[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6GNH4QFZP7H5KA12DR5PCNJ":3,"skill-same:01M6GNH4QFZP7H5KA12DR5PCNJ":75},{"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":30,"linkOnly":31,"markdownContent":32,"markdownBody":33,"frontMatter":34,"refName":36,"commitSha":37,"skillDirectory":38,"forksCount":39,"repositoryHtmlUrl":40,"repositoryHomepageUrl":6,"ownerHtmlUrl":41,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":42,"install":43,"resources":56},"01M6GNH4QFZP7H5KA12DR5PCNJ","research",null,"Use when finding sources through iterative web exploration and synthesis","unclejobs-ai\u002Fsecond-claude-code","unclejobs-ai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F150198827?v=4","01M6GNFYGQ7354R1N04YF1B269","Unclejobs","01M6GNFYNCQEKYTVX572HB0F26",41,"MIT","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000021","资料检索与研究","SCENARIO","使用场景",[5,24,25,26,27,28,29],"web","use","Synthesis","exploration","false","sources","2026-11-10 10:12:30",false,"---\nname: research\ndescription: \"Use when finding sources through iterative web exploration and synthesis\"\neffort: high\nuser-invocable: false\n---\n\n[Resolve runtime paths](..\u002Fruntime-paths.md) before operations.\n\n## Iron Law\n\n> **No guessing. If there is no data, collect it first.**\n\n## Red Flags\n\n- \"This one article covers everything\" → STOP. Single-source = blind spots. Depth minimums (3\u002F5\u002F10) exist for a reason.\n- \"I already know the answer, just need a citation\" → STOP. Confirmation bias produces cherry-picked sources.\n- \"Source is from 2022 but topic hasn't changed\" → STOP. Staleness must be verified, not assumed.\n- \"Skip the analyst step, just list links\" → STOP. Raw link dumps without gap analysis are rejected.\n- \"Page was blocked so I'll summarize from the URL title\" → STOP. Use fallback chain (Jina → Playwright → Gaps), never hallucinate from metadata.\n\n# Research\n\nAutonomous multi-round web research that produces structured Research Briefs.\n\n## When to Use\n\n- User asks to research a topic, trend, or question\n- Another skill (e.g., `\u002Fscc:write`, `\u002Fscc:analyze`) needs source material\n- User wants competitive intelligence, market data, or literature review\n\n### Web Engine\n\nThe researcher uses **Jina Search** (`s.jina.ai`) as the primary tool. Fallback chain: Jina Search → WebSearch + WebFetch → the **unblock chain** (`\u002Fscc:unblock`, zero-key 9-phase) → Playwright. See `references\u002Fjina-guide.md` and `docs\u002Fskills\u002Funblock.md`.\n\n## MMBridge Enhancement\n\nWhen mmbridge is detected (see `references\u002Fmmbridge-integration.md`), dispatch `mmbridge research` in parallel with the internal researcher.\n\n### Dispatch\n\nAt Step 1, also run via Bash:\n\n```bash\nmmbridge research \"\u003Ctopic>\" --type \u003Ctype> --json > \u002Ftmp\u002Fmmbridge-research-${RUN_ID}.json\n```\n\n- `--type code-aware`: when topic relates to the current codebase\n- `--type open`: for general topics unrelated to code\n- `--depth shallow`: skip mmbridge\n- `--depth medium|deep`: mmbridge enabled\n\n### Merge\n\nAt Step 3, parse the mmbridge JSON output file as supplemental source material. Count each distinct cited URL from mmbridge as one source and deduplicate overlaps during gap analysis.\n\n### Step-by-Step\n\n0. Auto-load `references\u002Fresearch-methodology.md` and `references\u002Fjina-guide.md`.\n1. Dispatch **researcher** (sonnet): run depth-appropriate Jina Search calls with varied query phrasings. If `$JINA_API_KEY` is unavailable, fall back to WebSearch + WebFetch. If mmbridge is available at medium or deep depth, dispatch it in parallel.