[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74J36RT5DH0VSSBXNVW7AS7":3,"skill-same:01M74J36RT5DH0VSSBXNVW7AS7":68},{"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":39,"commitSha":40,"skillDirectory":41,"forksCount":42,"repositoryHtmlUrl":43,"repositoryHomepageUrl":44,"ownerHtmlUrl":45,"ownerBlogUrl":6,"ownerDescription":46,"updatedAt":47,"install":48,"resources":61},"01M74J36RT5DH0VSSBXNVW7AS7","infocard-source-routing-decision-tree",null,"Route non-X infocard sources before subagent dispatch.","ccwq\u002Finfocard-pub","ccwq","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3853416?v=4","01M74HXVGA6GKM3KHHE63QGM4N","Dave C","01M74HXVGWDSB69XNYE41TBQ63",2,"MIT","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Finfocard\u002Finfocard-source-routing-decision-tree\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: infocard-source-routing-decision-tree\ndescription: \"Route non-X infocard sources before subagent dispatch.\"\nversion: 1.0.0\nauthor: Hermes Agent\nlicense: MIT\nmetadata:\n  hermes:\n    tags: [infocard, source-routing, decision-tree, youtube, video, web-article, paper]\n    related_skills: [infocard-publish-sop, infocard-mobile-verifier]\n---\n\n# Infocard source routing decision tree\n\n## Trigger\n\nUse this skill **before any Research A subagent dispatch or `publish-bundle.json` write**, whenever the user provides a source URL that is NOT `https:\u002F\u002Fx.com\u002Fi\u002Fstatus\u002F...`.\n\nAll infocard work stays in the current main checkout. This routing skill performs source classification only; it never creates, enters, removes, or recommends a Git worktree or clone. Authoring remains under `.docs\u002F\u003Crun-id>\u002F\u003Cslug>\u002F`, and any promotion or publication is owned by `infocard-publish-sop` and its Publisher flow.\n\nThe protocol's default entry path assumes an X post: status URL → author + timestamp + body extractable in one fetch. Non-X sources break that assumption in three places:\n\n1. **Topic detection is harder.** X post body text is captured immediately. A YouTube watch URL exposes only the page HTML; the orchestrator must first extract the title and decide if the topic matches any in-flight card. Failing this check is the 2026-08-02 incident: a 20-min YouTube video about Graph Engineering was misclassified as an AutoResearch follow-up because its `t=952s` timestamp was associated with an AutoResearch card earlier in the session.\n2. **Evidence shape is different.** X posts are first-class sources with display name, handle, timestamp, and media. YouTube videos produce a single content blob (title + description + transcript) without those structured fields.\n3. **Author \u002F timestamp attribution.** The `expected_public_identity` field in the v3 bundle normally names an X account. For YouTube it must name the channel.\n\n## Entry routing decision tree\n\n```\nSource URL provided\n│\n├─ x.com\u002Fi\u002Fstatus\u002F\u003Cid>          → existing X source flow, no change\n│\n├─ github.com\u002F\u003Cowner>\u002F\u003Crepo>    → existing source-fallback-github.md flow\n│\n├─ youtu.be \u002F youtube.com       → YouTube video flow (this file)\n│                                 1. Extract metadata from ytInitialPlayerResponse\n│                                 2. Topic-detect BEFORE author delegation\n│                                 3. Reject \"looks related to last card\" without\n│                                    explicit title \u002F description evidence\n│\n├─ arxiv.org\u002Fabs\u002F\u003Cid>           → Paper flow (TODO: future contribution)\n│\n├─ news.ycombinator.com\u002Fitem?id → HN discussion flow (TODO: future contribution)\n│\n└─ other web URL                → Generic web article flow\n                                   (blog post, podcast transcript, etc.)