类型档案（优势/短板用题库原文）、适配职业方向；4) 说明这份结果能用来做什么、不能替我决定什么。", "prompt_en": "Run the 44-item career MBTI and produce a report. Requirements: 1) load the bank and render all 44 items to collect answers; 2) tally and type with the scoring script's real rules, never by impression; 3) include per-pair tallies and verdicts, the type profile (strengths and cautions in the bank's original wording), and matched career directions; 4) state what this result can and cannot decide for me.", "skills": [ { "id": "career-personality", "name": "职业性格测评", "description": "44 题 MBTI 职业性格测评，四维计票输出 16 型报告", "skill_id": "skill_2089643348855885824" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 210, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "video-shot-remake", "title": "视频拉片复刻", "title_en": "Video Shot Remake", "subtitle": "拆解参考片的镜头语言，重做一支原创翡翠高级珠宝广告", "subtitle_en": "Reference-film grammar rebuilt as an original emerald high-jewelry campaign", "description": "想复用优秀广告的剪辑和运镜时，直接照搬画面容易侵权，也无法适配自己的产品主题。\n\n案例分析一支高级珠宝广告的景别、人物调度和金色时刻节奏，再以两位黑裙女性、意式庄园和翡翠戒指/项链/耳饰重建原创内容。\n\n适合：珠宝、时装、汽车、酒店等需要从参考片提取电影语法并更换主题的品牌内容。", "description_en": "Reusing the pacing and camera language of a strong commercial by copying imagery creates infringement risk and fails to fit a new product.\n\nThis case analyzes shot scale, blocking, camera motion, and golden-hour rhythm, then rebuilds them around two women in black dresses, an Italian-style estate, and emerald jewelry.\n\nBest for jewelry, fashion, automotive, and hospitality brands that need to extract film grammar from a reference while changing the subject.", "categories": [ "creation" ], "tags": [ "拉片复刻", "高级珠宝", "庄园广告" ], "tags_en": [ "shot remake", "high jewelry", "estate campaign" ], "artifact_type": "video", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.mp4", "artifacts": [ { "file": "output.mp4", "label": "成片视频" } ], "prompt": "请由「AI 视频拉片拆解与复刻」专家调用 Flova 的「视频拉片复刻」Skill。我会上传一支参考广告视频，以及翡翠戒指、项链和耳饰的产品图；请先逐镜分析参考视频的脚本、景别、人物调度、运镜、剪辑节奏、光线和声音结构，再围绕《翡翠庄园珠宝大片》生成一支约 45 秒、16:9 的原创广告。新主题限定为两位穿黑色礼服的成年女性在意式庄园与古城街巷相遇：马车抵达、露台对视、戒指微距、项链佩戴、金色时刻并肩行走，最后以三件产品英雄定帧收尾。先确认分析、全局规格和分镜，再确认人物/场景/产品元素与关键帧，批准后逐镜生成、配环境声与克制管弦乐并组装。保留抽象镜头语法与节奏，不复制参考片的品牌、Logo、文字、人物或具体画面。", "prompt_en": "Ask the “AI Video Breakdown and Remake” expert to invoke Flova’s “Video Shot Breakdown and Remake” Skill. I will upload a reference commercial plus product images of an emerald ring, necklace, and earrings. First analyze the reference shot by shot for script, shot size, blocking, camera movement, editing rhythm, lighting, and sound structure. Then create an original 16:9 commercial of about 45 seconds titled Emerald Estate Jewelry Film. The new theme features two adult women in black evening dresses meeting across an Italian-style estate and old-city streets: a carriage arrival, exchanged glances on a terrace, a macro ring shot, placing the necklace, walking together at golden hour, and a final hero frame of all three products. Confirm the analysis, global specification, and storyboard first, then obtain approval for the characters, locations, product elements, and keyframes. Once approved, generate each shot, add ambience and restrained orchestral music, and assemble the film. Preserve only abstract cinematic grammar and pacing; do not copy the reference brand, logo, text, people, or specific imagery.", "skills": [], "experts": [ { "id": "flova-shot-remake-master", "name": "AI 视频拉片拆解与复刻", "description": "支持视频、图片、剧本、短片、短剧、漫剧、电影、广告、商品 TVC 与设计等 AI 内容生成，覆盖文生图、文生视频、图生视频及视频拉片复刻。", "expertType": "agent", "source": "expert-marketplace", "promptFileSnapshot": "2465b48ad724", "expert_id": "ex_rxSUrDfuM0SF" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "avatar": "", "badge": "official" }, "sort_order": 210, "is_featured": false, "featured_order": 0, "quality_score": 0, "source": "official", "scenario": "", "scenario_use": "", "task_mode": "", "usage_count": 0, "created_at": "2026-08-28T16:00:00+08:00", "updated_at": "2026-08-28T16:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-mbti-enfp", "title": "MBTI：ENFP 竞选者，四个维度全部大比分胜出", "title_en": "MBTI: ENFP the Campaigner, Winning Every Dimension Decisively", "subtitle": "44 题给出的不只是四个字母，还有维度分布与职业清单", "subtitle_en": "Forty-four questions yield more than four letters", "description": "网上测 MBTI 只给你四个字母和一段星座式描述，看完并不知道这个结果稳不稳、能用来做什么。\n\n本案例用 44 题版真实跑分，输出 ENFP，类型清晰度 82 分。案例逐维度展示作答分布（四个维度全部 9:2 胜出）、题库原文的优势与短板描述，以及配套的 33 个职业方向，并说明维度差距为什么决定类型稳定性。\n\n适合：想认真理解自己性格倾向的学生、用 MBTI 辅助职业探索的人。", "description_en": "Online MBTI tests hand you four letters and a horoscope-style paragraph, leaving you unsure whether the result is stable or useful.\n\nThis case runs the 44-question version for real, returning ENFP with 82% type clarity. It shows the answer split on each dimension (all four won 9:2), the strengths and blind spots straight from the item bank, and the 33 matched career directions — explaining why the margin on each dimension determines how stable the type is.\n\nFor: students wanting a serious read on their disposition and anyone using MBTI for career exploration.", "categories": [ "education" ], "tags": [ "MBTI", "ENFP", "性格测评", "职业方向" ], "tags_en": [ "MBTI", "ENFP", "personality assessment", "career direction" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "MBTI ENFP 型测评结果报告", "label_en": "MBTI ENFP assessment report" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次 MBTI 职业性格测评。我作答时倾向于：更愿意向外表达和社交、关注可能性多过既有事实、决策时优先考虑他人感受、不喜欢被计划约束。要求：1) 用 44 题题库真实跑一遍评分脚本，给出人格类型和类型清晰度分数；2) 逐个维度对展示作答分布（各选了几题、百分比、判定结果与特征）；3) 完整给出题库中该类型的优势描述和需要留意的倾向，不要改写；4) 列出题库给出的匹配职业方向，并说明职业匹配是倾向提示而非能力判定；5) 解释维度差距大小与类型稳定性的关系。", "prompt_en": "Run an MBTI career-personality assessment. My answers lean outward and social, toward possibilities over established facts, weighing others' feelings in decisions, and resisting fixed plans. Requirements: 1) run the official 44-question scoring script and give the personality type and clarity score; 2) show the answer split for each dimension pair (counts, percentages, verdict and trait); 3) reproduce the strengths and blind spots for this type from the item bank verbatim; 4) list the matched career directions and note that these indicate inclination, not capability; 5) explain how the margin on each dimension relates to type stability.", "skills": [ { "id": "career-personality", "name": "职业性格测评", "description": "44 题 MBTI 测评，输出四维分布与 16 型人格档案", "skill_id": "skill_2089643348855885824" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 211, "is_featured": false, "featured_order": 0, "quality_score": 93, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "playwright-e2e-testing", "title": "电商下单流程 E2E 测试套件", "title_en": "E-Commerce Checkout E2E Test Suite", "subtitle": "Playwright 自动化测试的 Page Object 架构实战", "subtitle_en": "Playwright automation testing with Page Object Model in practice", "description": "每次上线前手动回归测试下单流程，费时费力还容易漏测？用全栈开发技能生成一套完整的电商核心流程 E2E 测试套件，采用 Page Object Model 架构，覆盖搜索→购物车→下单→支付全链路，还附带 CI 集成配置。\n\n本案例演示了 fullstack-dev 如何生成一套生产级的 Playwright E2E 测试方案：包含 POM 架构的 Page Object 封装、测试数据工厂、网络请求拦截 mock、截图对比、CI/CD 集成配置（GitHub Actions）、以及 HTML 测试报告生成。\n\n适合：QA 工程师、测试负责人、全栈开发者。", "description_en": "Manual regression testing before every release is slow and error-prone. Use the fullstack-dev skill to generate a complete E2E test suite for e-commerce checkout flow with Page Object Model architecture, covering search -> cart -> order -> payment, with CI integration config.\n\nThis case demonstrates how fullstack-dev generates a production-grade Playwright E2E test solution: POM page object encapsulation, test data factories, network request interception mocks, screenshot comparison, CI/CD integration (GitHub Actions), and HTML test report generation.\n\nPerfect for QA engineers, test leads, and fullstack devs.", "categories": [ "dev" ], "tags": [ "E2E测试", "Playwright", "自动化测试", "POM架构", "电商" ], "tags_en": [ "E2E Testing", "Playwright", "Test Automation", "POM Pattern", "E-Commerce" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "E2E测试套件方案" } ], "prompt": "帮我为一个电商网站的核心下单流程编写 Playwright E2E 测试套件。覆盖：商品搜索→加入购物车→填写收货地址→选择支付方式→确认订单→查看订单详情。包含：Page Object Model 架构、测试数据工厂、截图对比、网络请求拦截 mock、CI 集成配置、测试报告生成。请用全栈开发技能输出包含完整代码的 Markdown 文档。", "prompt_en": "Help me write a Playwright E2E test suite for the core checkout flow of an e-commerce website. Cover: product search → add to cart → fill in shipping address → select payment method → confirm order → view order details. Include: Page Object Model architecture, test data factories, screenshot comparison, network request interception mocks, CI integration configuration, and test report generation. Please use the fullstack development skill to output a Markdown document containing the complete code.", "skills": [ { "id": "fullstack-dev", "name": "全栈开发", "description": "全栈应用架构与开发指南", "description_en": "Full-stack architecture and development guide", "skill_id": "skill_2053082011994116096" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 211, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-mbti-estj", "title": "MBTI：ESTJ 总经理，和 ENFP 只差两个字母", "title_en": "MBTI: ESTJ the Executive, Two Letters From ENFP", "subtitle": "同一套题、同样 44 道，换一组作答就落到了相反的气质", "subtitle_en": "Same forty-four questions, opposite temperament", "description": "很多人拿到 MBTI 结果后最想问的是：这四个字母到底在测什么？光看一个类型的描述，很难有实感。\n\n本案例用同一套 44 题跑出 ESTJ，与 ENFP 那份并排看：共享 E，但在 S/N、T/F、J/P 三对上完全相反。案例展示逐维度分布、题库原文的优势与短板，并解释为什么两者的职业清单几乎没有重叠。\n\n适合：想理解 MBTI 维度含义的学生、给团队做性格差异科普的人。", "description_en": "The question people most want answered after an MBTI result is: what are these four letters actually measuring? Reading one type's description rarely settles it.\n\nThis case runs the same 44 questions to an ESTJ result and sets it beside the ENFP: they share E but sit opposite on S/N, T/F and J/P. It shows the dimension splits, the strengths and blind spots from the item bank, and explains why the two career lists barely overlap.\n\nFor: students wanting to understand what each dimension means, and anyone explaining personality differences to a team.", "categories": [ "education" ], "tags": [ "MBTI", "ESTJ", "维度对比", "性格测评" ], "tags_en": [ "MBTI", "ESTJ", "dimension comparison", "personality assessment" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "MBTI ESTJ 型测评结果报告", "label_en": "MBTI ESTJ assessment report" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次 MBTI 职业性格测评。我作答时倾向于：愿意主动社交、相信具体事实和经验多过直觉、决策时以规则和效率为准、喜欢提前计划并按计划执行。要求：1) 用 44 题题库真实跑一遍评分脚本，给出人格类型和类型清晰度分数；2) 逐个维度对展示作答分布与判定结果；3) 完整给出题库中该类型的优势与需要留意的倾向，不要改写；4) 列出匹配职业方向；5) 把这个结果与 ENFP 做对比，说明两者在哪几个维度上相反、这如何解释职业清单的差异。", "prompt_en": "Run an MBTI career-personality assessment. My answers lean toward active socialising, trusting concrete facts and experience over intuition, deciding by rules and efficiency, and planning ahead. Requirements: 1) run the official 44-question scoring script and give the type and clarity score; 2) show the answer split and verdict for each dimension pair; 3) reproduce the strengths and blind spots from the item bank verbatim; 4) list the matched career directions; 5) compare this result with ENFP, identifying which dimensions are opposite and how that explains the difference in career lists.", "skills": [ { "id": "career-personality", "name": "职业性格测评", "description": "44 题 MBTI 测评，输出四维分布与 16 型人格档案", "skill_id": "skill_2089643348855885824" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 212, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "retail-demand-forecast", "title": "零售门店需求预测与库存优化", "title_en": "Retail Store Demand Forecast & Inventory Optimization", "subtitle": "SKU 安全库存计算与智能补货策略", "subtitle_en": "SKU safety stock calculation and smart replenishment strategy", "description": "便利店缺货损失销售、积压占用资金，人工盘点补货全凭经验？用库存需求规划 Skill 输入门店和 SKU 数据，自动做 ABC/XYZ 分类、季节性分析、安全库存计算、补货触发逻辑，输出一套可落地的库存优化方案。适合连锁零售店长、供应链经理、采购主管。", "description_en": "Stockouts lose sales while overstock ties up capital, and manual replenishment relies on gut feeling. Use the Inventory Demand Planning Skill to auto-classify SKUs with ABC/XYZ analysis, calculate safety stock, and design replenishment triggers. Perfect for retail store managers, supply chain managers, and procurement leads.", "categories": [ "business", "data" ], "tags": [ "需求预测", "库存优化", "零售", "供应链", "安全库存" ], "tags_en": [ "Demand Forecast", "Inventory Optimization", "Retail", "Supply Chain", "Safety Stock" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "零售门店需求预测与库存优化" } ], "prompt": "我经营一家连锁便利店（3 家门店），有 200 个 SKU。请帮我制定一套需求预测与库存优化方案，包含：ABC/XYZ 分类分析、季节性商品管理策略、安全库存计算方法、自动补货触发逻辑、促销活动备货模型、供应商谈判框架。", "prompt_en": "I run a chain of convenience stores (3 locations) with 200 SKUs. Please help me develop a demand forecasting and inventory optimization plan, including: ABC/XYZ classification analysis, seasonal product management strategies, safety stock calculation methods, automatic replenishment trigger logic, a promotional event stocking model, and a supplier negotiation framework.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 212, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "chatlaw-divorce-property-custody", "title": "离婚财产抚养咨询", "title_en": "Divorce Property and Custody Consultation", "subtitle": "5 角色法律团队：房产分割、父母出资认定与子女抚养权一体分析，输出可执行离婚方案", "subtitle_en": "5-role legal team: integrated analysis of property division, parental down-payment contribution, and child custody with an actionable divorce strategy", "description": "离婚最难谈的，往往不是要不要离，而是房子到底归谁、父母出的首付款算不算个人贡献、孩子跟谁生活才更稳。尤其是婚后买房只写一方名字、孩子又长期由另一方照料时，很多人一边怕房子拿不住，一边又担心抚养权完全没机会。\n\n这个案例演示了中文法律咨询团如何完整处理一类高频婚姻家事问题：先把婚龄、房屋资金链、父母出资、车辆理财、孩子当前照料状态等事实拆清，再结合《民法典》婚姻家庭编、2025 年婚姻家庭编解释（二）以及北京和最高法的公开裁判趋势，分析哪些属于夫妻共同财产、哪些只能作为分割时倾斜因素、房屋归一方时补偿如何估算、以及 6 岁子女抚养权在实务上最看重什么。最终交付一份完整《法律咨询报告》，附结果区间、证据清单和 7 步行动建议。\n\n适合：准备离婚、正卡在房屋分割和孩子抚养安排上的夫妻；婚姻家事律师助理、家事调解员、企业法务；需要快速形成谈判底稿和诉前方案的人群。", "description_en": "The hardest part of divorce is often not whether to separate, but who keeps the home, whether a parental down payment counts as personal contribution, and which living arrangement is truly better for the child. This becomes even harder when the marital home was purchased during marriage but registered under only one spouse, while the child has long been cared for mainly by the other spouse.\n\nThis case shows how the ChatLaw Team handles a common family-law scenario end to end: first structuring the marriage timeline, housing payment chain, parental contribution, car and investment assets, and the child's current caregiving pattern; then applying the Civil Code marriage-and-family rules, the 2025 judicial interpretation (II), and public Beijing / Supreme People's Court case trends to analyze what counts as marital property, what may only influence allocation rather than ownership, how compensation works when the home is awarded to one side, and what courts practically care about most in custody disputes involving a 6-year-old child. The final output is a full legal consultation report with outcome ranges, evidence checklist, and a 7-step action plan.\n\nTarget audience: people preparing for divorce and stuck on property and custody issues; family-law legal assistants, mediators, and in-house legal teams; anyone needing a fast negotiation brief and pre-litigation strategy.", "categories": [ "business", "creation", "agent" ], "tags": [ "离婚财产分割", "抚养权", "父母出资购房", "法律咨询", "多Agent协作" ], "tags_en": [ "Divorce Property Division", "Child Custody", "Parental Down Payment", "Legal Consultation", "Multi-Agent Collaboration" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "离婚财产分割法律咨询报告（Markdown）" } ], "prompt": "我和妻子结婚 8 年，有 6 岁孩子。婚后共同买了一套房（房产证登记在我一人名下，首付 60 万是我父母出的，贷款双方共同还）、一辆车（登记在妻子名下）、还有些理财。现准备离婚，想要房子和孩子抚养权，愿意合理补偿。\n\n请使用中文法律咨询团完整流程处理：\n1. 先做信息采集，必要信息缺失时向我追问；\n2. 输出完整《法律咨询报告》Markdown，必须包含：案情摘要、适用法条、类似判例、法律定性、财产分割结果区间（协议离婚 vs 诉讼离婚）、孩子抚养权归属影响因素、推荐路径（协商→调解→诉讼）、风险提示与证据清单；\n3. 明确哪些属于夫妻共同财产、哪些可能视为个人财产或个人贡献；\n4. 说明若房屋归我，大致补偿思路如何测算；\n5. 给出下一步行动建议。", "prompt_en": "My wife and I have been married for 8 years and have a 6-year-old child. During the marriage we jointly bought an apartment (the property deed is registered solely under my name, the ¥600,000 down payment was paid by my parents, and we both repay the mortgage together), a car (registered under my wife's name), and we have some investments. We are now preparing to divorce, and I want the apartment and custody of the child, and I am willing to pay reasonable compensation.\n\nPlease handle this using the ChatLaw Team's complete workflow:\n1. First conduct information intake, and follow up with me when necessary information is missing;\n2. Output a complete Legal Consultation Report in Markdown, which must include: case summary, applicable statutes, similar precedents, legal characterization, property division outcome ranges (negotiated divorce vs. litigated divorce), factors influencing the custody decision, recommended path (negotiation → mediation → litigation), risk warnings, and an evidence checklist;\n3. Clearly identify which assets are marital property and which may be considered personal property or personal contributions;\n4. Explain, if the apartment is awarded to me, roughly how the compensation would be calculated;\n5. Provide next-step action recommendations.", "skills": [], "experts": [ { "id": "ChatLawTeam", "name": "中文法律咨询团", "description": "案情采集、法条研究、判例分析、建议撰写，为民事、婚姻、合同、劳动等高频场景出具专业法律咨询报告。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "2a70787f0a5a", "name_en": "ChatLaw Team", "description_en": "Case intake, statute research, precedent analysis, and advice drafting—delivering professional legal consultation reports for civil, marriage, contract, and labor cases.", "expert_id": "ex_pbny8vaVXPvg" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 213, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "event-planning-proposal", "title": "活动策划执行方案", "title_en": "Event Planning Proposal", "subtitle": "描述活动需求，自动生成含时间线、预算和应急预案的 Word 方案", "subtitle_en": "Describe event requirements, auto-generate a Word proposal with timeline, budget, and contingency plans", "description": "策划一场 200 人的公司庆典有多崩溃？场地、餐饮、节目、预算、人员分工、物料清单、应急预案……每一项都是大工程，传统方式靠 Excel + Word 来回切换，改一个预算数字就得连改五张表。\n\n这个案例演示了输入「公司五周年庆典」的核心需求（200 人、15 万预算、线下+直播），AI 自动生成一份完整的活动策划执行方案 Word 文档：包含精确到分钟的活动时间线、14 万+ 的详细预算明细表、RACI 人员分工矩阵、三轮抽奖方案、颁奖流程设计、物料采购清单、5 套应急预案、筹备倒推时间表，以及活动后复盘框架。生成正式 Word 格式，拿来就能用。\n\n适合：行政部门、HR 团队、市场部活动策划、项目经理、需要组织年会/庆典/团建的任何人。", "description_en": "How stressful is planning a 200-person company celebration? Venue, catering, entertainment, budget, staff assignments, materials list, contingency plans — each is a major project. The traditional approach of switching between Excel and Word means changing one budget number requires updating five spreadsheets.\n\nThis case demonstrates inputting core requirements for a \"Company 5th Anniversary Gala\" (200 people, 150K budget, offline + livestream), and AI auto-generates a complete event planning proposal as a Word document: minute-by-minute timeline, detailed budget breakdown, RACI staffing matrix, raffle plan, award ceremony design, materials checklist, 5 contingency plans, and post-event review framework.\n\nIdeal for: Admin departments, HR teams, marketing event planners, project managers, anyone organizing annual galas or team-building events.", "categories": [ "office", "business" ], "tags": [ "活动策划", "预算管理", "周年庆典", "项目管理", "执行方案" ], "tags_en": [ "Event Planning", "Budget Management", "Anniversary Gala", "Project Management", "Execution Plan" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "活动策划执行方案（Word）" } ], "prompt": "策划公司成立五周年庆典活动，生成一份完整的活动策划方案 Word 文档（output.docx）。\n\n公司是一家 200 人的互联网科技公司，总部在深圳。活动要求：\n1）形式：线下为主 + 线上直播（照顾异地同事）\n2）预算：15 万元以内\n3）时间：2026 年 6 月 14 日（周六）下午 14:00-21:00\n4）场地：公司附近可容纳 250 人的活动场地\n5）核心环节包括：开场 CEO 致辞与回顾视频、优秀员工/团队颁奖、各部门创意节目表演、抽奖互动环节（至少 3 轮）、自助晚宴 + 自由交流\n\n需要生成完整的活动策划方案，包含时间线、预算明细表、人员分工表、物料清单、应急预案。\n\n文档格式要求：包含封面页、目录、页眉页脚、多级标题编号、统一表格样式\n输出为正式 Word 文档（output.docx）", "prompt_en": "Plan a 5th-anniversary celebration event for the company, and generate a complete event proposal as a Word document (output.docx).\n\nThe company is a 200-person internet tech company headquartered in Shenzhen. Event requirements:\n1) Format: primarily in-person + live streaming (to accommodate colleagues in other locations)\n2) Budget: within RMB 150,000\n3) Time: Saturday, June 14, 2026, afternoon 14:00-21:00\n4) Venue: an event space near the company that can hold 250 people\n5) Core segments include: an opening CEO address and retrospective video, awards for outstanding employees/teams, creative performances from each department, lucky-draw interaction rounds (at least 3 rounds), and a buffet dinner + free mingling\n\nI need a complete event proposal that includes a timeline, a detailed budget table, a staffing assignment table, a materials checklist, and a contingency plan.