tomjiu
- 88 skills
- 0 followers
- 10 hours ago last updated
- ▌ Godot Dev · tomjiuUse when working on Godot 4.x projects (GDScript/C#, scenes, signals, export, debugging). Prefer local Godot skill packs under sources/packs before inventing patterns.
- ▌ Exa Search · tomjiuUse when doing AI-native web search or finding similar pages via Exa API (EXA_API_KEY). Prefer for research/code-oriented retrieval; use Tavily Hikari or Keenable for general agent search MCP.
- ▌ Keenable CLI · tomjiuInstall and use the Keenable CLI for fast web search and page fetching from the terminal. Use when you need to search the web, fetch a page as clean markdown, or configure Keenable as the search provider (MCP) for AI clients like Claude Code or Cursor.
- ▌ Source Index · tomjiuUse when choosing MCP servers, skills packs, tools, or GitHub orgs — look up local catalog and pre-cloned mirrors first; fall back to listed upstream repos only after local miss. Also use to answer "where is X installed locally".
- ▌ Dfir Workflow · tomjiuUse for digital forensics / incident response workflows — memory, disk, timeline. Defensive investigation. Prefer Volatility3, Velociraptor, Plaso, Autopsy from catalog URLs.
- ▌ Go Performance · tomjiuUse when optimizing Go services or diagnosing CPU/memory/latency issues — pprof, benchmarks, escape analysis, tracing. Measure first; prefer official Go diagnostics.
- ▌ TS Performance · tomjiuUse when optimizing TypeScript/Node or frontend performance — profiling, bundle size, async bottlenecks, React render costs. Measure with real tooling before rewriting.
- ▌ Rust Performance · tomjiuUse when optimizing Rust code or investigating slow builds/runtime — criterion benches, profiling, release flags, alloc pressure. Measure first; prefer official perf guidance.
- ▌ Firmware Analysis · tomjiuUse for firmware image analysis (router/IoT dumps) — binwalk extract, filesystem recovery, light emulation pointers. Authorized hardware/firmware only. Prefer official tool docs from local catalog.
- ▌ Repository Skills · tomjiuUse when a needed skill or MCP is not in local selected/packs mirrors — search listed upstream repos and awesome lists, download into sources/, review for malice and prompt injection before promoting to selected or enabling MCP.
- ▌
- ▌ Using Superpowers · tomjiu bundleUse only when explicitly requested (e.g. $using-superpowers) or when the user asks to follow the Superpowers methodology; do not auto-load for normal coding tasks
- ▌ Web Search Hikari · tomjiuUse when the agent needs web search, extract, crawl, map, or research via Tavily Hikari pool (tvly-hikari CLI or tavily_hikari MCP). Prefer this over inventing search APIs.
- ▌ Automation Workflow · tomjiuUse when building batch automation, agent pipelines, scheduled jobs, or multi-step unattended scripts. Prefer durable plans, checkpoints, and authorized scopes only.
- ▌ PDF · tomjiu bundleComprehensive PDF manipulation toolkit for extracting text and tables, creating new PDFs, merging/splitting documents, and handling forms. When you need to fill in a PDF form or programmatically process, generate, or analyze PDF documents at scale.
- ▌ DOCX · tomjiu bundleComprehensive document creation, editing, and analysis with support for tracked changes, comments, formatting preservation, and text extraction. When you need to work with professional documents (.docx files) for: (1) Creating new documents, (2) Modifying or editing content, (3) Working with tracked changes, (4) Adding comments, or any other document tasks
- ▌ PPTX · tomjiu bundleRead, create, and edit PowerPoint PPTX/PPT presentations. Use when the user explicitly mentions "ppt" or "PPT" (e.g. "做个PPT", "make a ppt"). If the user only says "演示文稿" or "presentation" without mentioning PPT, route to html-deck instead. Covers: parsing, summarizing, extracting content, inspecting themes/layouts, creating new decks with PptxGenJS, and editing existing PPTX while preserving formatting.
- ▌ XLSX · tomjiu bundleUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.
- ▌ Offensive Methodology · tomjiuUse when planning authorized penetration testing or offensive assessment methodology — phases, tool map (download/install/run pointers), engagement hygiene. Distilled from community awesome lists. Does not include exploit payloads, shellcode, or copy-paste attack strings. Prefer for knowing which tools/chains exist.
- ▌ Feedback · tomjiuUse only for explicit TRAE feedback submission intent: when the user explicitly invokes /feedback, the host injects Use Skill: feedback, or the user clearly asks to submit/send/record/report their own current TRAE issue as feedback. Do not activate merely because the user mentions feedback, bug, complaint, feedback page, feedback API, or user feedback data. If uncertain, do not activate.
