{"id":51609593,"url":"https://github.com/learnprompt/carl-skills","last_synced_at":"2026-07-12T06:02:29.628Z","repository":{"id":358244398,"uuid":"1240599492","full_name":"LearnPrompt/carl-skills","owner":"LearnPrompt","description":"Real-world AI workflows for creators.","archived":false,"fork":false,"pushed_at":"2026-05-16T11:47:25.000Z","size":14,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-16T13:37:14.804Z","etag":null,"topics":["agent-skills","ai-skills","ai-workflows","creator-tools","hermes-agent","prompt-engineering"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/LearnPrompt.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-16T10:31:11.000Z","updated_at":"2026-05-16T11:47:29.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/LearnPrompt/carl-skills","commit_stats":null,"previous_names":["learnprompt/carl-skills"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/LearnPrompt/carl-skills","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LearnPrompt%2Fcarl-skills","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LearnPrompt%2Fcarl-skills/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LearnPrompt%2Fcarl-skills/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LearnPrompt%2Fcarl-skills/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/LearnPrompt","download_url":"https://codeload.github.com/LearnPrompt/carl-skills/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LearnPrompt%2Fcarl-skills/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35383520,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-12T02:00:06.386Z","response_time":87,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent-skills","ai-skills","ai-workflows","creator-tools","hermes-agent","prompt-engineering"],"created_at":"2026-07-12T06:02:24.803Z","updated_at":"2026-07-12T06:02:29.616Z","avatar_url":"https://github.com/LearnPrompt.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\n# 🦞 Carl Skills\n\n#### 我自己真实跑通、反复用过的AI工作流，都收在这里\n\n[![License](https://img.shields.io/badge/License-MIT-3B82F6?style=for-the-badge)](./LICENSE)\n[![Skills](https://img.shields.io/badge/Skills-16-10B981?style=for-the-badge)](#-skills)\n[![Registry](https://img.shields.io/badge/Registry-catalog--first-F59E0B?style=for-the-badge)](./registry.json)\n[![First Star](https://img.shields.io/badge/First_Star-Humanize_PPT-8B5CF6?style=for-the-badge)](https://github.com/LearnPrompt/humanize-ppt)\n\n![Claude Code](https://img.shields.io/badge/Claude_Code-Skill-D97706?style=flat-square\u0026logo=anthropic\u0026logoColor=white)\n![Codex](https://img.shields.io/badge/Codex-Skill-10B981?style=flat-square\u0026logo=openai\u0026logoColor=white)\n![OpenCode](https://img.shields.io/badge/OpenCode-Skill-3B82F6?style=flat-square)\n![OpenClaw](https://img.shields.io/badge/OpenClaw-Skill-8B5CF6?style=flat-square)\n![Hermes](https://img.shields.io/badge/Hermes-Agent-EC4899?style=flat-square)\n\n\u003c/div\u003e\n\n我每天会试AI工具，也会把它们塞进真实工作里，选题、写稿、做评测、整理资料、改PPT、发文章、维护Obsidian，让Hermes/Codex/OpenClaw这类Agent真的帮上忙。\n\n很多东西不是想出来的，是一遍遍跑出来的。\n\n所以这个仓库只做一件事：把那些已经跑顺的AI工作流，整理成下次还能交给Agent继续用的skill。