\n\nDocument formatting requirements: include a cover page, table of contents, headers and footers, multi-level heading numbering, and consistent table styling\nOutput as a formal Word document (output.docx)", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 214, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-21T10:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "research-humanoid-robot-chain-2026", "title": "人形机器人产业链深度研究 2026", "title_en": "Humanoid Robot Industry Chain Deep Research 2026", "subtitle": "6 大环节 95+ 家代表公司 × 投资机会三层分级 × 产业链地图与融资趋势 SVG", "subtitle_en": "6 segments 95+ companies x 3-tier investment playbook, with industry map and funding trend SVG", "description": "2024 年 Figure 02 量产、NVIDIA GR00T 发布、Tesla Optimus V3 亮相，人形机器人赛道从「科研展示」真正跃迁为「千亿美金产业」。但一级二级投资人面对这个赛道最头疼的是：本体厂就有 18 家，零部件再按关节/传感器/控制/AI 大脑拆开，满屏都是代码，到底哪家值得买？哪家是假故事？\n\n这个案例演示了深度研究团队 13 分钟内交付的《人形机器人产业链深度研究报告 2026》：Markdown 主报告约 1.6 万字，按「本体 - 关节/执行器 - 感知 - 控制 - AI 大脑 - 应用场景」6 大环节全产业链拆解，每环节列出 8-18 家代表公司（A 股代码/港股/美股/非上市融资轮次全标注），合计覆盖 95+ 家公司；给出 2026-2030 五年产业链成熟度甘特图、投资机会三层分级（短期 8 家/中期 8 家/长期 8 家共 24 只推荐 + 买入逻辑）、4 类风险识别 13 条具体点。配套 HTML 可视化版含产业链地图（CSS Grid 6 环节×10 公司四色编码：国内上市红实/国内非上市红虚/海外上市蓝实/海外非上市蓝虚）、6 块代表公司矩阵卡（60 张 micro 卡片）、2020-2026 融资热度双轴 SVG 图（柱形事件数+折线总额，标注 2024 爆发拐点）。\n\n适合：VC / PE 投委会、公募基金新兴产业组、卖方分析师、FA、产业 CVC、做年度策略的买方研究总监。", "description_en": "With Figure 02 mass production, NVIDIA GR00T launch, and Tesla Optimus V3 unveil in 2024, humanoid robotics leapt from research demos to a hundred-billion-dollar industry. Yet investors face a brutal screening problem: 18 OEMs alone, then components split across joints/sensors/control/AI brains, who actually deserves capital and who is just a story?\n\nThis case shows the Deep Research Team delivering a complete Humanoid Robot Industry Chain Deep Research 2026 report in 13 minutes: a ~16k-character Markdown main report dissecting the full chain across 6 segments (OEM, joints/actuators, sensing, control, AI brain, applications) with 8-18 representative companies per segment (A-share/HK/US tickers, private-round history fully tagged), covering 95+ companies in total; a 2026-2030 five-year maturity Gantt chart; a 3-tier investment playbook (short 8 + mid 8 + long 8 = 24 picks with buy logic); and 13 concrete risk points across 4 categories. Companion HTML includes an industry map (CSS Grid, 6 segments x 10 companies, four-color coded: CN-listed solid red, CN-private dashed red, overseas-listed solid blue, overseas-private dashed blue), 6 matrix-card sections (60 micro cards), and a 2020-2026 funding trend SVG dual-axis chart (bars = deal count, line = total dollars, with 2024 inflection annotated).\n\nFit for: VC/PE investment committees, mutual-fund emerging-sector desks, sell-side analysts, FAs, corporate venture capital, and buy-side research heads running annual strategy.", "categories": [ "business", "data", "search" ], "tags": [ "投资主题", "人形机器人", "产业链研究", "投资机会", "估值分析" ], "tags_en": [ "Investment Theme", "Humanoid Robot", "Industry Chain", "Investment Opportunity", "Valuation" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "report.md", "artifacts": [ { "file": "report.md", "label": "Markdown 主报告（1.6 万字 · 95+ 家公司）" }, { "file": "visual-report.html", "label": "HTML 可视化版（产业链地图+融资趋势+矩阵卡）" } ], "prompt": "请面向一级/二级投资人产出一份《人形机器人产业链深度研究报告 2026》，使用「深度研究团队」5 阶段完整模式。\n\n产业链按 6 大环节拆解（本体、关节/执行器、感知、控制、AI 大脑、应用场景），每环节独立一章并列出代表公司（上市标注代码与市值，非上市标注融资轮次与估值，附核心技术与量产进度）。\n\n报告结构：引言、产业链全景与市场规模、六大环节、成熟度时间表、投资机会分级（短/中/长期，每层附买入逻辑）、风险识别、结论、参考文献。总字数 1.5-2 万字，2025-2026 数据标注「需核对」。\n\n另出 HTML 可视化版（单页内联、无外部依赖）：产业链地图、代表公司矩阵、融资热度趋势图。", "prompt_en": "Produce a \"Humanoid Robot Industry Chain Deep Research Report 2026\" for primary/secondary-market investors, using the Deep Research Team's full 5-phase mode.\n\nBreak the value chain into 6 segments (body/whole machine, joints/actuators, perception, control, AI brain, application scenarios). Give each its own chapter and list representative companies (listed ones with ticker and market cap; private ones with funding round and valuation; all with core tech and mass-production progress).\n\nReport structure: introduction, industry-chain landscape & market size, the six segments, a maturity timeline, tiered investment opportunities (short/mid/long term, each with buy thesis), risk identification, conclusion, references. Total 15,000–20,000 characters; mark 2025–2026 data as \"to be verified\".\n\nAlso produce an HTML visual version (single-page, inline, no external dependencies): industry-chain map, representative-company matrix, funding-heat trend chart.", "skills": [], "experts": [ { "id": "GPTResearcherTeam", "name": "深度研究团队", "description": "深度研究报告输出，7角色5阶段聚合多源信息，经审稿修订循环输出带引用的专业报告", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "a8da6f2dc1f7", "name_en": "Deep Research Team", "description_en": "Deep research report generation: 7 roles in 5 phases aggregate multi-source data through review-revise loop into cited professional reports", "expert_id": "ex_xdj3MuRUqbR0" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 215, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T03:35:00Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "python-automation-script", "title": "Python 数据处理自动化脚本", "title_en": "Python Data Processing Automation Script", "subtitle": "从数据读取到质量报告的全流程 ETL 管道", "subtitle_en": "Full ETL pipeline from data ingestion to quality report generation", "description": "手动处理数据文件费时费力，格式不统一、缺失值、异常值到处都是？用全栈开发技能生成一套完整的 Python 数据处理自动化脚本，覆盖从数据读取到质量报告输出的全流程 ETL 管道。\n\n脚本支持从 CSV、JSON、Excel、SQLite 四种数据源读取数据，内置数据清洗管道（去重、缺失值处理、类型转换、异常值检测），并提供分组统计、透视表、时间序列重采样等数据转换与聚合能力。运行结束后自动生成 HTML 格式的数据质量报告。\n\n适合：数据工程师、数据分析师、后端开发者、需要处理批量数据的运营人员。", "description_en": "Manual data processing is tedious with inconsistent formats, missing values, and outliers everywhere. Use the fullstack-dev skill to generate a complete Python data processing automation script covering the full ETL pipeline.\n\nThe script reads from CSV, JSON, Excel, and SQLite, includes data cleaning (deduplication, missing-value handling, type conversion, outlier detection), and provides group-by statistics, pivot tables, and time-series resampling. Auto-generates an HTML data quality report upon completion.\n\nPerfect for data engineers, data analysts, backend developers, and operations staff handling batch data.", "tags": [ "Python", "自动化", "数据处理", "ETL", "脚本" ], "tags_en": [ "Python", "Automation", "Data Processing", "ETL", "Script" ], "artifact_type": "code", "categories": [ "dev", "data" ], "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "Python 数据处理自动化脚本" } ], "prompt": "请生成一个 Python 数据处理自动化脚本，实现以下功能：1) 从多种数据源读取数据（CSV、JSON、Excel、SQLite）；2) 数据清洗管道（去重、缺失值处理、类型转换、异常值检测）；3) 数据转换与聚合（分组统计、透视表、时间序列重采样）；4) 生成 HTML 格式的数据质量报告（含统计摘要、数据分布图表）；5) 支持命令行参数配置（argparse）；6) 完整的日志记录和错误处理；7) 使用 dataclasses 定义配置模型。要求只使用 Python 标准库 + pandas + openpyxl。请用全栈开发技能输出包含完整代码的 Markdown 文档。", "prompt_en": "Generate a Python data processing automation script that implements the following features: 1) read data from multiple data sources (CSV, JSON, Excel, SQLite); 2) a data cleaning pipeline (deduplication, missing-value handling, type conversion, outlier detection); 3) data transformation and aggregation (group-by statistics, pivot tables, time-series resampling); 4) generate an HTML data quality report (with statistical summaries and data distribution charts); 5) support command-line argument configuration (argparse); 6) complete logging and error handling; 7) use dataclasses to define the configuration model. Requirement: use only the Python standard library + pandas + openpyxl. Please use the fullstack development skill to output a Markdown document containing the complete code.", "skills": [ { "id": "fullstack-dev", "name": "全栈开发", "description": "全栈应用架构与开发指南", "description_en": "Full-stack architecture and development guide", "skill_id": "skill_2053082011994116096" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "quality_score": 86, "sort_order": 217, "is_featured": false, "featured_order": 0, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "sales-pipeline-tracker", "title": "销售漏斗追踪表", "subtitle": "从线索到成交，全链路销售数据一表掌控", "description": "销售团队最头疼的问题：线索散落在微信、邮件、CRM 里，跟进状态靠口头汇报，月底才发现漏斗卡在了某个环节——等发现时，季度目标已经救不回来了。\n\n这个案例用 AI 一键生成一张完整的销售漏斗追踪 Excel：从线索录入、商机阶段推进、客户分级到成交预测，每个环节都有数据支撑。内置转化率自动计算、加权管线价值评估、月度/季度收入预测、客户 ABC 分层分析，打开就能看到哪个阶段在漏单、哪个销售在掉队。\n\n适合：销售主管、销售运营、BD 团队、创业公司 CEO、任何需要清晰掌控销售进展的人。", "categories": [ "data", "business" ], "tags": [ "Excel", "销售漏斗", "销售管理", "CRM", "数据分析", "收入预测" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "销售漏斗追踪表" } ], "prompt": "帮我生成一份专业的销售漏斗追踪 Excel 报表。\n\n**业务背景**：我们是一家 B2B SaaS 公司，销售团队 6 名销售代表，需要追踪从线索到成交的全链路漏斗。请生成 2026 年上半年的合理模拟数据。\n\n**报告结构（多 Sheet）**：\n- Sheet 1 漏斗总览：各阶段商机数量、金额、转化率（公式计算），及加权管线价值、平均销售周期、整体赢单率。\n- Sheet 2 线索库：约 50 条线索，代表性字段如线索ID、公司、来源渠道、状态、负责人。\n- Sheet 3 商机跟踪：约 30 条商机，代表性字段如商机ID、预计金额、成交概率、当前阶段、负责销售。\n- Sheet 4 收入预测：按月展示目标、已成交、加权预测及完成率，并按销售代表展示个人达成。\n- Sheet 5 客户分层：成交客户按 ABC 分层，统计各层数量、收入、续费率等。\n\n格式：专业商务风格，关键指标用条件格式分级，金额带 ¥，首行冻结并加自动筛选。输出为 .xlsx 文件。", "prompt_en": "Help me generate a professional sales-funnel tracking Excel report.\n\n**Business background**: We're a B2B SaaS company with a sales team of 6 reps, and we need to track the full funnel from lead to close. Please generate reasonable simulated data for the first half of 2026.\n\n**Report structure (multiple sheets)**:\n- Sheet 1 Pipeline Overview: per-stage deal count, amount, and conversion rate (formula-based), plus weighted pipeline value, average sales cycle, and overall win rate.\n- Sheet 2 Lead Database: ~50 leads, with representative fields like lead ID, company, source channel, status, owner.\n- Sheet 3 Deal Tracker: ~30 deals, with representative fields like deal ID, expected amount, close probability, current stage, owning rep.\n- Sheet 4 Revenue Forecast: monthly target, closed, weighted forecast, and completion rate, plus per-rep attainment.\n- Sheet 5 Customer Segmentation: closed customers in A/B/C tiers, with count, revenue, renewal rate, etc.\n\nFormatting: professional business style; conditional-format tiers for key metrics; amounts with ¥; freeze the top row and add auto-filter. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 218, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Sales Pipeline Tracker", "subtitle_en": "Track your entire sales funnel from leads to closed deals in one spreadsheet", "description_en": "The biggest headache for sales teams: leads scattered across WeChat, email, and CRM, follow-up status reported verbally, and by month-end you realize the funnel is stuck at some stage — by then, the quarterly target is already out of reach.\n\nThis case uses AI to generate a complete sales pipeline tracking Excel in one click: from lead entry, opportunity stage progression, customer tiering to deal forecasting, every step is data-driven. Built-in conversion rate calculations, weighted pipeline value assessment, monthly/quarterly revenue forecasts, and ABC customer segmentation — open it and instantly see which stage is leaking deals and which rep is falling behind.\n\nBest for: Sales managers, sales ops, BD teams, startup CEOs, anyone who needs clear visibility into sales progress.", "tags_en": [ "Excel", "Sales Pipeline", "Sales Management", "CRM", "Data Analysis", "Revenue Forecast" ], "is_recommended": false, "recommended_order": 0 }, { "id": "freelancer-contract", "title": "自由职业者服务合同", "title_en": "Freelancer Service Contract", "subtitle": "设计/开发外包合同 Word 模板，保障双方权益", "subtitle_en": "Design/development outsourcing contract Word template protecting both parties", "description": "接外包项目没有正式合同？口头约定容易导致需求变更扯皮、尾款收不回来，项目交付后知识产权归属不清。\n\n用 Word 文档生成技能生成一份专业的自由职业者服务合同 Word 文档，覆盖项目范围、付款条款、知识产权、保密协议、违约责任等关键条款，格式规范可直接打印签署。\n\n适用于自由设计师、独立开发者、外包项目甲方、创业公司等需要签订服务合同的场景。", "description_en": "Taking on a freelance project without a formal contract? Verbal agreements lead to scope creep disputes, unpaid invoices, and unclear IP ownership after delivery.\n\nUse the Word Document Generation skill to produce a professional freelancer service contract as a Word document, covering project scope, payment terms, IP rights, NDAs, and breach penalties — formatted and ready for printing and signing.\n\nPerfect for freelance designers, independent developers, outsourcing clients, and startups needing service contracts.", "categories": [ "business", "office" ], "tags": [ "合同模板", "自由职业", "外包合同", "法务", "知识产权" ], "tags_en": [ "Contract Template", "Freelance", "Outsourcing", "Legal", "IP Rights" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "自由职业者服务合同（Word）" } ], "prompt": "我是一名自由设计师，即将接一个 App UI 设计外包项目，预算 3 万元，工期 2 个月。请帮我生成一份专业的自由职业者服务合同 Word 文档（output.docx），需要包含：项目范围定义、里程碑付款条款（分 3 期）、修改次数限制、知识产权归属、保密条款、违约责任、争议解决方式。\n\n文档格式要求：\n- 合同要正式严谨，包含封面页、目录、条款编号\n- 包含甲乙方信息区、签字盖章区\n- 页眉页脚、页码\n- 可直接打印签署\n- 输出为正式 Word 文档（output.docx）", "prompt_en": "I'm a freelance designer about to take on an outsourced App UI design project, with a budget of ¥30,000 and a timeline of 2 months. Please help me generate a professional freelancer service contract as a Word document (output.docx). It should include: definition of project scope, milestone payment terms (in 3 installments), a limit on the number of revisions, intellectual property ownership, a confidentiality clause, liability for breach of contract, and the dispute-resolution method.\n\nDocument formatting requirements:\n- The contract should be formal and rigorous, including a cover page, table of contents, and numbered clauses\n- Include information areas for Party A and Party B, and a signature/seal area\n- Headers/footers and page numbers\n- Ready to print and sign as-is\n- Output as a formal Word document (output.docx)", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 219, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-22T06:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "ai-weekly-report-prd", "title": "AI 智能周报功能完整 PRD", "title_en": "Complete PRD for AI Smart Weekly Report", "subtitle": "5 人产品专家团队从用研到路线图的一站式 PRD 交付", "subtitle_en": "End-to-end PRD delivery by a 5-member product specialist team", "description": "写周报是中层管理者每周最抗拒的事——翻遍系统找痕迹、对着空白页发呆、汇总 10 个人的周报再翻译成领导听得懂的话。这个案例演示了产品战略团队如何为「AI 智能周报」功能完成从用户研究（8 大痛点、5 大洞察）、竞品分析（7 家对标）、数据指标框架（北极星指标+16 个埋点事件）、完整 PRD（7 个功能模块 F1-F7）到 RICE 优先级评估与 3 批次发布路线图的全链路交付，产出的 Markdown 文档可直接提交工程与设计评审。\n\n适合：产品经理、产品负责人、需要从零撰写功能 PRD 的团队、希望系统化产品战略规划的创业者。", "description_en": "Weekly reports are the most dreaded task for middle managers — digging through systems for traces, staring at blank pages, and summarizing 10 people's reports into language leadership understands. This case demonstrates how a Product Strategy Team delivers end-to-end for an AI Smart Weekly Report feature: user research (8 pain points, 5 insights), competitive analysis (7 competitors), data metrics framework (North Star + 16 tracking events), complete PRD (7 feature modules F1-F7), RICE prioritization and 3-batch release roadmap. The Markdown output is ready for engineering and design review.\n\nSuitable for: Product managers, product leads, teams writing feature PRDs from scratch, entrepreneurs seeking systematic product strategy planning.", "categories": [ "creation", "business" ], "tags": [ "PRD", "产品战略", "AI周报", "竞品分析", "路线图" ], "tags_en": [ "PRD", "Product Strategy", "AI Weekly Report", "Competitive Analysis", "Roadmap" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI 智能周报完整 PRD 交付包" } ], "prompt": "请为「AI 智能周报」功能撰写完整 PRD，面向企业协作 SaaS，目标用户为中层管理者。\n\n要求：\n1. 用户研究综合：采集周报痛点、期望场景，输出用户洞察和旅程地图\n2. 竞品分析：对标飞书、钉钉、Notion、MS Teams、Slack、Weekly10 等产品的周报功能实现，输出功能矩阵和差异化方向\n3. 数据指标基线：周报使用率、完成率、阅读率等行业基线，以及 AI 功能的北极星指标和埋点方案\n4. 完整 PRD：包含产品目标与非目标、用户故事、功能规格（F1-F7）、交互细节（页面线框图、核心流程）、数据埋点、权限与安全、异常处理、待确认问题\n5. 优先级评估：RICE 评分、MoSCoW 分类、分批次发布策略、风险调整时间线\n\n输出格式：标准化 Markdown，可直接提交工程与设计评审。", "prompt_en": "Please write a complete PRD for an \"AI Smart Weekly Report\" feature, targeting enterprise collaboration SaaS, with middle managers as the target users.\n\nRequirements:\n1. User research synthesis: gather weekly-report pain points and desired scenarios, and output user insights and a journey map\n2. Competitive analysis: benchmark the weekly-report feature implementations of products such as Feishu, DingTalk, Notion, MS Teams, Slack, and Weekly10, and output a feature matrix and differentiation directions\n3. Data metric baselines: industry baselines for weekly-report adoption rate, completion rate, read rate, etc., plus the North Star metric and event-tracking plan for the AI feature\n4. Complete PRD: including product goals and non-goals, user stories, feature specifications (F1-F7), interaction details (page wireframes, core flows), data tracking, permissions and security, exception handling, and open questions\n5. Priority assessment: RICE scoring, MoSCoW classification, a phased release strategy, and a risk-adjusted timeline\n\nOutput format: standardized Markdown, ready to submit directly for engineering and design review.", "skills": [], "experts": [ { "id": "ProductStrategyTeam", "name": "产品战略团队", "description": "由产品总监领导的 5 人产品专家团队：需求分析师（PRD/功能规格书）、用户研究员（调研综合分析）、竞品分析师（竞争情报）、数据分析师（指标追踪）和路线图规划师（路线图管理/迭代规划）。覆盖从构思到上线的完整产品生命周期。", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "", "name_en": "Product Strategy Team", "description_en": "A team of 5 product specialists led by a Product Director: Requirement Analyst (PRD/feature spec writing), User Researcher (research synthesis), Competitive Analyst (competitive intelligence), Data Analyst (metrics tracking), and Roadmap Planner (roadmap management/sprint planning). Covers the full product lifecycle from ideation to launch.", "expert_id": "ex_LqCvsVsLOp3J" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 220, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T15:10:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "hmc-health-content-boundary-review", "title": "健康科普软文「科普-广告」边界审查", "title_en": "Health-Content Education-vs-Ad Boundary Review", "subtitle": "私域科普文过审，专查借科普夹带药品/器械推荐、恐谎标题、个案当普遍", "subtitle_en": "Review private-domain science popularization — flags disguised drug/device promotion, fear-mongering headlines, case-as-universal", "description": "私域里很多「科普文章」其实夹带产品推荐——用患教外衣做软广，最容易被认定为「变相医疗广告」，编辑还以为只是科普。\n\n本灵感把一篇健康科普/患教软文交给医疗营销合规专家，按健康内容营销合规边界审查：抓取借科普夹带具体药品/器械推荐、恐惧营销标题（「不看后悔」）、个案当普遍、未标注「广告/推广」、医生个人账户违规荐药等边界问题，给出「科普留、广告删」的切分建议与标注规范。\n\n适合: 药企/互联网医院内容团队、私域运营、医生个人IP运营、品牌公关。", "description_en": "Many science popularization articles in private domains actually embed product recommendations — using patient-education packaging as soft advertising, most easily judged disguised medical advertising, while editors believe they are just educating.\n\nThis inspiration hands a health-science / patient-education soft article to the Healthcare Marketing Compliance Expert for a health-content marketing boundary review: it catches disguised specific drug / device recommendations under the guise of education, fear-mongering headlines (regret if you don't read), case-as-universal, missing advertisement / promotion labels, and physician personal accounts illegally recommending drugs, then provides keep education, cut ads split suggestions and labeling norms.\n\nSuitable for: pharma / internet-hospital content teams, private-domain operations, physician personal-IP operations, brand PR.", "categories": [ "business" ], "tags": [ "科普边界", "软文合规", "变相广告" ], "tags_en": [ "Education Boundary", "Soft-Ad Compliance", "Disguised Advertising" ], "artifact_type": "html", "preview": "output.html", "prompt": "我运营一个健康类公众号，最近开始接商单。这篇稿子前半段是正经科普，后半段带了产品链接，编辑说这样\"软一点\"不算广告。我不太确定——这种写法到底算不算变相广告？会不会被平台限流甚至处罚？帮我把这篇拆开看，哪部分是干净的科普可以留、哪部分已经越线必须删或者明示\"广告\"。\n\n稿件全文：\n\n---\n\n# 糖尿病患者到底该怎么吃主食？\n\n很多糖友一确诊就把米饭馒头全戒了，结果饿得头晕、血糖反而更乱。其实主食不是敌人，关键在选什么、吃多少。\n\n## 一、为什么不能完全不吃主食\n\n碳水是大脑主要能量来源。长期极低碳水饮食可能导致低血糖、酮症、肌肉流失。中国居民膳食指南建议，糖尿病患者碳水供能比控制在 45%–60% 之间为宜。\n\n## 二、看懂 GI 和 GL\n\n- **GI（升糖指数）**：反映食物升高血糖的速度。白米饭 GI 约 83，属高 GI；燕麦约 55，属中低 GI。\n- **GL（血糖负荷）**：同时考虑速度和分量，比 GI 更实用。\n\n## 三、主食替换的三个原则\n\n1. **粗细搭配**：白米里掺三分之一杂豆、糙米\n2. **控制总量**：每餐主食生重 50–75 克为参考\n3. **调整顺序**：先吃蔬菜和蛋白，最后吃主食，餐后血糖更平稳\n\n## 四、常见误区\n\n- ❌ \"无糖食品可以随便吃\" —— 无糖只是不加蔗糖，淀粉照样升糖\n- ❌ \"喝粥养胃对糖友好\" —— 粥糊化程度高，升糖比干饭快\n- ❌ \"水果一律不能吃\" —— 低 GI 水果两餐之间少量可以吃\n\n## 五、说点实在的\n\n道理都懂，可真到做饭的时候还是难。我们医院营养科很多医生都推荐一款**「稳糖膳」控糖主食面**——低 GI、高膳食纤维，煮起来跟普通面条一样方便。\n\n我自己吃了两个月，**吃着吃着血糖就稳了**，比天天吃药舒服多了。很多糖友反映餐后血糖比吃白米饭低不少。\n\n**现在下单还有优惠，糖友专属价 79 元 3 包，点下方链接就能买**：\n👉 【立即购买 稳糖膳控糖主食面】\n\n买过的糖友都说好，回购率超高。放心买，我们只推荐自己吃过的。\n\n---\n\n我想清楚三件事：一是\"我们医院营养科医生推荐\"这种说法能不能写；二是食品说\"血糖就稳了\"\"比吃药舒服\"是不是踩线；三是如果我在文章开头标个\"广告\"，后半段是不是就合法了。