\n2. Validate sources: reject login walls, blocked pages, empty content. For blocked URLs, run the unblock chain (`node \"\u003Cplugin-root>\u002Fskills\u002Funblock\u002Fengine\u002Fcli.mjs\" \"\u003CURL>\" --json`); read the trace before any retry.\n3. Dispatch **analyst** (sonnet): merge internal and mmbridge findings, identify gaps, and apply conflict-resolution rules from `references\u002Fresearch-methodology.md`.\n4. Optional second round: only when depth allows and the analyst found critical gaps.\n5. Dispatch **writer** (sonnet): synthesize the final brief with conflict annotations.\n6. Verify actual search counts before output. If counts exceed the depth cap, discard excess results and re-synthesize.\n\n## Options\n\n| Flag | Values | Default | Effect |\n|------|--------|---------|--------|\n| `--depth` | `shallow\\|medium\\|deep` | `medium` | HARD limit on search rounds — see Depth Behavior |\n| `--sources` | `web\\|academic\\|news` | `web` | Constrains search domain |\n| `--lang` | `ko\\|en\\|auto` | `auto` | Output language. When called from another skill (write, analyze), inherits the caller's `--lang` value. |\n| `--interactive` | flag | off | Force Playwright for all URL fetches (useful for SPAs, dashboards) |\n| `--engine` | `jina\\|legacy` | `jina` | `jina`: use Jina Search API. `legacy`: use WebSearch + WebFetch |\n\n### Depth Behavior\n\n- **shallow**: EXACTLY 3 Jina Search calls. No deep reads. No gap analysis round.\n- **medium**: EXACTLY 5 Jina Search calls plus up to 2 Jina Reader deep reads.\n- **deep**: 10+ Jina Search calls, unlimited Jina Reader deep reads, and at most 3 gap-fill rounds before documenting remaining gaps.\n\nWhen `--engine legacy` is set or `$JINA_API_KEY` is unavailable, replace \"Jina Search\" with \"WebSearch\" and \"Jina Reader\" with \"WebFetch\" in the above counts.\n\nCoverage requirements and conflict resolution rules are in `references\u002Fresearch-methodology.md`.\n\n### Source Domain (`--sources`)\n\n- **web** (default): General web search. No domain restrictions.\n- **academic**: Bias at least half the calls toward Google Scholar or arXiv.\n- **news**: Add `after:30d` and prefer established editorial sources over blogs.\n\n## Fallback Chain\n\n```\nJina Search (s.jina.ai) ─── primary: search + content in one call\n  └─ blocked\u002Fempty → Jina Reader (r.jina.ai) with X-Engine: browser\n       └─ still fails → Playwright (browser_navigate + browser_snapshot)\n            └─ unavailable → note in Gaps & Limitations\n```\n\nWhen `--engine legacy` or `$JINA_API_KEY` is unavailable:\n```\nWebSearch → WebFetch → Playwright → Gaps & Limitations\n```\n\nWhen `--interactive` is set, skip Jina\u002FWebFetch and use Playwright for every URL.\n\n**Cost controls**: Max 3 Playwright navigations per round. Exceeding the cap means skip the rest and note the gap.\n\n**Graceful degradation**: Both Jina and Playwright are optional. If `$JINA_API_KEY` is missing, fall back silently. If Playwright MCP is unavailable, proceed without it and note unreachable URLs.\n\n## Auto-Save\n\n- Path: `.captures\u002Fresearch-{slug}-{YYYY-MM-DD}.md`\n- `{slug}` = topic lowercased, spaces to hyphens, max 40 chars\n- Write the full brief using the Write tool and tell the user the saved path.