\n\n- **Skills**，Agent能直接加载的结构化工作流，安装后可以在Claude Code、Codex、OpenCode、OpenClaw、Hermes里复用\n- **Registry**，给Agent和脚本读的机器目录，记录每个skill从哪个canonical repo安装\n- **Catalog**，给人看的入口，告诉你每个skill适合干什么、不适合干什么\n\n---\n\n## 📋 目录\n\n| 名字 | Star | 一句话 | 安装来源 |\n|---|---|---|---|\n| 🛰️ [**ai-news-radar**](#ai-news-radar) | [![](https://img.shields.io/github/stars/LearnPrompt/ai-news-radar?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/ai-news-radar) | 24小时AI/科技信息雷达，持续追踪高信号更新 | [canonical](https://github.com/LearnPrompt/ai-news-radar) |\n| 🪚 [**鲁班 luban**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/luban-skill?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/luban-skill) | 把能用的skill打磨成能被装、能传播、能验证的公共资产 | [canonical](https://github.com/LearnPrompt/luban-skill) |\n| 🎯 [**humanize-ppt**](#humanize-ppt) | [![](https://img.shields.io/github/stars/LearnPrompt/humanize-ppt?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/humanize-ppt) | 先把资料变成人愿意听的PPT主线，再交给下游工具生成页面 | [canonical](https://github.com/LearnPrompt/humanize-ppt) |\n| 🧩 [**cc-harness-skills**](#cc-harness-skills) | [![](https://img.shields.io/github/stars/LearnPrompt/cc-harness-skills?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/cc-harness-skills) | 一套Agent工作底座，记忆、压缩、协调、验证、主动模式一起用 | [canonical](https://github.com/LearnPrompt/cc-harness-skills) |\n| 🏘️ [**skillrush-town**](#skillrush-town) | [![](https://img.shields.io/github/stars/LearnPrompt/skillrush-town?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/skillrush-town) | 淘金小镇，追踪ClawHub Top100和潜力Skill变化 | [canonical](https://github.com/LearnPrompt/skillrush-town) |\n| 📚 [**carl-weread**](#carl-weread) | [![](https://img.shields.io/github/stars/LearnPrompt/carl-weread?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/carl-weread) | 微信读书行动型阅读教练，按当前问题推荐今天读的一小节 | [canonical](https://github.com/LearnPrompt/carl-weread) |\n| 🔪 [**庖丁 paoding**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/paoding-skill?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/paoding-skill) | 零API拆解任何博主的爆款打法，蒸馏成可安装的内容教练 | [canonical](https://github.com/LearnPrompt/paoding-skill) |\n| 🤝 [**搭子 dazi**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/partner-skill?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/partner-skill) | 让Claude和Codex结对开发，分工、互查、合并一条线 | [canonical](https://github.com/LearnPrompt/partner-skill) |\n| ✍️ [**x-article-publisher**](#x-article-publisher) | [![](https://img.shields.io/github/stars/LearnPrompt/x-article-publisher-skill?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/x-article-publisher-skill) | 把飞书或本地Markdown文章发布到X Articles草稿 | [canonical](https://github.com/LearnPrompt/x-article-publisher-skill) |\n| 🔁 [**skill-sync**](#skill-sync) | [![](https://img.shields.io/github/stars/LearnPrompt/skill-sync?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/skill-sync) | 把多端Agent skills整理成一个可信来源 | [canonical](https://github.com/LearnPrompt/skill-sync) |\n| 🏮 [**阿福 afu**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/afu-llm-todo?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/afu-llm-todo) | Obsidian收件箱管家，Inbox到Wiki到待办到周历一条线 | [canonical](https://github.com/LearnPrompt/afu-llm-todo) |\n| 📜 [**蔡伦 cailun**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/cailun-skill?