[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6C5JS2RVVYTKK1VTXX7H69A":3,"skill-same:01M6C5JS2RVVYTKK1VTXX7H69A":79},{"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":23,"publishedAt":43,"linkOnly":44,"markdownContent":45,"markdownBody":46,"frontMatter":47,"refName":48,"commitSha":49,"skillDirectory":50,"forksCount":14,"repositoryHtmlUrl":51,"repositoryHomepageUrl":6,"ownerHtmlUrl":52,"ownerBlogUrl":53,"ownerDescription":6,"updatedAt":54,"install":55,"resources":68},"01M6C5JS2RVVYTKK1VTXX7H69A","prompt-router",null,"Route a classified request to the right skills, agent strategy, Context7 directives, and output format. Use this skill whenever you need to decide which skills to load for a task, choose between solo agent vs subagent dispatch vs tmux agent team, determine if Context7 library docs are needed, or select the right output format for a prompt. Trigger on: any request that has been classified and needs routing, when someone asks 'what skills should I use for this', 'should I use subagents', 'do I need tmux mode', or any prompt generation workflow that needs skill selection and strategy planning.","CoralShades\u002Facm-ai","CoralShades","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F220297585?v=4","01M6C5211QSJHZ8HXSDDWCR507","Coral Shades AI Agency","01M6C52161FNQDRQ0FRR8DAA9D",0,"MIT","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fprompt-router\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42],"prompt router","router","prompt","select","dispatch","use","subagents","strategy","decide","choose","docs","format","subagent","team","library","planning","routing","workflow","tmux","2026-11-10 10:12:30",false,"---\nname: prompt-router\ndescription: \"Route a classified request to the right skills, agent strategy, Context7 directives, and output format. Use this skill whenever you need to decide which skills to load for a task, choose between solo agent vs subagent dispatch vs tmux agent team, determine if Context7 library docs are needed, or select the right output format for a prompt. Trigger on: any request that has been classified and needs routing, when someone asks 'what skills should I use for this', 'should I use subagents', 'do I need tmux mode', or any prompt generation workflow that needs skill selection and strategy planning.\"\n---\n\n# Prompt Router\n\nMaps a classified request to the optimal skill set, agent strategy, Context7 directives, and output format. Outputs a `PromptPlan` JSON consumed by the downstream prompt-generator skill.\n\n## When to Use\n\n- After running `\u002Frequest-classifier` and you have a `RequestClassification` object\n- When a user asks \"what skills should I use for this task?\"\n- When a user asks \"should I use subagents or a tmux team?\"\n- When a user asks \"do I need Context7 for this?\"\n- As Step 3 in the full `\u002Fgenerate-prompt` pipeline\n\n---\n\n## Quick Routing Guide\n\n| If you have... | Route to... |\n|---|---|\n| `feature` + complexity > 6 | Tmux team + planning-with-files + parallel dispatch |\n| `feature` + complexity 4-6 | Subagent dispatch + planning-with-files |\n| `feature` + complexity 1-3 | Solo agent + domain skills only |\n| `bug-fix` (any complexity) | Solo focused agent + systematic-debugging |\n| `research` (any complexity) | Parallel subagents + planning-with-files + Context7 |\n| `improvement` + complexity > 6 | Subagent dispatch with gates + verification-before-completion |\n| `improvement` + complexity 1-6 | Solo agent + verification-before-completion |\n| `pipeline` (any complexity) | Tmux team + langgraph-fundamentals + acm-observability + Context7 |\n| `frontend` + complexity > 6 | Tmux team + frontend skills + maybe Context7 |\n| `frontend` + complexity 1-6 | Solo agent + frontend skills |\n| `quick-task` | Solo agent, minimal or no extra skills |\n| `documentation` | Solo agent, no extra skills |\n\n---\n\n## Routing Steps\n\n### Step 1: Read the classification\n\nAccept a `RequestClassification` object from `\u002Frequest-classifier` or extract it inline from the user's request:\n\n```json\n{\n  \"type\": \"feature|bug-fix|research|improvement|pipeline|frontend|quick-task|documentation\",\n  \"complexity\": 1-10,\n  \"plan_mode\": true|false,\n  \"domain_signals\": [\"extraction\", \"graph\", \"model\", \"debug\", ...],\n  \"scope\": \"single-file|multi-file|cross-cutting\",\n  \"estimated_files\": 0\n}\n```\n\nIf no classification is available, run the classification heuristics from `\u002Frequest-classifier` taxonomy before proceeding.