\n\n## Gotchas\n\n| Failure Mode | Mitigation |\n|-------------|------------|\n| Stops after 1 search | researcher MUST meet depth minimums: shallow=3, medium=5, deep=10. Fewer = restart. |\n| Lists links without analysis | analyst step is required. Raw link dumps are rejected. |\n| Hallucinated sources | Every URL must come from an actual Jina Search\u002FWebSearch result. writer cannot invent URLs. |\n| Duplicate queries | researcher must vary phrasing with synonyms and different angles per query. |\n| Jina API key missing | Fall back to WebSearch + WebFetch silently. Do not error. |\n\n## Subagents\n\n```yaml\nresearcher: { model: sonnet, tools: [Bash, WebSearch, WebFetch, browser_navigate, browser_snapshot], constraint: \"Jina Search when $JINA_API_KEY set, else WebSearch+WebFetch; meet depth minimums, vary phrasing, validate content, flag staleness; Playwright only when Jina+WebFetch both fail; max 3 navigations\" }\nanalyst: { model: sonnet, tools: [], constraint: \"produce gap list, flag data conflicts, verify coverage requirements\" }\nwriter: { model: sonnet, tools: [], constraint: \"every claim needs a source, no invented URLs, include conflict annotations\" }\n```\n\n## Output & Integration\n\nSee `references\u002Fresearch-methodology.md` for output format template, extended gotchas table, and integration notes.\n","[Resolve runtime paths](..\u002Fruntime-paths.md) before operations.\n\n## Iron Law\n\n> **No guessing. If there is no data, collect it first.**\n\n## Red Flags\n\n- \"This one article covers everything\" → STOP. Single-source = blind spots. Depth minimums (3\u002F5\u002F10) exist for a reason.\n- \"I already know the answer, just need a citation\" → STOP. Confirmation bias produces cherry-picked sources.\n- \"Source is from 2022 but topic hasn't changed\" → STOP. Staleness must be verified, not assumed.\n- \"Skip the analyst step, just list links\" → STOP. Raw link dumps without gap analysis are rejected.\n- \"Page was blocked so I'll summarize from the URL title\" → STOP. Use fallback chain (Jina → Playwright → Gaps), never hallucinate from metadata.\n\n# Research\n\nAutonomous multi-round web research that produces structured Research Briefs.\n\n## When to Use\n\n- User asks to research a topic, trend, or question\n- Another skill (e.g., `\u002Fscc:write`, `\u002Fscc:analyze`) needs source material\n- User wants competitive intelligence, market data, or literature review\n\n### Web Engine\n\nThe researcher uses **Jina Search** (`s.jina.ai`) as the primary tool. Fallback chain: Jina Search → WebSearch + WebFetch → the **unblock chain** (`\u002Fscc:unblock`, zero-key 9-phase) → Playwright. See `references\u002Fjina-guide.md` and `docs\u002Fskills\u002Funblock.md`.\n\n## MMBridge Enhancement\n\nWhen mmbridge is detected (see `references\u002Fmmbridge-integration.md`), dispatch `mmbridge research` in parallel with the internal researcher.\n\n### Dispatch\n\nAt Step 1, also run via Bash:\n\n```bash\nmmbridge research \"\u003Ctopic>\" --type \u003Ctype> --json > \u002Ftmp\u002Fmmbridge-research-${RUN_ID}.json\n```\n\n- `--type code-aware`: when topic relates to the current codebase\n- `--type open`: for general topics unrelated to code\n- `--depth shallow`: skip mmbridge\n- `--depth medium|deep`: mmbridge enabled\n\n### Merge\n\nAt Step 3, parse the mmbridge JSON output file as supplemental source material. Count each distinct cited URL from mmbridge as one source and deduplicate overlaps during gap analysis.\n\n### Step-by-Step\n\n0. Auto-load `references\u002Fresearch-methodology.md` and `references\u002Fjina-guide.md`.\n1. Dispatch **researcher** (sonnet): run depth-appropriate Jina Search calls with varied query phrasings. If `$JINA_API_KEY` is unavailable, fall back to WebSearch + WebFetch. If mmbridge is available at medium or deep depth, dispatch it in parallel.\n2. Validate sources: reject login walls, blocked pages, empty content. For blocked URLs, run the unblock chain (`node \"\u003Cplugin-root>\u002Fskills\u002Funblock\u002Fengine\u002Fcli.mjs\" \"\u003CURL>\" --json`); read the trace before any retry.\n3. Dispatch **analyst** (sonnet): merge internal and mmbridge findings, identify gaps, and apply conflict-resolution rules from `references\u002Fresearch-methodology.md`.