最后给我一张\"哪些留、哪些删、哪些改\"的对照清单，我照着动刀。", "prompt_en": "I run a health-focused WeChat public account and recently started taking sponsored posts. The first half of this draft is legitimate health education; the second half carries a product link. My editor says framing it this way makes it \"softer\" so it doesn't count as advertising. I'm not convinced — does this actually constitute disguised advertising? Could we get throttled or even penalized by the platform? Please break the piece apart: which portions are clean education worth keeping, and which have crossed the line and must be deleted or explicitly labeled \"Advertisement.\"\n\nFull draft:\n\n---\n\n# How Should Diabetes Patients Actually Eat Staples?\n\nMany patients cut out rice and steamed buns entirely upon diagnosis, only to end up dizzy from hunger with even more erratic glucose. Staples aren't the enemy — what matters is which ones and how much.\n\n## 1. Why You Shouldn't Cut Staples Entirely\n\nCarbohydrates are the brain's primary energy source. Prolonged very-low-carb eating can cause hypoglycemia, ketosis, and muscle loss. Chinese dietary guidelines suggest diabetes patients keep carbohydrate energy contribution between 45% and 60%.\n\n## 2. Understanding GI and GL\n\n- **GI (glycemic index)**: how fast a food raises blood glucose. White rice is around 83 (high GI); oats around 55 (medium-low).\n- **GL (glycemic load)**: accounts for both speed and portion — more practical than GI alone.\n\n## 3. Three Principles for Swapping Staples\n\n1. **Mix refined with whole**: add one-third mixed beans or brown rice to white rice\n2. **Control total volume**: 50–75 g dry weight of staples per meal as a reference\n3. **Change the order**: vegetables and protein first, staples last — steadier post-meal glucose\n\n## 4. Common Misconceptions\n\n- ❌ \"Sugar-free foods can be eaten freely\" — sugar-free only means no added sucrose; starch still raises glucose\n- ❌ \"Congee is gentle on the stomach and good for diabetics\" — congee is highly gelatinized and raises glucose faster than steamed rice\n- ❌ \"No fruit at all\" — small portions of low-GI fruit between meals are fine\n\n## 5. Let's Get Practical\n\nKnowing the theory is one thing; actually cooking is another. Many doctors in our hospital's nutrition department recommend a product called **\"Wentang Shan\" glucose-friendly staple noodles** — low GI, high dietary fiber, and as easy to cook as ordinary noodles.\n\nI've eaten them for two months and **my blood sugar just steadied out as I went**, far more comfortable than taking pills every day. Many patients report notably lower post-meal glucose than with white rice.\n\n**There's a discount right now — a diabetes-community price of ¥79 for 3 packs, just tap the link below**:\n👉 【Buy Now: Wentang Shan Glucose-Friendly Staple Noodles】\n\nEveryone who's bought them says they're great, with a very high repurchase rate. Buy with confidence — we only recommend what we eat ourselves.\n\n---\n\nI want clarity on three things: first, whether \"doctors in our hospital's nutrition department recommend it\" is permissible; second, whether a food claiming \"blood sugar just steadied\" and \"more comfortable than pills\" crosses the line; third, if I label the article \"Advertisement\" at the top, does that make the second half lawful? Finish with a keep/delete/revise comparison checklist I can cut against.", "skills": [], "experts": [ { "id": "nges-healthcare-compliance-lite", "name": "腾讯健康NGES合规专家", "description": "医疗营销合规专家，审核药品广告、医美种草、保健品声称，识别合规风险并支持改写为合规版本。", "expertType": "agent", "source": "expert-marketplace", "expert_id": "ex_ytn7WGJKoya8", "promptFileSnapshot": "f4a36dc03d8d" } ], "mcps": [], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 220, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-08-23T12:30:00+08:00", "updated_at": "2026-08-30T16:30:00+08:00", "artifacts": [ { "file": "output.html", "label": "健康科普软文边界审核报告" }, { "file": "output.md", "label": "健康科普软文边界审核报告（Markdown 版）" } ], "cover_image": "cover.png", "cover_mode": "landscape", "theme_label": "琥珀橙 · 科普软文边界审查", "theme_color": "#d97a16", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-career-anchor", "title": "职业锚测评：安全型主导", "title_en": "Career Anchors: Security-Dominant Profile", "subtitle": "识别「让你放弃也不肯放弃的东西」，比兴趣更底层", "subtitle_en": "Find what you refuse to give up — deeper than interests", "description": "兴趣会变，价值观很难变。职业锚测的是后者：让你在取舍时死守不放的那个东西。\n\n本案例跑完 40 题职业锚测评（每题 1–6 分），八维得分算出锚型 SE+TF+SV：安全型 30 分主导，技术职能型与服务型并列 25 分，报告给出三个维度的原文解读与「稳定平台上的专业岗」这一适配方向。\n\n适合：在体制内外之间摇摆的人、想搞清楚自己「图什么」的职场新人。", "description_en": "Interests shift; values rarely do. Career anchors measure the latter — the thing you won't trade away when forced to choose.\n\nThis case completes the 40-item anchor assessment (1–6 points each); the eight-dimension scores yield SE+TF+SV: Security dominant at 30, Technical-Functional and Service tied at 25, with verbatim readings of all three and a matched direction of \"specialist roles on stable platforms\".\n\nFor: people torn between public and private sectors and early-career workers clarifying what they're really after.", "categories": [ "education" ], "tags": [ "职业锚", "职业价值观", "安全型", "测评" ], "tags_en": [ "career anchors", "work values", "security type", "assessment" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "职业锚 SE+TF+SV 报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做 40 题职业锚测评并出报告。要求：1) 读题库渲染 40 题，每题按 1-6 分作答；2) 用评分脚本按真实规则计算八维得分与锚型；3) 报告含八维得分条形对比、前三维度的原文解读、适配方向；4) 结合我在纠结体制内外，给出这份锚型的落地建议并标注哪些是解读。", "prompt_en": "Run the 40-item career-anchor assessment and produce a report. Requirements: 1) load the bank and render all 40 items on a 1-6 scale; 2) compute the eight-dimension scores and anchor type with the scoring script's real rules; 3) include a bar comparison of all eight, verbatim readings of the top three, and matched directions; 4) given I'm weighing public versus private sector, add grounded suggestions clearly labelled as interpretation.", "skills": [ { "id": "career-anchor", "name": "职业锚测评", "description": "40 题评估八类职业价值取向，输出职业锚类型解读", "skill_id": "skill_2089337733810339840" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 220, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-anchor-autonomy", "title": "职业锚：自主型 29 分居首，挑战型紧随其后", "title_en": "Career Anchors: Autonomy Leads at 29, Challenge Close Behind", "subtitle": "40 道量表题算出的不是标签，而是一张取舍优先级表", "subtitle_en": "Forty scale items produce a priority order for trade-offs, not a label", "description": "拿到两个 offer 却选不出来，通常不是因为信息不够，而是因为没想清楚自己最不愿意放弃的是什么。\n\n本案例用 40 题职业锚测评跑出 AU+CH+TF 组合：自主型 29 分居首，与安全型 9 分相差 20 分——八个维度中跨度最大的一对。案例展示八维得分分布、前三锚的脚本原文释义、发展方向建议，并给出一套用职业锚做取舍的方法。\n\n适合：在多个 offer 之间做选择的人、考虑是否进体制的求职者。", "description_en": "Being stuck between two offers usually isn't an information problem — it's not having worked out what you're least willing to give up.\n\nThis case runs the 40-item career-anchor assessment to an AU+CH+TF result: autonomy leads at 29, twenty points clear of security at 9 — the widest gap among the eight dimensions. It shows the full score distribution, the top three anchors as the script defines them, the suggested directions, and a method for using anchors to make trade-offs.\n\nFor: anyone choosing between offers or weighing a public-sector career.", "categories": [ "education" ], "tags": [ "职业锚", "自主型", "价值取向", "职业选择" ], "tags_en": [ "career anchor", "autonomy", "values", "career choice" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "职业锚 AU+CH+TF 结果报告", "label_en": "Career anchor report: AU+CH+TF" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次职业锚测评。我作答时：对「按自己的方式和节奏工作」打分最高，对「解决复杂问题、迎接新挑战」也打得很高，对「安定稳定比自由更重要」打分很低。要求：1) 用 40 题题库真实跑一遍评分脚本，给出职业锚组合和八个维度的得分；2) 按得分高低展示八个维度分布；3) 完整给出前三名职业锚的脚本释义，不要改写；4) 列出脚本给出的发展方向建议，并说明这些方向的共同点；5) 给出一套用职业锚结果做取舍的方法，并说明它测的是价值取向而不是能力。", "prompt_en": "Run a career-anchor assessment. In my answers I rate \"working my own way at my own pace\" highest, \"solving complex problems and taking on new challenges\" nearly as high, and \"stability matters more than freedom\" very low. Requirements: 1) run the official 40-item scoring script and give the anchor combination and all eight dimension scores; 2) show the distribution ranked by score; 3) reproduce the script's definitions of the top three anchors verbatim; 4) list the suggested directions and what they have in common; 5) give a method for using anchor results to make trade-offs, noting that this measures values rather than ability.", "skills": [ { "id": "career-anchor", "name": "职业锚测评", "description": "40 题李克特量表，输出八维职业锚得分与前三组合", "skill_id": "skill_2089337733810339840" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 221, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "python-cli-scaffold", "title": "Python CLI 工具脚手架", "title_en": "Python CLI Tool Scaffold", "subtitle": "生产级命令行工具，开箱即用", "subtitle_en": "Production-grade CLI tool, ready to use out of the box", "description": "构建一个功能完整、用户体验良好的命令行工具涉及大量的基础设施代码：参数解析、彩色输出、日志系统、配置管理、错误处理等。这些脚手架工作重复且耗时，但又不可或缺。\n\n本案例演示了如何利用 AI 一次性生成一个生产级的 Python CLI 工具脚手架——文件批量重命名工具。支持正则替换、序号添加、日期前缀、大小写转换等多种模式，并内置预览、撤销、配置文件支持和完整的日志系统。\n\n适合 Python 开发者、DevOps 工程师、系统管理员，以及需要快速构建命令行工具的任何开发人员使用。", "description_en": "Building a complete CLI tool involves substantial infrastructure code for argument parsing, colored output, logging, config management, and error handling.\n\nThis case demonstrates using AI to generate a production-grade Python CLI scaffold — a batch file renaming tool with regex, sequencing, date prefix, and case conversion modes, plus preview, undo, and config support.\n\nIdeal for Python developers, DevOps engineers, and anyone needing to quickly build CLI tools.", "categories": [ "dev" ], "tags": [ "Python", "CLI工具", "脚手架", "开发效率" ], "tags_en": [ "Python", "CLI Tool", "Scaffold", "Dev Productivity" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.py", "prompt": "请生成一个生产级 Python CLI 工具脚手架：一个文件批量重命名工具（支持正则替换、序号添加、日期前缀、大小写转换）。技术要求：1) Click 多级命令（rename/preview/undo）；2) Rich 彩色输出和进度条；3) 日志系统（文件+控制台）；4) YAML 配置文件；5) 操作历史（支持 undo）；6) 完整错误处理和用户确认；7) 类型注解和 docstring。输出单文件版本完整 CLI。\n\n代码注释和用户可见字符串用简体中文，标识符保留英文。", "prompt_en": "Please generate a production-grade Python CLI tool scaffold: a batch file renaming tool (supporting regex replacement, adding sequence numbers, date prefixes, and case conversion). Technical requirements: 1) Click multi-level commands (rename/preview/undo); 2) Rich colored output and progress bars; 3) a logging system (file + console); 4) a YAML configuration file; 5) operation history (with undo support); 6) complete error handling and user confirmation; 7) type annotations and docstrings. Output a complete single-file version of the CLI.\n\nCode comments and user-visible strings should be in Simplified Chinese; identifiers should remain in English.", "skills": [], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 221, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.py", "label": "CLI 工具代码" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-anchor-management", "title": "职业锚：管理型 29 分居首，与自主型画像正好相反", "title_en": "Career Anchors: Management Leads at 29, the Mirror of the Autonomy Profile", "subtitle": "同一套 40 题，第一锚与最低锚几乎完全对调", "subtitle_en": "Same forty items — top and bottom anchors almost exactly swapped", "description": "技术路线还是管理路线？这个问题在职业早期反复出现，而答案往往取决于你从哪种成就感里获得满足。\n\n本案例用同一套 40 题跑出 GM+EC+TF 组合：管理型 29 分、创造型 24 分居前，生活型只有 9 分。与自主型那份并排看，两者的第一锚与最低锚几乎完全对调，而题目一字未改——这正是职业锚测评的价值所在。\n\n适合：面临技术转管理选择的人、评估是否愿意为责任让渡生活空间的求职者。", "description_en": "Technical track or management track? The question recurs throughout an early career, and the answer usually turns on which kind of achievement satisfies you.\n\nThis case runs the same 40 items to a GM+EC+TF result: management at 29 and entrepreneurship at 24 lead, lifestyle trails at 9. Set beside the autonomy profile, the top and bottom anchors are nearly swapped — with the questions unchanged. That contrast is the point of the instrument.\n\nFor: anyone weighing a move into management, or how much personal time they'd trade for responsibility.", "categories": [ "education" ], "tags": [ "职业锚", "管理型", "技术转管理", "价值取向" ], "tags_en": [ "career anchor", "management", "technical to management", "values" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "职业锚 GM+EC+TF 结果报告", "label_en": "Career anchor report: GM+EC+TF" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次职业锚测评。我作答时：对「整合并管理他人的工作很有成就感」打分最高，对「创造属于自己的事业」也打得很高，对「工作要为个人生活让路」打分很低。要求：1) 用 40 题题库真实跑一遍评分脚本，给出职业锚组合和八维得分；2) 按得分高低展示八个维度分布；3) 完整给出前三名职业锚的脚本释义，不要改写；4) 列出脚本给出的发展方向建议并说明其共同点；5) 与自主型主导的结果做对比，指出两者在第一锚和最低锚上的差异说明了什么。", "prompt_en": "Run a career-anchor assessment. In my answers I rate \"finding fulfilment in integrating and managing others' work\" highest, \"building something of my own\" nearly as high, and \"work should give way to personal life\" very low. Requirements: 1) run the official 40-item scoring script and give the anchor combination and all eight scores; 2) show the ranked distribution; 3) reproduce the script's definitions of the top three anchors verbatim; 4) list the suggested directions and what they share; 5) compare with an autonomy-dominant result and explain what the swapped top and bottom anchors reveal.", "skills": [ { "id": "career-anchor", "name": "职业锚测评", "description": "40 题李克特量表，输出八维职业锚得分与前三组合", "skill_id": "skill_2089337733810339840" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 222, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "student-grade-analysis", "title": "学生成绩分析表", "subtitle": "多科目成绩录入到排名分析，教学数据一表看清", "description": "班主任和教务主任最头疼的事情：每次考试后，手动算平均分、排名次、统计及格率，一科一科在 Excel 里拉公式，还要给每个学生做学情分析——一个班60个人、8门课，期中期末加起来上千条数据，光整理就要大半天。\n\n这个案例用 AI 一键生成一张完整的学生成绩分析 Excel：60名学生、8个科目（语文/数学/英语/物理/化学/生物/历史/地理）的期中和期末成绩全部自动录入，并内置总分排名、单科排名、班级平均分、标准差、及格率、优秀率等统计指标。还为每位学生生成雷达图数据和学科优劣势分析，一眼看出偏科情况，帮助教师精准辅导。\n\n适合：班主任、年级主任、教务处管理者、教育培训机构教学负责人。", "categories": [ "data", "education" ], "tags": [ "Excel", "成绩分析", "学生管理", "教学数据", "排名统计", "雷达图" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "学生成绩分析表" } ], "prompt": "帮我生成一份专业的学生成绩分析 Excel 报表。\n\n**业务背景**：我是高中班主任，班级 60 名学生，需要对期中、期末两次考试、8 个科目的成绩做全面分析。请生成合理的模拟成绩数据。\n\n**报告结构（多 Sheet）**：\n- Sheet 1 成绩录入：60 名学生完整记录，代表性字段如学号、姓名、各科期中/期末成绩，并用公式算出总分。\n- Sheet 2 排名表：期末总分及单科排名，对比期中名次给出进退步分析。\n- Sheet 3 统计分析：按科目统计平均分、标准差、及格率、优秀率等。\n- Sheet 4 学生画像：每位学生的排名、优劣势科目、雷达图数据与综合评价。\n- Sheet 5 班级对比：与同年级其他班对比各科平均分与成绩分布。\n\n格式：教育风格配色，成绩与排名变化用条件格式区分，首行首列冻结并加自动筛选。输出为 .xlsx 文件。", "prompt_en": "Help me generate a professional student grade analysis Excel report.\n\n**Business background**: I'm a high school homeroom teacher with 60 students. I need a comprehensive analysis of midterm and final scores across 8 subjects. Please generate reasonable simulated score data.\n\n**Report structure (multiple sheets)**:\n- Sheet 1 Grade Entry: complete records for 60 students, with representative fields like student ID, name, and each subject's midterm/final scores; use formulas for totals.\n- Sheet 2 Ranking: final total and per-subject rankings, with a progress/regression analysis vs. midterm.\n- Sheet 3 Statistics: per-subject average, standard deviation, pass rate, excellence rate, etc.\n- Sheet 4 Student Profile: each student's rank, strongest/weakest subjects, radar-chart data, and overall rating.\n- Sheet 5 Class Comparison: compare per-subject averages and score distribution against other classes in the grade.\n\nFormatting: education-style palette; use conditional formatting for scores and rank changes; freeze the first row and column and add auto-filter. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 222, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Student Grade Analysis", "subtitle_en": "From multi-subject grade entry to ranking analysis — all teaching data at a glance", "description_en": "The biggest headache for homeroom teachers and academic directors after every exam: manually calculating averages, rankings, and pass rates — pulling formulas subject by subject in Excel, then preparing individual student analyses. With 60 students and 8 subjects, midterm and final exams add up to over a thousand data points, and just organizing them takes half a day.\n\nThis case uses AI to generate a complete student grade analysis Excel in one click: 60 students, 8 subjects (Chinese, Math, English, Physics, Chemistry, Biology, History, Geography) with both midterm and final exam scores automatically entered. Built-in total score rankings, per-subject rankings, class averages, standard deviation, pass rates, and excellence rates. It also generates radar chart data and subject strength/weakness analysis for each student, making it easy to spot imbalances and help teachers provide targeted guidance.\n\nBest for: Homeroom teachers, grade directors, academic affairs administrators, and education training institution heads.", "tags_en": [ "Excel", "Grade Analysis", "Student Management", "Teaching Data", "Ranking Statistics", "Radar Chart" ], "is_recommended": false, "recommended_order": 0 }, { "id": "csv-data-cleaning-report", "title": "CSV 数据清洗分析报告", "subtitle": "脏数据一键清洗，数据质量评分自动生成", "description": "数据分析师的日常噩梦：拿到一份电商交易 CSV，打开一看——重复订单、缺失字段、日期格式五花八门、金额出现负数和离群值，手动逐行排查不仅耗时，还容易漏掉隐藏问题，最终导致报表数据失真、决策依据不可靠。\n\n这个案例用 AI 一键完成 CSV 数据清洗全流程：自动识别重复记录、填补缺失值、修正格式不一致、标记并处理异常离群值，同时生成完整的清洗日志记录每一步操作。更关键的是，它会从完整性、一致性、准确性、时效性四个维度自动计算数据质量评分，并输出清洗前后的统计汇总对比——让你对数据质量一目了然。\n\n适合：数据分析师、BI 工程师、数据运营、ETL 开发者、任何需要处理脏数据并确保数据质量的团队。", "categories": [ "data" ], "tags": [ "Excel", "数据清洗", "数据质量", "CSV", "数据分析", "ETL" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "CSV 数据清洗分析报告" } ], "prompt": "帮我生成一份完整的 CSV 数据清洗分析报告 Excel。\n\n**业务背景**：电商平台导出了一份约 1000 条交易记录的 CSV，存在多种数据质量问题，需要系统性清洗和分析。\n\n**第一步**：（提供待清洗数据）。\n...\n\n**报告结构（多 Sheet）**：\n- Sheet 1 原始数据：展示样本，问题单元格用底色标记，冻结首行 + 自动筛选\n- Sheet 2 清洗日志：逐条记录问题类型、处理方法、影响行数、处理前后示例\n- Sheet 3 数据质量评分：完整性/一致性/准确性/时效性四维度评分，含清洗前后对比\n- Sheet 4 统计汇总：数值字段描述性统计 + 分类字段频次 + 清洗前后关键指标对比\n- Sheet 5 清洗后数据：展示清洗后的干净数据样本\n\n专业商务风格，金额带千分位。输出为 .xlsx 文件。", "prompt_en": "Help me generate a complete CSV data cleaning and analysis report in Excel.\n\n**Business background**: An e-commerce platform exported a CSV of about 1,000 transaction records with various data quality issues that need systematic cleaning and analysis.\n\n**Step 1**: (Provide the data to be cleaned).\n...\n\n**Report structure (multiple sheets)**:\n- Sheet 1 Raw Data: show a sample, mark problem cells with background colors, freeze the first row + add auto-filter\n- Sheet 2 Cleaning Log: record each step — problem type, handling method, rows affected, before/after examples\n- Sheet 3 Quality Score: score across four dimensions (completeness/consistency/accuracy/timeliness), with before-vs-after comparison\n- Sheet 4 Summary Stats: descriptive statistics for numeric fields + frequency for categorical fields + before-vs-after key-metric comparison\n- Sheet 5 Cleaned Data: show a sample of the cleaned data\n\nProfessional business style with amounts shown with thousands separators. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 223, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "CSV Data Cleaning & Analysis Report", "subtitle_en": "One-click dirty data cleanup with automated quality scoring", "description_en": "A data analyst's daily nightmare: you receive an e-commerce transaction CSV, open it up, and find duplicate orders, missing fields, wildly inconsistent date formats, negative amounts, and outliers. Manual row-by-row inspection is not only time-consuming but error-prone, ultimately leading to inaccurate reports and unreliable decision-making.\n\nThis case uses AI to complete the entire CSV data cleaning workflow in one click: automatically identify duplicate records, fill missing values, correct format inconsistencies, flag and handle outliers, while generating a complete cleaning log documenting every operation. More importantly, it automatically calculates data quality scores across four dimensions — completeness, consistency, accuracy, and timeliness — and outputs a before-and-after statistical comparison so you can see data quality at a glance.\n\nBest for: Data analysts, BI engineers, data operations, ETL developers, and any team that needs to handle dirty data and ensure data quality.", "tags_en": [ "Excel", "Data Cleaning", "Data Quality", "CSV", "Data Analysis", "ETL" ], "is_recommended": false, "recommended_order": 0 }, { "id": "ecommerce-product-catalog", "title": "电商商品管理目录表", "subtitle": "SKU 管理、多规格价格矩阵一表搞定", "description": "管几百个 SKU 的电商运营最懂那种痛：商品信息散落在各平台后台，价格改了淘宝忘了京东，库存见底才发现没补货，每次大促前通宵理货——全靠人肉表格硬扛，出错是早晚的事。\n\n这个案例用 AI 一键生成一张完整的电商商品管理目录 Excel：200 个 SKU 横跨数码、服饰、美妆、食品、家居五大品类，每件商品都有多规格变体（尺码/颜色/容量），零售价/批发价/VIP 价三套定价体系自动计算利润率，库存预警一目了然，淘宝/京东/抖音三平台上架状态实时对照，品类汇总表直接给老板看。\n\n适合：电商运营、商品经理、采购、品类负责人、任何需要管理多平台多规格商品的人。", "categories": [ "data", "business" ], "tags": [ "Excel", "电商", "SKU管理", "商品目录", "价格矩阵", "库存管理" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "电商商品管理目录表" } ], "prompt": "帮我生成一份专业的电商商品管理目录 Excel（output.xlsx）。\n\n业务背景：一家在淘宝、京东、抖音三平台经营的电商公司，商品覆盖数码、服饰、美妆、食品、家居五大品类，共约 200 个 SKU。