ai-skill

通过持续、逐条追问对齐方案/决策的每个关键节点，走遍决策树的每一个分支，直到达成共识后再动手。适用于开始执行前先把问题定义清楚。触发词：grill me、追问我、盘问我、帮我理一下这个方案、我们先对齐一下、深度盘问。

Downloads YouTube video transcripts/subtitles and cover images by URL or video ID. Supports multiple languages, translation, chapters, and speaker identification. Caches raw data for fast re-formatting. Use when user asks to "get YouTube transcript", "download subtitles", "get captions", "YouTube字幕", "YouTube封面", "视频封面", "video thumbnail", "video cover image", or provides a YouTube URL and wants the transcript/subtitle text or cover image extracted.

生成高质量 draw.io 图表并导出 PNG/SVG/PDF，含自检循环（最多5轮）和321个 AI/LLM 品牌 Logo。触发词：drawio图、导出PNG、导出PDF、AI架构图（带品牌Logo）、ML模型图、神经网络图、Transformer图、CNN图、代码结构图、代码转图、自定义样式图、精品图。注意：普通架构图/流程图/时序图请优先用 uml skill；网络拓扑→network；云架构→cloud。需要 draw.io 桌面版（brew install --cask drawio）。

【备用搜索】多引擎联合搜索，无需 API Key，支持17个引擎（Google、Bing、Scholar、知乎等）。Tavily 未配置时的降级方案。触发词：搜索、查一下、帮我找、search、找资料、查数据、查文献。注意：Tavily skill 配置后请优先用 tavily。

【首选搜索工具】Tavily AI 搜索，专为 AI Agent 优化，返回高质量摘要和来源，搜索结果比 web-search/multi-search-engine 更准确、更快。触发词：搜索、查一下、帮我找、search、找资料、查数据、查文献、最新消息、新闻、调研、research、查询、检索。注意：需要 API Key（tvly-xxxx）。未配置时降级使用 multi-search-engine。

用 PlantUML 生成云架构图，含 AWS/Azure/GCP/阿里云官方服务图标。触发词：云架构图、云架构、AWS架构、Azure架构、GCP架构、阿里云架构、云服务图、云迁移图、cloud architecture、cloud diagram。

咨询专属图表，输出独立 HTML 文件。仅处理 PlantUML 不擅长的咨询框架类型。触发词：象限图、2x2矩阵、四象限、金字塔图、漏斗图、韦恩图、venn、层堆栈、咨询图表、editorial、品牌风格图。注意：架构图/流程图/序列图/类图等请用 uml skill；网络拓扑→network；云架构→cloud。

输出可编辑格式图表：Draw.io（.drawio）/ Mermaid（.md）/ Excalidraw（.excalidraw），适合需要后续手动编辑、嵌入文档或版本管理的图表。触发词：drawio、mermaid、excalidraw、可编辑图表、可下载图表、ER图、数据库关系图。注意：UML/云架构/网络拓扑等有专用PlantUML skill时优先用那些skill；数据图表用vega；信息图用infographic；咨询框架图用diagram-design。

当明确需要 SVG/PNG 文件输出时使用，或 AI/Agent 专用图（RAG管道、多Agent协作）PlantUML 不擅长时使用。触发词：生成PNG、生成SVG、高清图、RAG图、AI架构图、Agent架构图、多Agent图、组织架构图、脑图、mind map。注意：普通架构图/流程图/时序图请优先用 uml skill。

用 PlantUML 生成网络拓扑图，含 Cisco/Citrix 等真实设备图标，支持物理/逻辑拓扑。触发词：网络拓扑、网络图、拓扑图、LAN、WAN、数据中心网络、网络架构图、路由器交换机、network topology、network diagram。

【首选画图工具】用 PlantUML 生成各类技术图表，自动布局、效果好、输出 Markdown 文件可直接用 Markdown Viewer 渲染。覆盖绝大多数图表需求。触发词：架构图、流程图、时序图、序列图、类图、组件图、状态图、活动图、用例图、部署图、系统设计图、技术图、画图、diagram、UML、PlantUML。不处理：网络拓扑→network；云架构→cloud；企业架构→archimate；业务流程→bpmn；象限图/泳道图/咨询图→diagram-design；需要PNG文件→fireworks-tech-graph。