\n4. Optional second round: only when depth allows and the analyst found critical gaps.\n5. Dispatch **writer** (sonnet): synthesize the final brief with conflict annotations.\n6. Verify actual search counts before output. If counts exceed the depth cap, discard excess results and re-synthesize.\n\n## Options\n\n| Flag | Values | Default | Effect |\n|------|--------|---------|--------|\n| `--depth` | `shallow\\|medium\\|deep` | `medium` | HARD limit on search rounds — see Depth Behavior |\n| `--sources` | `web\\|academic\\|news` | `web` | Constrains search domain |\n| `--lang` | `ko\\|en\\|auto` | `auto` | Output language. When called from another skill (write, analyze), inherits the caller's `--lang` value. |\n| `--interactive` | flag | off | Force Playwright for all URL fetches (useful for SPAs, dashboards) |\n| `--engine` | `jina\\|legacy` | `jina` | `jina`: use Jina Search API. `legacy`: use WebSearch + WebFetch |\n\n### Depth Behavior\n\n- **shallow**: EXACTLY 3 Jina Search calls. No deep reads. No gap analysis round.\n- **medium**: EXACTLY 5 Jina Search calls plus up to 2 Jina Reader deep reads.\n- **deep**: 10+ Jina Search calls, unlimited Jina Reader deep reads, and at most 3 gap-fill rounds before documenting remaining gaps.\n\nWhen `--engine legacy` is set or `$JINA_API_KEY` is unavailable, replace \"Jina Search\" with \"WebSearch\" and \"Jina Reader\" with \"WebFetch\" in the above counts.\n\nCoverage requirements and conflict resolution rules are in `references\u002Fresearch-methodology.md`.\n\n### Source Domain (`--sources`)\n\n- **web** (default): General web search. No domain restrictions.\n- **academic**: Bias at least half the calls toward Google Scholar or arXiv.\n- **news**: Add `after:30d` and prefer established editorial sources over blogs.\n\n## Fallback Chain\n\n```\nJina Search (s.jina.ai) ─── primary: search + content in one call\n  └─ blocked\u002Fempty → Jina Reader (r.jina.ai) with X-Engine: browser\n       └─ still fails → Playwright (browser_navigate + browser_snapshot)\n            └─ unavailable → note in Gaps & Limitations\n```\n\nWhen `--engine legacy` or `$JINA_API_KEY` is unavailable:\n```\nWebSearch → WebFetch → Playwright → Gaps & Limitations\n```\n\nWhen `--interactive` is set, skip Jina\u002FWebFetch and use Playwright for every URL.\n\n**Cost controls**: Max 3 Playwright navigations per round. Exceeding the cap means skip the rest and note the gap.\n\n**Graceful degradation**: Both Jina and Playwright are optional. If `$JINA_API_KEY` is missing, fall back silently. If Playwright MCP is unavailable, proceed without it and note unreachable URLs.\n\n## Auto-Save\n\n- Path: `.captures\u002Fresearch-{slug}-{YYYY-MM-DD}.md`\n- `{slug}` = topic lowercased, spaces to hyphens, max 40 chars\n- Write the full brief using the Write tool and tell the user the saved path.\n\n## Gotchas\n\n| Failure Mode | Mitigation |\n|-------------|------------|\n| Stops after 1 search | researcher MUST meet depth minimums: shallow=3, medium=5, deep=10. Fewer = restart. |\n| Lists links without analysis | analyst step is required. Raw link dumps are rejected. |\n| Hallucinated sources | Every URL must come from an actual Jina Search\u002FWebSearch result. writer cannot invent URLs. |\n| Duplicate queries | researcher must vary phrasing with synonyms and different angles per query. |\n| Jina API key missing | Fall back to WebSearch + WebFetch silently. Do not error. |\n\n## Subagents\n\n```yaml\nresearcher: { model: sonnet, tools: [Bash, WebSearch, WebFetch, browser_navigate, browser_snapshot], constraint: \"Jina Search when $JINA_API_KEY set, else WebSearch+WebFetch; meet depth minimums, vary phrasing, validate content, flag staleness; Playwright only when Jina+WebFetch both fail; max 3 navigations\" }\nanalyst: { model: sonnet, tools: [], constraint: \"produce gap list, flag data conflicts, verify coverage requirements\" }\nwriter: { model: sonnet, tools: [], constraint: \"every claim needs a source, no invented URLs, include conflict annotations\" }\n```\n\n## Output & Integration\n\nSee `references\u002Fresearch-methodology.md` for output format template, extended gotchas table, and integration notes.