\n\n请包含以下 Sheet：\n1. 商品主表：SKU 编码、商品名称、品类、品牌、规格、产地、上架日期、状态等\n2. 价格矩阵：成本价、零售价、批发价、VIP 价、毛利率、毛利额，派生字段用公式计算\n3. 库存状态：当前库存、安全库存、补货点、月均销量、库存预警\n4. 平台上架：三平台各自的上架状态与链接 ID\n5. 品类汇总：各品类 SKU 数、均价、平均毛利率、库存金额、上架率，用公式引用\n\n格式：专业商务风格，标题行高亮，金额规范显示，关键指标加条件格式，冻结首行并加自动筛选。输出为 .xlsx 文件。", "prompt_en": "Help me generate a professional e-commerce product-management catalog in Excel (output.xlsx).\n\nBusiness context: an e-commerce company operating on Taobao, JD, and Douyin, with products across five categories (Digital, Apparel, Beauty, Food, Home), about 200 SKUs total.\n\nPlease include these sheets:\n1. Product Master: SKU code, product name, category, brand, specs, origin, listing date, status, etc.\n2. Pricing Matrix: cost price, retail price, wholesale price, VIP price, gross margin, gross profit; calculate derived fields with formulas\n3. Stock Status: current stock, safety stock, reorder point, average monthly sales, stock alerts\n4. Platform Listing: listing status and link ID per platform\n5. Category Summary: SKU count, average price, average margin, inventory value, listing rate per category, referenced via formulas\n\nFormat: professional business style, highlighted header rows, properly formatted amounts, conditional formatting on key metrics, freeze the top row and add auto-filter. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 224, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "E-commerce Product Catalog", "subtitle_en": "Manage SKUs and multi-variant pricing matrix in one spreadsheet", "description_en": "Every e-commerce operator managing hundreds of SKUs knows the pain: product info scattered across platform backends, price updated on Taobao but forgotten on JD, stock running out before anyone notices, pulling all-nighters before every big sale — relying on manual spreadsheets until something inevitably breaks.\n\nThis case uses AI to generate a complete e-commerce product catalog Excel in one click: 200 SKUs across five categories (electronics, apparel, beauty, food, home), each with multi-variant specs (size/color/volume), three pricing tiers (retail/wholesale/VIP) with auto-calculated margins, at-a-glance inventory alerts, cross-platform listing status for Taobao/JD/Douyin, and a category summary sheet ready for management review.\n\nBest for: E-commerce operators, product managers, merchandisers, category leads, anyone managing multi-platform multi-variant products.", "tags_en": [ "Excel", "E-commerce", "SKU Management", "Product Catalog", "Pricing Matrix", "Inventory Management" ], "is_recommended": false, "recommended_order": 0 }, { "id": "inventory-management", "title": "库存进销存管理表", "subtitle": "入库出库盘点全记录，库存预警零遗漏", "description": "手工记库存最怕什么？入库漏登、出库错记、盘点对不上账——月底一盘点，系统数和实际数差了一大截，到底是哪个环节出了问题，翻遍纸质单据也理不清。\n\n这个案例用 AI 一键生成一套完整的库存进销存管理 Excel：覆盖 100+ SKU、3 个仓库、3 个月的入库出库流水，自动计算实时库存、安全库存预警、库存周转率，并按 ABC 分类法对商品进行价值分层。打开就能看到哪些商品该补货、哪些积压严重、哪个品类周转最快。\n\n适合：仓库管理员、电商运营、供应链团队、中小企业主、任何需要精细化管理库存的人。", "categories": [ "data", "business" ], "tags": [ "Excel", "库存管理", "进销存", "安全库存", "ABC分析", "周转率" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "库存进销存管理表" } ], "prompt": "帮我生成一份专业的库存进销存管理 Excel（output.xlsx）。\n\n业务背景：一家中型电商公司，经营 3C 数码、家居日用、食品饮料三大品类，3 个仓库（华东/华南/华北），管理 100+ SKU。请生成一份合理的模拟数据（2026 年 1-3 月）。\n\n包含以下 Sheet：商品主数据、入库记录、出库记录、库存总览（含安全库存线与库存状态自动判断）、ABC 分析、周转率分析。代表性字段如 SKU 编码、品类、仓库、数量、单价、库存状态等。\n\n格式：专业商务风格，标题行深色背景，金额带千分位，库存状态用条件格式标色，冻结首行 + 自动筛选。", "prompt_en": "Generate a professional inventory management Excel workbook (output.xlsx).\n\nBackground: a mid-sized e-commerce company with three categories (3C/electronics, home goods, food & beverage), three warehouses (East/South/North China), and 100+ SKUs. Generate a reasonable set of mock data (Jan–Mar 2026).\n\nInclude these sheets: product master data, inbound records, outbound records, stock overview (with safety-stock line and auto-judged stock status), ABC analysis, turnover analysis. Representative fields such as SKU code, category, warehouse, quantity, unit price, stock status.\n\nStyle: professional business look, dark header row, thousands-separated amounts, conditional-format coloring for stock status, freeze top row + auto-filter.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 225, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Inventory Management Tracker", "subtitle_en": "Complete stock-in/out tracking with smart inventory alerts", "description_en": "The biggest fear of manual inventory tracking? Missing inbound entries, wrong outbound records, stocktake numbers that don't match — at month-end reconciliation, system counts are way off from physical counts, and figuring out which step went wrong means digging through piles of paper receipts.\n\nThis case uses AI to generate a complete inventory management Excel in one click: covering 100+ SKUs across 3 warehouses with 3 months of inbound/outbound transaction records, automatic real-time stock calculation, safety stock alerts, inventory turnover rates, and ABC classification for product value segmentation. Open it and instantly see which products need restocking, which are overstocked, and which categories turn over fastest.\n\nBest for: Warehouse managers, e-commerce operators, supply chain teams, SMB owners, anyone who needs precise inventory control.", "tags_en": [ "Excel", "Inventory Management", "Stock Control", "Safety Stock", "ABC Analysis", "Turnover Rate" ], "is_recommended": false, "recommended_order": 0 }, { "id": "meeting-agenda-planner", "title": "会议议程与时间规划", "title_en": "Meeting Agenda & Time Planner", "subtitle": "2 小时需求评审会，6 个议题完整议程 Word 文档一键生成", "subtitle_en": "2-hour requirements review meeting — 6 agenda items as Word document", "description": "开会最怕没准备：议题不清楚、时间分配随意、该谁发言没人知道，两小时开完发现什么都没定下来——更别提会后没有行动项跟进，等于白开。\n\n这个案例演示了如何用 Word 文档生成技能生成一份 v3.5 版本需求评审会的完整议程 Word 文档。覆盖 6 个议题（复盘→需求评审→技术评估→排期→风险→行动项），每个议题明确时间分配、负责人、讨论内容、预期产出和会前准备材料。包含 12 个待评审需求的优先级矩阵、里程碑时间表模板、风险登记表和行动项清单——格式规范可直接发送。\n\n适合：产品经理、项目经理、技术 Leader、Scrum Master、任何需要组织高效会议的人。", "description_en": "Meetings without preparation are meetings wasted: unclear topics, random time allocation, nobody knows who speaks when — two hours later nothing is decided, and without action items there's no follow-through.\n\nThis case shows how the Word Document Generation skill generates a complete agenda Word document for a v3.5 requirements review meeting with 6 agenda items, each with clear time allocation, owners, discussion points, expected outputs, and pre-meeting preparation materials.\n\nBest for: Product managers, project managers, tech leads, Scrum Masters, anyone who needs to run efficient meetings.", "categories": [ "office" ], "tags": [ "会议议程", "需求评审", "时间管理", "项目管理", "办公效率" ], "tags_en": [ "Meeting Agenda", "Requirements Review", "Time Management", "Project Management", "Office Productivity" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "v3.5 需求评审会完整议程（Word）" } ], "prompt": "产品部要召开 v3.5 版本的需求评审会议，预计 2 小时。请生成完整的会议议程 Word 文档（output.docx）。\n\n参会人员：产品经理（主持人）、前端负责人、后端负责人、设计师、QA 负责人、运营代表。\n\n议题包括：\n1）v3.4 上线复盘（15分钟）\n2）v3.5 需求池 Review，共 12 个需求待评审（45分钟）\n3）技术可行性讨论（20分钟）\n4）排期与里程碑确认（15分钟）\n5）风险识别与应对（10分钟）\n6）Q&A 与行动项确认（15分钟）\n\n需要生成完整的会议议程文档，包含每个议题的时间分配、负责人、讨论内容、预期产出和所需准备材料。\n\n文档格式要求：\n- 包含封面页、页眉页脚、多级标题\n- 格式专业清晰，适合直接分享\n- 输出为正式 Word 文档（output.docx）", "prompt_en": "The product department is holding a requirements review meeting for version v3.5, expected to last 2 hours. Please generate a complete meeting agenda as a Word document (output.docx).\n\nAttendees: Product Manager (host), Front-end Lead, Back-end Lead, Designer, QA Lead, Operations Representative.\n\nTopics include:\n1) v3.4 launch retrospective (15 minutes)\n2) v3.5 backlog review, 12 requirements to be reviewed (45 minutes)\n3) Technical feasibility discussion (20 minutes)\n4) Schedule and milestone confirmation (15 minutes)\n5) Risk identification and mitigation (10 minutes)\n6) Q&A and action-item confirmation (15 minutes)\n\nI need a complete meeting agenda document that includes, for each topic, the time allocation, the owner, the discussion content, the expected deliverables, and the materials to prepare.\n\nDocument formatting requirements:\n- Include a cover page, headers and footers, and multi-level headings\n- Professional and clear formatting, suitable for sharing directly\n- Output as a formal Word document (output.docx)", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 226, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "official", "usage_count": 0, "created_at": "2026-04-21T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "openspec-doc-clinical-t2dm-guideline", "title": "糖尿病基层诊疗规范", "title_en": "Primary Care T2DM Guideline", "subtitle": "8 章完整规范·跨章审计闭环·基层临床送审初稿", "subtitle_en": "An 8-chapter primary-care guideline with cross-chapter audit closure and review-ready delivery", "description": "基层临床长文档最怕的不是篇幅长，而是关键口径前后打架：诊断阈值一章一个版本、SGLT-2i 的 eGFR 下限写乱、低血糖分级把 2 级和 3 级混在一起、老年 HbA1c 目标前后漂移。对于 2 型糖尿病这种横跨诊断、药物、随访、并发症筛查、特殊人群和转诊的规范文件，只要一处口径失真，就可能直接变成医疗安全风险。\n\n这个案例基于 openspec-doc-team 专家团的真实协作产物，围绕《2 型糖尿病基层诊疗规范》完成 8 章长文档交付，并保留了项目参数卡与跨章审核报告。完整稿整合了国家基层糖尿病防治管理指南（2025）、CDS 2024、医保目录 2024、基药目录 2018 与基本公卫第三版的核心口径，覆盖疾病定义、诊断标准、药物治疗路径、非药物干预、随访与并发症筛查、特殊人群管理、双向转诊、附录工具与医嘱模板；同时通过跨章审计修正了 SGLT-2i eGFR 阈值、老年 HbA1c 分层、低血糖三级分类、CKD 用药矩阵和 16.7 mmol/L 紧急阈值等高风险问题。\n\n适合：社区医院和乡镇卫生院的全科医师、家庭医生团队、基层慢病管理负责人、医疗内容产品经理、临床规范编辑，以及想把 AI 用作“规范初稿 + 审核闭环”助手的医疗机构。", "description_en": "The hardest part of writing long clinical documents is not the length — it is keeping every high-risk threshold consistent. One chapter uses one diagnostic cutoff, another drifts on SGLT-2 inhibitor eGFR thresholds, hypoglycemia levels get mixed up, and elderly HbA1c targets start contradicting each other. In a T2DM guideline that spans diagnosis, drugs, follow-up, complications, special populations, and referrals, a single drift can become a patient-safety issue.\n\nThis case packages a real multi-agent delivery produced by the openspec-doc-team around a full Primary Care T2DM Guideline, with the final 8-chapter document, the project card, and the cross-chapter audit record preserved together. The deliverable consolidates the 2025 National Primary Diabetes Prevention and Management Guideline, CDS 2024, the 2024 NRDL, the 2018 Essential Medicines List, and the National Basic Public Health Service Specification. It covers disease definition, diagnosis, drug pathways, non-drug interventions, follow-up and complication screening, special populations, referral rules, and appendix tools/templates, while explicitly fixing high-risk inconsistencies around SGLT-2i eGFR thresholds, elderly HbA1c stratification, hypoglycemia classification, CKD drug adjustments, and the 16.7 mmol/L emergency threshold.\n\nBest for community hospitals, township clinics, primary-care physicians, family doctor teams, chronic disease program leads, medical content product managers, and clinical guideline editors who want AI to function as a draft-plus-audit-loop assistant rather than a free-form text generator.", "categories": [ "creation", "business" ], "tags": [ "临床规范", "糖尿病", "基层医疗", "专家团协作", "长文档" ], "tags_en": [ "Clinical Guideline", "Diabetes", "Primary Care", "Expert Team", "Long Document" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "2 型糖尿病基层诊疗规范完整稿" }, { "file": "PROJECT_CARD.md", "label": "项目参数卡（跨章一致性锚点）" }, { "file": "cross_chapter_audit.md", "label": "跨章一致性审核报告" } ], "prompt": "请基于权威指南与目录，整理一份《2 型糖尿病基层诊疗规范》，面向社区医院、乡镇卫生院、基层全科与家庭医生团队。\n\n交付完整 Markdown 长文档，至少 8 章：疾病定义与流行病学、诊断与分层管理、药物治疗路径、非药物干预、随访与并发症筛查、特殊人群管理、双向转诊、附录。\n\n要求：\n- 所有药物剂量、诊断阈值、分类与条款均须有指南出处，不确定处保留 [待核对] 等占位符\n- 重点统一跨章关键临床口径\n- 文末附「发布前核对清单」与「复核说明」\n- 适合作为基层规范化诊疗初稿与内部审核材料\n\n用「专业文档生成团队」专家团协作产出。", "prompt_en": "Based on authoritative guidelines and catalogs, please compile a \"Primary-Care Diagnosis and Treatment Protocol for Type 2 Diabetes\" for community hospitals, township health centers, primary-care general practitioners, and family doctor teams.\n\nDeliver a complete long-form Markdown document with at least 8 chapters: disease definition and epidemiology, diagnosis and stratified management, pharmacological treatment pathways, non-pharmacological interventions, follow-up and complication screening, special-population management, two-way referral, and appendices.\n\nRequirements:\n- All drug dosages, diagnostic thresholds, classifications, and provisions must have a guideline source; where uncertain, keep placeholders such as [to be verified]\n- Focus on unifying key cross-chapter clinical parameters\n- Attach a \"pre-publication checklist\" and a \"review note\" at the end\n- Suitable as an initial draft of a standardized protocol and internal review material for primary care\n\nProduce it collaboratively with the \"Professional Document Generation Team\" expert team.", "skills": [], "experts": [ { "id": "OpenSpecDocTeam", "name": "专业文档生成团队", "description": "企业级长文档生成，4角色协作完成深度调研、大纲规划、内容撰写与合规审核全流程", "expertType": "team", "source": "expert-marketplace", "promptFileSnapshot": "b9f926245b20", "name_en": "Professional Document Generation Team", "description_en": "Enterprise long document generation: 4 roles collaborate on deep research, outline planning, content writing, and compliance review", "expert_id": "ex_EqCLDSDp4c5O" } ], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 227, "is_featured": false, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-26T16:24:04Z", "updated_at": "2026-06-16T23:30:00Z", "featured_order": 0, "is_recommended": false, "recommended_order": 0 }, { "id": "brainstorm-content-strategy", "title": "内容营销策略头脑风暴", "title_en": "Content Marketing Strategy Brainstorm", "subtitle": "从品牌定位到选题日历，一次头脑风暴搞定内容规划", "subtitle_en": "From brand positioning to content calendar — one brainstorm session for complete content planning", "description": "内容团队每周选题会都在「写什么」上纠结半天，没有系统化的内容策略，导致产出零散、效果难衡量。\n\n本案例演示如何用营销技能库为一家 SaaS 创业公司做完整内容营销规划——从目标人群画像到内容支柱、选题库、发布日历，一次会话输出可直接落地的策略方案。\n\n适用于内容运营、市场总监、自媒体创业者、品牌经理等需要系统化内容规划的角色。", "description_en": "Content teams waste hours each week debating 'what to write' without a systematic strategy, leading to scattered output and unmeasurable results.\n\nThis case demonstrates using the marketing skills suite to build a complete content marketing plan for a SaaS startup — from target personas to content pillars, topic libraries, and publishing calendars in a single session.\n\nIdeal for content operations, marketing directors, self-media entrepreneurs, and brand managers who need systematic content planning.", "categories": [ "business", "productivity" ], "tags": [ "内容营销", "选题策划", "品牌策略", "自媒体", "内容日历" ], "tags_en": [ "Content Marketing", "Topic Planning", "Brand Strategy", "Self-media", "Content Calendar" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "内容营销策略方案" } ], "prompt": "我是一家 SaaS 创业公司的内容负责人，请帮我做一次系统化的内容营销策略规划。包括：目标人群画像分析、3-5 个内容支柱方向、每个方向下的选题库（至少 10 个选题）、月度内容日历模板、内容分发渠道策略。我们的产品是面向中小企业的项目管理工具。", "prompt_en": "I'm the head of content at a SaaS startup. Please help me do a systematic content marketing strategy plan. It should include: target audience persona analysis, 3-5 content pillar directions, a topic library under each direction (at least 10 topics each), a monthly content calendar template, and a content distribution channel strategy. Our product is a project management tool aimed at small and medium-sized businesses.", "skills": [ { "id": "marketing-skills", "name": "营销技能库", "description": "23个营销模块：CRO、SEO、文案、分析、投放、产品发布、社交媒体，含检查清单和即用模板", "description_en": "23 marketing playbooks: CRO, SEO, copywriting, ads, social media — checklists and ready-to-use templates", "skill_id": "skill_2053084540171059200" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "is_featured": false, "sort_order": 228, "featured_order": 0, "quality_score": 92, "usage_count": 0, "created_at": "2026-04-23", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "unit-test-generation", "title": "TypeScript 单元测试自动生成", "title_en": "TypeScript Unit Test Auto-Generation", "subtitle": "购物车模块一键生成 Jest 测试，覆盖率 > 90%", "subtitle_en": "One-click Jest test generation for a shopping cart module — 90%+ coverage", "description": "手动写单元测试费时且容易遗漏边界情况？用全栈开发技能为 TypeScript 购物车模块自动生成完整的 Jest 单元测试。\n\n源代码包含 CartService 类（含 addItem、removeItem、updateQuantity、getTotal、applyCoupon、checkout 六个核心方法）、CartItem 接口与 CouponType 枚举。生成的测试覆盖所有公共方法的正常路径、边界条件与异常场景，使用 describe/it 组织结构，mock 外部支付网关与库存服务依赖，并通过 beforeEach 管理测试状态。\n\n适合：TypeScript 开发者、前端工程师、全栈开发者。", "description_en": "Writing unit tests manually is time-consuming and often misses edge cases. Use the fullstack-dev skill to auto-generate comprehensive Jest unit tests for a TypeScript shopping cart module.\n\nThe source includes a CartService class with six core methods, CartItem interface, and CouponType enum. Generated tests cover happy paths, boundary conditions, and error scenarios using describe/it blocks, mock external dependencies, and manage state via beforeEach.\n\nPerfect for TypeScript developers, frontend engineers, and fullstack devs.", "categories": [ "dev" ], "tags": [ "TypeScript", "单元测试", "Jest", "TDD", "测试驱动" ], "tags_en": [ "TypeScript", "Unit Test", "Jest", "TDD", "Test-Driven" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "购物车模块单元测试代码" } ], "prompt": "请为以下 TypeScript 购物车模块生成完整的 Jest 单元测试。源代码包含：CartService 类（addItem、removeItem、updateQuantity、getTotal、applyCoupon、checkout 方法），CartItem 接口，CouponType 枚举。要求：1) 覆盖所有公共方法的正常和边界情况；2) 使用 describe/it 组织测试结构；3) 包含 mock 外部依赖（支付网关、库存服务）；4) 测试异常场景（库存不足、优惠券过期、支付失败）；5) 使用 beforeEach 管理测试状态；6) 覆盖率目标 > 90%。请用全栈开发技能输出包含完整测试代码的 Markdown 文档。", "prompt_en": "Please generate complete Jest unit tests for the following TypeScript shopping cart module. The source code includes: a CartService class (addItem, removeItem, updateQuantity, getTotal, applyCoupon, checkout methods), a CartItem interface, and a CouponType enum. Requirements: 1) cover the normal and edge cases of all public methods; 2) use describe/it to organize the test structure; 3) include mocking of external dependencies (payment gateway, inventory service); 4) test exception scenarios (insufficient inventory, expired coupon, payment failure); 5) use beforeEach to manage test state; 6) coverage target > 90%. Please use the full-stack development skill to output a Markdown document containing the complete test code.", "skills": [ { "id": "fullstack-dev", "name": "全栈开发", "description": "全栈应用架构与开发指南", "description_en": "Full-stack architecture and development guide", "skill_id": "skill_2053082011994116096" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 229, "is_featured": false, "featured_order": 0, "quality_score": 88, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-interest-test", "title": "兴趣测评：霍兰德 SEC 型", "title_en": "Interests: Holland Code SEC", "subtitle": "六维得分落到具体岗位方向，顺风逆风一目了然", "subtitle_en": "Six dimension scores mapped to concrete role families", "description": "霍兰德测完给个三字母代码就完了？代码要能落到岗位上才有用。\n\n本案例跑完 60 题霍兰德测评：六维得分 R1/I2/A3/S8/E8/C7，代码 SEC。报告不止给代码——三高三低的形态说明动手操作和独立钻研方向可直接排除；对照到体制内，综合管理与教育服务类顺风、技术操作与纯研究类逆风，E 与 C 并存的张力也给出了解释框架。\n\n适合：纠结考编方向的学生、给学生做生涯解读的老师。", "description_en": "A Holland test that ends at a three-letter code isn't much use — the code has to land on actual roles.\n\nThis case completes the 60-item inventory: dimension scores R1/I2/A3/S8/E8/C7, code SEC. The report goes beyond the code — the three-high/three-low shape rules out hands-on and solo-research directions outright; mapped to the public sector, general administration and education/service roles are tailwinds while technical operations and pure research are headwinds, with a framework for the E-versus-C tension.\n\nFor: students weighing public-sector directions and advisors interpreting results.", "categories": [ "education" ], "tags": [ "霍兰德", "职业兴趣", "RIASEC", "测评" ], "tags_en": [ "Holland", "career interests", "RIASEC", "assessment" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "霍兰德 SEC 型报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做 60 题霍兰德职业兴趣测评并出报告。要求：1) 说明三段题目结构后渲染作答；2) 我的作答序列（题号 1-60，Y/N）：NNNNNNNYNNNYNNYYYYNYYYNYYYNYYYYNNYNNNNNYYYNYYNYYNNNNYNYNYYNY；3) 用评分脚本算出六维得分与兴趣代码；4) 报告含六维条形图、优势维度的典型职业原文，并对照体制内岗位说明顺风/逆风方向，标注哪些是解读。", "prompt_en": "Run the 60-item Holland inventory and produce a report. Requirements: 1) explain the three-part structure, then render for answering; 2) my response sequence (items 1-60, Y/N): NNNNNNNYNNNYNNYYYYNYYYNYYYNYYYYNNYNNNNNYYYNYYNYYNNNNYNYNYYNY; 3) compute the six scores and interest code with the scoring script; 4) include a bar chart, verbatim typical occupations for the dominant dimensions, and tailwind/headwind mapping to public-sector roles with interpretation clearly labelled.", "skills": [ { "id": "holland-career-assessment", "name": "职业兴趣测评", "description": "60 题评估 RIASEC 六类职业兴趣，生成兴趣代码与解读", "skill_id": "skill_2089338086439067648" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 230, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "travel-expense-split", "title": "旅行 AA 分账表", "title_en": "Travel Expense Splitting Sheet", "subtitle": "输入旅行花费明细，自动计算每人应付金额与转账清算", "subtitle_en": "Input travel expenses, auto-calculate each person's share and settlement transfers", "description": "朋友一起旅行最尴尬的时刻就是算账——谁垫付了酒店、谁请了那顿饭、有人提前走了两天该怎么算？