- ▌ Authorized Analysis Scope · tomjiuUse before reverse engineering, security testing, firmware, mobile instrumentation, or automation against live systems — write explicit authorization and scope; refuse unclear targets.
- ▌ HTML Deck · tomjiu bundleCreate stunning, animation-rich HTML slide decks from scratch. Outputs a self-contained HTML file that opens in any browser, agent-friendly, visually impressive and pixel-perfect across all platforms. Use when the user says "演示文稿", "presentation", "slides", or "deck" WITHOUT explicitly mentioning "PPT" or "ppt" — if the user says "PPT"/"ppt", route to the `pptx` skill instead. Covers building a brand-new presentation from a topic, outline, document, or Markdown — i.e. no existing .pptx file is involved. The deliverable is always HTML.
- ▌ Dynamic UI · tomjiu bundleShow visual content inline alongside your text response — diagrams, charts, interactive demos, comparisons. Use only when a compact visual makes the answer clearer. Not for websites, apps, reports, dashboards, or slides.
- ▌ HTML Report · tomjiu bundleCreate any self-contained HTML deliverable except slides — research reports, whitepapers, PRDs, dashboards, portfolios, resumes, email templates, data visualizations, and more. Use when the user wants to produce a polished, visually designed HTML page or multi-page site for any purpose. The final deliverable is a self-contained directory (<name>.html + assets + _shared) deployed to the user's workspace with zero external dependencies.
- ▌ Solo Design · tomjiu bundleDesign website pages, UI screens, prototypes, page-level visual systems, and existing `.design` project mutations. Use for websites, landing pages, UI prototypes, redesigns, theme changes, 1:1 restoration, editable static layouts, and complex canvas projects. Route bitmap image edits to solo-image-edit, bitmap-first poster/banner/KV/cover generation to solo-graphic-generation, and reusable Design Library / Design System / token architecture creation to design-library-creator.
- ▌ Trae Doctor · tomjiu bundleScenario-indexed TRAE product manual for interpreting verified TRAE evidence during environment diagnosis. Currently supports disk diagnosis only, including storage ownership, directory purpose, accounting, and cleanup safety on local products and Remote SSH hosts. Always assess the whole disk first and rank all major contributors neutrally. Use TRAE-specific facts only after independent evidence establishes TRAE as an actual analysis target under the same criteria as other contributors, or when the user's natural-language request explicitly asks for TRAE-specific analysis. A product-generated TRAE-doctor slash command alone is not such a request. Do not prioritize TRAE merely because this Skill was selected or a TRAE path exists. This Skill supplies product facts only; it does not choose diagnostic targets, prioritize or excuse TRAE, or define conclusions or response format.
- ▌ Reverse Engineering Workflow · tomjiuUse for reverse engineering, malware analysis, Android APK static analysis, or tool selection (Ghidra, Frida, JADX, radare2). Index-first; load domain packs on demand. Authorized analysis only.
- ▌ Skill Creator · tomjiuMANDATORY tool for creating SKILLs - MUST be invoked IMMEDIATELY when user wants to create/add any skill
- ▌ Research Guide · tomjiu bundlePrimary research & analysis scenario Skill. Activate when the user's intent involves searching, looking up facts, researching, investigating, analyzing, comparing products or technologies, conducting competitive analysis, writing a research report, or producing any evidence-based deliverable. Provides source-hierarchy rules, cross-validation methodology, search paradigms, and sub-scenario routing. For deliverable generation (research reports, competitive analyses), routes to specialized references in the references/ directory.
- ▌ Solo Image Edit · tomjiu bundleEdit existing bitmap images with image-to-image generation and place every kept result on a `.design` canvas. Use for structured Design image comments or explicit pixel-level edits with an original image.
- ▌ Trae Browseruse · tomjiuBrowser automation guide. Invoke when user wants to browse websites, access URLs, scrape web content, test frontend UI, perform any browser interaction, or navigate to a specific URL and perform multi-step actions on it (click, verify elements, fill forms).
- ▌ Doc Writing Guide · tomjiu bundlePrimary skill for document and content writing. Triggers on PRD, product requirements, MVP spec, feature spec, tech proposal, research report, competitive analysis, manuals, and any structured written deliverable. Governs intent interpretation, genre & format selection, writing-style calibration, content structuring, visual generation guidance, and sub-scenario routing to specialized references (e.g., PRD routes to references/prd-document.md).