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/cailun-skill) | 把对话里聊出来的结论，3秒造成一页能传阅的单文件纸 | [canonical](https://github.com/LearnPrompt/cailun-skill) |\n| ⛰️ [**愚公 yugong**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/loop-engineering?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/loop-engineering) | Loop工程方法论，把模糊目标改造成带验证门的自动循环 | [canonical](https://github.com/LearnPrompt/loop-engineering) |\n| 🎨 [**Irasutoya配图**](#banmen-family) | [![](https://img.shields.io/github/stars/LearnPrompt/carl-irasutoya-illustrations?style=flat\u0026label=%E2%98%85\u0026color=555)](https://github.com/LearnPrompt/carl-irasutoya-illustrations) | 给中文文章配会吐槽的Irasutoya反应人物正文配图 | [canonical](https://github.com/LearnPrompt/carl-irasutoya-illustrations) |\n\n---\n\n## 📦 安装方式\n\n### 装一个skill\n\n比如只装Humanize PPT：\n\n```bash\nhermes skills install https://raw.githubusercontent.com/LearnPrompt/humanize-ppt/main/SKILL.md --yes\n```\n\n### 装这个目录里的全部可安装skill\n\n```bash\ngit clone https://github.com/LearnPrompt/carl-skills.git\ncd carl-skills\npython3 scripts/install_all_hermes_skills.py --dry-run\npython3 scripts/install_all_hermes_skills.py --yes\n```\n\n`--dry-run`只打印安装命令，不改本机环境。确认没问题后再执行`--yes`。\n\n---\n\n## 🧠 这个仓库的逻辑\n\nCarl Skills现在是**catalog-first**。\n\n也就是说，这里不再把每个`SKILL.md`复制一份做镜像。每个skill都有自己的canonical repo，源码、README、demo、issue、更新都在那里维护。Carl Skills只负责把它们放进一个好找、好读、Agent也能读的目录里。\n\n这样有两个好处：\n\n- 更新Humanize PPT这类独立skill时，只需要改它自己的主仓库\n- Agent批量安装时，仍然可以通过`registry.json`找到全部canonical install URL\n\n如果以后有只属于Carl Skills合集的skill，再把`SKILL.md`直接放在这个仓库里。\n\n---\n\n## ✨ Skills\n\n\u003ca id=\"-skills\"\u003e\u003c/a\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"ai-news-radar\"\u003e\u003c/a\u003e\n\n### 🛰️ ai-news-radar\n\n\u003e *\"AI圈一天发太多东西，真正有用的信号得有人替你扫出来。\"*\n\nAI News Radar是一个24小时AI/科技信息雷达。它把RSS、OPML、GitHub feed、来源健康检查、GitHub Actions和网页展示串起来，用来持续追踪高信号更新。\n\n它不是普通资讯收藏夹，更像是给内容创作者和研究型Agent准备的上游信号源。\n\n**适合**\n\n- 每天需要找AI工具、模型、产品、论文的新动向\n- 想维护自己的信息源池，而不是只刷平台推荐\n- 做选题前需要先看到上游发生了什么\n\n**不适合**\n\n- 只想临时查一条新闻\n- 不准备维护来源质量\n- 只看中文二手总结\n\n[![Repo](https://img.shields.io/badge/GitHub-ai--news--radar-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/ai-news-radar)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/ai-news-radar/master/skills/ai-news-radar/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/ai-news-radar) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/ai-news-radar/master/skills/ai-news-radar/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"humanize-ppt\"\u003e\u003c/a\u003e\n\n### 🎯 humanize-ppt\n\n\u003e *\"PPT不是资料容器，是观众状态改变器。\"*\n\n很多AI工具都能生成PPT，但它们经常只是把资料塞进页面里。页数不少，信息不少，听众却不知道为什么要继续听。\n\nHumanize PPT做的是更前面的那一步：先把原始资料整理成一条人愿意听下去的演示路径。它会先处理听众、场景、状态转移、叙事张力、页级意图，再把结构交给下游工具生成页面。\n\n**适合**\n\n- 已经有一堆资料，但PPT主线很散\n- 想把文章、报告、产品介绍改成演讲型deck\n- 做AI生成PPT前，先把观众路径和页面意图定住\n\n**不适合**\n\n- 只想一键生成漂亮模板\n- 只需要改字体、配色、版式\n- 原始资料还完全没有方向\n\n[![Repo](https://img.shields.io/badge/GitHub-humanize--ppt-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/humanize-ppt)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/humanize-ppt/main/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/humanize-ppt) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/humanize-ppt/main/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"cc-harness-skills\"\u003e\u003c/a\u003e\n\n### 🧩 CC Harness Skills\n\n\u003e *\"这不是六个散装skill，是一套Agent工作底座。