微信群里丢一堆截图和数字，越算越乱，最后要么有人吃亏要么伤感情。\n\n这个案例生成完整的旅行 AA 分账 Excel：费用明细表（每笔消费的付款人/参与人/金额）、个人汇总表（每人应付总额/已付总额/差额）、最优转账方案（最少转账次数完成清算）。公式自动计算分摊，支持部分人参与的费用按实际人头平摊。\n\n适合：朋友出游、公司团建、合租室友分摊、同事聚餐。", "description_en": "The most awkward moment of a group trip is splitting the bill — who fronted the hotel, who paid for that dinner, how to calculate when someone left two days early? Screenshots and numbers flying around in group chats just make it worse, and someone always ends up short-changed or resentful.\n\nThis case generates a complete travel expense splitting Excel: expense detail sheet (payer/participants/amount per transaction), personal summary (total owed/total paid/balance per person), and optimal settlement plan (minimum transfers to clear all debts). Formulas auto-calculate splits, supporting partial participation expenses divided by actual headcount.\n\nBest for: Friend trips, company team-building, roommate expense sharing, colleague dinner splits.", "categories": [ "travel", "data" ], "tags": [ "AA分账", "旅行费用", "Excel", "分摊计算", "生活工具" ], "tags_en": [ "Expense Splitting", "Travel Expenses", "Excel", "Cost Sharing", "Life Tools" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "prompt": "请为一次 5 人成都旅行生成一份完整的 AA 分账 Excel 表格。\n\n旅行信息：成员小明、小红、阿杰、Lisa、大伟；目的地成都；2026 年 5 月 1–4 日（4 天 3 晚）；大伟 5 月 3 日下午提前离开，不参与最后一天费用。\n\n请生成以下工作表：\n1. 费用明细表：代表性字段如日期、项目、金额、付款人、参与人、人均分摊、备注；包含 20+ 笔消费，部分消费非全员参与（按实际人头平摊）。\n2. 个人汇总表：用公式汇总每人应付总额、已垫付总额、差额。\n3. 清算方案：用最少转账次数完成所有人之间的清算，含转账人、收款人、金额。\n\n格式：金额保留两位小数，差额用条件格式标色。输出为 .xlsx 文件。", "prompt_en": "Please generate a complete AA (split-the-bill) expense Excel sheet for a 5-person trip to Chengdu.\n\nTrip info: members Xiao Ming, Xiao Hong, A Jie, Lisa, Da Wei; destination Chengdu; May 1-4, 2026 (4 days, 3 nights); Da Wei leaves early on the afternoon of May 3 and does not share the last day's expenses.\n\nPlease generate the following worksheets:\n1. Expense Detail Sheet: representative fields like date, item, amount, payer, participants, per-person share, notes; include 20+ expenses, with some not involving everyone (split by actual headcount).\n2. Personal Summary Sheet: use formulas to summarize each person's total owed, total fronted, and balance.\n3. Settlement Plan: clear all debts with the minimum number of transfers, including payer, payee, and amount.\n\nFormatting: amounts to 2 decimal places; use conditional formatting to color the balance. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 230, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.xlsx", "label": "旅行 AA 分账表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "insp-holland-ria", "title": "霍兰德：RIA 型，动手、研究、创意三条线同时亮起", "title_en": "Holland Codes: RIA — Hands-On, Investigative and Creative", "subtitle": "60 题分三部分问「感兴趣 / 擅长 / 喜欢这个职业」，六维分别计分", "subtitle_en": "Sixty items across three sections, scored on six dimensions", "description": "兴趣测评最容易被当成算命——真正有用的读法是看六个维度的分布形状，尤其是得分为 0 的那几个。\n\n本案例用 60 题版跑出 RIA 型，兴趣显著度 46 分。六维分布极度两极化：R/I/A 拿下 10/9/8 分，而 E（企业型）和 C（常规型）双双为 0。案例给出题库原文的维度特征与典型职业，并强调兴趣不等于适合。\n\n适合：不确定自己喜欢什么类型工作的学生、做专业与职业匹配的人。", "description_en": "Interest inventories are easily mistaken for fortune-telling — the useful reading is the shape of the six-dimension distribution, especially the zeros.\n\nThis case runs the 60-item version to an RIA result with 46% interest salience. The distribution is sharply polarised: R/I/A take 10/9/8 while enterprising and conventional both score zero. It reproduces the item bank's dimension characteristics and typical occupations, and stresses that interest is not the same as fit.\n\nFor: students unsure what kind of work appeals, and anyone matching major to career.", "categories": [ "education" ], "tags": [ "霍兰德", "职业兴趣", "RIA", "兴趣分布" ], "tags_en": [ "Holland codes", "vocational interest", "RIA", "interest profile" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "霍兰德 RIA 型测评结果报告", "label_en": "Holland RIA assessment report" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次霍兰德职业兴趣测评。我作答时对动手操作类、研究探索类、创意表达类的活动都答「是」，对销售管理类和事务性工作基本答「否」。要求：1) 用 60 题题库真实跑一遍评分脚本，给出霍兰德代码和总分；2) 展示六个维度的得分分布；3) 完整给出代码对应三个维度的共同特征和典型职业（题库原文，不要改写）；4) 指出得分最低的几个维度是什么，说明低分维度同样重要的原因；5) 解释总分是怎么算出来的、它代表兴趣的广度还是集中度，以及为什么兴趣不等于适合。", "prompt_en": "Run a Holland vocational-interest assessment. I answer yes to hands-on, investigative and creative activities, and largely no to sales, management and clerical work. Requirements: 1) run the official 60-item scoring script and give the Holland code and interest salience score; 2) show the six-dimension score distribution; 3) reproduce the item bank's characteristics and typical occupations for the three code dimensions verbatim; 4) identify the lowest-scoring dimensions and explain why low scores matter too; 5) explain what the salience score means and why interest is not the same as fit.", "skills": [ { "id": "holland-career-assessment", "name": "职业兴趣测评", "description": "60 题三段式测评，输出六维兴趣得分与霍兰德代码", "skill_id": "skill_2089338086439067648" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 231, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "meeting-notes-doc", "title": "会议纪要自动整理", "title_en": "Auto Meeting Notes Organizer", "subtitle": "粘贴讨论记录，秒变结构化纪要 Word 文档", "subtitle_en": "Paste meeting transcript, get structured minutes as Word document instantly", "description": "开完会还在手动整理纪要？把讨论记录丢给 AI，自动按「议题→讨论→决议→待办」结构输出为正式 Word 文档，方案对比生成表格，待办自动标注负责人和优先级。\n\n本案例演示了一场产品评审会的完整纪要 Word 文档生成：三项需求的讨论要点、设计方案决策、优先级排序，以及带截止日期的待办清单，一键搞定。\n\n适合产品经理、项目经理、团队 Leader，以及所有被会议纪要折磨过的打工人。", "description_en": "Still writing meeting notes by hand after every meeting? Paste the raw discussion into AI and get structured minutes as a formal Word document instantly — organized by agenda, discussion, decisions, and action items, with comparison tables and prioritized to-dos.\n\nThis case demonstrates a full product review meeting: three feature discussions, design decisions, priority rankings, and a complete action item list with owners and deadlines — all generated as a Word document in one shot.\n\nPerfect for product managers, project managers, team leads, and anyone who's ever dreaded writing up meeting notes.", "categories": [ "office", "productivity" ], "tags": [ "会议纪要", "办公效率", "结构化文档", "自动整理" ], "tags_en": [ "Meeting Notes", "Productivity", "Structured Docs", "Auto-organize" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "结构化会议纪要（Word）" } ], "prompt": "请根据以下会议讨论记录，生成一份结构化的会议纪要 Word 文档（output.docx）。\n\n会议主题 / 参会人 / 时间：（填写）\n\n讨论记录（粘贴你的会议逐字记录或讨论要点）：\n...\n\n文档要求：\n- 会议摘要（3 句话概括）\n- 按议题逐条整理讨论要点与结论\n- 待办事项表格（负责人、截止日期、优先级）\n- 下次会议安排\n- 含封面页、目录、页眉页脚\n- 输出为正式 Word 文档（output.docx）", "prompt_en": "Based on the following meeting discussion notes, generate a structured meeting-minutes Word document (output.docx).\n\nMeeting topic / attendees / time: (fill in)\n\nDiscussion notes (paste your meeting transcript or discussion points):\n...\n\nDocument requirements:\n- A meeting summary (3 sentences)\n- Discussion points and conclusions organized by agenda item\n- An action-item table (owner, deadline, priority)\n- Next meeting arrangements\n- Include a cover page, table of contents, headers and footers\n- Output as a formal Word document (output.docx)", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 231, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-21T14:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "brainstorm-product-feature", "title": "产品功能头脑风暴", "title_en": "Product Feature Brainstorming", "subtitle": "AI 辅助需求拆解与优先级排序，从混乱到清晰", "subtitle_en": "AI-powered requirement breakdown and prioritization", "description": "产品需求评审会上七嘴八舌，讨论了两小时还是一团乱麻，需求多但资源有限不知道先做什么？用 FBS-BookWriter 长文写作能力从一个模糊的产品方向出发，系统化发散 50+ 功能点，用 RICE 模型打分排序，输出一份结构化的需求池文档。\n\n本案例以「智能健身 App」为产品方向，演示了 AI 如何从核心训练、饮食管理、社交互动、数据分析、商业化五个维度发散功能点，用 RICE 框架（Reach/Impact/Confidence/Effort）进行优先级排序，划分 P0-P3 四个层级，推荐 MVP 功能组合，并识别关键技术风险与外部依赖。\n\n适合：产品经理、创业者、项目负责人、技术 Leader。", "description_en": "Product review meetings go nowhere after two hours of debate — too many ideas, limited resources, and no clear priorities. Use FBS-BookWriter's long-form writing capability to systematically brainstorm 50+ feature ideas from a vague product direction, score them with the RICE model, and output a structured requirements backlog.\n\nThis case uses a \"Smart Fitness App\" as the product direction, demonstrating how AI brainstorms features across five dimensions (core training, diet management, social interaction, data analytics, monetization), prioritizes with RICE scoring, classifies into P0-P3 tiers, recommends an MVP feature set, and identifies key technical risks and dependencies.\n\nBest for: Product managers, founders, project leads, tech leads.", "categories": [ "business", "office" ], "tags": [ "产品需求", "头脑风暴", "优先级排序", "RICE模型", "需求分析" ], "tags_en": [ "Product Requirements", "Brainstorming", "Prioritization", "RICE Model", "Requirements Analysis" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "功能需求池与优先级排序" } ], "prompt": "我正在规划一款「智能健身 App」产品，定位是 AI 私教 + 社交 + 健康管理。请用 FBS-BookWriter 的长文写作能力，帮我做一次全面的产品功能头脑风暴，输出结构化 Markdown 文档（output.md，含表格与分级列表）。\n\n1. 功能发散：从核心训练、饮食管理、社交互动、数据分析、商业化五个维度，发散出 50+ 功能点，每个含简要描述\n2. RICE 评分：筛选 Top 20 功能，用 RICE 框架打分\n3. 优先级排列：按得分分为 P0-P3 四个层级，说明分级理由\n4. MVP 推荐：推荐 8-10 个功能组成 MVP，给出理由与大致开发周期\n5. 风险与依赖：列出关键技术风险、外部依赖与潜在阻塞点", "prompt_en": "I'm planning a \"Smart Fitness App,\" positioned as AI personal trainer + social + health management. Please use FBS-BookWriter's long-form writing capability to run a comprehensive product feature brainstorm, output as a structured Markdown document (output.md, with tables and tiered lists).\n\n1. Feature brainstorming: across five dimensions—core training, diet management, social interaction, data analytics, and monetization—brainstorm 50+ feature ideas, each with a brief description\n2. RICE scoring: select the Top 20 features and score them with the RICE framework\n3. Prioritization: sort features into four tiers P0-P3 by score, explaining the rationale\n4. MVP recommendation: recommend 8-10 features to form the MVP, with rationale and a rough development timeline\n5. Risks and dependencies: list key technical risks, external dependencies, and potential blockers", "skills": [ { "id": "fbs-bookwriter", "name": "fbs-bookwriter", "description": "福帮手出品 | 高质量长文档手稿工具链：书籍、手册、白皮书、行业指南、长篇报道、深度专题；支持联网查证（宿主允许时启用，离线自动降级）、S/P/C/B 分层审校、中文排版与 MD/HTML 交付。触发词：福帮手、福帮手写书skill、福帮手写书、写书、出书、写长篇、写手册、写白皮书、写行业指南、协作写书、定大纲、写章节、封面、插图、排版构建、导出、去AI味、质量自检、图文书、写报道、写深度稿、写特稿、写专题、写调查报道、写长文、拆书改写、海外本地化改写、爆款结构改写、激活原料、原料盘点、整理素材", "description_en": "Dual-channel long-form writing workflow for books, manuals, whitepapers, guides and reports with layered QC, 7-lock anti-drift anchors, and offline fallback.", "skill_id": "skill_2053083113738002432" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 232, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-holland-sea", "title": "霍兰德：SEA 型，兴趣落在「人」这一侧", "title_en": "Holland Codes: SEA — Interests on the People Side", "subtitle": "换一组作答，代码从 RIA 移到 SEA，几乎处于六边形的对角", "subtitle_en": "A different answer set moves the code from RIA to SEA, nearly diagonal on the hexagon", "description": "霍兰德六边形模型的核心分野是「和物打交道」还是「和人打交道」，但光看一个结果，感受不到这条线在哪。\n\n本案例用同一套 60 题跑出 SEA 型，与 RIA 那份并排看：两者在六边形上几乎处于对角位置，共享的只有 A（艺术型）。案例说明相邻三角代码为什么比跨边代码更稳定，并给出各维度的题库原文释义。\n\n适合：偏好与人打交道的学生、想理解霍兰德模型结构的人。", "description_en": "The core divide in Holland's hexagon is working with things versus working with people — but a single result doesn't show you where that line falls.\n\nThis case runs the same 60 items to an SEA result and sets it beside RIA: the two sit almost diagonally opposite on the hexagon, sharing only A. It explains why adjacent-triangle codes are more stable than codes spanning the hexagon, with each dimension's definition from the item bank.\n\nFor: students drawn to people-facing work and anyone learning the model's structure.", "categories": [ "education" ], "tags": [ "霍兰德", "职业兴趣", "SEA", "六边形模型" ], "tags_en": [ "Holland codes", "vocational interest", "SEA", "hexagon model" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "霍兰德 SEA 型测评结果报告", "label_en": "Holland SEA assessment report" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次霍兰德职业兴趣测评。我作答时对助人服务类、组织说服类、创意表达类的活动答「是」，对机械操作类和事务性工作基本答「否」。要求：1) 用 60 题题库真实跑一遍评分脚本，给出霍兰德代码和总分；2) 展示六个维度的得分分布；3) 完整给出代码三个维度的共同特征和典型职业（题库原文）；4) 与 RIA 型的结果做对比，说明两者在六边形上的位置关系、共享哪个维度；5) 解释相邻三角代码与跨边代码在稳定性上的区别。", "prompt_en": "Run a Holland vocational-interest assessment. I answer yes to helping, persuading and creative activities, and largely no to mechanical and clerical work. Requirements: 1) run the official 60-item scoring script and give the Holland code and salience score; 2) show the six-dimension distribution; 3) reproduce the item bank's characteristics and typical occupations for the three code dimensions; 4) compare with an RIA result, describing their positions on the hexagon and which dimension they share; 5) explain how adjacent-triangle codes differ in stability from codes spanning the hexagon.", "skills": [ { "id": "holland-career-assessment", "name": "职业兴趣测评", "description": "60 题三段式测评，输出六维兴趣得分与霍兰德代码", "skill_id": "skill_2089338086439067648" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 232, "is_featured": false, "featured_order": 0, "quality_score": 90, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "wordbox-app-delivery", "title": "单词盒子 App 从0到1完整交付", "title_en": "Wordbox App End-to-End Delivery", "subtitle": "React Native + AI 对话学单词，PRD到代码测试全链路", "subtitle_en": "React Native + AI Chat Vocabulary App, PRD to Code & Tests", "description": "独立开发者和产品团队常常卡在「有想法但不知从何下手」——PRD怎么写、架构怎么定、代码怎么起步。本案例演示 software-company 专家团（PM/架构师/工程师/QA/TL）如何接过一句话需求，一次性产出完整的 PRD、系统架构、任务拆解、React Native 工程代码（含艾宾浩斯 SM-2 算法、AI 对话抽象层、本地存储）、59个测试用例（覆盖率 90%+）。适合想学习标准软件交付流程、或需要快速搭建 MVP 的产品经理、创业者和开发者。", "description_en": "Indie developers and product teams often get stuck at 'I have an idea but don't know where to start' — how to write a PRD, define architecture, kickstart code. This case demonstrates how the software-company expert team (PM/Architect/Engineer/QA/TL) takes a one-sentence requirement and produces a complete PRD, system architecture, task breakdown, React Native engineering code (with Ebbinghaus SM-2 algorithm, AI chat abstraction, local storage), and 59 test cases (90%+ coverage). Ideal for PMs, founders, and developers who want to learn standard software delivery workflow or rapidly build an MVP.", "tags": [ "软件工程", "从0到1", "React Native", "AI应用", "完整交付" ], "tags_en": [ "Software Engineering", "Zero to One", "React Native", "AI App", "Full Delivery" ], "categories": [ "dev" ], "artifact_type": "code", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "DELIVERY_SUMMARY.md", "artifacts": [ { "file": "DELIVERY_SUMMARY.md", "label": "交付汇总说明" }, { "file": "PRD.md", "label": "产品需求文档" }, { "file": "ARCHITECTURE.md", "label": "系统架构设计" }, { "file": "TASK_BREAKDOWN.md", "label": "任务拆解清单" }, { "file": "TEST_REPORT.md", "label": "测试报告（59/59通过）" }, { "file": "src/", "label": "React Native 工程代码（16文件）" }, { "file": "tests/", "label": "单元与集成测试（7文件）" } ], "prompt": "我想做一款叫「单词盒子」的移动端 App（前端 React Native），目标用户是准备考研/出国的年轻人。核心诉求：每天 15 分钟通过 AI 对话学单词，支持艾宾浩斯复习曲线、生词本、每日打卡、打卡日历、学习统计。\n\n请完成从 0 到 1 的完整交付：\n1. 完整 PRD（目标、用户故事、竞品分析、功能清单与优先级、待确认问题）；\n2. 系统架构设计（技术选型、模块划分、核心数据结构、接口定义、时序图）；\n3. 任务拆解清单（按依赖排序、每个文件的逻辑分析、共享知识、API Spec）；\n4. 可运行的工程代码（前端页面 + 本地存储 + AI 对话接口对接层）；\n5. 完整的单元/集成测试用例及测试报告。\n\n以工程目录 + Markdown 文档形式交付，最终汇总一份项目交付说明。", "prompt_en": "I want to build a mobile app called \"Wordbox\" (frontend in React Native), with a target audience of young people preparing for graduate school entrance exams / studying abroad. Core need: learn vocabulary through AI conversation for 15 minutes a day, supporting the Ebbinghaus review curve, a vocabulary notebook, daily check-ins, a check-in calendar, and study statistics.\n\nPlease complete the full end-to-end delivery from 0 to 1:\n1. A complete PRD (goals, user stories, competitor analysis, feature list with priorities, open questions to confirm);\n2. System architecture design (technology choices, module breakdown, core data structures, interface definitions, sequence diagrams);\n3. A task breakdown list (sorted by dependency, logical analysis of each file, shared knowledge, API Spec);\n4. Runnable engineering code (frontend pages + local storage + an integration layer for the AI conversation interface);\n5. Complete unit/integration test cases and a test report.\n\nDeliver in the form of an engineering directory + Markdown documents, and finally summarize a project delivery summary.", "skills": [], "experts": [ { "id": "SoftwareCompany", "name": "软件开发团队", "description": "软件研发全流程协作，产品经理定义需求、架构师设计系统、项目经理协调进度、工程师实现代码、QA保障质量", "expertType": "team", "source": "expert-marketplace", "name_en": "Software Company", "description_en": "Software development collaboration: PM defines requirements, architect designs systems, PM coordinates schedules, engineers implement code, QA ensures quality", "expert_id": "ex_jPdPfIC04OKD" } ], "quality_score": 90, "source": "official", "sort_order": 233, "featured_order": 0, "usage_count": 0, "is_featured": false, "created_at": "2026-04-27T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "is_recommended": false, "recommended_order": 0 }, { "id": "user-review-sentiment", "title": "用户评价情感分析", "title_en": "User Review Sentiment Analysis", "subtitle": "电商产品用户评价情感分析与洞察报告", "subtitle_en": "E-commerce Product User Review Sentiment Analysis & Insights Report", "description": "电商运营最大的盲区之一，就是只看评分不看评价内容。4.5 分的产品可能隐藏着严重的售后问题，3.8 分的产品也许只是物流拖了后腿。手动翻几百条评价效率太低，关键信息容易遗漏，更别提跨产品维度的系统性分析了。\n\n这个案例基于 500 条真实用户评价数据，自动完成从情感分类到深度洞察的全链路分析：整体情感分布统计、五大产品维度（外观、性能、价格、售后、物流）的情感热力图、高频关键词词云与情感关联、负面评价 Top 10 问题归类、按月情感趋势变化追踪、竞品情感得分横向对比，最终输出改进建议优先级矩阵，帮你把用户的声音转化为可执行的产品决策。\n\n适合：电商运营、产品经理、品牌管理、用户研究员、市场分析师。无论是日常评价监控还是季度产品复盘，这份报告都能帮你从海量评价中快速提取核心洞察。", "description_en": "One of the biggest blind spots in e-commerce operations is looking only at ratings while ignoring review content. A 4.5-star product might be hiding serious after-sales issues, while a 3.8-star product might just be dragged down by logistics. Manually reading hundreds of reviews is inefficient, key insights get missed, and systematic cross-dimension analysis is nearly impossible.\n\nThis case analyzes 500 real user reviews through a complete pipeline from sentiment classification to deep insights: overall sentiment distribution, sentiment heatmap across five product dimensions (appearance, performance, price, after-sales, logistics), high-frequency keyword word cloud with sentiment correlation, Top 10 negative issue categorization, monthly sentiment trend tracking, competitor sentiment score benchmarking, and a final improvement priority matrix that turns user voices into actionable product decisions.\n\nBest for: E-commerce operators, product managers, brand managers, user researchers, and market analysts. Whether for daily review monitoring or quarterly product retrospectives, this report helps you extract core insights from massive review volumes quickly.", "categories": [ "search", "data" ], "tags": [ "情感分析", "用户评价", "电商数据", "NLP", "消费者洞察" ], "tags_en": [ "sentiment analysis", "user reviews", "e-commerce data", "NLP", "consumer insights" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "用户评价情感分析" } ], "prompt": "帮我做一份电商产品用户评价的情感分析报告。数据包含 500 条用户评价，需要分析：1）整体情感分布（正面/中性/负面比例）；2）按产品维度的情感热力图（外观、性能、价格、售后、物流）；3）高频关键词词云与情感关联；4）负面评价的 Top 10 问题归类；5）情感趋势变化（按月）；6）竞品对比情感得分；7）改进建议优先级矩阵。输出为结构化的 Markdown 分析报告。", "prompt_en": "Help me create a sentiment analysis report for e-commerce product user reviews. The data contains 500 user reviews, and I need to analyze: 1) the overall sentiment distribution (positive/neutral/negative proportions); 2) a sentiment heatmap by product dimension (appearance, performance, price, after-sales, logistics); 3) a word cloud of high-frequency keywords and their sentiment associations; 4) a Top 10 categorization of issues in negative reviews; 5) sentiment trend changes (by month); 6) a competitor comparison of sentiment scores; 7) a priority matrix for improvement recommendations. Output as a structured Markdown analysis report.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 234, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-email-writer", "title": "商务邮件一键生成", "subtitle": "输入关键信息，自动生成措辞得体、没有 AI 味的商务邮件", "description": "写商务邮件是每天都要做的事，但总在措辞上纠结半天——太正式怕显得生硬，太随意又怕不专业。