- ▌ Trae Computer Use · tomjiuControl local apps through Computer Use. Use for tasks that require reading or operating app UI by clicking, typing, scrolling, dragging, pressing keys, or setting values.
- ▌ Trae Computer Use Ptc · tomjiuComputer Use Guide for Windows. You MUST invoke the skill first and strictly follow its instructions when operate app UI, inspect app UI state, or perform desktop interactions through Computer Use; Prefer purpose-built MCPs (like browseruse) or CLIs when available unless the user explicitly requires Computer Use.
- ▌ Design Library Creator · tomjiu bundleCreate, expand, and refine professional Design Libraries / Design Systems with structured token architecture. ONLY invoke when user EXPLICITLY mentions design-system terminology: 'Design Library', 'Design System', '设计系统', '设计库', 'token system', 'design tokens', '组件库规范', or provides a parsed design-spec bundle (ZIP/Figma export). Also covers creating a new theme / color theme as a design-system deliverable (创建主题/新建暖色主题/深色主题等 — tokens + palette + component styling). Do NOT route here for: '素材库', '全套素材', '素材合集', '物料库', '生成素材', '做一套设计', '做几张图', '设计一套页面', '帮我出图', page/poster/landing-page creation, or any visual-asset generation intent where the user does not demonstrate knowledge of design-system concepts. Those belong to solo-design.
- ▌ Digital Avatar Creator · tomjiuMANDATORY tool for creating digital avatar skills that run as autonomous SubAgents. MUST be invoked IMMEDIATELY when user wants to create a digital avatar, persona, virtual role, or specialized AI assistant.
- ▌ Trae Product Knowledge · tomjiuUse this skill for TRAE brand identity and official product knowledge questions, including: who you are, what TRAE is, product differences, TraeCode / TraeWork / TraeCode CLI / TraeCode Plugin entry points, supported capabilities, MCP, Skills, official documentation, and official product links. Do not use it for ordinary coding questions or repository-specific implementation work.
- ▌ Solo Graphic Generation · tomjiu bundleGenerate bitmap-first static visual assets such as posters, banners, KV/key visuals, covers, invitations, social cards, illustrations, product images, and image batches without HTML. Confirm the output size before generation; if an exact size is unsupported, expose the constraint and wait for the user to choose a replacement instead of substituting one.
- ▌ Trae Browseruse External · tomjiuAutomate tasks in the user's own browser (Chrome on their machine). Invoke when the user says things like 'use my browser', 'open in my Chrome', 'use the browser on my computer', or wants to browse/interact/test web pages in their local Chrome.
- ▌ Trae Code Mode Orchestrator · tomjiu bundleCode Mode (Exec) usage patterns and applicable scenarios. Covers parallel fan-out, pipeline with JS transforms, conditional branching, loop-until-condition, and multi-source aggregation. Trigger when the task benefits from orchestrating multiple tool calls in a single JavaScript script rather than sequential direct calls.
- ▌ Skill · tomjiu bundle将生成内容发布为在线链接,或取消发布已上线的应用。适合网站、小游戏等交互式内容,支持任意可作为单端口 HTTP 服务运行的项目(静态站点 / PDF / Node.js / Python / Go)。当用户想把本地项目部署、发布、上线、生成分享链接、在云端预览或下线分享链接时使用。
- ▌ Skill 2 · tomjiu bundle当要写/整理在线文档、建数据表增删改查、导入 CSV·Excel、做看板/dashboard/运营页/汇报页、md 转网页或演示 HTML 发布、建目录、上传下载网盘文件、审阅修订、分享协作,以及提到资料库/知识库/网盘/空间/workbuddy.cn/space 链接时使用。资料库是 WorkBuddy 原生的内容管理、分享协作与轻发布模块,覆盖在线文档(doc)、网页链接(link)、结构化数据表(database)、数据/汇报页面(page)、网盘文件(drive)与空间管理(manage);凡说资料库/知识库或出现 space 链接、未显式点名腾讯文档/飞书/Notion/乐享/金山文档/iWiki 等外部产品时一律走本 skill,默认落原生库、不反问存到哪个产品。
- ▌ Ardot Poster · tomjiu bundleUse this skill for Ardot canvas design tasks whose deliverable is a visual poster — posters, flyers, billboards, banners, event posters, promotional single-page graphics, e-commerce banners, brand visuals, logos, icons & illustrations rendered on the Ardot canvas. Trigger phrases: design a poster, create a banner, make a flyer, event poster, product launch poster, e-commerce banner, 设计海报, 海报设计, 宣传单, 横幅, banner, 封面, 双十一Banner, 活动海报, 图标设计, 品牌设计, 插画, Logo 设计, VI 设计, 视觉识别系统. Pairs with ardot-design-core (injected alongside). NOT for full UI screens (ardot-ui-design), slides (ardot-slides), or PowerPoint .pptx files (pptx skill).