\n\n### Step 2: Read the routing matrix\n\nOpen `references\u002Frouting-rules.md` and locate the row matching `type + complexity_band`:\n- **Simple**: complexity 1-3\n- **Medium**: complexity 4-6\n- **Complex**: complexity 7-10\n\n### Step 3: Match classification to routing matrix row\n\nFind the exact row. If the type is ambiguous (e.g., a feature with pipeline signals), use the more specific type (`pipeline` beats `feature`).\n\nPriority order when multiple types match:\n1. `pipeline` — most specific, always wins\n2. `frontend` — specific domain\n3. `bug-fix` — has fixed strategy\n4. `feature` \u002F `improvement` \u002F `research` — use complexity band\n5. `quick-task` \u002F `documentation` — fallback\n\n### Step 4: Apply domain skill selection logic\n\nCheck `domain_signals` array from the classification and append additional skills:\n\n| Signal | Additional Skills |\n|---|---|\n| \"extraction\", \"pipeline\", \"graph\", \"node\" | `\u002Flanggraph-fundamentals`, `\u002Facm-observability` |\n| \"agent\", \"tool\", \"chain\" | `\u002Flangchain-fundamentals` |\n| \"model\", \"schema\", \"pydantic\", \"validation\" | `\u002Fpydantic-models-py` |\n| \"debug\", \"error\", \"trace\", \"failing\" | `\u002Fsystematic-debugging`, `\u002Facm-observability` |\n| \"component\", \"page\", \"UI\", \"React\", \"css\" | `\u002Freact-best-practices`, `\u002Fnext-best-practices` |\n| \"streaming\", \"SSE\", \"websocket\" | `\u002Fsse-streaming` |\n| \"test\", \"coverage\", \"pytest\", \"playwright\" | `\u002Ftest-driven-development`, `\u002Fverification-before-completion` |\n| No specific signal | No additional skills |\n\nDeduplicate the final skill list (routing matrix skills + domain signal skills).\n\n### Step 5: Build Context7 directives if needed\n\nCheck the `context7` column of the matched routing row. If yes (or conditional and condition is met), build directives using the templates in `references\u002Frouting-rules.md`:\n\n- For each library relevant to the request, add one directive:\n  `resolve-library-id for \"{library}\" → query-docs for \"{topic from request}\"`\n\nAlways include Context7 for:\n- `pipeline` type → LangGraph + LangChain\n- `research` type → any library mentioned in the request\n- `feature` with explicit library version in request → that library\n\n### Step 6: Select agent strategy template\n\nUse `references\u002Fagent-strategies.md` and select the template matching the routing matrix `agent_strategy` column:\n- `solo` → Template A\n- `subagent-dispatch` → Template B\n- `tmux-team` → Template C\n\nFill in the `{placeholder}` variables from the classification data.\n\n### Step 7: Determine output format and path\n\n| Output Format | When | Path |\n|---|---|---|\n| `prompt-pack` (markdown file) | Complex requests, plan mode ON, tmux or subagent strategy | `docs\u002Fsprint-artifacts\u002Fprompt-packs\u002FYYYY-MM-DD-{slug}.md` |\n| `copy-paste` (terminal print) | Medium requests, solo or subagent, plan mode optional | Print to terminal |\n| `terminal` (inline response) | Simple requests, quick-task, documentation | Respond inline |\n\nIf user explicitly requested \"save\" or \"prompt-pack\" → always use `prompt-pack` format.\nIf user requested \"no plan\" → downgrade from `prompt-pack` to `copy-paste` unless still complex.\n\n### Step 8: Output the PromptPlan JSON\n\nAssemble and return the complete `PromptPlan` object:\n\n```json\n{\n  \"classification\": {\n    \"type\": \"...\",\n    \"complexity\": 0,\n    \"plan_mode\": true,\n    \"domain_signals\": [],\n    \"scope\": \"...