\n```\n\n## YouTube source extraction recipe\n\n```bash\n# 1. Fetch the watch page with a normal desktop User-Agent\ncurl -sL \"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=$VIDEO_ID\" \\\n  -A \"Mozilla\u002F5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit\u002F537.36 (KHTML, like Gecko) Chrome\u002F120.0.0.0 Safari\u002F537.36\"\n\n# 2. Extract title, duration, publish date from the embedded JSON blob\npython3 -c \"\nimport sys, re, json\nh = sys.stdin.read()\nm = re.search(r'var ytInitialPlayerResponse = (\\{.+?\\});', h)\nd = json.loads(m.group(1))\nmicro = d.get('microformat',{}).get('playerMicroformatRenderer',{})\nprint('title:', micro.get('title',{}).get('simpleText',''))\nprint('lengthSeconds:', micro.get('lengthSeconds',''))\nprint('publishDate:', micro.get('publishDate',''))\nprint('description:', (micro.get('description',{}).get('simpleText','') or '')[:2000])\n\"\n```\n\n### Caption retrieval (optional but useful for chapter detection)\n\n```bash\n# 3a. Find the caption base URL (it is signed and time-limited)\npython3 -c \"\nimport sys, re, json\nh = sys.stdin.read()\nm = re.search(r'var ytInitialPlayerResponse = (\\{.+?\\});', h)\nd = json.loads(m.group(1))\nfor tr in d.get('captions',{}).get('playerCaptionsTracklistRenderer',{}).get('captionTracks',[]):\n    print(tr.get('languageCode'), tr.get('baseUrl','')[:300])\n\"\n\n# 3b. Fetch the captions\ncurl -sL \"\u003Ccaption-baseUrl>&fmt=srv3&lang=zh-Hans\" -A \"Mozilla\u002F5.0\"\n```\n\n**Pitfall**: caption URLs include a signed `signature=` parameter and an `expire` timestamp. After a few minutes the URL returns 404. If `caption fetch 404`, **do not retry the same URL** — re-fetch the watch page and extract a fresh caption URL.\n\n## Topic detection gate (the lesson from 2026-08-02)\n\n**Symptom**: User supplied `https:\u002F\u002Fm.youtube.com\u002Fwatch?v=8RedSkw1UjE&t=952s` (a 20-min Graph Engineering video by 大飞说科技 \u002F 最佳拍档). I read the URL and associated the `t=952s` timestamp with an AutoResearch card from earlier in the session, dispatched Research A + worktree + bundle, and only discovered the actual title 「什么是图工程」several minutes into the run.\n\n**Hard rule**: Before any research delegation, verify the current main checkout, then fetch the source and report the title + description + topic in one line. Worktrees and temporary clones are prohibited. If the title disagrees with the assumed topic, **stop and re-discuss with the user** — do not let momentum push you into an off-topic run.\n\n**Concrete check that fits in a single turn**:\n\n```bash\ncurl -sL \"$URL\" -A \"Mozilla\u002F5.0\" | python3 -c \"\nimport sys, re\nh = sys.stdin.read()\nm = re.search(r'\u003Cmeta property=\\\"og:title\\\" content=\\\"([^\\\"]+)\\\"', h) or \\\n    re.search(r'\u003Cmeta name=\\\"title\\\" content=\\\"([^\\\"]+)\\\"', h) or \\\n    re.search(r'\u003Ctitle>([^\u003C]+)\u003C\u002Ftitle>', h)\nprint('TITLE:', m.group(1) if m else 'NONE')\nm = re.search(r'\u003Cmeta property=\\\"og:description\\\" content=\\\"([^\\\"]+)\\\"', h)\nprint('OG_DESC:', m.group(1)[:300] if m else 'NONE')\n\"\n```\n\nIf `TITLE: NONE`, the URL did not fetch successfully — fix the fetch before continuing. If the title is empty \u002F generic (\"YouTube\"), the request was routed to a different domain.