- ▌ Ardot Slides · tomjiu bundleUse this skill for Ardot canvas design tasks whose deliverable is a slide deck / presentation design (NOT a PowerPoint .pptx file). Covers presentations, decks, pitch decks, keynote-style slides, launch-event slides, roadshow decks, proposal/pitch slides rendered as Ardot design canvases. Trigger phrases: design a presentation, create a deck, generate slides, design slides, create a pitch deck, design a keynote, slide deck mockup, presentation mockup, pitch deck design, 幻灯片设计稿, 演示文稿设计, 发布会幻灯片, 路演 PPT 设计稿, keynote 设计, 提案稿, 宣讲稿, 制作 PPT, 做一份幻灯片. Pairs with ardot-design-core (injected alongside). IMPORTANT boundary: this skill OWNS slide/deck/幻灯片 requests whose output is a DESIGN (Ardot canvas, 设计稿, mockup). Defer to the pptx skill ONLY when the user explicitly requires a PowerPoint .pptx file (导出 pptx / 生成 PPT 文件 / references an existing .pptx). Words like 'slides', '幻灯片', 'PPT' alone do NOT trigger pptx.
- ▌ Skill Creator 2 · tomjiu bundleGuide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends CodeBuddy's capabilities with specialized knowledge, workflows, or tool integrations.
- ▌ Expert Manager · tomjiu bundle专家包的全生命周期运营:转化(从开源仓库/本地项目创建专家包)、修改已有专家、合规检查、批量更新、质量审查。 触发词:创建专家、转化专家、转成专家、生成专家包、导入专家、convert expert、修改专家、编辑专家、更新专家、modify expert、检查专家、审查专家包、专家合规、专家运营、expert ops。
- ▌ Ardot UI Design · tomjiu bundleUse this skill for Ardot canvas design tasks whose deliverable is a UI / interface design — web pages, web apps, dashboards, landing pages, marketing sites, official sites, homepages, mobile app screens (iOS / Android / 小程序), forms, tables, components, and design systems / design tokens / component libraries. Trigger phrases: generate/create/design a page, design a screen, create a landing page, make a dashboard, design a login screen, build a UI, generate homepage, design a form, design an App, WebApp, 设计页面, 创建界面, 生成页面, 生成网站, 设计App, 做一个页面, 画一个页面, 设计后台, 组件库规范, 设计系统, Design System, Design Token, 移动端, 手机App, 小程序. Pairs with ardot-design-core (injected alongside). NOT for slides/posters/design-to-code (separate Ardot skills) and NOT for PowerPoint .pptx files (pptx skill).
- ▌ Wb Finance Skill · tomjiu bundle金融/投资/股票/基金/ETF/板块/指数/宏观/外汇/大宗商品/财报/估值/持仓/交易/量化/选股/期权/衍生品/投行建模/技术指标/行情等——金融场景总入口,优先级高于其他金融skill。涉及任一上述领域和相关金融场景时(包括字面未明说但本质相关),都务必**第一时间优先加载本skill**。
- ▌ Ardot Design Core · tomjiu bundleFoundational workflow and hard rules shared by ALL Ardot canvas design tasks (UI screens, slides, posters, livestream posters, design systems, design-to-code). This skill carries the canvas schema, editing principles, the standard step-by-step workflow, visual style guide, and composite effects. It is injected automatically alongside a domain-specific Ardot skill (ardot-ui-design / ardot-slides / ardot-poster / livestream-poster / ardot-design-to-code). It does NOT cover any single deliverable type on its own — always pair it with the matching domain skill. Not for PowerPoint .pptx file output (that is the pptx skill).
- ▌ Livestream Poster · tomjiu bundle生成直播宣传海报:支持风格文件/视觉大类和用户风格参考图两种工作流,收集人物、Logo、文案与比例后,用 ImageGen 生成底图并在 Ardot 画布完成排版。 当用户提到「直播海报」「直播宣传图」「直播预告海报」「线上直播海报」「livestream poster」,或引用直播海报风格文件时使用。
- ▌ Recommend Experts · tomjiu bundleWhen a task needs professional judgment, deep research, a specialist role, or multi-role collaboration and no Expert is selected in the current session, search for real Experts or Expert Teams and recommend them with an inline card.