\",\n    \"estimated_files\": 0\n  },\n  \"selected_skills\": [\"\u002Fplanning-with-files\", \"\u002Flanggraph-fundamentals\"],\n  \"agent_strategy\": \"tmux-team|subagent-dispatch|solo\",\n  \"agent_config\": {\n    \"panes\": [],\n    \"subagents\": [],\n    \"solo\": true\n  },\n  \"context7_directives\": [],\n  \"output_format\": \"prompt-pack|copy-paste|terminal\",\n  \"output_path\": \"docs\u002Fsprint-artifacts\u002Fprompt-packs\u002F\",\n  \"plan_mode\": true,\n  \"plan_type\": \"full|debug|research|refactor|none\",\n  \"verification_items\": [\n    \"uv run ruff check .\",\n    \"uv run pytest tests\u002F\",\n    \"cd frontend && npm run build\"\n  ]\n}\n```\n\nThe `PromptPlan` JSON is passed directly to the `\u002Fprompt-generator` skill (S4) as its primary input.\n\n---\n\n## Verification Items by Type\n\nAlways include these base verification items in `verification_items`:\n\n| Type | Verification Commands |\n|---|---|\n| Backend only | `uv run ruff check .`, `uv run pytest tests\u002F` |\n| Frontend only | `cd frontend && npm run lint`, `cd frontend && npm run build` |\n| Full stack | `uv run ruff check .`, `uv run pytest tests\u002F`, `cd frontend && npm run build` |\n| Pipeline \u002F LangGraph | All backend + `uv run pytest tests\u002Ftest_extraction*` |\n| Quick-task | `uv run ruff check .` (lint only) |\n| Documentation | None required |\n\n---\n\n## Common Routing Scenarios\n\n**\"Add a new extraction provider for MinerU v3\"**\n→ type=`pipeline`, complexity=7, skills=[`\u002Flanggraph-fundamentals`, `\u002Facm-observability`, `\u002Fplanning-with-files`, `\u002Fpydantic-models-py`], strategy=tmux-team, Context7=LangGraph+LangChain, format=prompt-pack\n\n**\"Fix the timeout error in the building extraction graph\"**\n→ type=`bug-fix`, complexity=5, skills=[`\u002Fsystematic-debugging`, `\u002Facm-observability`, `\u002Flanggraph-fundamentals`], strategy=solo, Context7=conditional (LangGraph if API-related), format=copy-paste\n\n**\"Rename the extract_all_rows function\"**\n→ type=`quick-task`, complexity=1, skills=[], strategy=solo, Context7=no, format=terminal\n\n**\"Investigate why correction LLM calls are spiking\"**\n→ type=`research`, complexity=6, skills=[`\u002Facm-observability`, `\u002Fplanning-with-files`, `\u002Flanggraph-fundamentals`], strategy=subagent-dispatch (parallel research panes), Context7=LangGraph, format=prompt-pack\n\n**\"Add the building summary panel to the source detail page\"**\n→ type=`frontend`, complexity=5, skills=[`\u002Freact-best-practices`, `\u002Fnext-best-practices`], strategy=solo, Context7=no, format=copy-paste\n\n**\"Refactor all pre-extraction stages to reduce LLM calls\"**\n→ type=`improvement`, complexity=8, skills=[`\u002Fplanning-with-files`, `\u002Fsubagent-driven-development`, `\u002Fverification-before-completion`, `\u002Fsystematic-debugging`], strategy=subagent-dispatch-with-gates, Context7=conditional (LangGraph if graph patterns change), format=prompt-pack\n","# Prompt Router\n\nMaps a classified request to the optimal skill set, agent strategy, Context7 directives, and output format. Outputs a `PromptPlan` JSON consumed by the downstream prompt-generator skill.\n\n## When to Use\n\n- After running `\u002Frequest-classifier` and you have a `RequestClassification` object\n- When a user asks \"what skills should I use for this task?\"\n- When a user asks \"should I use subagents or a tmux team?\"\n- When a user asks \"do I need Context7 for this?\"\n- As Step 3 in the full `\u002Fgenerate-prompt` pipeline\n\n---\n\n## Quick Routing Guide\n\n| If you have... | Route to... |\n|---|---|\n| `feature` + complexity > 6 | Tmux team + planning-with-files + parallel dispatch |\n| `feature` + complexity 4-6 | Subagent dispatch + planning-with-files |\n| `feature` + complexity 1-3 | Solo agent + domain skills only |\n| `bug-fix` (any complexity) | Solo focused agent + systematic-debugging |\n| `research` (any complexity) | Parallel subagents + planning-with-files + Context7 |\n| `improvement` + complexity > 6 | Subagent dispatch with gates + verification-before-completion |\n| `improvement` + complexity 1-6 | Solo agent + verification-before-completion |\n| `pipeline` (any complexity) | Tmux team + langgraph-fundamentals + acm-observability + Context7 |\n| `frontend` + complexity > 6 | Tmux team + frontend skills + maybe Context7 |\n| `frontend` + complexity 1-6 | Solo agent + frontend skills |\n| `quick-task` | Solo agent, minimal or no extra skills |\n| `documentation` | Solo agent, no extra skills |\n\n---\n\n## Routing Steps\n\n### Step 1: Read the classification\n\nAccept a `RequestClassification` object from `\u002Frequest-classifier` or extract it inline from the user's request:\n\n```json\n{\n  \"type\": \"feature|bug-fix|research|improvement|pipeline|frontend|quick-task|documentation\",\n  \"complexity\": 1-10,\n  \"plan_mode\": true|false,\n  \"domain_signals\": [\"extraction\", \"graph\", \"model\", \"debug\", ...],\n  \"scope\": \"single-file|multi-file|cross-cutting\",\n  \"estimated_files\": 0\n}\n```\n\nIf no classification is available, run the classification heuristics from `\u002Frequest-classifier` taxonomy before proceeding.\n\n### Step 2: Read the routing matrix\n\nOpen `references\u002Frouting-rules.md` and locate the row matching `type + complexity_band`:\n- **Simple**: complexity 1-3\n- **Medium**: complexity 4-6\n- **Complex**: complexity 7-10\n\n### Step 3: Match classification to routing matrix row\n\nFind the exact row. If the type is ambiguous (e.g., a feature with pipeline signals), use the more specific type (`pipeline` beats `feature`).\n\nPriority order when multiple types match:\n1. `pipeline` — most specific, always wins\n2. `frontend` — specific domain\n3. `bug-fix` — has fixed strategy\n4. `feature` \u002F `improvement` \u002F `research` — use complexity band\n5. `quick-task` \u002F `documentation` — fallback\n\n### Step 4: Apply domain skill selection logic\n\nCheck `domain_signals` array from the classification and append additional skills:\n\n| Signal | Additional Skills |\n|---|---|\n| \"extraction\", \"pipeline\", \"graph\", \"node\" | `\u002Flanggraph-fundamentals`, `\u002Facm-observability` |\n| \"agent\", \"tool\", \"chain\" | `\u002Flangchain-fundamentals` |\n| \"model\", \"schema\", \"pydantic\", \"validation\" | `\u002Fpydantic-models-py` |\n| \"debug\", \"error\", \"trace\", \"failing\" | `\u002Fsystematic-debugging`, `\u002Facm-observability` |\n| \"component\", \"page\", \"UI\", \"React\", \"css\" | `\u002Freact-best-practices`, `\u002Fnext-best-practices` |\n| \"streaming\", \"SSE\", \"websocket\" | `\u002Fsse-streaming` |\n| \"test\", \"coverage\", \"pytest\", \"playwright\" | `\u002Ftest-driven-development`, `\u002Fverification-before-completion` |\n| No specific signal | No additional skills |\n\nDeduplicate the final skill list (routing matrix skills + domain signal skills).\n\n### Step 5: Build Context7 directives if needed\n\nCheck the `context7` column of the matched routing row. If yes (or conditional and condition is met), build directives using the templates in `references\u002Frouting-rules.md`:\n\n- For each library relevant to the request, add one directive:\n  `resolve-library-id for \"{library}\" → query-docs for \"{topic from request}\"`\n\nAlways include Context7 for:\n- `pipeline` type → LangGraph + LangChain\n- `research` type → any library mentioned in the request\n- `feature` with explicit library version in request → that library\n\n### Step 6: Select agent strategy template\n\nUse `references\u002Fagent-strategies.md` and select the template matching the routing matrix `agent_strategy` column:\n- `solo` → Template A\n- `subagent-dispatch` → Template B\n- `tmux-team` → Template C\n\nFill in the `{placeholder}` variables from the classification data.\n\n### Step 7: Determine output format and path\n\n| Output Format | When | Path |\n|---|---|---|\n| `prompt-pack` (markdown file) | Complex requests, plan mode ON, tmux or subagent strategy | `docs\u002Fsprint-artifacts\u002Fprompt-packs\u002FYYYY-MM-DD-{slug}.md` |\n| `copy-paste` (terminal print) | Medium requests, solo or subagent, plan mode optional | Print to terminal |\n| `terminal` (inline response) | Simple requests, quick-task, documentation | Respond inline |\n\nIf user explicitly requested \"save\" or \"prompt-pack\" → always use `prompt-pack` format.\nIf user requested \"no plan\" → downgrade from `prompt-pack` to `copy-paste` unless still complex.