更要命的是，如果直接让 AI 写，收件人一眼就能看出是 AI 生成的：每段结构一模一样、用词过于漂亮、开头永远是「希望这封邮件找到你一切安好」。\n\n这个案例演示了如何借助「去 AI 味」Skill，在 AI 生成邮件初稿后做一轮深度润色——拆掉千篇一律的句式、干掉空洞的客套话、换上真正像人写的表达。最终输出的邮件自然得体，收件人读起来就像你本人写的。\n\n适合：所有需要频繁写商务邮件的职场人、跨国协作团队、商务拓展。", "categories": [ "office", "productivity" ], "tags": [ "邮件", "商务沟通", "去AI味", "办公效率", "职场" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "去 AI 味商务邮件" } ], "prompt": "帮我写一封商务邮件，写完后请用「去 AI 味」的方式润色，让邮件读起来像真人写的。\n\n背景信息：\n- 收件人：张总（某合作方公司的产品负责人）\n- 目的：跟进上周五会议讨论的 API 对接方案，确认技术细节和排期\n- 核心内容：\n 1. 感谢上周五的会议，讨论很有成效\n 2. 我方已完成接口文档的整理（附件会另发）\n 3. 希望对方本周内确认 3 个技术问题：鉴权方式、数据格式、回调地址\n 4. 建议下周三安排一次 30 分钟的技术对齐会\n 5. 如有其他问题随时联系\n- 语气：正式但友好，不要太生硬\n- 署名：Vernon Liao / WorkBuddy 产品团队\n\n请生成完整邮件，包含主题行。生成后做一轮去 AI 味润色，确保没有典型 AI 痕迹（如过度使用破折号、三段式排比、空洞修饰词等）。", "prompt_en": "Help me write a business email, and after it's done, please polish it to \"remove the AI flavor\" so it reads like it was written by a real person.\n\nBackground:\n- Recipient: Mr. Zhang (the product lead at a partner company)\n- Purpose: Follow up on the API integration plan discussed in last Friday's meeting, confirming technical details and the schedule\n- Key content:\n 1. Thank them for last Friday's meeting; the discussion was very productive\n 2. We have finished compiling the interface documentation (the attachment will be sent separately)\n 3. We hope they can confirm 3 technical questions within this week: authentication method, data format, and callback address\n 4. Propose scheduling a 30-minute technical alignment meeting next Wednesday\n 5. Feel free to reach out anytime with other questions\n- Tone: formal but friendly, not too stiff\n- Signature: Vernon Liao / WorkBuddy Product Team\n\nPlease generate the complete email, including a subject line. After generating it, do a round of \"de-AI\" polishing to ensure there are no typical AI tells (e.g., overuse of em dashes, three-part parallel structures, hollow filler adjectives, etc.).", "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 235, "is_featured": false, "featured_order": 0, "quality_score": 89, "source": "official", "usage_count": 0, "created_at": "2026-04-13T14:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "title_en": "Business Email One-Click Generator", "subtitle_en": "Input key details, auto-generate well-worded, natural business emails", "description_en": "Writing business emails is a daily task, yet you always agonize over wording — too formal feels stiff, too casual feels unprofessional. Worse, if you let AI write it directly, recipients spot it instantly: every paragraph structured identically, overly polished vocabulary, always opening with \"I hope this email finds you well.\"\n\nThis case shows how the Humanizer Skill helps AI-generated business emails sound genuinely human — varying sentence structures, using natural transitions, and maintaining appropriate formality without the robotic feel.\n\nBest for: Professionals, sales teams, managers, anyone who sends business emails daily.", "tags_en": [ "Email", "Business Communication", "De-AI", "Office Productivity", "Workplace" ], "is_recommended": false, "recommended_order": 0 }, { "id": "employee-salary-roster", "title": "员工薪酬花名册", "title_en": "Employee Salary Roster", "subtitle": "基本工资到到手工资，薪酬核算全自动", "subtitle_en": "Automate payroll from base salary to take-home pay", "description": "每月发薪日前最怕什么？HR 手动核算五十号人的基本工资、绩效奖金、五险一金、个税……一个公式拉错，要么多发要么少发，员工投诉不断、财务对账头大。更别说遇到社保基数调整、个税累计预扣法变化，Excel 里一堆嵌套公式改到怀疑人生。\n\n这个案例演示如何用 AI 一键生成一套完整的薪酬核算 Excel：覆盖 5 个部门 50 名员工，从基本工资、岗位津贴、绩效奖金出发，自动计算养老保险、医疗保险、失业保险、工伤保险、生育保险和住房公积金（五险一金），再按最新累进税率表计算个人所得税，最终得出每人实发工资。同时生成部门薪酬汇总分析，一眼看清各部门人力成本结构。\n\n适合：HR 专员、薪酬主管、财务会计、中小企业老板、任何需要批量处理薪酬核算的人。", "description_en": "What's the biggest dread before payday? HR manually calculating base salary, performance bonuses, social insurance, and income tax for dozens of employees — one wrong formula means overpayment or underpayment, triggering employee complaints and accounting nightmares. It gets worse when social insurance bases adjust or tax withholding rules change, leaving you drowning in nested Excel formulas.\n\nThis case demonstrates how AI can generate a complete payroll calculation Excel in one click: covering 50 employees across 5 departments, starting from base salary, position allowances, and performance bonuses, automatically computing pension, medical, unemployment, work injury, and maternity insurance plus housing fund (five insurances and one fund), then calculating individual income tax using the latest progressive tax brackets, and finally arriving at each employee's take-home pay. It also generates department-level payroll summaries for instant visibility into labor cost structures.\n\nIdeal for: HR specialists, compensation managers, accountants, SMB owners, and anyone who needs to process batch payroll calculations.", "categories": [ "data", "office" ], "tags": [ "Excel", "薪酬核算", "五险一金", "个税计算", "花名册", "HR" ], "tags_en": [ "Excel", "Payroll", "Social Insurance", "Income Tax", "Roster", "HR" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "员工薪酬花名册" } ], "prompt": "帮我生成一份专业的员工薪酬花名册 Excel（output.xlsx）。\n\n业务背景：一家科技公司，50 名员工，覆盖 5 个部门，核算月份 2026 年 3 月。\n\n请包含以下 Sheet：\n1. 员工信息：工号、姓名、部门、岗位、职级、入职日期等\n2. 薪酬明细：基本工资、岗位津贴、绩效奖金、应发合计、五险一金扣除、应纳税所得额、个人所得税、实发工资，金额用 Excel 公式计算\n3. 五险一金明细\n4. 个税计算：按累进税率展示计算过程\n5. 部门汇总：各部门人数、薪酬总额、人力成本、人均薪酬\n\n格式：专业商务风格，标题行高亮，金额规范显示，冻结首行并加自动筛选。输出为 .xlsx 文件。", "prompt_en": "Help me generate a professional employee salary roster in Excel (output.xlsx).\n\nBusiness context: a tech company with 50 employees across 5 departments, payroll month March 2026.\n\nPlease include these sheets:\n1. Employee Info: employee ID, name, department, position, job level, date of joining, etc.\n2. Salary Detail: base salary, position allowance, performance bonus, gross pay, social-insurance & fund deductions, taxable income, individual income tax, net pay; calculate amounts with Excel formulas\n3. Social Insurance Detail\n4. Tax Calculation: show the calculation using progressive tax rates\n5. Department Summary: headcount, total compensation, labor cost, and average pay per department\n\nFormat: professional business style, highlighted header rows, properly formatted amounts, freeze the top row and add auto-filter. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 236, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-15T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "interview-preparation-guide", "title": "面试模拟与准备", "title_en": "Interview Preparation Guide", "subtitle": "目标岗位 + 候选人背景，一键生成面试全攻略报告", "subtitle_en": "Target role + candidate background, instantly generate a comprehensive interview prep report", "description": "面试准备是求职中最焦虑的环节——不知道面几轮、不知道会问什么、准备了一堆却总在追问时卡壳。网上搜的面经碎片化，缺乏针对性，更别说根据你的背景做个性化匹配。\n\n这个案例演示了如何用教育 Skill 为一位 3 年 B 端 SaaS 产品经理生成字节跳动高级产品经理面试全攻略：包含完整面试流程解析（4-5 轮）、20 道高频面试题（产品设计/数据分析/行为面试/商业分析四类）、每道题的参考回答框架与得分要点、面试官常见追问方向、以及面试前一周的准备清单和时间规划表。\n\n适合：求职者、产品经理、运营、设计师——任何正在准备大厂面试的职场人。", "description_en": "Interview prep is the most stressful part of any job search — you don't know how many rounds, what questions to expect, and even when prepared, you freeze under follow-up questions. Interview tips found online are scattered and generic, let alone tailored to your specific background.\n\nThis case demonstrates using the Education Skill to create a comprehensive interview prep guide for a product manager with 3 years of B2B SaaS experience targeting a senior PM role at ByteDance: complete interview process breakdown (4-5 rounds), 20 high-frequency questions (product design, data analysis, behavioral STAR, business analysis), reference answer frameworks with scoring tips, common follow-up directions, and a one-week preparation checklist with daily schedule.\n\nPerfect for: job seekers, product managers, operators, designers — anyone preparing for interviews at major tech companies.", "categories": [ "education" ], "tags": [ "面试", "求职", "产品经理", "职场", "面试题" ], "tags_en": [ "Interview", "Job Hunting", "Product Manager", "Career", "Interview Questions" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "面试准备全攻略" } ], "prompt": "为一位准备面试字节跳动高级产品经理岗位的候选人生成面试准备全攻略。候选人背景：3 年 B 端 SaaS 产品经验，擅长数据分析和用户增长。面试准备内容需包含：1）字节跳动产品经理面试流程解析（一般 4-5 轮：业务面→交叉面→HR 面→总监面）；2）高频面试题库（20 道），分为四类：a）产品设计题 b）数据分析题 c）行为面试题（STAR 框架）d）商业分析题；3）每道题的参考回答框架和得分要点；4）面试官常见的追问方向；5）面试前一周的准备清单和时间规划表。", "prompt_en": "Generate a complete interview preparation guide for a candidate preparing for a Senior Product Manager position at ByteDance. Candidate background: 3 years of B2B SaaS product experience, strong in data analysis and user growth. The interview preparation content should include: 1) an analysis of ByteDance's product manager interview process (typically 4-5 rounds: business interview → cross-functional interview → HR interview → director interview); 2) a high-frequency question bank (20 questions), divided into four categories: a) product design questions b) data analysis questions c) behavioral interview questions (STAR framework) d) business analysis questions; 3) a reference answer framework and scoring key points for each question; 4) common follow-up directions interviewers take; 5) a preparation checklist and time-planning schedule for the week before the interview.", "skills": [ { "id": "education", "name": "Education", "description": "学习助手：生成学习计划、测验、抽认卡、复习材料并跟踪进度", "description_en": "Study assistant: generate plans, quizzes, flashcards, review materials and track progress", "skill_id": "skill_2053082003609702400" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 237, "is_featured": false, "featured_order": 0, "quality_score": 87, "source": "official", "usage_count": 0, "created_at": "2026-04-21T10:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "prd-doc-generator", "title": "产品需求文档 PRD 一键生成", "title_en": "One-Click PRD Generator", "subtitle": "从产品想法到完整 PRD Word 文档，AI 帮你搞定全部章节", "subtitle_en": "From product idea to complete PRD Word document — AI handles every section", "description": "写 PRD 是产品经理最头疼的日常：搭框架、填内容、调格式，一份文档少说半天。尤其是功能需求表、数据字段、风险评估这些结构化章节，重复劳动太多。\n\n这个案例演示了如何用一段产品描述，让 AI 自动生成一份结构完整的 PRD Word 文档。涵盖产品概述、功能需求（含用户故事和验收标准）、非功能需求、数据模型、交互设计、里程碑计划、风险评估等核心章节，表格规范、层级清晰，输出为正式 Word 格式拿到手就能发评审。\n\n适合产品经理、产品实习生、创业团队——凡是需要快速输出专业需求文档的场景，都能用上。", "description_en": "Writing PRDs is every product manager's headache: building frameworks, filling content, adjusting formatting — a single document easily takes half a day. Especially structured sections like feature requirements tables, data models, and risk assessments involve too much repetitive work.\n\nThis case demonstrates how a product description can generate a complete, well-structured PRD as a Word document via AI. It covers product overview, functional requirements (with user stories and acceptance criteria), non-functional requirements, data models, interaction design, milestone planning, and risk assessment — all formatted professionally and ready for stakeholder review.\n\nPerfect for product managers, PM interns, and startup teams — anyone who needs to quickly produce professional requirement documents.", "categories": [ "business", "productivity" ], "tags": [ "PRD", "产品文档", "需求分析", "产品经理" ], "tags_en": [ "PRD", "Product Docs", "Requirements", "Product Manager" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "完整产品需求文档 PRD（Word）" } ], "prompt": "请撰写一份完整的产品需求文档（PRD）Word 文档（output.docx）。\n\n产品背景（填入你的产品）：说明目标用户、核心问题与产品定位。\n\n请包含以下章节：\n1. 文档信息\n2. 产品概述\n3. 功能需求（至少 5 个核心功能，用表格呈现）\n4. 非功能需求\n5. 数据需求\n6. 交互设计要求\n7. 里程碑计划（3 个阶段）\n8. 风险评估（至少 3 项）\n9. 附录\n\n格式：含封面、自动目录、页眉页脚、多级标题编号，输出正式 Word 文档。\n用「Word 文档生成」技能输出。", "prompt_en": "Please write a complete Product Requirements Document (PRD) as a Word document (output.docx).\n\nProduct background (fill in your product): describe target users, the core problem, and product positioning.\n\nPlease include the following sections:\n1. Document Information\n2. Product Overview\n3. Functional Requirements (at least 5 core features, presented in tables)\n4. Non-functional Requirements\n5. Data Requirements\n6. Interaction Design Requirements\n7. Milestone Plan (3 phases)\n8. Risk Assessment (at least 3 items)\n9. Appendix\n\nFormat: include a cover page, automatic table of contents, headers/footers, and multi-level heading numbering; output a formal Word document.\nUse the \"Word Document Generation\" skill to output it.", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 238, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-21T14:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "brand-visual-guideline", "title": "品牌视觉规范手册", "title_en": "Brand Visual Guideline", "subtitle": "从 Logo 到界面的全方位品牌视觉标准", "subtitle_en": "Comprehensive brand visual standards from logo to UI", "description": "品牌视觉不统一是很多成长期公司的通病——设计师各自发挥、外包供应商理解偏差、不同渠道物料风格割裂，导致用户对品牌印象模糊，信任感难以建立。\n\n本案例演示为一家名为「CloudMind」的 AI 生产力工具公司生成完整的品牌视觉规范手册。涵盖品牌概述、Logo 使用规范、色彩系统（含 RGB/HEX/CMYK 全色值）、字体规范、图片风格指引、排版栅格系统，以及名片/PPT/社交媒体等真实应用场景示例，形成一份即查即用的品牌「圣经」。\n\n适合品牌设计师、市场部门负责人、创业公司联合创始人，以及任何需要建立或升级品牌视觉体系的团队。即便没有专业设计背景，也能通过这份规范快速对齐团队视觉标准。", "description_en": "Brand visual inconsistency is a common problem for growing companies — designers freelancing their own styles, outsourced vendors misinterpreting the brief, and marketing materials across channels looking disjointed, leaving users with a blurry brand impression and undermining trust.\n\nThis case demonstrates generating a complete brand visual guideline for an AI productivity tool company called 'CloudMind'. It covers brand overview, logo usage rules (standard/horizontal/icon versions with safe zones and misuse examples), color system (primary #4F46E5 with RGB/HEX/CMYK values and usage ratios), typography specs (Source Han Sans for Chinese, Inter for English), imagery style guides, layout grid system (8px baseline), and real-world application examples including business cards, PPT templates, social media headers, and app UI screenshots.\n\nIdeal for brand designers, marketing leads, startup co-founders, and any team needing to establish or upgrade their brand visual system — even without a professional design background.", "categories": [ "business", "creation" ], "tags": [ "品牌设计", "视觉规范", "VI手册", "品牌标准" ], "tags_en": [ "Brand Design", "Visual Guidelines", "VI Manual", "Brand Standards" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "prompt": "请为一家名为「CloudMind」的 AI 生产力工具公司生成品牌视觉规范手册。包含：1) 品牌概述（使命/愿景/品牌个性）；2) Logo 使用规范（标准版/横版/图标版的使用场景和安全距离，错误示例说明）；3) 色彩系统（主色 #4F46E5/辅助色/中性色/语义色，含 RGB/HEX/CMYK 值和使用比例）；4) 字体规范（中文思源黑体/英文 Inter，标题/正文/注释的字号行高）；5) 图片风格（摄影风格指引、插画风格、图标设计原则）；6) 排版栅格（8px 基准网格、间距系统、卡片组件规范）；7) 应用示例（名片/PPT 模板/社交媒体头图/App 界面截图的品牌应用）。", "prompt_en": "Please create a brand visual guidelines handbook for an AI productivity-tool company called \"CloudMind.\" Include: 1) Brand overview (mission/vision/brand personality); 2) Logo usage guidelines (use cases and clear space for the standard, horizontal, and icon versions, with examples of incorrect usage); 3) Color system (primary color #4F46E5/secondary colors/neutrals/semantic colors, with RGB/HEX/CMYK values and usage proportions); 4) Typography (Source Han Sans for Chinese, Inter for English; font sizes and line heights for headings/body/captions); 5) Imagery style (photography style guidelines, illustration style, icon design principles); 6) Layout grid (8px base grid, spacing system, card component specs); 7) Application examples (brand application on business cards, PPT templates, social media header images, and App UI screenshots).", "skills": [], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 91, "sort_order": 239, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-25T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "artifacts": [ { "file": "output.md", "label": "品牌视觉规范手册" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "scene-ability-test", "title": "能力测评：17 维能力画像", "title_en": "Abilities: A 17-Dimension Profile", "subtitle": "85 题算出优势带与潜力带，长板怎么用短板要不要补", "subtitle_en": "85 items yield strength and potential bands — and what to do with each", "description": "「我能力怎么样」太笼统——能力是 17 个维度的分布，不是一个分数。\n\n本案例跑完 85 题职业能力测评（天赋能力 + 职业通用能力两部分）：语言、人际交往、组织协调三项满分构成优势带，逻辑思维等 75 分档是潜力带，空间与审美 25 分。报告给出全景条形图、Top5 原文解读，以及「对文职序列短板不必补、长板拉满更划算」的使用建议。\n\n适合：写简历找不到亮点的学生、准备结构化面试自我认知题的考生。", "description_en": "\"How capable am I\" is too vague — ability is a distribution across 17 dimensions, not one number.\n\nThis case completes the 85-item assessment (innate aptitudes plus general professional abilities): perfect scores in language, interpersonal and coordination form the strength band; logic and others at 75 form the potential band; spatial and aesthetic sit at 25. The report gives a full bar panorama, verbatim Top-5 readings, and the advice that for desk-role tracks you maximise strengths rather than patch irrelevant gaps.\n\nFor: students hunting résumé highlights and candidates prepping self-awareness questions.", "categories": [ "education" ], "tags": [ "职业能力", "能力测评", "优势分析" ], "tags_en": [ "professional abilities", "assessment", "strengths analysis" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "17 维能力画像报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做 85 题职业能力倾向测评并出报告。要求：1) 说明两部分结构后渲染作答；2) 用评分脚本按真实规则算出 17 维得分与总分；3) 报告含 17 维全景条形图、Top5 优势维度的原文解读、潜力带清单；4) 结合公考文职方向给出「长板怎么用、短板要不要补」的建议并标注哪些是解读。", "prompt_en": "Run the 85-item career-ability assessment and produce a report. Requirements: 1) explain the two-part structure, then render for answering; 2) compute the 17 dimension scores and total with the scoring script's real rules; 3) include the full 17-bar panorama, verbatim readings of the top five, and the potential band; 4) for a public-sector desk-role direction, advise how to use strengths and whether to patch gaps, labelling interpretation clearly.", "skills": [ { "id": "career-ability-assessment", "name": "职业能力测评", "description": "85 题 17 维评估天赋与职业通用能力，输出能力结构解读", "skill_id": "skill_2089337587766607872" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 240, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "website-traffic-analysis", "title": "网站流量数据分析", "title_en": "Website Traffic Data Analysis", "subtitle": "CSV 数据一键生成趋势洞察报告", "subtitle_en": "One-click trend insights report from CSV data", "description": "每月导出 Google Analytics 数据，面对几万行 CSV 不知如何下手，手动做透视表耗时半天。用 CSV 数据分析 Skill 导入网站流量数据，自动清洗、统计关键指标、发现流量趋势，输出一份包含数据摘要和关键洞察的分析报告。适合运营分析师、市场经理、独立站站长、SEO 专员。", "description_en": "Exporting Google Analytics data every month but drowning in tens of thousands of CSV rows? Manual pivot tables eat up half a day. This case shows how to import website traffic data, auto-clean it, calculate key metrics, and spot traffic trends — producing a comprehensive analysis report with data summaries and actionable insights.\n\nBest for: operations analysts, marketing managers, independent site owners, and SEO specialists.", "categories": [ "data" ], "tags": [ "数据分析", "CSV", "网站流量", "趋势分析", "运营分析" ], "tags_en": [ "Data Analysis", "CSV", "Web Traffic", "Trend Analysis", "Operations" ], "artifact_type": "table", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "网站流量数据分析" } ], "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "我有一份网站近 30 天的流量数据（CSV 格式），包含日期、UV、PV、跳出率、平均停留时长、来源渠道、设备类型等字段。请帮我分析这份数据，生成一份流量分析报告，包括：关键指标汇总、流量趋势分析（日/周维度）、渠道分布对比、设备分布、Top 10 着陆页、异常波动检测、以及 3-5 条可执行的优化建议。", "prompt_en": "I have website traffic data for the past 30 days (in CSV format), containing fields such as date, UV, PV, bounce rate, average session duration, source channel, and device type. Please analyze this data and generate a traffic analysis report, including: key metric summary, traffic trend analysis (daily/weekly dimensions), channel distribution comparison, device distribution, Top 10 landing pages, anomaly detection, and 3-5 actionable optimization recommendations.", "quality_score": 93, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "sort_order": 240, "experts": [], "is_featured": false, "featured_order": 0, "usage_count": 0, "mcps": [], "scenario": "", "task_mode": "", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-ability-quantitative", "title": "能力测评：数理逻辑四项 90 分的尖峰型画像", "title_en": "Aptitude Assessment: A Spiked Profile With Four Quantitative Strengths at 90", "subtitle": "85 题覆盖 17 个维度，这份结果的形状非常「尖」", "subtitle_en": "Eighty-five items across seventeen dimensions, producing a sharply peaked shape", "description": "能力测评常被当成「测我行不行」，其实它测的是形状——哪几项突出、哪几项塌陷，比总分有用得多。\n\n本案例用 85 题版跑出一份尖峰型画像：数学能力、逻辑思维、研究分析、细节洞察四项并列 90 分，而动觉、审美、人际交往、商业思维并列垫底 35 分，总分 62。案例给出 17 维分布、题库原文的高分描述，并说明尖峰型画像的优势与迁移风险。\n\n适合：想找技术类岗位的学生、判断自己适合走专精还是通才路线的人。", "description_en": "Aptitude tests get read as \"am I good enough\", when what they actually measure is shape — which dimensions spike and which collapse matters far more than the total.