- ▌ Ardot Design Router · tomjiu bundleDispatcher for WorkBuddy's Ardot design assistant. Use this when a design task is in Craft/design mode but the specific deliverable type is not yet determined from UI signals. It classifies intent (UI screen / slides / poster / livestream poster / design-to-code) and tells you which domain Ardot skill to load, then defers to that skill plus ardot-design-core. Carries the always-on hard rules for canvas design. Not for PowerPoint .pptx file output (pptx skill).
- ▌ Ardot Design To Code · tomjiu bundleUse this skill for Ardot canvas tasks that convert a design into frontend code, or extract a design system / style guide from a website. Covers: design-to-code, design → HTML/CSS/JS, export as webpage, pixel-perfect reproduction, generate an Application from a design, slide transitions, responsive scaling; and website → design-guide / design-token extraction. Trigger phrases: convert design to code, design to HTML, export as webpage, pixel-perfect reproduction, design to App, generate Application, create design system from website, extract design tokens, 设计稿转代码, 转为前端代码, 生成HTML, 导出为网页, 一比一还原, 复刻设计稿, 设计稿出码, 设计稿转应用, 转应用, 网站风格转设计稿, 提取设计风格, 生成设计指南, 提取设计 token. Pairs with ardot-design-core (injected alongside).
- ▌ Recommend Connectors · tomjiu bundleWhen a user task requires an external app, service, API, MCP server, authorization, or third-party data and no connected tool covers it, search for real Connectors and recommend them with an inline card.
- ▌ Tencent Docs Routing · tomjiu bundleLoad this Skill before handling local Office/WPS files such as doc/docx/dot/wps/wpt, xls/xlsx/xlt/csv/tsv, or ppt/pptx/pps/pot, and before creating a new local document, spreadsheet, or presentation. Load first whenever the user uploads, mentions, previews, edits, or asks to create such a local file, to decide whether to use the local office editing Skill, the local spreadsheet agent, the spreadsheet generation Skill, or the PPT generation Skill. Text-based outputs such as speech scripts, talking points, copywriting, summaries, reports, or meeting minutes should default to local DOCX unless the user explicitly specifies another local format such as markdown or plain text. Other file types are out of scope.
- ▌ Tencent DOCX · tomjiu bundle专业 Word 文档(.docx)创作与美化助手。用于生成研报、论文、公文、合同等垂类专业 Word 文档,或对已有 .docx 进行排版美化。当用户需要写文档、生成 Word、排版或美化 .docx 时调用。
- ▌ Geo Map Compliance Guard · tomjiu bundleALWAYS TRIGGER map compliance skill for any map generation, visualization, routing, or location service request to enforce strict China map data compliance rules, including approved data-source whitelist, key safety controls, territorial integrity requirements, privacy protection, and Tencent Maps GL JS scenario-based integration guidelines.
- ▌ Tencent Local Office Edit · tomjiu bundle通过本地 editor_sdk 实时读写本机磁盘上的 Office/WPS 类型文件——文件打开后用户在编辑器中实时可见,编辑所见即所得,保存用 save_file 即可、不要主动 close_file。适用于本地 doc/docx/dot/wps/wpt、xls/xlsx/xlt/csv/tsv、ppt/pptx/pps/pot 等文件的打开、编辑、保存与新建;含宏格式可打开但宏不执行,et/ett/dps/dpt 不支持。按 doc/sheet/slide 三类调用对应工具,调用任一编辑工具前先用 edsdk.py schema 查它单个的参数、不要凭摘要猜参数。所有操作经本目录 edsdk.py(python3 edsdk.py list/schema/call)封装,端点自动发现、上层无需关心底层细节。触发词:本地文档、本机 docx/xlsx/pptx/wps、Office/WPS 文件、editor_sdk、edsdk.py、本地编辑、实时编辑。
- ▌ 3d · tomjiu bundle3D模型生成和基于模板的视频特效能力。仅支持:文生3D模型、图生3D模型、基于模板的图片视频特效(video-fx)。 适用于用户需要 AI 创作 3D 模型或对图片应用模板特效(如拥抱、变身、万物归尘等动效模板)的场景。 当用户提出以下意图时触发:生成/制作3D模型、对图片应用特效模板或动效模板。 注意:本技能不处理图片生成、视频生成任务。生成图片请用 ImageGen 工具,生成视频请用 VideoGen 工具。
- ▌ Marketplace Skill Installer · tomjiu bundle在 WorkBuddy 对话中通过一句话从推荐市场(BuiltinMarket)搜索并安装 Skill。 触发词:安装 skill、安装技能、install skill、添加技能、装个 X 技能、帮我安装 X、find skill、install marketplace skill、search skill。
- ▌ Weixinpay Pay · tomjiu【微信支付官方】微信支付AI专属卡支付能力。普通支付在商户下单后由系统自动触发工程化支付卡片,无需智能体介入;本技能主要指导「重新支付」:用户取消/关闭支付后想对同一笔订单再付一次时,智能体须先询问并确认订单,再用上次相同凭据(payUrls 或 WeixinPay-Required 的值)调用 mcp__weixinpay__weixinpay_retry_pay 重新触发支付,无需重新下单。关键词:重新支付、再次支付、重新付款、重付、取消后再付一次、微信AI支付、微信Agent支付。
- ▌ Browser Automation · tomjiuThis skill should be used when the user asks about browser automation, testing web pages, scraping content, filling forms, taking screenshots, or monitoring console/network activity. Activates for E2E testing, web scraping, form automation, or debugging web applications.