歌华有线专属 PPT 生成器。根据用户提供的内容文档和页面规划，生成符合歌华视觉规范的 .pptx 演示文稿（32cm×18cm，白底，品牌蓝

文稿快校。通读后直接输出实际存在的问题，按优先级排列，不走空流程。触发词：校对、审阅、检查文稿、QA、review、帮我看看、逻辑有没有问题、文稿质量、这篇报告怎么样、帮我改一下、看看有没有问题。

头脑风暴与创意探索，将模糊想法转化为清晰方案。适用于产品设计、功能构思、咨询方案设计、研究框架规划等场景。触发词：我有个想法、帮我想想、设计一个方案、brainstorm、我想做一个、帮我构思、idea探索、怎么设计。

管理 cc-connect 本地桥接服务，将 Claude Code / Codex / OpenClaw 三个 Agent 与 Telegram / 飞书等消息平台打通。触发词：启动/停止 cc-connect、查看会话、切换 Agent、Telegram 连不上、配置 Bot、cc-connect 状态、Telegram Bot 管理、消息推送、手机发指令给 AI、远程控制 Claude、Bot 不响应。

写作引擎，带着观点出发在写作过程中把它想透，适合文章、报告、分析稿、观点文。触发词：帮我写、写一篇、写文章、写报告、起稿、写一个分析、帮我起草、写稿、我想写一篇。

管理对话记忆和历史记录。当用户询问过去的对话内容、个人偏好、历史信息，或需要存储/检索记忆时使用此 skill。触发词："你还记得..."、"我们之前讨论过..."、"帮我找之前说的..."、"记住这个"、"do you remember"、"we talked about"。

将英文技术内容翻译为地道中文，模仿宝玉（@dotey）三步翻译法，去除 AI 腔，保持技术准确性。触发词：翻译、translate、把这段英文翻译、技术文章翻译、帮我翻译、翻译一下、中文翻译。

制定多步骤任务的详细执行计划，将复杂任务拆解为可操作步骤。适用于编码项目、咨询方案规划、研究项目设计、工作流设计等。触发词：帮我规划、制定计划、拆解任务、制定方案步骤、怎么推进、任务拆解、项目计划。

访谈音频转写全流程系统：选择 ASR 路线运行转写、LLM 纠错、结构化阅读版生成、词库维护、评分表生成。当用户提到"转写"、"访谈转写"、"ASR"、"音频转文字"、"纠错"、"添加词库"、"结构化阅读"、"评分表"、"formal_transcript"、"corrected_transcript"时使用此 skill。

Content caster (铸). Transforms content into PNG visuals. Six molds: -l (default) long reading card, -i infograph, -m multi-card reading cards (1080x1440), -v visual sketchnote, -c comic (manga-style B&W), -w whiteboard (marker-style board layout). Output to ~/Downloads/. Use when user says '铸', 'cast', '做成图', '做成卡片', '做成信息图', '做成海报', '视觉笔记', 'sketchnote', '漫画', 'comic', 'manga', '白板', 'whiteboard'. Replaces ljg-cards and ljg-infograph.

OpenCLI — Make any website or Electron App your CLI. Zero risk, AI-powered, reuse Chrome login.

Search and download ebooks via Telegram Z-Lib Bot channel. Trigger automatically when the user: (1) inputs a book title or author name, (2) says "download book", "find book", "get ebook", "下载书", "找书", "搜书", "帮我下载", "下载电子书", "epub", "pdf 下载", or provides a book name with author. Default format is epub. Saves to ~/Downloads/books/.

YC Office Hours — two modes. Startup mode: six forcing questions that expose demand reality, status quo, desperate specificity, narrowest wedge, observation, and future-fit. Builder mode: design thinking brainstorming for side projects, hackathons, learning, and open source. Saves a design doc. Use when asked to "brainstorm this", "I have an idea", "help me think through this", "office hours", or "is this worth building". Proactively suggest when the user describes a new product idea or is exploring whether something is worth building — before any code is written. Use before /plan-ceo-review or /plan-eng-review.