\n\n### Step 8: Output the PromptPlan JSON\n\nAssemble and return the complete `PromptPlan` object:\n\n```json\n{\n  \"classification\": {\n    \"type\": \"...\",\n    \"complexity\": 0,\n    \"plan_mode\": true,\n    \"domain_signals\": [],\n    \"scope\": \"...\",\n    \"estimated_files\": 0\n  },\n  \"selected_skills\": [\"\u002Fplanning-with-files\", \"\u002Flanggraph-fundamentals\"],\n  \"agent_strategy\": \"tmux-team|subagent-dispatch|solo\",\n  \"agent_config\": {\n    \"panes\": [],\n    \"subagents\": [],\n    \"solo\": true\n  },\n  \"context7_directives\": [],\n  \"output_format\": \"prompt-pack|copy-paste|terminal\",\n  \"output_path\": \"docs\u002Fsprint-artifacts\u002Fprompt-packs\u002F\",\n  \"plan_mode\": true,\n  \"plan_type\": \"full|debug|research|refactor|none\",\n  \"verification_items\": [\n    \"uv run ruff check .\",\n    \"uv run pytest tests\u002F\",\n    \"cd frontend && npm run build\"\n  ]\n}\n```\n\nThe `PromptPlan` JSON is passed directly to the `\u002Fprompt-generator` skill (S4) as its primary input.\n\n---\n\n## Verification Items by Type\n\nAlways include these base verification items in `verification_items`:\n\n| Type | Verification Commands |\n|---|---|\n| Backend only | `uv run ruff check .`, `uv run pytest tests\u002F` |\n| Frontend only | `cd frontend && npm run lint`, `cd frontend && npm run build` |\n| Full stack | `uv run ruff check .`, `uv run pytest tests\u002F`, `cd frontend && npm run build` |\n| Pipeline \u002F LangGraph | All backend + `uv run pytest tests\u002Ftest_extraction*` |\n| Quick-task | `uv run ruff check .` (lint only) |\n| Documentation | None required |\n\n---\n\n## Common Routing Scenarios\n\n**\"Add a new extraction provider for MinerU v3\"**\n→ type=`pipeline`, complexity=7, skills=[`\u002Flanggraph-fundamentals`, `\u002Facm-observability`, `\u002Fplanning-with-files`, `\u002Fpydantic-models-py`], strategy=tmux-team, Context7=LangGraph+LangChain, format=prompt-pack\n\n**\"Fix the timeout error in the building extraction graph\"**\n→ type=`bug-fix`, complexity=5, skills=[`\u002Fsystematic-debugging`, `\u002Facm-observability`, `\u002Flanggraph-fundamentals`], strategy=solo, Context7=conditional (LangGraph if API-related), format=copy-paste\n\n**\"Rename the extract_all_rows function\"**\n→ type=`quick-task`, complexity=1, skills=[], strategy=solo, Context7=no, format=terminal\n\n**\"Investigate why correction LLM calls are spiking\"**\n→ type=`research`, complexity=6, skills=[`\u002Facm-observability`, `\u002Fplanning-with-files`, `\u002Flanggraph-fundamentals`], strategy=subagent-dispatch (parallel research panes), Context7=LangGraph, format=prompt-pack\n\n**\"Add the building summary panel to the source detail page\"**\n→ type=`frontend`, complexity=5, skills=[`\u002Freact-best-practices`, `\u002Fnext-best-practices`], strategy=solo, Context7=no, format=copy-paste\n\n**\"Refactor all pre-extraction stages to reduce LLM calls\"**\n→ type=`improvement`, complexity=8, skills=[`\u002Fplanning-with-files`, `\u002Fsubagent-driven-development`, `\u002Fverification-before-completion`, `\u002Fsystematic-debugging`], strategy=subagent-dispatch-with-gates, Context7=conditional (LangGraph if graph patterns change), format=prompt-pack\n",{"name":5,"description":7},"main","1ca81b0a7de199942400f965edd4d7544077ac5d",".claude\u002Fskills\u002Fprompt-router","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai","https:\u002F\u002Fgithub.com\u002FCoralShades","coralshades.com.au","2026-04-18 00:45:07",{"source":8,"skillName":5,"commands":56,"prompt":66,"zipAvailable":67},[57,60,63],{"tool":58,"command":59},"npx","npx skills add CoralShades\u002Facm-ai --skill prompt-router",{"tool":61,"command":62},"bunx","bunx skills add CoralShades\u002Facm-ai --skill prompt-router",{"tool":64,"command":65},"pnpm","pnpm dlx skills add CoralShades\u002Facm-ai --skill