\n\n## Bundle shape adjustments for video sources\n\n```json\n{\n  \"identity\": {\n    \"slug\": \"graph-engineering\",\n    \"title\": \"图工程（Graph Engineering）：AI 智能体架构从 Loop 到 Graph 的范式迁移\",\n    \"expected_public_identity\": \"Graph Engineering (视频解读)\"\n  },\n  \"requirements\": {\n    \"source_type\": \"youtube-video\",\n    \"source_url\": \"https:\u002F\u002Fm.youtube.com\u002Fwatch?v=8RedSkw1UjE&t=952s\",\n    \"channel\": \"大飞说科技 \u002F 最佳拍档\",\n    \"publish_date\": \"2026-08-02\",\n    \"duration_seconds\": 1211,\n    \"key_timestamps\": {\"952s\": \"视频后段 · Graph 拓扑 \u002F 验证器\"}\n  }\n}\n```\n\nThe meta.yaml `source` field should be `youtube-video` (not `x-post`). Card section 01 should be re-labelled `VIDEO IDENTITY` instead of `PROJECT IDENTITY`, with 频道 \u002F 时长 \u002F 发布日 \u002F 关键时间戳 as the four headline cards.\n\n## When to recommend \"skip the card\"\n\nAfter metadata extraction, refuse to publish when the source is:\n\n- A 3-minute intro video with no real content\n- A paywalled article with no abstract\n- A paper with no accessible PDF\n\nDo not auto-publish. Report the access boundary in one line and ask the user to provide an alternative source. Do not pad with L3 community speculation.\n\n## Web-article fallback (when the source is neither X, GitHub, nor YouTube)\n\nSame pre-flight gate applies:\n\n```bash\n# Fetch the article\ncurl -sL \"$URL\" -A \"Mozilla\u002F5.0\" > \u002Ftmp\u002Fsource.html\n# Extract: title, author byline (look for \u003Cmeta name=\"author\">),\n# published date (look for \u003Cmeta property=\"article:published_time\">),\n# description (og:description)\npython3 -c \"\nimport re\nh = open('\u002Ftmp\u002Fsource.html').read()\nfor prop in ['og:title', 'twitter:title', 'title']:\n    m = re.search(rf'\u003Cmeta (?:property|name)=\\\"{prop}\\\" content=\\\"([^\\\"]+)\\\"', h)\n    if m: print(f'{prop}:', m.group(1)); break\nfor prop in ['og:description', 'twitter:description', 'description']:\n    m = re.search(rf'\u003Cmeta (?:property|name)=\\\"{prop}\\\" content=\\\"([^\\\"]+)\\\"', h)\n    if m: print(f'{prop}:', m.group(1)[:300]); break\nm = re.search(r'\u003Cmeta name=\\\"author\\\" content=\\\"([^\\\"]+)\\\"', h)\nif m: print('author:', m.group(1))\nm = re.search(r'\u003Cmeta property=\\\"article:published_time\\\" content=\\\"([^\\\"]+)\\\"', h)\nif m: print('published:', m.group(1))\n\"\n```\n\nFor arXiv papers: prefer `https:\u002F\u002Farxiv.org\u002Fabs\u002F\u003Cid>` (HTML, easier to parse) over the PDF. `og:title` is usually the full title, `\u003Cmeta name=\"citation_author\" content=\"...\">` repeats for each author.\n\n## References\n\n- `references\u002Fnon-x-source-extraction-recipes.md` — concrete curl + python snippets for YouTube, arXiv, podcast transcripts, and paywall-aware news sites\n\n## Deprecated compatibility entry\n\n此旧入口仅作兼容转发，替代入口为 `infocard-source-and-content`。来源分类结果必须写入统一 source pack，不得自行分发 Author 或发布。\n","# Infocard source routing decision tree\n\n## Trigger\n\nUse this skill **before any Research A subagent dispatch or `publish-bundle.json` write**, whenever the user provides a source URL that is NOT `https:\u002F\u002Fx.com\u002Fi\u002Fstatus\u002F...`.\n\nAll infocard work stays in the current main checkout. This routing skill performs source classification only; it never creates, enters, removes, or recommends a Git worktree or clone. Authoring remains under `.docs\u002F\u003Crun-id>\u002F\u003Cslug>\u002F`, and any promotion or publication is owned by `infocard-publish-sop` and its Publisher flow.\n\nThe protocol's default entry path assumes an X post: status URL → author + timestamp + body extractable in one fetch. Non-X sources break that assumption in three places:\n\n1. **Topic detection is harder.** X post body text is captured immediately. A YouTube watch URL exposes only the page HTML; the orchestrator must first extract the title and decide if the topic matches any in-flight card. Failing this check is the 2026-08-02 incident: a 20-min YouTube video about Graph Engineering was misclassified as an AutoResearch follow-up because its `t=952s` timestamp was associated with an AutoResearch card earlier in the session.