Weekly engineering retrospective. Analyzes commit history, work patterns, and code quality metrics with persistent history and trend tracking. Team-aware: breaks down per-person contributions with praise and growth areas. Use when asked to "weekly retro", "what did we ship", or "engineering retrospective". Proactively suggest at the end of a work week or sprint.

Pre-landing PR review. Analyzes diff against the base branch for SQL safety, LLM trust boundary violations, conditional side effects, and other structural issues. Use when asked to "review this PR", "code review", "pre-landing review", or "check my diff". Proactively suggest when the user is about to merge or land code changes.

Use this skill to query your Google NotebookLM notebooks directly from Claude Code for source-grounded, citation-backed answers from Gemini. Browser automation, library management, persistent auth. Drastically reduced hallucinations through document-only responses.

Use local PDF-native tools plus the local GLM-OCR install on this machine to work with PDF files. Trigger when the user asks for PDF reading, text extraction, table extraction, form inspection or filling, merge, split, rotate, or OCR on scanned/image-only PDFs. Route native-text PDFs to `scripts/pdf_native.py` first, and only use GLM-OCR when the PDF has no reliable text layer or the user explicitly needs OCR-style output.

Inspect or interact with pages already open in the local Chrome browser via Chrome DevTools Protocol. Use when the user wants to debug, inspect, navigate, click, type, take screenshots, or run JavaScript in their existing Chrome session instead of opening a separate browser.

High-quality prompt optimization for turning vague or unstable prompts into reliable, testable instructions with explicit goals, scope, constraints, rewrite boundaries, and output formats. Use when asked to optimize/refine/rewrite prompts, improve output stability, reduce ambiguity, prevent missing items, enforce extraction-vs-rewrite boundaries, or adapt prompts for models such as DeepSeek, Qwen, ChatGPT, and Claude.

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.

Create and edit presentation slide decks (`.pptx`) with PptxGenJS, bundled layout helpers, and render/validation utilities. Use when tasks involve building a new PowerPoint deck, recreating slides from screenshots/PDFs/reference decks, modifying slide content while preserving editable output, adding charts/diagrams/visuals, or diagnosing layout issues such as overflow, overlaps, and font substitution.

Convert WeChat public account article links (mp.weixin.qq.com) into Markdown files with local image downloads. Use when the user asks to transform a WeChat article URL into .md, keep/retain inline images, batch-convert multiple links, or regenerate cleaner Markdown from a WeChat page.

Run Apify Actors from Codex with a repeatable pipeline (discover actor, run sync/async, fetch dataset, and save outputs). Use when the user asks to scrape or collect structured data and wants files saved locally.

Inspect or interact with pages already open in the local Chrome browser via Chrome DevTools Protocol. Use when the user wants to debug, inspect, navigate, click, type, take screenshots, or run JavaScript in their existing Chrome session instead of opening a separate browser.

Converts X (Twitter) tweets and articles to markdown with YAML front matter. Uses reverse-engineered API requiring user consent. Use when user mentions "X to markdown", "tweet to markdown", "save tweet", or provides x.com/twitter.com URLs for conversion.

Formats plain text or markdown files with frontmatter, titles, summaries, headings, bold, lists, and code blocks. Use when user asks to "format markdown", "beautify article", "add formatting", or improve article layout. Outputs to {filename}-formatted.md.

Translates articles and documents between languages with three modes - quick (direct), normal (analyze then translate), and refined (analyze, translate, review, polish). Supports custom glossaries and terminology consistency via EXTEND.md. Use when user asks to "translate", "翻译", "精翻", "translate article", "translate to Chinese/English", "改成中文", "改成英文", "convert to Chinese", "localize", "本地化", or needs any document translation. Also triggers for "refined translation", "精细翻译", "proofread translation", "快速翻译", "快翻", "这篇文章翻译一下", or when a URL or file is provided with translation intent.

Fetch any URL and convert to markdown using Chrome CDP. Supports two modes - auto-capture on page load, or wait for user signal (for pages requiring login). Use when user wants to save a webpage as markdown.

Use when the task involves reading, creating, or editing `.docx` documents, especially when formatting or layout fidelity matters; prefer `python-docx` plus the bundled `scripts/render_docx.py` for visual checks.