\n\nThis case runs the 85-item version to a sharply peaked profile: mathematical ability, logical thinking, research analysis and attention to detail all tie at 90, while kinaesthetic, aesthetic, interpersonal and commercial thinking tie at 35, for a total of 62. It shows all seventeen dimensions with the item bank's high-score descriptions, and discusses the strengths and transfer risk of a spiked profile.\n\nFor: students targeting technical roles and anyone deciding between specialising and generalising.", "categories": [ "education" ], "tags": [ "能力测评", "数理逻辑", "尖峰型", "优势识别" ], "tags_en": [ "aptitude assessment", "quantitative reasoning", "spiked profile", "strengths" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "职业能力倾向尖峰型结果报告", "label_en": "Aptitude report: spiked profile" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次职业能力倾向测评。我作答时对数学、逻辑推理、研究分析、细节核对类的描述基本都选「完全符合」或「比较符合」，对动手操作、审美设计、人际社交、商业谈判类的描述多选「比较不符合」。要求：1) 用 85 题题库真实跑一遍评分脚本，给出总分和 17 个维度的得分；2) 展示得分最高的 10 个维度；3) 完整给出前五名维度的题库高分描述，不要改写；4) 指出有没有并列同分的维度，说明同分该怎么解读；5) 分析这种尖峰型分布的优势和风险，并说明这是自评量表、测的是自我判断而非客观能力。", "prompt_en": "Run a career-aptitude assessment. My answers mark mathematics, logical reasoning, research analysis and attention to detail as strongly matching, and hands-on work, aesthetic design, interpersonal and commercial negotiation as largely not matching. Requirements: 1) run the official 85-item scoring script and give the total and all seventeen dimension scores; 2) show the top ten dimensions; 3) reproduce the item bank's high-score descriptions for the top five verbatim; 4) note any tied dimensions and how ties should be read; 5) analyse the strengths and risks of a spiked profile, noting this is a self-report instrument measuring self-perception rather than objective ability.", "skills": [ { "id": "career-ability-assessment", "name": "职业能力测评", "description": "85 题自评量表，输出 17 个能力维度得分与优势排序", "skill_id": "skill_2089337587766607872" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 241, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "restaurant-opening-sop", "title": "餐饮门店开业 SOP", "title_en": "Restaurant Opening SOP", "subtitle": "从选址到试营业，60天开店全流程 Word 标准化手册", "subtitle_en": "From site selection to soft opening — 60-day standardized Word manual", "description": "第一次开餐饮店，流程一片空白，不知道要办什么证、装修注意什么、开业前要准备什么？这个案例展示了如何用 Word 文档生成技能一键输出一份完整的餐饮门店开业标准操作手册 Word 文档，覆盖选址评估、证照办理、装修施工、人员招聘、供应链搭建、试营业 6 大阶段共 50+ 检查项。\n\n手册从 Week 1 的商圈调研和人流量评估开始，逐步推进到证照办理全流程、装修施工管控、团队组建与培训、供应链搭建，最终在 Week 8 完成试营业准备。附录还包含关键文档模板清单和常见踩坑提醒。正式 Word 格式，打印出来就能用。\n\n适合：餐饮创业者、连锁品牌拓店经理、加盟商、餐饮顾问。", "description_en": "Opening your first restaurant with no roadmap? This case demonstrates using the Word Document Generation skill to generate a comprehensive restaurant opening SOP as a Word document covering 6 phases with 50+ checkpoints: site selection, licensing, renovation, hiring & training, supply chain setup, and soft opening.\n\nThe handbook walks through each week from location scouting to soft opening preparation. Generated as a formal Word document, print-ready for use.\n\nBest for: Restaurant entrepreneurs, chain expansion managers, franchisees, and F&B consultants.", "categories": [ "business" ], "tags": [ "餐饮开业", "SOP", "标准流程", "创业开店", "运营管理" ], "tags_en": [ "Restaurant Opening", "SOP", "Standard Process", "Startup", "Operations" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "餐饮开业SOP手册（Word）" } ], "prompt": "请帮我创建一份餐饮门店开业 SOP（标准操作流程）Word 文档（output.docx），覆盖从选址到试营业的全流程。\n\n要求：\n1. 时间跨度约 60 天（8 周）\n2. 包含 6 大阶段：选址与市场调研、证照办理、装修施工、人员招聘与培训、供应链搭建、试营业\n3. 每个阶段需包含：目标、负责人角色、详细检查项、时间节点、风险提示\n4. 证照办理部分要列出每个证照的办理流程和所需材料\n5. 装修阶段要包含厨房设备清单和验收标准\n6. 人员部分要有岗位配置表和培训大纲\n7. 供应链部分要有供应商筛选标准和成本控制方法\n8. 试营业部分要有完整的软开业 checklist 和应急预案\n9. 附录包含关键文档模板清单、常见踩坑提醒、预算估算参考\n\n文档格式要求：包含封面页、目录、页眉页脚、多级标题编号、统一表格样式\n输出为正式 Word 文档（output.docx）", "prompt_en": "Please help me create a restaurant opening SOP (Standard Operating Procedure) Word document (output.docx) covering the full process from site selection to soft opening.\n\nRequirements:\n1. A time span of about 60 days (8 weeks)\n2. Include 6 major phases: site selection & market research, license processing, renovation & construction, staff recruitment & training, supply chain setup, and soft opening\n3. Each phase should include: objectives, responsible role, detailed checklist items, time milestones, and risk warnings\n4. The license processing section should list the application process and required materials for each license\n5. The renovation phase should include a kitchen equipment list and acceptance criteria\n6. The staffing section should have a position configuration table and a training outline\n7. The supply chain section should have supplier screening criteria and cost control methods\n8. The soft opening section should have a complete soft-opening checklist and contingency plans\n9. The appendix should include a checklist of key document templates, common-pitfall reminders, and a budget estimate reference\n\nDocument formatting requirements: include a cover page, table of contents, headers and footers, multi-level heading numbering, and a unified table style.\nOutput as a formal Word document (output.docx).", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 241, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "business-plan-doc", "title": "商业计划书 (BP)", "title_en": "Business Plan (BP)", "subtitle": "从执行摘要到财务预测的完整融资 BP", "subtitle_en": "Complete fundraising BP from executive summary to financial projections", "description": "写 BP 是创业者最头疼的事之一——投资人要看市场规模、竞品分析、财务模型、团队背景，每个板块都有讲究，结构稍乱就丢印象分。很多创始人花两周写出来的 BP，投资人 30 秒就翻完了，因为重点不突出、数据没说服力。\n\n本案例以一家 AI 智能客服 SaaS 创业公司的 Pre-A 轮融资为场景，演示了 AI 如何生成一份完整的商业计划书：涵盖执行摘要、公司概况与里程碑、TAM/SAM/SOM 市场分析、产品架构与技术壁垒、竞品矩阵对比、SaaS 商业模式与单位经济模型、核心团队介绍、三年财务预测（含盈亏平衡分析）、融资计划与资金用途、以及风险评估矩阵。\n\n适合：创业者、融资顾问、投资经理、商业分析师。", "description_en": "Writing a business plan is one of the most daunting tasks for founders — investors want to see market sizing, competitive analysis, financial models, and team backgrounds. Each section has its own standards, and a poorly structured BP loses credibility in seconds. Many founders spend two weeks crafting a BP that investors skim in 30 seconds because key points aren't prominent and the data isn't compelling.\n\nThis case uses an AI-powered customer service SaaS startup's Pre-A fundraising round as the scenario, demonstrating how AI generates a complete business plan: covering executive summary, company overview with milestones, TAM/SAM/SOM market analysis, product architecture and technical moats, competitor matrix comparison, SaaS business model with unit economics, core team profiles, three-year financial projections with break-even analysis, fundraising plan and use of funds, and risk assessment matrix.\n\nBest for: Founders, fundraising advisors, investment managers, business analysts.", "categories": [ "business" ], "tags": [ "商业计划书", "BP", "融资", "创业" ], "tags_en": [ "Business Plan", "BP", "Fundraising", "Startup" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "AI 智能客服 SaaS 商业计划书（Pre-A 轮）" } ], "prompt": "请用 FBS-BookWriter 的长文写作能力，为一家 AI 智能客服 SaaS 创业公司撰写一份面向 Pre-A 轮融资的商业计划书（3 万字以上）。\n\n公司概况（可替换为你自己的信息）：成立约 1.5 年、团队近 20 人、数十家付费客户、MRR 数十万且稳定月增；产品为基于大模型的多轮对话客服机器人；本轮拟融数百万至千万级人民币。\n...\n\n要求生成完整 BP，包含：执行摘要、市场分析（TAM/SAM/SOM）、产品介绍、商业模式、竞争分析、团队介绍、三年财务预测、融资计划与资金用途。", "prompt_en": "Using FBS-BookWriter's long-form writing capability, write a Pre-A fundraising business plan (30,000+ characters) for an AI customer-service SaaS startup.\n\nCompany snapshot (replace with your own): ~1.5 years old, team of nearly 20, dozens of paying customers, MRR in the hundreds of thousands with steady monthly growth; product is an LLM-based multi-turn customer-service chatbot; this round raises several million to ~ten million RMB.\n...\n\nProduce a complete BP including: executive summary, market analysis (TAM/SAM/SOM), product, business model, competitive analysis, team, three-year financial projection, and fundraising plan & use of funds.", "skills": [ { "id": "fbs-bookwriter", "name": "fbs-bookwriter", "description": "福帮手出品 | 高质量长文档手稿工具链：书籍、手册、白皮书、行业指南、长篇报道、深度专题；支持联网查证（宿主允许时启用，离线自动降级）、S/P/C/B 分层审校、中文排版与 MD/HTML 交付。触发词：福帮手、福帮手写书skill、福帮手写书、写书、出书、写长篇、写手册、写白皮书、写行业指南、协作写书、定大纲、写章节、封面、插图、排版构建、导出、去AI味、质量自检、图文书、写报道、写深度稿、写特稿、写专题、写调查报道、写长文、拆书改写、海外本地化改写、爆款结构改写、激活原料、原料盘点、整理素材", "description_en": "Dual-channel long-form writing workflow for books, manuals, whitepapers, guides and reports with layered QC, 7-lock anti-drift anchors, and offline fallback.", "skill_id": "skill_2053083113738002432" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 242, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-21T00:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "insp-ability-balanced", "title": "能力测评：总分 77 却九项并列的均衡型画像", "title_en": "Aptitude Assessment: A Balanced Profile Scoring 77 With Nine Dimensions Tied", "subtitle": "总分更高，但没有一项拔尖——这样的结果该怎么用", "subtitle_en": "A higher total with no standout — how to read that", "description": "有人测完发现「哪项都还行，但哪项都不突出」，第一反应是「我是不是没有特长」——这其实是对自评量表的误读。\n\n本案例跑出一份均衡型画像：总分 77（高于尖峰型的 62），但前五名全部并列 80 分，另有 4 个维度同样 80 分，实际是九项并列。案例解释这种分布的真实含义，并建议改用职业锚和霍兰德来定位方向。\n\n适合：测评结果没有明显优势项的人、纠结于「我到底擅长什么」的学生。", "description_en": "People whose results come back \"decent everywhere, outstanding nowhere\" tend to conclude they have no strengths — which misreads what a self-report instrument does.\n\nThis case produces a balanced profile: a total of 77 (above the spiked profile's 62), but the top five all tie at 80, with four more dimensions also at 80 — nine tied in all. It explains what that distribution actually means and recommends using career anchors and Holland codes to find direction instead.\n\nFor: anyone whose results show no clear peak, and students stuck on \"what am I actually good at\".", "categories": [ "education" ], "tags": [ "能力测评", "均衡型", "并列同分", "结果解读" ], "tags_en": [ "aptitude assessment", "balanced profile", "tied scores", "interpretation" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "职业能力倾向均衡型结果报告", "label_en": "Aptitude report: balanced profile" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我做一次职业能力倾向测评。我作答时大部分题都选「比较符合」，少数几题选「完全符合」，整体比较平均，没有明显偏向。要求：1) 用 85 题题库真实跑一遍评分脚本，给出总分和 17 个维度的得分；2) 展示得分最高的 10 个维度；3) 完整给出前五名维度的题库高分描述；4) 指出有多少个维度并列同分，说明并列意味着什么、脚本是按什么规则排序的；5) 分析均衡型画像该怎么用，如果它无法帮我定方向，还应该做哪些测评。", "prompt_en": "Run a career-aptitude assessment. My answers are mostly \"somewhat matching\" with a few \"strongly matching\" — fairly even, with no clear lean. Requirements: 1) run the official 85-item scoring script and give the total and all seventeen dimension scores; 2) show the top ten dimensions; 3) reproduce the item bank's high-score descriptions for the top five; 4) report how many dimensions are tied, what ties mean, and the rule the script uses to order them; 5) analyse how to use a balanced profile, and which other assessments to run if it can't point to a direction.", "skills": [ { "id": "career-ability-assessment", "name": "职业能力测评", "description": "85 题自评量表，输出 17 个能力维度得分与优势排序", "skill_id": "skill_2089337587766607872" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 242, "is_featured": false, "featured_order": 0, "quality_score": 91, "usage_count": 0, "created_at": "2026-08-27T00:00:00+08:00", "updated_at": "2026-08-27T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "wedding-budget-planner", "title": "婚礼预算规划表", "title_en": "Wedding Budget Planner", "subtitle": "输入婚礼需求，生成详细的预算规划与费用追踪表格", "subtitle_en": "Input wedding requirements, generate detailed budget planning and expense tracking spreadsheet", "description": "筹备婚礼最崩溃的不是选婚纱、不是定酒店，而是预算——婚庆公司报价水分多大？酒席到底按什么标准算？零零碎碎加起来到底花了多少？等到婚礼结束一算账，发现比预期超支了 30% 以上的新人不在少数。\n\n这个案例演示了输入婚礼基本信息（宾客人数、预算上限、婚礼风格），AI 自动生成一份完整的婚礼预算规划 Excel 表格：总览仪表盘（总预算/已确认/已支付/待付款）、分类明细表（场地、餐饮、婚纱摄影、婚庆布置、喜糖回礼、交通住宿等 8 大类）、供应商对比表、付款时间线、宾客礼金预估。每项都有预算金额、实际金额、差异、备注，公式自动汇总。\n\n适合：准新人、婚礼策划师、父母帮忙筹备婚礼的家庭。", "description_en": "The most stressful part of wedding planning isn't choosing the dress or the venue — it's the budget. Vendor quotes are inflated, banquet costs are confusing, and miscellaneous expenses add up invisibly. Many couples end up 30%+ over budget by the end.\n\nThis case demonstrates inputting basic wedding info (guest count, budget cap, wedding style), with AI generating a complete wedding budget planning Excel spreadsheet: overview dashboard (total budget/confirmed/paid/pending), categorized expense breakdown (venue, catering, photography, decorations, gifts, transportation, etc. across 8 categories), vendor comparison sheet, payment timeline, and gift money estimation. Each item includes budgeted amount, actual amount, variance, and notes, with automatic formula summation.\n\nBest for: Engaged couples, wedding planners, families helping with wedding preparation.", "categories": [ "data" ], "tags": [ "婚礼预算", "费用追踪", "Excel", "生活规划", "预算管理" ], "tags_en": [ "Wedding Budget", "Expense Tracking", "Excel", "Life Planning", "Budget Management" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "prompt": "请为一对准新人生成一份完整的婚礼预算规划 Excel（output.xlsx）。\n\n基本信息：婚礼日期 2026 年 10 月 3 日，宾客约 200 人（20 桌），总预算上限 25 万元，户外草坪 + 室内晚宴，城市杭州。\n\n请生成以下工作表：\n1. 总览仪表盘：总预算、已确认、已支付、待付款，各分类占比与使用进度\n2. 费用明细：约 8 大分类（场地、餐饮、婚纱摄影、布置、服饰造型、喜糖回礼、交通住宿等），每项含预算金额、实际金额、差异、供应商、状态、备注\n3. 供应商对比：对比几家婚庆公司与酒店的报价和服务\n4. 付款时间线：各笔大额支出的定金 / 尾款节点，按月排列现金流\n5. 宾客礼金预估：按关系分类预估礼金及总收入\n\n要求：用公式自动汇总各类小计与总计，超预算标红、节省标绿，状态列用下拉，金额规范显示。输出为 .xlsx 文件。", "prompt_en": "Please generate a complete wedding budget planning Excel (output.xlsx) for a soon-to-be-married couple.\n\nBasic info: wedding date October 3, 2026, about 200 guests (20 tables), total budget cap ¥250,000, outdoor lawn ceremony + indoor evening banquet, in Hangzhou.\n\nPlease generate these worksheets:\n1. Overview Dashboard: total budget, confirmed, paid, pending, each category's share and usage progress\n2. Expense Detail: about 8 categories (venue, catering, photography, decoration, attire & styling, candy & gifts, transport & accommodation, etc.), each item with budgeted amount, actual amount, variance, supplier, status, notes\n3. Supplier Comparison: compare quotes and services of several wedding companies and hotels\n4. Payment Timeline: deposit / final-payment milestones for large expenses, cash flow by month\n5. Guest Gift Money Estimate: estimate gift money by relationship and total income\n\nRequirements: auto-sum subtotals and totals with formulas, mark over-budget in red and savings in green, dropdown for the status column, properly formatted amounts. Output as an .xlsx file.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 243, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "artifacts": [ { "file": "output.xlsx", "label": "婚礼预算规划表" } ], "is_recommended": false, "recommended_order": 0 }, { "id": "saas-contract-redline-docx", "title": "SaaS 合同修订意见稿", "title_en": "SaaS Contract Redline Memo", "subtitle": "生成带风险分级、条款对照与签批区的正式 Word 审查稿", "subtitle_en": "Generate a formal Word review memo with risk levels, clause comparisons, and sign-off areas", "description": "审一份 SaaS 合同最怕的是条款分散、结论不统一，商务、法务和业务方各看各的版本，最后谁都说不清到底哪些地方必须改。\n\n这个案例演示了如何用 Word 文档生成技能生成一份正式的 Word 合同修订意见稿：封面、目录、关键商务条款摘要、风险矩阵、逐条修订建议、谈判优先级清单、签批区全部排版完成，适合直接进入内部审批流。\n\n适合：法务、采购、商务负责人、企业服务销售、需要输出正式审查文档的项目经理。", "description_en": "Reviewing a SaaS contract becomes painful when clauses are scattered, conclusions are inconsistent, and business, legal, and procurement teams are all reading different versions.\n\nThis case shows how to use the Word Document Generation skill to generate a formal Word contract redline memo with a cover page, table of contents, commercial term summary, risk matrix, clause-by-clause revisions, negotiation priorities, and sign-off areas — ready for internal approval workflows.\n\nBest for: Legal teams, procurement managers, business owners, enterprise sales teams, and project managers who need polished review documents.", "categories": [ "office", "business", "productivity" ], "tags": [ "合同审查", "SaaS", "Word", "风险分级", "商务谈判" ], "tags_en": [ "Contract Review", "SaaS", "Word", "Risk Matrix", "Negotiation" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "SaaS 合同修订意见稿（Word）" } ], "prompt": "请生成一份正式的 Word 文档（output.docx），标题《SaaS 合同修订意见稿》，可直接进入审批流。\n\n场景（可替换为你自己的）：甲方为云服务商，乙方为连锁零售企业，拟签多年期门店运营 SaaS 合同。\n...\n\n格式：含封面、自动目录、页眉页脚页码、多级标题编号、统一表格样式、审批签字区。\n\n内容至少包含：文档信息、审查结论摘要、关键商务条款摘要表、风险总览与谈判优先级矩阵、逐条修订意见（原条款/风险说明/建议改法/建议话术）、签署前检查清单、审批与确认。重点审查单方调整、续约定价、SLA 补偿、数据导出与归属、争议管辖等高风险条款。语言专业克制，适合法务/采购/业务三方共用。", "prompt_en": "Generate a formal Word document (output.docx) titled \"SaaS Contract Redline Proposal\", ready to enter an approval flow.\n\nScenario (replace with your own): Party A is a cloud service provider, Party B is a retail chain, negotiating a multi-year store-operations SaaS contract.\n...\n\nFormat: cover page, auto table of contents, header/footer with page numbers, multi-level heading numbering, consistent table styling, an approval-signature area.\n\nContent at minimum: document info, review-conclusion summary, key commercial-terms summary table, risk overview & negotiation-priority matrix, clause-by-clause redlines (original clause / risk note / suggested wording / negotiation talking points), a pre-signing checklist, and approval & confirmation. Focus on high-risk clauses such as unilateral change, renewal pricing, SLA compensation, data export & ownership, and dispute jurisdiction. Professional, restrained language usable by legal, procurement, and business together.", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 92, "sort_order": 244, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:03:53Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "career-transition-roadmap", "title": "职业转型路线图", "title_en": "Career Transition Roadmap", "subtitle": "AI 定制化职业转型学习路径与技能规划", "subtitle_en": "AI-Customized Career Transition Learning Path & Skill Planning", "description": "想从传统软件开发转型 AI/ML 工程师，却不知道从哪里开始？技能差距有多大、该学什么、按什么顺序学、学到什么程度才能面试——这些问题困扰着每一个想要转型的工程师。\n\n这个案例生成了一份完整的职业转型路线图，包含现有技能评估与差距分析、6个月分阶段学习计划（从基础数学到面试准备）、每阶段推荐学习资源、8个由浅入深的实战项目、技能树可视化、里程碑检查点与自测题，以及薪资预期与市场需求分析。\n\n适合：正在考虑转型 AI/ML 方向的软件工程师、希望系统性提升 AI 技能的开发者、需要制定学习规划的职场人士、职业规划咨询师和培训机构", "description_en": "Want to transition from traditional software development to AI/ML engineering but don't know where to start? How big is the skill gap, what should you learn, in what order, and to what level before interviewing — these questions plague every engineer considering a career pivot.\n\nThis case generates a comprehensive career transition roadmap including existing skill assessment and gap analysis, a 6-month phased learning plan (from foundational math to interview prep), recommended learning resources per phase, 8 progressive hands-on projects, a visual skill tree, milestone checkpoints with self-assessment quizzes, and salary expectations with market demand analysis.\n\nIdeal for: Software engineers considering AI/ML transition, developers seeking systematic AI skill improvement, professionals needing structured learning plans, career consultants and training institutions", "categories": [ "education" ], "tags": [ "职业转型", "学习路径", "技能规划", "职业发展", "终身学习" ], "tags_en": [ "career transition", "learning path", "skill planning", "career development", "lifelong learning" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "职业转型路线图" } ], "prompt": "帮我制定一份从传统软件开发工程师转型为 AI/ML 工程师的职业转型路线图。需要包含：1）现有技能评估与差距分析（编程、数学、ML框架、数据处理）；2）6 个月分阶段学习计划（基础数学→ML基础→深度学习→工程实践→项目实战→面试准备）；3）每阶段推荐学习资源（课程、书籍、开源项目）；4）实战项目清单（从简单到复杂 8 个项目）；5）技能树可视化（已掌握/学习中/待学习）；6）里程碑检查点与自测题；7）薪资预期与市场需求分析。输出为结构化 Markdown 路线图文档。", "prompt_en": "Help me put together a career-transition roadmap for moving from a traditional software development engineer to an AI/ML engineer. It should include: 1) an assessment of current skills and a gap analysis (programming, math, ML frameworks, data processing); 2) a 6-month phased learning plan (foundational math → ML fundamentals → deep learning → engineering practice → hands-on projects → interview prep); 3) recommended learning resources for each phase (courses, books, open-source projects); 4) a list of hands-on projects (8 projects from simple to complex); 5) a skill tree visualization (mastered / learning / to learn); 6) milestone checkpoints and self-assessment questions; 7) salary expectations and market-demand analysis. Output as a structured Markdown roadmap document.", "skills": [ { "id": "education", "name": "Education", "description": "学习助手：生成学习计划、测验、抽认卡、复习材料并跟踪进度", "description_en": "Study assistant: generate plans, quizzes, flashcards, review materials and track progress", "skill_id": "skill_2053082003609702400" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 93, "sort_order": 245, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2025-07-18", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "resume-optimizer", "title": "个人简历智能优化", "subtitle": "粘贴现有简历，AI 优化后去 AI 味，输出 HR 爱看的专业简历", "description": "简历是求职的第一道门面，但大多数人的简历都有同样的问题：成果没量化、措辞太口语化、重点不突出。