\n2. **Evidence shape is different.** X posts are first-class sources with display name, handle, timestamp, and media. YouTube videos produce a single content blob (title + description + transcript) without those structured fields.\n3. **Author \u002F timestamp attribution.** The `expected_public_identity` field in the v3 bundle normally names an X account. For YouTube it must name the channel.\n\n## Entry routing decision tree\n\n```\nSource URL provided\n│\n├─ x.com\u002Fi\u002Fstatus\u002F\u003Cid>          → existing X source flow, no change\n│\n├─ github.com\u002F\u003Cowner>\u002F\u003Crepo>    → existing source-fallback-github.md flow\n│\n├─ youtu.be \u002F youtube.com       → YouTube video flow (this file)\n│                                 1. Extract metadata from ytInitialPlayerResponse\n│                                 2. Topic-detect BEFORE author delegation\n│                                 3. Reject \"looks related to last card\" without\n│                                    explicit title \u002F description evidence\n│\n├─ arxiv.org\u002Fabs\u002F\u003Cid>           → Paper flow (TODO: future contribution)\n│\n├─ news.ycombinator.com\u002Fitem?id → HN discussion flow (TODO: future contribution)\n│\n└─ other web URL                → Generic web article flow\n                                   (blog post, podcast transcript, etc.)\n```\n\n## YouTube source extraction recipe\n\n```bash\n# 1. Fetch the watch page with a normal desktop User-Agent\ncurl -sL \"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=$VIDEO_ID\" \\\n  -A \"Mozilla\u002F5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit\u002F537.36 (KHTML, like Gecko) Chrome\u002F120.0.0.0 Safari\u002F537.36\"\n\n# 2. Extract title, duration, publish date from the embedded JSON blob\npython3 -c \"\nimport sys, re, json\nh = sys.stdin.read()\nm = re.search(r'var ytInitialPlayerResponse = (\\{.+?\\});', h)\nd = json.loads(m.group(1))\nmicro = d.get('microformat',{}).get('playerMicroformatRenderer',{})\nprint('title:', micro.get('title',{}).get('simpleText',''))\nprint('lengthSeconds:', micro.get('lengthSeconds',''))\nprint('publishDate:', micro.get('publishDate',''))\nprint('description:', (micro.get('description',{}).get('simpleText','') or '')[:2000])\n\"\n```\n\n### Caption retrieval (optional but useful for chapter detection)\n\n```bash\n# 3a. Find the caption base URL (it is signed and time-limited)\npython3 -c \"\nimport sys, re, json\nh = sys.stdin.read()\nm = re.search(r'var ytInitialPlayerResponse = (\\{.+?\\});', h)\nd = json.loads(m.group(1))\nfor tr in d.get('captions',{}).get('playerCaptionsTracklistRenderer',{}).get('captionTracks',[]):\n    print(tr.get('languageCode'), tr.get('baseUrl','')[:300])\n\"\n\n# 3b. Fetch the captions\ncurl -sL \"\u003Ccaption-baseUrl>&fmt=srv3&lang=zh-Hans\" -A \"Mozilla\u002F5.0\"\n```\n\n**Pitfall**: caption URLs include a signed `signature=` parameter and an `expire` timestamp. After a few minutes the URL returns 404. If `caption fetch 404`, **do not retry the same URL** — re-fetch the watch page and extract a fresh caption URL.\n\n## Topic detection gate (the lesson from 2026-08-02)\n\n**Symptom**: User supplied `https:\u002F\u002Fm.youtube.com\u002Fwatch?v=8RedSkw1UjE&t=952s` (a 20-min Graph Engineering video by 大飞说科技 \u002F 最佳拍档). I read the URL and associated the `t=952s` timestamp with an AutoResearch card from earlier in the session, dispatched Research A + worktree + bundle, and only discovered the actual title 「什么是图工程」several minutes into the run.