[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77YZDNSRZ8F7AVWZPS2W1BA":3,"skill-same:01M77YZDNSRZ8F7AVWZPS2W1BA":57},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":34,"commitSha":35,"skillDirectory":36,"forksCount":37,"repositoryHtmlUrl":38,"repositoryHomepageUrl":39,"ownerHtmlUrl":40,"ownerBlogUrl":6,"ownerDescription":41,"updatedAt":42,"install":43,"resources":56},"01M77YZDNSRZ8F7AVWZPS2W1BA","Skill Router — 智能技能路由引擎",null,"两阶段路由：embedding粗筛之后LLM精排，977技能里挑top5。中文query自动翻成英文再匹配，外加三重拒选防线，纯信息类查询直接踢到L0工具，别浪费LLM token。","Lord1Egypt\u002Fawesome-skill-forge","Lord1Egypt","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F99746359?v=4","01M7592YDNT0XK9WJ0T2MQVPDN","Mohamed Mounir Ali AL Keshita","01M75GRFNHC8VZNWEB8DKDE433",3,"MIT","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fy\u002Fyan-skill-router\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: \"Skill Router — 智能技能路由引擎\"\ndescription: \"两阶段路由：embedding粗筛之后LLM精排，977技能里挑top5。中文query自动翻成英文再匹配，外加三重拒选防线，纯信息类查询直接踢到L0工具，别浪费LLM token。\"\nversion: \"2.1.2\"\nauthor: \"炎哥\"\nhomepage: https:\u002F\u002Fgithub.com\u002Fawoo129\u002Fwangyan\nsource: https:\u002F\u002Fgithub.com\u002Fawoo129\u002Fwangyan\u002Ftree\u002Fmain\u002Fexperiments\u002Fskill-router-p0\ntags: [\"tool-selection\", \"skill-router\", \"routing\", \"intent-classification\", \"embedding\"]\n---\n\n# Skill Router\n\nskill库大到977个以后，让LLM一条条翻着看太傻也太贵。这个router用两段路由解决：先embedding粗筛，再LLM精排。\n\n## 怎么用\n\n```python\nfrom skill_router import SkillRouter\n\nrouter = SkillRouter()\nresult = router.route(\"做一篇新生儿低血糖的系统综述\")\n# → top1: literature-search (96%)\n\nprint(router.route_formatted(\"帮我画个森林图\"))\n```\n\n## 干了什么\n\n```\nQuery → BM25 + 翻译桥(双通道embedding) → RRF融合 → top5候选 → 拒选防线 → LLM精排\n                                                                  ↓\n                                                          拒选的走L0工具\n```\n\n### 四个块\n\n- **BM25关键词匹配** — 对英文skill名\u002F描述做关键词硬匹配\n- **中英翻译桥** — 中文query先翻成英文，0.3zh + 0.7en加权揉一起\n- **RRF排序** — 加权Reciprocal Rank Fusion，BM25×0.5 + Embed×2.0\n- **三重拒选** — 信息查询\u002F任务重叠\u002F分数阈值，三道坎\n\n### 拒选规则\n\n三道防线：\n\n1. **信息查询** — query里只有\"天气\"\u002F\"百科\"\u002F\"用药\"这类词，没有强技能关键词 → 不走LLM，直接L0工具\n2. **任务重叠** — 根本看不出要干什么 → 打回噪声\n3. **分数阈值** — RRF top1 \u003C 0.008 或 gap \u003C 0.06 → 太模糊，拒了\n\n## 什么时候用\n\nskill超过50个以后，传统一条条审查的方案token成本就扛不住了。特别是中英文混着用的场景——用户说中文但skill库全是英文名。还有就是明确不想让LLM碰的信息查询类请求，路由器帮你拦外面。\n\n## 要求\n\n- Python 3.10+\n- sentence-transformers (all-MiniLM-L6-v2)\n- numpy\n- 第一次跑要联网下embedding模型，大概80MB\n\n## 隐私\n\n路由引擎本身全本地跑：skill索引存在本地，embedding模型本地加载，query不走外部API。\n\nL0工具层（技能全拒后的保底）可能调用外部API：PubMed、Semantic Scholar、TinyFish等。\n这些工具的API key通过环境变量注入，详见tool_catalog.json中的`${VAR_NAME}`占位符。\n\n拒选防线保证纯查询类请求不会进LLM上下文。\n","# Skill Router\n\nskill库大到977个以后，让LLM一条条翻着看太傻也太贵。这个router用两段路由解决：先embedding粗筛，再LLM精排。\n\n## 怎么用\n\n```python\nfrom skill_router import SkillRouter\n\nrouter = SkillRouter()\nresult = router.route(\"做一篇新生儿低血糖的系统综述\")\n# → top1: literature-search (96%)\n\nprint(router.route_formatted(\"帮我画个森林图\"))\n```\n\n## 干了什么\n\n```\nQuery → BM25 + 翻译桥(双通道embedding) → RRF融合 → top5候选 → 拒选防线 → LLM精排\n                                                                  ↓\n                                                          拒选的走L0工具\n```\n\n### 四个块\n\n- **BM25关键词匹配** — 对英文skill名\u002F描述做关键词硬匹配\n- **中英翻译桥** — 中文query先翻成英文，0.3zh + 0.7en加权揉一起\n- **RRF排序** — 加权Reciprocal Rank Fusion，BM25×0.5 + Embed×2.0\n- **三重拒选** — 信息查询\u002F任务重叠\u002F分数阈值，三道坎\n\n### 拒选规则\n\n三道防线：\n\n1. **信息查询** — query里只有\"天气\"\u002F\"百科\"\u002F\"用药\"这类词，没有强技能关键词 → 不走LLM，直接L0工具\n2. **任务重叠** — 根本看不出要干什么 → 打回噪声\n3. **分数阈值** — RRF top1 \u003C 0.008 或 gap \u003C 0.06 → 太模糊，拒了\n\n## 什么时候用\n\nskill超过50个以后，传统一条条审查的方案token成本就扛不住了。