而如果直接让 AI 重写，简历又会充满「具有丰富的 XX 经验」「在 XX 领域有着深刻的理解」这类一看就是 AI 的话。HR 每天看几百份简历，AI 味太重直接扣分。\n\n这个案例演示了两步走：第一步让 AI 按照标准简历格式重新组织内容、量化成果；第二步用「去 AI 味」Skill 把措辞调自然，去掉 AI 常用的空洞修饰词和千篇一律的句式。最终输出的简历既专业又像你本人写的。\n\n适合：正在求职的职场人、准备跳槽的同学、应届毕业生、任何想提升简历质量的人。", "categories": [ "office", "productivity" ], "tags": [ "简历", "求职", "去AI味", "职场", "HR" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "去 AI 味优化简历" } ], "prompt": "帮我优化简历，优化后用「去 AI 味」处理，确保读起来像真人写的、不会被 HR 一眼看出是 AI 生成。最终输出 Markdown（output.md）。\n\n目标岗位：高级产品经理（B 端 SaaS 方向）\n\n现有信息（粘贴你的简历素材）：\n...\n\n要求：\n1. 按标准简历格式输出\n2. 工作成果尽量量化\n3. 措辞专业但不浮夸，自然真实\n4. 突出 B 端产品与从 0 到 1 经验\n5. 控制在一页 A4 的内容量", "prompt_en": "Help me optimize my resume, then apply a \"de-AI\" treatment so it reads like it was written by a real person and won't be instantly flagged by HR as AI-generated. Output as Markdown (output.md).\n\nTarget position: Senior Product Manager (B2B SaaS).\n\nExisting information (paste your resume material):\n...\n\nRequirements:\n1. Output in a standard resume format\n2. Quantify achievements as much as possible\n3. Keep the wording professional but not exaggerated, natural and genuine\n4. Highlight B2B product and 0-to-1 experience\n5. Keep it to the length of one A4 page", "skills": [ { "id": "humanizer", "name": "去AI味", "description": "去除文本中的 AI 写作痕迹", "description_en": "Remove AI writing patterns from text", "skill_id": "skill_2053082097175687168" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 246, "is_featured": false, "featured_order": 0, "quality_score": 91, "source": "official", "usage_count": 0, "created_at": "2026-04-13T14:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "title_en": "Smart Resume Optimizer", "subtitle_en": "Paste your resume, AI optimizes and de-AIs it for a professional output HR loves", "description_en": "Your resume is the first impression in job hunting, but most resumes share the same problems: achievements aren't quantified, wording is too casual, key points don't stand out. And if you let AI rewrite it directly, it fills up with \"rich experience in XX\" and \"deep understanding of XX\" — obvious AI language. HR reviews hundreds of resumes daily; heavy AI traces mean instant minus points.\n\nThis case demonstrates a two-step approach: first, AI restructures and enhances the content; then the Humanizer Skill removes AI traces to produce a professional resume that reads naturally.\n\nBest for: Job seekers, career changers, fresh graduates, anyone updating their resume.", "tags_en": [ "Resume", "Job Hunting", "De-AI", "Workplace", "HR" ], "is_recommended": false, "recommended_order": 0 }, { "id": "weekly-data-dashboard-excel", "title": "项目周数据看板 Excel 报表", "title_en": "Weekly Project Data Dashboard Excel", "subtitle": "核心指标一目了然，数据驱动周会汇报", "subtitle_en": "Key metrics at a glance — data-driven weekly reporting", "description": "每周手动整理数据做看板，耗时费力还容易出错？\n\n这个案例演示如何用 AI 一键生成专业的项目周数据看板 Excel 报表。包含 3 个 Sheet：核心指标 Dashboard（DAU/注册/转化/NPS 四大卡片 + 7天明细 + 环比公式）、渠道分析（5大渠道 ROI 排名）、功能使用热度 Top10（数据条可视化）。深蓝商务风格，条件格式自动标红标绿，打开即用。\n\n适合：产品经理、数据分析师、运营负责人、SaaS 团队 Leader 等需要定期制作周报看板的人。", "description_en": "Tired of manually assembling weekly dashboards — slow, tedious, and error-prone?\n\nThis case shows how AI generates a professional weekly data dashboard Excel in one shot. It includes 3 sheets: a KPI Dashboard (DAU/signups/conversion/NPS cards + 7-day breakdown + WoW formulas), Channel Analysis (5 channels with ROI ranking), and Feature Usage Heatmap Top 10 (with data bar visualization). Dark blue business theme, automatic conditional formatting in green/red, ready to present.\n\nBest for: Product managers, data analysts, ops leads, SaaS team leaders — anyone who builds weekly report dashboards regularly.", "categories": [ "data", "office" ], "tags": [ "数据看板", "Excel", "周报", "产品数据", "SaaS" ], "tags_en": [ "Dashboard", "Excel", "Weekly Report", "Product Data", "SaaS" ], "artifact_type": "table", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.xlsx", "artifacts": [ { "file": "output.xlsx", "label": "项目周数据看板 Excel 报表" } ], "prompt": "请创建一份项目周数据看板 Excel（output.xlsx），供产品经理每周向管理层汇报核心指标。填入你的产品名与汇报周期。\n\n工作表骨架：\n1. 周数据看板：顶部核心指标卡（含周环比）、7 天多指标明细表、趋势区\n2. 渠道分析：各获客渠道的转化与 ROI 排名\n3. 功能使用热度：Top 10 功能使用情况，用数据条可视化\n\n要求：商务风格，增长/下降用条件格式区分，百分比与环比用公式计算。\n用「Excel 文件处理」技能输出。", "prompt_en": "Please create a weekly project data dashboard Excel (output.xlsx) for a product manager to report core metrics to management each week. Fill in your product name and reporting period.\n\nWorksheet skeleton:\n1. Weekly Dashboard: top metric cards (with WoW change), a 7-day multi-metric detail table, and a trend area\n2. Channel Analysis: conversion and ROI ranking across acquisition channels\n3. Feature Usage Heatmap: Top 10 features, visualized with data bars\n\nRequirements: business style, use conditional formatting to distinguish growth/decline, and use formulas for percentages and WoW changes.\nUse the \"Excel File Processing\" skill to output it.", "skills": [ { "id": "minimax-xlsx", "name": "Excel 文件处理", "description": "Excel 文件创建与分析", "description_en": "Create, edit, and analyze Excel spreadsheets", "skill_id": "skill_2053082401071824896" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "source": "official", "quality_score": 94, "sort_order": 247, "is_featured": false, "featured_order": 0, "usage_count": 0, "created_at": "2026-04-23T12:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "sop-document-generator", "title": "SOP 标准操作流程文档", "title_en": "SOP Standard Operating Procedure", "subtitle": "描述业务流程，自动生成可落地的 Word 标准操作手册", "subtitle_en": "Describe a business process, auto-generate an actionable SOP as Word document", "description": "团队规模一大，最怕的就是「每个人做同一件事的方式都不一样」。客户投诉处理，有人先道歉再查原因，有人先查原因让客户等半天；新人入职，不同 mentor 教的流程五花八门。没有标准化的 SOP，团队效率就是一笔糊涂账。\n\n这个案例演示了描述一个「客户投诉处理」业务流程，AI 自动生成一份完整的 SOP Word 文档：包含目的与范围、角色与职责矩阵（RACI）、分步操作流程、判断分支与升级路径、异常情况处理预案、质量检查清单、以及 KPI 度量指标。生成正式 Word 格式，开箱即用，打印出来就能培训新人。\n\n适合：运营经理、客服主管、质量管理人员、创业公司搭建标准化流程、任何需要把经验变成文档的团队。", "description_en": "When teams grow, the biggest problem is \"everyone does the same thing differently.\" Without standardized SOPs, team efficiency is a mess. But writing SOPs is tedious — mapping the full process, detailing every step and exception, then iterating.\n\nThis case demonstrates describing a \"Customer Complaint Handling\" process and having AI auto-generate a complete SOP as a Word document: purpose and scope, RACI responsibility matrix, step-by-step procedures, decision branches and escalation paths, exception handling, quality checklists, and KPI metrics. Generated as a formal Word document, print-ready for training.\n\nBest for: Operations managers, customer service leads, quality managers, startups building standardized processes.", "categories": [ "office", "productivity" ], "tags": [ "SOP", "流程管理", "标准化", "运营管理" ], "tags_en": [ "SOP", "Process Management", "Standardization", "Operations" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "客户投诉处理SOP（Word）" } ], "prompt": "请为「客户投诉处理」业务流程生成一份完整的 SOP 标准操作文档（output.docx）。\n\n背景：一家约 200 人的中型 SaaS 公司，客服团队 15 人，日均投诉 50-80 单，渠道含在线客服、电话、邮件、社交媒体；现有处理标准不统一、升级路径不清晰。\n\n请按以下结构输出完整 SOP：\n1. 文档信息（编号、版本、生效日期、审批人、修订记录）\n2. 目的与范围\n3. 术语定义（含 P0-P3 投诉分级）\n4. 角色与职责（RACI 矩阵）\n5. 投诉分级标准与响应时限\n6. 标准操作流程\n7. 升级路径\n8. 异常情况处理\n9. 质量检查清单\n10. KPI 指标\n11. 附录\n\n格式：含封面页、目录、页眉页脚、多级标题编号、统一表格样式。输出为正式 Word 文档（output.docx）。", "prompt_en": "Please generate a complete SOP document (output.docx) for the \"Customer Complaint Handling\" business process.\n\nBackground: a mid-sized SaaS company of about 200 people with a 15-person customer service team, 50-80 complaints daily, channels including online chat, phone, email, and social media; current handling standards are inconsistent and escalation paths unclear.\n\nPlease output a complete SOP with this structure:\n1. Document info (number, version, effective date, approver, revision history)\n2. Purpose and scope\n3. Terminology (including P0-P3 complaint grading)\n4. Roles and responsibilities (RACI matrix)\n5. Complaint grading standards and response time limits\n6. Standard operating procedure\n7. Escalation paths\n8. Exception handling\n9. Quality inspection checklist\n10. KPI metrics\n11. Appendices\n\nFormat: cover page, table of contents, headers/footers, multi-level heading numbering, consistent table styling. Output as a formal Word document (output.docx).", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 248, "is_featured": false, "featured_order": 0, "quality_score": 90, "source": "official", "usage_count": 0, "created_at": "2026-04-16T10:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "cross-border-ecommerce-plan", "title": "跨境电商商业计划书", "title_en": "Cross-Border E-Commerce Business Plan", "subtitle": "东南亚出海全案，从市场分析到财务预测", "subtitle_en": "Southeast Asia expansion — from market analysis to financial projections", "description": "想做跨境电商但不知道目标市场怎么选、物流怎么搭、合规怎么做，写商业计划书更是无从下手——东南亚六国市场差异大，Shopee/Lazada/TikTok Shop 各有生态，物流有直发、海外仓、本地仓三种模式，税务清关政策各国不同，一份靠谱的 BP 需要覆盖太多维度。\n\n本案例演示了如何用 FBS-BookWriter 长文写作能力生成一份完整的东南亚跨境电商 BP，覆盖市场分析（六国电商市场规模与增长率对比）、目标市场选择（印尼+泰国 PEST 分析）、竞争格局（三大平台差异化定位）、商业模式（B2C 自营+平台混合+供应链架构）、运营策略（选品/定价/TikTok 达人+Facebook 广告+本地 KOL）、物流方案（三种模式成本测算）、合规与风险（税务/清关/知识产权/支付）、财务预测（三年 P&L + Unit Economics + BEP 分析）、团队与里程碑（核心四人团队+12 个月时间线）共 9 大模块。\n\n适合：跨境电商创业者、出海业务负责人、投资人、供应链企业。", "description_en": "Want to launch a cross-border e-commerce business but unsure how to choose target markets, set up logistics, handle compliance, or even write a business plan? Southeast Asia's six major markets differ widely — Shopee, Lazada, and TikTok Shop each have their own ecosystems, logistics options range from direct shipping to overseas warehouses to local warehouses, and tax and customs policies vary by country. A solid BP needs to cover an overwhelming number of dimensions.\n\nThis case demonstrates how to use FBS-BookWriter's long-form writing capability to generate a comprehensive Southeast Asia cross-border e-commerce BP covering: market analysis (six-country e-commerce market size and growth rate comparison), target market selection (Indonesia + Thailand PEST analysis), competitive landscape (differentiated positioning across three major platforms), business model (B2C direct + platform hybrid + supply chain architecture), operations strategy (product selection, pricing, TikTok influencers + Facebook ads + local KOLs), logistics solutions (cost analysis for three models), compliance and risk (tax, customs, IP, payments), financial projections (three-year P&L + Unit Economics + BEP analysis), and team & milestones (four-person core team + 12-month timeline) — nine major modules in total.\n\nBest for: Cross-border e-commerce entrepreneurs, overseas expansion leads, investors, supply chain companies.", "categories": [ "business" ], "tags": [ "跨境电商", "商业计划书", "东南亚出海", "BP", "市场分析" ], "tags_en": [ "Cross-border E-commerce", "Business Plan", "Southeast Asia", "BP", "Market Analysis" ], "artifact_type": "md", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.md", "artifacts": [ { "file": "output.md", "label": "跨境电商商业计划书" } ], "prompt": "为一家计划进军东南亚市场的跨境电商创业公司撰写完整的商业计划书。公司定位：家居日用品垂直品类跨境电商，首批目标市场为印尼和泰国。商业模式：B2C自营+Shopee/Lazada/TikTok Shop平台店铺混合模式，供应链架构为国内仓→海外仓→Last mile。营销策略：以TikTok达人带货+Facebook广告+本地KOL为主要获客渠道。融资需求：500万元人民币用于首批市场拓展。要求生成完整BP，覆盖执行摘要、东南亚六国市场分析（规模/增长率/互联网渗透率对比）、目标市场PEST分析、竞争格局分析、商业模式与供应链、运营策略、物流方案对比、合规与风险、三年财务预测（含Unit Economics和BEP分析）、核心团队与12个月里程碑。请用 FBS-BookWriter 的长文写作能力输出一份详尽的商业计划书。", "prompt_en": "Write a complete business plan for a cross-border e-commerce startup planning to enter the Southeast Asian market. Company positioning: a cross-border e-commerce business in the home & daily-goods vertical, with Indonesia and Thailand as the first batch of target markets. Business model: a hybrid of B2C self-operation and Shopee/Lazada/TikTok Shop platform stores; the supply chain architecture is domestic warehouse → overseas warehouse → last mile. Marketing strategy: customer acquisition primarily through TikTok influencer livestream selling + Facebook ads + local KOLs. Funding needs: RMB 5 million for the first wave of market expansion. The requirement is to produce a complete BP covering the executive summary, market analysis of six Southeast Asian countries (comparing size / growth rate / internet penetration), a PEST analysis of the target markets, competitive landscape analysis, business model and supply chain, operations strategy, logistics options comparison, compliance and risk, a three-year financial forecast (including Unit Economics and BEP analysis), the core team, and a 12-month milestone plan. Please use FBS-BookWriter's long-form writing capability to output a detailed business plan.", "skills": [ { "id": "fbs-bookwriter", "name": "fbs-bookwriter", "description": "福帮手出品 | 高质量长文档手稿工具链：书籍、手册、白皮书、行业指南、长篇报道、深度专题；支持联网查证（宿主允许时启用，离线自动降级）、S/P/C/B 分层审校、中文排版与 MD/HTML 交付。触发词：福帮手、福帮手写书skill、福帮手写书、写书、出书、写长篇、写手册、写白皮书、写行业指南、协作写书、定大纲、写章节、封面、插图、排版构建、导出、去AI味、质量自检、图文书、写报道、写深度稿、写特稿、写专题、写调查报道、写长文、拆书改写、海外本地化改写、爆款结构改写、激活原料、原料盘点、整理素材", "description_en": "Dual-channel long-form writing workflow for books, manuals, whitepapers, guides and reports with layered QC, 7-lock anti-drift anchors, and offline fallback.", "skill_id": "skill_2053083113738002432" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 249, "is_featured": false, "featured_order": 0, "quality_score": 93, "source": "official", "usage_count": 0, "created_at": "2026-04-22T00:00:00Z", "updated_at": "2026-06-16T18:10:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "scene-resume-diagnosis", "title": "简历诊断：46 分是怎么算出来的", "title_en": "Resume Diagnosis: How a 46 Is Computed", "subtitle": "复刻后端流水线逐字段扫描，三类诊断 + 六维评分", "subtitle_en": "A field-by-field pipeline scan: three diagnosis types, six scoring dimensions", "description": "「简历帮我看看」得到的往往是几句感觉式点评。诊断应该像体检：逐项扫、给依据、能复查。\n\n本案例把一份虚构的会计学学生简历（李慕华）跑过 1:1 复刻高顿后端的诊断流水线：52 个字段逐一判定，产出 21 项空白提醒（NO_FILL）、1 个缺失模块（NO_DATA）、3 条描述改进建议，六维评分合成总分 46、打败 50% 求职者，并给出修改优先级。\n\n适合：投递前想过一遍机检的学生、批量看简历的就业老师。", "description_en": "\"Take a look at my resume\" usually earns a few impressionistic comments. Diagnosis should work like a physical: scan every item, cite evidence, be repeatable.\n\nThis case runs a fictional accounting student's resume (Li Muhua) through a pipeline that replicates Gaodun's backend 1:1: all 52 fields judged one by one, producing 21 blank-field alerts (NO_FILL), 1 missing module (NO_DATA), 3 description suggestions, six-dimension scores totalling 46 — beating 50% of applicants — plus a fix-priority list.\n\nFor: students pre-screening before applying and careers staff reviewing at volume.", "categories": [ "education" ], "tags": [ "简历诊断", "简历评分", "求职" ], "tags_en": [ "resume diagnosis", "resume scoring", "job hunting" ], "artifact_type": "html", "preview": "output.html", "artifacts": [ { "file": "output.html", "label": "简历诊断 46 分报告" } ], "cover_image": "cover.png", "cover_mode": "landscape", "prompt": "帮我诊断这份简历（上传 PDF/DOCX）。要求：1) 按诊断流水线逐字段解析，产出 NO_DATA/NO_FILL/DESCRIPTION_SUGGEST 三类结果；2) 六维评分（完整度/经历描述/学习能力/专业能力/创新领导力/职场能力）用脚本真实计算，经历描述按量化规则分档；3) 报告含总分、打败百分比、逐模块诊断明细；4) 给出修改优先级：先补什么最回分。", "prompt_en": "Diagnose this resume (PDF/DOCX upload). Requirements: 1) parse field by field through the diagnosis pipeline, producing NO_DATA / NO_FILL / DESCRIPTION_SUGGEST results; 2) compute the six dimension scores (completeness, experience description, learning, professional, innovation-leadership, workplace) with the real script, banding descriptions by the quantification rule; 3) report the total, beat percentage, and per-module details; 4) end with fix priorities — what recovers the most points first.", "skills": [ { "id": "resume-diagnosis", "name": "简历诊断", "description": "1:1 复刻后端诊断流水线：3 类诊断 + 6 维评分", "skill_id": "skill_2089529938557247488" } ], "experts": [], "mcps": [], "creator": { "name": "上海高顿教育科技有限公司", "avatar": "", "badge": "official" }, "source": "official", "sort_order": 250, "is_featured": false, "featured_order": 0, "quality_score": 0, "usage_count": 0, "created_at": "2026-08-18T00:00:00+08:00", "updated_at": "2026-08-18T00:00:00+08:00", "is_recommended": false, "recommended_order": 0 }, { "id": "tech-outsourcing-contract", "title": "技术外包服务合同", "title_en": "Tech Outsourcing Service Contract", "subtitle": "输入合同要素，AI 自动生成结构完整的 Word 合同模板", "subtitle_en": "Enter contract elements, AI auto-generates a well-structured Word contract template", "description": "找律师起草一份技术外包合同动辄大几千，小公司和创业团队根本承受不起。自己写又担心条款不严谨——付款节点怎么设？验收标准写多细？知识产权归属有没有坑？违约金比例定多少合理？\n\n这个案例演示了输入一份技术外包合同的核心要素（甲乙方信息、项目内容、金额 38 万三期付款、4 个月开发周期、验收标准、质保条款等），AI 自动生成一份结构完整的正式合同 Word 文档：包含项目概况与技术要求、分期付款方案、里程碑交付计划、验收标准、12 个月质保条款、知识产权归属、保密条款、违约责任、争议解决等十二章完整条款。可直接作为谈判底稿使用。\n\n适合：创业公司签外包合同、技术负责人管理供应商、采购经理起草服务合同、自由职业者签项目协议。", "description_en": "Hiring a lawyer to draft a tech outsourcing contract costs thousands — unaffordable for small companies and startups. Writing it yourself risks incomplete clauses: How should payment milestones be structured? How detailed should acceptance criteria be?\n\nThis case demonstrates entering core contract elements (party information, project scope, 380K in three payment installments, 4-month development timeline, acceptance criteria, warranty terms) and having AI auto-generate a complete formal contract as a Word document: covering project overview, staged payments, milestone delivery, acceptance standards, 12-month warranty, IP ownership, confidentiality, breach penalties, dispute resolution — twelve complete chapters ready as a negotiation draft.\n\nBest for: Startups signing outsourcing contracts, tech leads managing vendors, procurement managers drafting service agreements, freelancers signing project agreements.", "categories": [ "office", "productivity" ], "tags": [ "合同模板", "技术外包", "法律文书", "协议生成" ], "tags_en": [ "Contract Template", "Tech Outsourcing", "Legal Document", "Agreement Generator" ], "artifact_type": "doc", "cover_image": "cover.png", "cover_mode": "landscape", "preview": "output.docx", "artifacts": [ { "file": "output.docx", "label": "技术外包服务合同（Word）" } ], "prompt": "生成一份技术外包服务合同 Word 文档（output.docx）。甲方为一家互联网公司（深圳市创新科技有限公司），乙方为一家软件开发外包商（广州极客软件有限公司）。\n\n项目概况：甲方委托乙方开发一套内部 OA 审批系统（请假、报销、采购三个模块），需对接企业微信和飞书。\n\n合同要素：\n1）总金额 ¥38 万元，分三期付款（签约 30%、中期验收 40%、终验 30%）\n2）开发周期 4 个月（2026.05.01 - 2026.08.31）\n3）验收标准：功能测试通过率 ≥ 98%，500 并发响应时间 ≤ 2s\n4）质保期 12 个月，期间 bug 免费修复\n5）知识产权归甲方所有\n6）保密条款：合同期及终止后 3 年内有效\n7）违约责任：延期交付每日按合同金额 0.5% 计违约金，上限 10%\n\n格式：含封面页、目录、页眉页脚、多级标题编号、甲乙方信息区与签字盖章区，条款结构严谨可直接打印签署。输出为正式 Word 文档（output.docx）。", "prompt_en": "Generate a technology outsourcing service contract Word document (output.docx). Party A is an internet company (Shenzhen Innovation Technology Co., Ltd.), Party B is a software outsourcing vendor (Guangzhou Geek Software Co., Ltd.).\n\nProject overview: Party A commissions Party B to develop an internal OA approval system (leave, reimbursement, procurement modules), integrating with WeCom and Feishu.\n\nContract elements:\n1) Total amount ¥380,000, paid in three installments (30% on signing, 40% on mid-term acceptance, 30% on final acceptance)\n2) Development period 4 months (2026.05.01 - 2026.08.31)\n3) Acceptance criteria: functional test pass rate >= 98%, response time <= 2s under 500 concurrent users\n4) Warranty period 12 months, with free bug fixes during it\n5) Intellectual property belongs to Party A\n6) Confidentiality: effective during the contract term and for 3 years after termination\n7) Breach liability: liquidated damages of 0.5% of the contract amount per day of late delivery, capped at 10%\n\nFormat: cover page, table of contents, headers/footers, multi-level heading numbering, Party A/B information section and signature/seal section, rigorous clause structure ready to print and sign. Output as a formal Word document (output.docx).", "skills": [ { "id": "minimax-docx", "name": "Word 文档生成", "description": "Word 文档生成与编辑", "description_en": "Generate and edit Word documents from scratch", "skill_id": "skill_2053082396193849344" } ], "experts": [], "mcps": [], "scenario": "", "task_mode": "", "creator": { "name": "WorkBuddy 官方", "name_en": "WorkBuddy Official", "avatar": "", "badge": "official" }, "sort_order": 250, "is_featured": false, "featured_order": 0, "quality_score": 92, "source": "official", "usage_count": 0, "created_at": "2026-04-21T10:00:00Z", "updated_at": "2026-06-16T23:30:00Z", "is_recommended": false, "recommended_order": 0 }, { "id": "email-newsletter-template", "title": "营销邮件模板", "title_en": "Marketing Email Newsletter Template", "subtitle": "兼容主流邮件客户端，打开即用的精美邮件", "subtitle_en": "Email-client compatible, ready-to-use beautiful newsletter template", "description": "做邮件营销最头疼的就是 HTML 兼容性：Outlook 不支持 flexbox，Gmail 会剥离