\"*\n\nCC Harness Skills来自同一个仓库，适合一起用。它们处理的不是某个具体内容任务，而是Agent长期工作时绕不开的底层问题：怎么记忆、怎么压缩上下文、怎么协调多Agent、怎么验证完成声明、怎么做轻量主动模式。\n\n所以在Carl Skills里统一加`cc-harness-`前缀，并放在同一个suite里。\n\n| 名字 | 用来干什么 |\n|---|---|\n| `cc-harness-dream-memory` | 把近期日志、会话和记忆文件整理成可持续使用的主题记忆 |\n| `cc-harness-kairos-lite` | 构建轻量主动模式，包含定时检查、睡眠间隔和过期保护 |\n| `cc-harness-memory-extractor` | 从近期对话里提取长期记忆，避免把临时状态写成永久事实 |\n| `cc-harness-structured-context-compressor` | 把长会话压缩成可续接摘要，保留当前工作和下一步 |\n| `cc-harness-swarm-coordinator` | 拆分多Agent工作，让协调者专注集成而不是淹没在探索里 |\n| `cc-harness-verification-gate` | 做只读验证，检查完成声明和测试结果是不是真的 |\n\n**适合**\n\n- 长时间跑Agent任务\n- 多Agent协作\n- 需要跨会话延续上下文\n- 不想让Agent越跑越乱\n\n**不适合**\n\n- 只想完成一次简单问答\n- 不需要记忆、压缩、验证和调度\n\n[![Repo](https://img.shields.io/badge/GitHub-cc--harness--skills-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/cc-harness-skills)\n[![Suite](https://img.shields.io/badge/Suite-6_skills-8B5CF6?style=flat-square)](./registry.json)\n\n→ [canonical repo](https://github.com/LearnPrompt/cc-harness-skills) · [registry suite](./registry.json)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"skillrush-town\"\u003e\u003c/a\u003e\n\n### 🏘️ skillrush-town\n\n\u003e *\"不是所有skill都值得装，但值得装的东西应该被更早发现。\"*\n\nSkillrush Town，淘金小镇，用来追踪ClawHub Top100下载快照和潜力AI Skill变化。它把每天的市场变化沉淀下来，方便看出哪些skill在涨、哪些只是昙花一现。\n\n这更像一个AI Skill市场雷达，而不是单纯榜单页面。\n\n**适合**\n\n- 观察ClawHub生态和skill增长趋势\n- 给自己的Skill产品找参考对象\n- 做AI Agent生态选题和案例研究\n\n**不适合**\n\n- 只想找一个马上能用的单点工具\n- 不关心skill市场变化\n\n[![Repo](https://img.shields.io/badge/GitHub-skillrush--town-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/skillrush-town)\n[![Pages](https://img.shields.io/badge/GitHub_Pages-live-3B82F6?style=flat-square)](https://learnprompt.github.io/skillrush-town/)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/skillrush-town/main/skills/skillrush-town/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/skillrush-town) · [GitHub Pages](https://learnprompt.github.io/skillrush-town/) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/skillrush-town/main/skills/skillrush-town/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"carl-weread\"\u003e\u003c/a\u003e\n\n### 📚 carl-weread\n\n\u003e *\"把微信读书从「读了多少」改成「今天哪个问题可以被读懂一点」。\"*\n\ncarl-weread是一个微信读书行动型阅读教练。它根据你当前卡住的问题，交叉书架、笔记和章节，推荐一本书里今天读的一小节，并把阅读收束成一张读后行动卡，Markdown回流，周阅读行动复盘。\n\n它解决的不是「读什么书」，而是「今天读哪一小节、读完做什么」。\n\n**适合**\n\n- 微信读书里囤了很多书，但阅读和手头问题脱节\n- 想让每次阅读都落成一个可执行动作\n- 希望Agent帮你做书架、笔记、章节的交叉推荐\n\n**不适合**\n\n- 只想要一串书单推荐\n- 不用微信读书\n- 不想登录微信读书API\n\n```bash\nhermes skills install https://raw.githubusercontent.com/LearnPrompt/carl-weread/main/SKILL.md\n```\n\n[![Repo](https://img.shields.io/badge/GitHub-carl--weread-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/carl-weread)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/carl-weread/main/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/carl-weread) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/carl-weread/main/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"x-article-publisher\"\u003e\u003c/a\u003e\n\n### ✍️ x-article-publisher\n\n\u003e *\"文章写完，不该卡在复制、粘贴、丢格式这一步。\"*\n\nX Article Publisher负责把飞书或本地Markdown文章发布到X Articles草稿里，尽量保留富文本结构、封面图和媒体位置。\n\n它解决的是发布链路里最烦人的那段：内容已经写完了，但平台编辑器不听话。