特别是中英文混着用的场景——用户说中文但skill库全是英文名。还有就是明确不想让LLM碰的信息查询类请求，路由器帮你拦外面。\n\n## 要求\n\n- Python 3.10+\n- sentence-transformers (all-MiniLM-L6-v2)\n- numpy\n- 第一次跑要联网下embedding模型，大概80MB\n\n## 隐私\n\n路由引擎本身全本地跑：skill索引存在本地，embedding模型本地加载，query不走外部API。\n\nL0工具层（技能全拒后的保底）可能调用外部API：PubMed、Semantic Scholar、TinyFish等。\n这些工具的API key通过环境变量注入，详见tool_catalog.json中的`${VAR_NAME}`占位符。\n\n拒选防线保证纯查询类请求不会进LLM上下文。\n",{"name":5,"tags":24,"author":30,"source":31,"version":32,"homepage":33,"description":7},[25,26,27,28,29],"tool-selection","skill-router","routing","intent-classification","embedding","炎哥","https:\u002F\u002Fgithub.com\u002Fawoo129\u002Fwangyan\u002Ftree\u002Fmain\u002Fexperiments\u002Fskill-router-p0","2.1.2","https:\u002F\u002Fgithub.com\u002Fawoo129\u002Fwangyan","master","4aa004da461d820f4ea5499e58eb8d196cf99016","community\u002Fclawhub\u002Fy\u002Fyan-skill-router",0,"https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge","https:\u002F\u002Fskill-explorer-rho.vercel.app","https:\u002F\u002Fgithub.com\u002FLord1Egypt","Unbreakable. Unstoppable. Unmatched⚡.\r\n Speaker of machine tongues 🤖  | World-builder of decentralized dreams 🌐 | Defender of the digital realm🛡️.","2026-06-10 08:21:01",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add Lord1Egypt\u002Fawesome-skill-forge --skill \"Skill Router — 智能技能路由引擎\"",{"tool":49,"command":50},"bunx","bunx skills add Lord1Egypt\u002Fawesome-skill-forge --skill \"Skill Router — 智能技能路由引擎\"",{"tool":52,"command":53},"pnpm","pnpm dlx skills add Lord1Egypt\u002Fawesome-skill-forge --skill \"Skill Router — 智能技能路由引擎\"","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fy\u002Fyan-skill-router\u002FSKILL.md",true,[],[58,65,72,79,86,93],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":19},"01M785WKJB99G3STR1NF9NRTJW","youtube-content","YouTube transcripts to summaries, threads, blogs.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fyoutube-content\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M785WEFHB0G2S9YA6EPR3H0E","xurl","X\u002FTwitter via xurl CLI: post, search, DM, media, v2 API.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fxurl\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M785W98EDN4EFJXJPMQEHCRC","touchdesigner-mcp","Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftouchdesigner-mcp\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M785W5S9FP3QBH2V191E025P","test-driven-development","TDD: enforce RED-GREEN-REFACTOR, tests before code.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M785W2AS08NSSH9YATCV6FJB","teams-meeting-pipeline","Operate the Teams meeting summary pipeline via Hermes CLI — summarize meetings, inspect pipeline status, replay jobs, manage Microsoft Graph subscriptions.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fteams-meeting-pipeline\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M785VXKW4T34BTP1SCHW1R7F","systematic-debugging","4-phase root cause debugging: understand bugs before fixing.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[]]