- ▌ Doc Typeset · tomjiu bundle消费 design tokens + 文档内容,输出排版美化的 HTML。 内置 7 种垂类模板(合同/学术论文/公文/商务报告/会议纪要/研报/年报)和 4 个装饰组件,所有样式通过 CSS 变量引用 Token,禁止裸值。 在 doc-formatter 流水线中每当需要将文本内容转化为结构化美化 HTML 时,必须使用此 skill; 不要自己写 HTML 或内联 CSS——doc-typeset 是内容→HTML 的唯一标准转换路径。
- ▌ HTML Review · tomjiu bundleHTML 质量门禁 skill(路线 A + 路线 B 共享)。对 doc-typeset(路线 A)或 doc-edit(路线 B)输出的 HTML 进行 5 维度质量检测(design-token 合规性、结构完整性、排版合理性、文体契合度、装饰使用合理性),输出结构化检测报告,不通过则打回上游做一次定向修正(不循环)。 在 doc-formatter 路线 A 和路线 B 流水线中,每次 HTML 输出后必须调用此 skill; 当需要检查 HTML 是否存在裸样式值、标题跳级、缺少必需文体元素、装饰过度等问题时,都应使用此 skill 而非手动检查。
- ▌ Design Token · tomjiu bundle按文档类型(genre)选择主题并输出标准化设计令牌(Design Tokens),驱动 doc-typeset skill 的所有样式决策。 支持公文、学术论文、商务报告、创意营销、通用五大场景,输出含 typography/color/spacing/layout 的 W3C DTCG 格式 JSON 和 CSS 变量映射。 当 doc-formatter 流水线需要确定排版风格、需要字体/颜色/间距/页边距决策时,必须调用此 skill; 不要跳过它直接写裸值——所有样式决策的唯一入口就是这里。
- ▌ HTML To DOCX · tomjiu bundle将 HTML 字符串或文件高保真转换为 Microsoft Word (.docx) 文档。 支持 CSS 变量预处理、10+ 种 HTML 元素精确映射、14 种 CSS 属性映射、中文字体、4 种装饰组件、TOC 目录、页眉页脚、图片嵌入。 提供 CLI(subprocess)和 Python API 双模式调用;在需要将 HTML 文档转为可编辑 Word 文件时使用此 skill。
- ▌ Tencent PPTX · tomjiu bundle创建专业的 PowerPoint 演示文稿。适用于根据主题、大纲、文档、数据或参考材料生成完整 .pptx;在新建演示文稿时参考上传PPTX 的视觉风格;也可基于材料或旧 PPT 内容重新生成一版演示文稿。
- ▌ Shared Styles · tomjiuBind/unbind shared styles via textStyleId/fillStyleId/strokeStyleId/effectStyleId (bare nodeId or Style:key,ver; never `$`)
- ▌ Brief Compose · tomjiu短篇文档创作(目标篇幅 <1000 字):从零撰稿 → 生成 HTML → html-to-docx 转 .docx → present_files 预览。
- ▌ Tencent Docs Sheet Generation · tomjiu bundle从零生成 Excel/xlsx 工作簿。当用户请求"创建/生成/新建/做一份/from scratch" 一个 XLSX 文件、且**没有源 .xlsx/.csv 文件**时使用。支持纯文字需求,也支持 以 pdf/docx/pptx 附件作为内容参考(由 `extract.py` 抽取为 markdown)。 本 skill 引导主代理读设计准则后直接写 openpyxl 脚本生成本地 xlsx,并用 recalc 校验公式零错误。