\n\n**Hard rule**: Before any research delegation, verify the current main checkout, then fetch the source and report the title + description + topic in one line. Worktrees and temporary clones are prohibited. If the title disagrees with the assumed topic, **stop and re-discuss with the user** — do not let momentum push you into an off-topic run.\n\n**Concrete check that fits in a single turn**:\n\n```bash\ncurl -sL \"$URL\" -A \"Mozilla\u002F5.0\" | python3 -c \"\nimport sys, re\nh = sys.stdin.read()\nm = re.search(r'\u003Cmeta property=\\\"og:title\\\" content=\\\"([^\\\"]+)\\\"', h) or \\\n    re.search(r'\u003Cmeta name=\\\"title\\\" content=\\\"([^\\\"]+)\\\"', h) or \\\n    re.search(r'\u003Ctitle>([^\u003C]+)\u003C\u002Ftitle>', h)\nprint('TITLE:', m.group(1) if m else 'NONE')\nm = re.search(r'\u003Cmeta property=\\\"og:description\\\" content=\\\"([^\\\"]+)\\\"', h)\nprint('OG_DESC:', m.group(1)[:300] if m else 'NONE')\n\"\n```\n\nIf `TITLE: NONE`, the URL did not fetch successfully — fix the fetch before continuing. If the title is empty \u002F generic (\"YouTube\"), the request was routed to a different domain.\n\n## Bundle shape adjustments for video sources\n\n```json\n{\n  \"identity\": {\n    \"slug\": \"graph-engineering\",\n    \"title\": \"图工程（Graph Engineering）：AI 智能体架构从 Loop 到 Graph 的范式迁移\",\n    \"expected_public_identity\": \"Graph Engineering (视频解读)\"\n  },\n  \"requirements\": {\n    \"source_type\": \"youtube-video\",\n    \"source_url\": \"https:\u002F\u002Fm.youtube.com\u002Fwatch?v=8RedSkw1UjE&t=952s\",\n    \"channel\": \"大飞说科技 \u002F 最佳拍档\",\n    \"publish_date\": \"2026-08-02\",\n    \"duration_seconds\": 1211,\n    \"key_timestamps\": {\"952s\": \"视频后段 · Graph 拓扑 \u002F 验证器\"}\n  }\n}\n```\n\nThe meta.yaml `source` field should be `youtube-video` (not `x-post`). Card section 01 should be re-labelled `VIDEO IDENTITY` instead of `PROJECT IDENTITY`, with 频道 \u002F 时长 \u002F 发布日 \u002F 关键时间戳 as the four headline cards.\n\n## When to recommend \"skip the card\"\n\nAfter metadata extraction, refuse to publish when the source is:\n\n- A 3-minute intro video with no real content\n- A paywalled article with no abstract\n- A paper with no accessible PDF\n\nDo not auto-publish. Report the access boundary in one line and ask the user to provide an alternative source. Do not pad with L3 community speculation.\n\n## Web-article fallback (when the source is neither X, GitHub, nor YouTube)\n\nSame pre-flight gate applies:\n\n```bash\n# Fetch the article\ncurl -sL \"$URL\" -A \"Mozilla\u002F5.0\" > \u002Ftmp\u002Fsource.html\n# Extract: title, author byline (look for \u003Cmeta name=\"author\">),\n# published date (look for \u003Cmeta property=\"article:published_time\">),\n# description (og:description)\npython3 -c \"\nimport re\nh = open('\u002Ftmp\u002Fsource.html').read()\nfor prop in ['og:title', 'twitter:title', 'title']:\n    m = re.search(rf'\u003Cmeta (?:property|name)=\\\"{prop}\\\" content=\\\"([^\\\"]+)\\\"', h)\n    if m: print(f'{prop}:', m.group(1)); break\nfor prop in ['og:description', 'twitter:description', 'description']:\n    m = re.search(rf'\u003Cmeta (?:property|name)=\\\"{prop}\\\" content=\\\"([^\\\"]+)\\\"', h)\n    if m: print(f'{prop}:', m.group(1)[:300]); break\nm = re.search(r'\u003Cmeta name=\\\"author\\\" content=\\\"([^\\\"]+)\\\"', h)\nif m: print('author:', m.group(1))\nm = re.search(r'\u003Cmeta property=\\\"article:published_time\\\" content=\\\"([^\\\"]+)\\\"', h)\nif m: print('published:', m.group(1))\n\"\n```\n\nFor arXiv papers: prefer `https:\u002F\u002Farxiv.org\u002Fabs\u002F\u003Cid>` (HTML, easier to parse) over the PDF. `og:title` is usually the full title, `\u003Cmeta name=\"citation_author\" content=\"...\">` repeats for each author.