prompt-router","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fprompt-router\u002FSKILL.md",true,[69,75],{"path":70,"resourceType":71,"resourceTypeDesc":72,"sizeBytes":73,"sourceHtmlUrl":74},"references\u002Fagent-strategies.md","REFERENCE","参考文档",10694,"https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fprompt-router\u002Freferences\u002Fagent-strategies.md",{"path":76,"resourceType":71,"resourceTypeDesc":72,"sizeBytes":77,"sourceHtmlUrl":78},"references\u002Frouting-rules.md",14028,"https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fprompt-router\u002Freferences\u002Frouting-rules.md",[80,101,114,130,151,161],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":89,"publishedAt":43},"01M6C63HKXA890HHV47WBAX0TB","uncodixfy","Prevents generic AI\u002FCodex UI patterns when generating frontend code. Use this skill whenever generating HTML, CSS, React, Vue, Svelte, or any frontend UI code to enforce clean, human-designed aesthetics inspired by Linear, Raycast, Stripe, and GitHub instead of typical AI-generated UI.","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Funcodixfy\u002FSKILL.md",[86],{"categoryNo":87,"categoryName":88,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",[82,90,91,92,93,94,95,96,97,98,99,29,100],"github","frontend","css","codex","react","code","linear","ui","vue","generated","html",{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":105,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":43},"01M6C63FJ2GEBBGZ5XSMV40KXB","runpodctl","Runpod CLI to manage your GPU workloads.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Frunpodctl\u002FSKILL.md",[],[103,109,110,111,112,113],"macos","cli","linux","gpu","Bash",{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":120,"publishedAt":43},"01M6C63DPAV46TFVQ01SFYPJ7X","langgraph-persistence","INVOKE THIS SKILL when your LangGraph needs to persist state, remember conversations, travel through history, or configure subgraph checkpointer scoping. Covers checkpointers, thread_id, time travel, Store, and subgraph persistence modes.","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Flanggraph-persistence\u002FSKILL.md",[],[121,122,123,124,125,126,127,128,129],"langgraph persistence","langgraph","persistence","travel","history","scoping","remember","state","configure",{"skillNo":131,"name":132,"nameZh":6,"description":133,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":134,"categories":135,"keywords":141,"publishedAt":43},"01M6C63BNS9CQEBTB8RNSSBFPA","langgraph-human-in-the-loop","INVOKE THIS SKILL when implementing human-in-the-loop patterns, pausing for approval, or handling errors in LangGraph. Covers interrupt(), Command(resume=...), approval\u002Fvalidation workflows, and the 4-tier error handling strategy.","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Flanggraph-human-in-the-loop\u002FSKILL.md",[136],{"categoryNo":137,"categoryName":138,"taxonomyType":139,"taxonomyTypeDesc":140},"01JBSEEDCAT000000000000038","人力资源","PROFESSION","适用职业",[142,143,122,144,145,146,147,148,31,149,150],"langgraph human in the loop","loop","workflows","error handling","validation","human-in-the-loop","command","approval","resume",{"skillNo":152,"name":153,"nameZh":6,"description":154,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":155,"categories":156,"keywords":157,"publishedAt":43},"01M6C639GVYZFX3NFPV5FBDR4F","langgraph-fundamentals","INVOKE THIS SKILL when writing ANY LangGraph code. Covers StateGraph, state schemas, nodes, edges, Command, Send, invoke, streaming, and error handling.","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Flanggraph-fundamentals\u002FSKILL.md",[],[158,122,145,128,159,160,148,95],"langgraph fundamentals","streaming","writing",{"skillNo":162,"name":163,"nameZh":6,"description":164,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":165,"categories":166,"keywords":167,"publishedAt":43},"01M6C6375N34YX2GZ3SD3R7XR2","langchain-rag","INVOKE THIS SKILL when building ANY retrieval-augmented generation (RAG) system. Covers document loaders, RecursiveCharacterTextSplitter, embeddings (OpenAI), and vector stores (Chroma, FAISS, Pinecone).","https:\u002F\u002Fgithub.com\u002FCoralShades\u002Facm-ai\u002Fblob\u002Fmain\u002F.qwen\u002Fskills\u002Flangchain-rag\u002FSKILL.md",[],[168,169,170,171,172,173,174,175,176,177],"rag","langchain rag","langchain","embeddings","vector","openai","chroma","faiss","vector stores","system"]