- ▌ Format Extract · tomjiu bundle把 .docx 文档转为语义化 HTML + 提取内嵌图片(标题层级、表格样式、段落缩进、字体颜色等结构信息),供调用方做后续格式分析;也可独立评估 HTML 转换质量。适用场景(**技术性底层调用**):"docx 转 HTML"、"提取 docx 结构为 HTML"、"HTML 转换质量评估"。本 Skill 是底层格式提取工具,**不自行判断是否应被触发**、**不猜测输出位置**——是否调用、何时调用、产物落在哪一律由上层调用方(如 doc-formatter)决定,调用时必须显式传入 `--output-dir`。用户关于"文档排版 / 套模板 / 参考这份 docx 生成"等需求,应由 orchestrator 统一编排入口接管,**不要旁路直点本 Skill**。
- ▌ General Writer · tomjiu bundleL1 通用写作兜底专家。覆盖公文、周报、方案、邮件、文案、散文、新媒体稿件等场景,采用 7 维质量评分框架(事实/逻辑/结构/语言/风格/受众/洞察)与 10 种文体适配矩阵。当 L0 路由未命中任何 L2 专家时作为兜底触发。适用于没有特定领域严格规范的通用写作任务。
- ▌ Weixinpay Feedback · tomjiu【微信支付官方】问题反馈。当用户使用微信AI支付/AI专属卡的开通/绑定或支付/管理过程遇到异常或错误(尤其是同一环节连续/反复报错、重试仍失败),主动引导用户上报问题。开通/绑定过程的问题通过反馈收集表链接反馈;支付/管理问题或用户直接想要反馈时通过工具反馈。关键词:微信AI支付反馈、Agent支付反馈、AI专属卡反馈、微信支付反馈。
- ▌ Weixinpay Register · tomjiu【微信支付官方】用于用户开通/绑定微信AI支付或者AI专属卡功能,或查询开通/绑定状态。用于用户要开通、绑定、激活在 AI 对话中使用微信支付的能力。也用于回答用户"我是否已开通、如何查看或管理微信AI支付"等开通状态咨询。关键词:微信支付AI专属卡、微信AI支付、开通微信支付、绑定微信支付、激活AI支付、开通/绑定AI专属卡、绑定智能体钱包、是否开通AI支付、是否开通AI专属卡、查询开通状态、管理AI支付、管理AI专属卡。
- ▌ Design Variables · tomjiuBind/unbind design variables (tokens) on node props with `$:<Set>:<Name>` or `$<id>`; create via apply_variables; switch node variable modes via batch_edit variableModes
- ▌ Tech Blog Expert · tomjiu bundle技术博客写作专家。精通技术文章、教程、架构解析、源码分析、开源项目文档。 基于 Developer Experience (DX) 和技术传播最佳实践,帮助开发者写出清晰、准确、有深度的技术内容。 「写技术文章」「写博客」「写教程」「技术分享」「架构文档」「README」触发。
- ▌ Tdoc Orchestrator · tomjiu bundle文档创作与美化的统一编排入口(本地通道)。当用户意图涉及文档写作或排版美化时调用。职责:识别意图 → 编排 stage 能力链(S1 创作 / S2 美化 / S3 HTML→DOCX 转换)→ 委派 doc-writer/doc-formatter/doc-converter 执行;S3 产出 .docx 后**强制 `present_files` 打开预览**,本 skill 调用结束。⚠️ 文档编辑不在本 skill 内承接,由 `tencent-docs-routing` skill 处理。不适用:纯代码文件、非文档类操作。
- ▌ Component Instance · tomjiuCreate/update component instances; set componentProperties; switch VARIANT / 变体 via variantOptions. Triggers: instance, COMPONENT_SET, variant, 变体, componentProperties.