\n\n## References\n\n- `references\u002Fnon-x-source-extraction-recipes.md` — concrete curl + python snippets for YouTube, arXiv, podcast transcripts, and paywall-aware news sites\n\n## Deprecated compatibility entry\n\n此旧入口仅作兼容转发，替代入口为 `infocard-source-and-content`。来源分类结果必须写入统一 source pack，不得自行分发 Author 或发布。\n",{"name":5,"author":24,"license":15,"version":25,"metadata":26,"description":7},"Hermes Agent","1.0.0",{"hermes":27},{"tags":28,"related_skills":36},[29,30,31,32,33,34,35],"infocard","source-routing","decision-tree","youtube","video","web-article","paper",[37,38],"infocard-publish-sop","infocard-mobile-verifier","main","4ea40456a3623e494105429a3ae13c3ff61f17f5",".agents\u002Fskills\u002Finfocard\u002Finfocard-source-routing-decision-tree",1,"https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub","https:\u002F\u002Fccwq.github.io\u002Finfocard-pub\u002F","https:\u002F\u002Fgithub.com\u002Fccwq","丐帮37代iphone","2026-11-12 20:21:30",{"source":8,"skillName":5,"commands":49,"prompt":59,"zipAvailable":60},[50,53,56],{"tool":51,"command":52},"npx","npx skills add ccwq\u002Finfocard-pub --skill infocard-source-routing-decision-tree",{"tool":54,"command":55},"bunx","bunx skills add ccwq\u002Finfocard-pub --skill infocard-source-routing-decision-tree",{"tool":57,"command":58},"pnpm","pnpm dlx skills add ccwq\u002Finfocard-pub --skill infocard-source-routing-decision-tree","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Finfocard\u002Finfocard-source-routing-decision-tree\u002FSKILL.md",true,[62],{"path":63,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":66,"sourceHtmlUrl":67},"references\u002Fnon-x-source-extraction-recipes.md","REFERENCE","参考文档",6662,"https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Finfocard\u002Finfocard-source-routing-decision-tree\u002Freferences\u002Fnon-x-source-extraction-recipes.md",[69,76,83,90,97,104],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M74J5D6V3BJYVJXMMXVJ8H30","authorized-infocard-execution","Use for authorized infocard runs after tool failures.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fpublishing\u002Fauthorized-infocard-execution\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M74J5A599BH65TWJZDMTF07S","web-capture","Use when a web page needs screenshots or device simulation. Capture desktop, mobile, and tablet screenshots via agent-browser and return auditable image paths.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fproductivity\u002Fweb-capture\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M74J50DVMHVKYQHC112CYR4K","infocard-x-content-tracing","X content to upstream project\u002Fauthor tracing.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fproductivity\u002Finfocard-x-content-tracing\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":19},"01M74J4WVB5QEJ23QKXKHWT6DK","infocard-visual-pass-loop","Screenshot-and-fix loop when 200 passes but theme is broken.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fproductivity\u002Finfocard-visual-pass-loop\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":19},"01M74J4SF3YN9G53QCRXNNE4CV","infocard-topic-selection","Use when selecting or evaluating information-card topics from the user's profile or explicitly supplied themes; judge value and visual expression potential without choosing a concrete infocard theme or publishing.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fproductivity\u002Finfocard-topic-selection\u002FSKILL.md",[],[],{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":19},"01M74J4KRQYFGM4ZJRRGRV90VE","infocard-theme-validation","Use when changing infocard theme CSS or pushing docs.","https:\u002F\u002Fgithub.com\u002Fccwq\u002Finfocard-pub\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fproductivity\u002Finfocard-theme-validation\u002FSKILL.md",[],[]]