\n\n**适合**\n\n- 飞书或Markdown文章已经定稿，要发到X Articles\n- 文章里有封面、图片、视频、分割线等结构\n- 希望Agent帮你把发布动作跑完，而不是只给一段文本\n\n**不适合**\n\n- 没有X Premium/Articles权限\n- 还没登录X或不想使用浏览器自动化\n- 只是发一条普通短推\n\n[![Repo](https://img.shields.io/badge/GitHub-x--article--publisher-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/x-article-publisher-skill)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/x-article-publisher-skill/main/skills/x-article-publisher/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/x-article-publisher-skill) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/x-article-publisher-skill/main/skills/x-article-publisher/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"skill-sync\"\u003e\u003c/a\u003e\n\n### 🔁 skill-sync\n\n\u003e *\"Skill多了以后，最大的问题不是没有工具，是到处都有一份。\"*\n\nSkill Sync用来审计Codex、Claude、OpenClaw、OpenCode、本地workspace和共享目录里的skills，把重复、冲突、过期安装整理成一个可信来源。\n\n它适合在你已经装了一堆skill之后，用来做一次大扫除。\n\n**适合**\n\n- 多个Agent环境里都有skills，已经分不清谁是最新\n- 想把本地skills整理成统一源头\n- 迁移或公开前，需要先做一次清点和去重\n\n**不适合**\n\n- 你只有一两个skill\n- 只是临时安装，不准备长期维护\n\n[![Repo](https://img.shields.io/badge/GitHub-skill--sync-111827?style=flat-square\u0026logo=github)](https://github.com/LearnPrompt/skill-sync)\n[![Install](https://img.shields.io/badge/Install-raw_SKILL.md-10B981?style=flat-square)](https://raw.githubusercontent.com/LearnPrompt/skill-sync/main/SKILL.md)\n\n→ [canonical repo](https://github.com/LearnPrompt/skill-sync) · [raw SKILL.md](https://raw.githubusercontent.com/LearnPrompt/skill-sync/main/SKILL.md)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n---\n\n## 🗂 Registry\n\n`registry.json`是这个仓库的机器可读目录。它记录：\n\n- 哪些skill已经可安装\n- 每个skill的canonical repo在哪里\n- 同一套skill是否属于同一个`suite`\n- 应该从哪个`raw_skill_url`安装\n- 当前索引到哪个`source_commit`\n\n如果你只装一个skill，不需要关心registry。直接装它的raw `SKILL.md`就行。\n\n如果你想让Agent理解「Carl Skills里到底有什么」，或者想批量安装，就看registry。\n\n\u003ctable\u003e\n\u003ctr\u003e\u003ctd\u003e\n\n\u003ca id=\"banmen-family\"\u003e\u003c/a\u003e\n\n### 🏯 班门家族\n\n\u003e *\"名字背故事，定位背痛苦。装一个，就知道这一家子的脾气。\"*\n\n一组以中国工匠命名的方法论skills，各自有独立的canonical repo，共享同一套班规（疑问句不构成授权、宁空不摆拍、commit即push）：\n\n- 🪚 **鲁班**，把\"能用的skill\"打磨成能被装、能传播、能验证的公共资产 → [luban-skill](https://github.com/LearnPrompt/luban-skill)\n- 🔪 **庖丁**，零API拆解任何博主的爆款打法，蒸馏成可安装的内容教练 → [paoding-skill](https://github.com/LearnPrompt/paoding-skill)\n- 🤝 **搭子**，让Claude和Codex结对开发，分工、互查、合并一条线 → [partner-skill](https://github.com/LearnPrompt/partner-skill)\n- 🏮 **阿福**，站在Obsidian收件箱门口的管家，Inbox到周历一条线 → [afu-llm-todo](https://github.com/LearnPrompt/afu-llm-todo)\n- 📜 **蔡伦**，把对话里聊出来的结论，3秒造成一页零依赖单文件的纸 → [cailun-skill](https://github.com/LearnPrompt/cailun-skill)\n- ⛰️ **愚公**，把模糊目标改造成带验证门、会自己挖的自动循环 → [loop-engineering](https://github.com/LearnPrompt/loop-engineering)\n- 🎨 **Irasutoya配图**，给文章里最值钱的判断配一张会吐槽的图 → [carl-irasutoya-illustrations](https://github.com/LearnPrompt/carl-irasutoya-illustrations)\n\n**安装**（任选其一）：\n\n```bash\nnpx skills add LearnPrompt/luban-skill -g        # 鲁班\nnpx skills add LearnPrompt/paoding-skill -g      # 庖丁\nnpx skills add LearnPrompt/partner-skill -g      # 搭子\nnpx skills add LearnPrompt/afu-llm-todo -g       # 阿福\nnpx skills add LearnPrompt/cailun-skill -g       # 蔡伦\nnpx skills add LearnPrompt/loop-engineering      # 愚公\nnpx skills add LearnPrompt/carl-irasutoya-illustrations -g  # Irasutoya配图\n```\n\n[![luban](https://img.shields.io/github/stars/LearnPrompt/luban-skill?style=flat-square\u0026label=鲁班)](https://github.com/LearnPrompt/luban-skill)\n[![paoding](https://img.shields.io/github/stars/LearnPrompt/paoding-skill?style=flat-square\u0026label=庖丁)](https://github.com/LearnPrompt/paoding-skill)\n[![dazi](https://img.shields.io/github/stars/LearnPrompt/partner-skill?style=flat-square\u0026label=搭子)](https://github.com/LearnPrompt/partner-skill)\n[![afu](https://img.shields.io/github/stars/LearnPrompt/afu-llm-todo?style=flat-square\u0026label=阿福)](https://github.com/LearnPrompt/afu-llm-todo)\n[![cailun](https://img.shields.io/github/stars/LearnPrompt/cailun-skill?style=flat-square\u0026label=蔡伦)](https://github.com/LearnPrompt/cailun-skill)\n[![yugong](https://img.shields.io/github/stars/LearnPrompt/loop-engineering?style=flat-square\u0026label=愚公)](https://github.com/LearnPrompt/loop-engineering)\n[![irasutoya](https://img.shields.io/github/stars/LearnPrompt/carl-irasutoya-illustrations?style=flat-square\u0026label=Irasutoya)](https://github.com/LearnPrompt/carl-irasutoya-illustrations)\n\n\u003c/td\u003e\u003c/tr\u003e\n\u003c/table\u003e\n\n---\n\n## 🛣 Roadmap\n\n- [x] 建立catalog-first registry\n- [x] 收录LearnPrompt下已公开且带`SKILL.md`的skill项目\n- [x] 将Humanize PPT安装入口改回canonical repo，避免collection mirror版本同步问题\n- [x] 将CC Harness Skills按suite统一分组\n- [x] 收录班门家族（鲁班/庖丁/蔡伦/阿福/Irasutoya配图）为banmen-family suite\n- [ ] 给每个active skill补真实案例截图和更具体的使用入口\n- [ ] 如果未来出现只属于Carl Skills合集的skill，再在本仓库内放置collection-native `SKILL.md`\n\n---\n\n## 🌟 关于\n\n我是Carl，日常主要做AI工具实测、内容创作、工作流搭建和Agent协作。\n\n这个仓库不会追求「技能数量很多」，更在意一件事：这些skill是不是真的在真实工作里跑过，能不能让下一个Agent少走一点弯路。\n\n如果你也在把AI从「聊天窗口」推进到真实工作流里，欢迎直接clone、安装、改造。\n\n---\n\n\u003cdiv align=\"center\"\u003e\n\n[MIT License](./LICENSE) · Real-world AI workflows for creators\n\n\u003c/div\u003e\n\n---\n\n\u003cdiv align=\"center\"\u003e\n\n**更多好用 Skill · More Skills** → [learnprompt.pro/skills](https://learnprompt.pro/skills/)\n\n[鲁班·Skill打磨](https://github.com/LearnPrompt/luban-skill) · [庖丁·博主蒸馏](https://github.com/LearnPrompt/paoding-skill) · [蔡伦·对话造纸](https://github.com/LearnPrompt/cailun-skill) · [阿福·LLM Todo](https://github.com/LearnPrompt/afu-llm-todo) · [愚公·Loop工程](https://github.com/LearnPrompt/loop-engineering) · [搭子·结对开发](https://github.com/LearnPrompt/partner-skill) · [AI雷达·零API资讯](https://github.com/LearnPrompt/ai-news-radar)\n\n[淘金小镇·ClawHub日榜](https://github.com/LearnPrompt/skillrush-town) · [Irasutoya·正文配图](https://github.com/LearnPrompt/carl-irasutoya-illustrations) · [Humanize PPT·演讲系统](https://github.com/LearnPrompt/humanize-ppt) · [CC Harness·六件套](https://github.com/LearnPrompt/cc-harness-skills) · [微信读书教练](https://github.com/LearnPrompt/carl-weread) · [X Article发布](https://github.com/LearnPrompt/x-article-publisher-skill)\n\n\u003csub\u003e**[LearnPrompt](https://github.com/LearnPrompt) 出品** · 公众号「卡尔的AI沃茨」 · [X @aiwarts](https://x.com/aiwarts)\u003c/sub\u003e\n\n\u003c/div\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flearnprompt%2Fcarl-skills","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Flearnprompt%2Fcarl-skills","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flearnprompt%2Fcarl-skills/lists"}