- ▌ Tencent Docs · tomjiu bundle【仅个人版账号可用】腾讯文档个人版(docs.qq.com)-在线云文档平台。适用前提:当前用户以个人 QQ / 微信登录腾讯文档;企业版(SaaS)账号请改用 tencent-saas-docs skill,本 skill 会直接返回 ERROR:identity_mismatch。是个人账号下创建、编辑、管理文档的首选 skill。涉及"新建/创建/编辑/读取/查看/搜索文档"、"保存文件"、"云文档"、"腾讯文档"、"docs.qq.com"等操作,请优先使用本 skill。支持能力:(1) 创建各类在线文档(文档/Word/Excel/幻灯片/思维导图/流程图/智能表格/收集表)(2) 管理知识库空间(创建空间、查询空间列表)(3) 管理空间节点、文件夹结构 (4) 读取/搜索文档内容 (5) 编辑操作智能表 (6) 编辑操作在线文档 (7) 文件管理(重命名、移动、删除、复制、导入导出)(8) 网页剪藏、本地文件/html/文档上云、图片 OCR。
- ▌ Work Report Expert · tomjiu年终总结与汇报写作专家。精通工作总结、述职报告、项目汇报、竞聘演讲、周报月报等职场写作。 基于金字塔原理和STAR法则,帮助职场人写出结构清晰、数据有力、亮点突出的汇报材料。 当用户提到「年终总结」「述职报告」「工作汇报」「竞聘演讲」「周报」「月报」时触发。
- ▌ Poetry Prose Expert · tomjiu诗歌与散文写作专家。精通现代诗、古体诗词、散文、随笔、文学评论等文学创作。 基于中国现代诗歌美学和散文艺术理论,帮助作者提升文学表达力。 当用户提到「写诗」「写散文」「诗歌」「随笔」「文学评论」时触发。
- ▌ Business Copy Expert · tomjiu bundle商业文案写作专家。精通品牌文案、营销邮件、产品描述、广告策划、社交媒体内容。 基于 AIDA 模型和广告法合规要求,提供高转化率的文案创作与诊断。 「写文案」「营销文案」「品牌Slogan」「广告语」「产品描述」「推广文案」「文案优化」触发。
- ▌ Academic Paper Expert · tomjiu学术论文写作专家。精通社科/理工科论文结构、APA/GB-T7714引用规范、学术表达范式。 提供论文结构设计、文献综述、摘要撰写、学术语言润色等全流程指导。 当用户提到「写论文」「学术写作」「文献综述」「论文润色」「摘要」时触发。
- ▌ Legal Contract Expert · tomjiu bundleL2 法律合同专家。覆盖各类合同、协议、条款、契约的起草与审查,包含必备条款完整性检查(标的/价款/违约/争议解决/生效条件)、权利义务对称性审核、高风险点防范(不可抗力/知识产权归属/保密期限)。采用严格 Critic 审查,涉及法律责任任务**强制**走 per-section 模式分章审查。触发关键词:合同、契约、协议、条款、甲乙方、违约金、不可抗力、争议解决。
- ▌ Science Writing Expert · tomjiu科普写作专家。精通面向大众的科学解释、科技评测、深度报道、科普文章等。 基于科学传播理论和费曼学习法,将复杂科学概念转化为人人能懂的精彩内容。 当用户提到「科普文章」「科学解释」「科技评测」「深度报道」「科普」时触发。
- ▌ Tencent Saas Docs · tomjiu bundle【仅企业版账号可用】腾讯文档企业版(saas.docs.qq.com)-在线云文档平台。适用前提:当前用户已开通腾讯文档企业版(SaaS / OneID);个人 QQ / 微信账号请改用 tencent-docs skill,本 skill 会直接返回 ERROR:identity_mismatch。是企业账号下创建、编辑、管理文档的首选 skill。涉及"新建/创建/编辑/读取/查看/搜索文档"、"企业文档"、"团队文档"、"saas.docs.qq.com"等操作,请优先使用本 skill。支持能力:(1) 创建各类在线文档(文档/Word/Excel/幻灯片/思维导图/流程图/智能表格/收集表)(2) 管理知识库空间(创建空间、查询空间列表)(3) 管理空间节点、文件夹结构 (4) 读取/搜索文档内容 (5) 编辑操作智能表 (6) 编辑操作在线文档 (7) 文件管理(重命名、移动、删除、复制、导入导出)(8) 网页剪藏、本地文件/文档上云。
- ▌ Stock Research Report Expert · tomjiu bundleL2 证券/行业研究报告专家。生成专业的行业深度报告、个股研究、动态点评等金融研究文档。内置标准化报告结构框架,覆盖深度研报、常规点评、商业计划书、咨询交付物四档体量。触发关键词:行业研究、深度报告、个股研究、研报、券商报告、产业分析、行业跟踪、投资分析。
- ▌ Generate Fillable Contract HTML · tomjiu生成适于 HTML 转 DOCX 的中文待填合同、报价单和授权委托书 HTML。用户要求创建待填业务文档模板、合同填空或可按书签填写的 HTML 时使用。