\n",{"name":5,"effort":35,"description":7,"user-invocable":31},"high","main","3e2938e2698dc6fc652571205aad03aa1205ec3d","skills\u002Fresearch",9,"https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code","https:\u002F\u002Fgithub.com\u002Funclejobs-ai","2026-11-10 10:15:09",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add unclejobs-ai\u002Fsecond-claude-code --skill research",{"tool":49,"command":50},"bunx","bunx skills add unclejobs-ai\u002Fsecond-claude-code --skill research",{"tool":52,"command":53},"pnpm","pnpm dlx skills add unclejobs-ai\u002Fsecond-claude-code --skill research","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002FSKILL.md",true,[57,63,67,71],{"path":58,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":61,"sourceHtmlUrl":62},"gotchas.md","REFERENCE","参考文档",940,"https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Fgotchas.md",{"path":64,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":65,"sourceHtmlUrl":66},"references\u002Fjina-guide.md",4771,"https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Freferences\u002Fjina-guide.md",{"path":68,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":69,"sourceHtmlUrl":70},"references\u002Fplaywright-guide.md",3715,"https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Freferences\u002Fplaywright-guide.md",{"path":72,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":73,"sourceHtmlUrl":74},"references\u002Fresearch-methodology.md",5644,"https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Freferences\u002Fresearch-methodology.md",[76,87,102,114,125,138],{"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":30},"01M6GNHTFEKCG6X5Q0VCH0T2C1","write","Use when producing newsletters, articles, reports, shorts, or social content","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fwrite\u002FSKILL.md",[],[78,83,84,25,85,86,28],"content","social content","reports","social",{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":96,"publishedAt":30},"01M6GNHP58KN60BQ8T7QXCC43D","workflow","Use when chaining multiple \u002Fscc commands into a reusable workflow","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002FSKILL.md",[93],{"categoryNo":94,"categoryName":95,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[89,97,98,25,99,28,100,101],"reusable","chaining","commands","scc","true",{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":30},"01M6GNHHQZJG8AVTCYPYPFYTNE","translate","Use when translating content between English and Korean with formatting preservation","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Ftranslate\u002FSKILL.md",[],[104,25,109,110,28,111,112,113,101,83],"korean","formatting","preservation","english","with",{"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":30},"01M6GNHD52AQE5YCZKRK6D71ZX","soul","Use when observing user patterns and synthesizing a persistent identity profile","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Fsoul\u002FSKILL.md",[],[116,101,121,122,123,25,124,28],"identity","user","profile","patterns",{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":131,"publishedAt":30},"01M6GNH8AX5EGA8BWF2JTK5HR0","review","Use when benchmarking content, strategy, or code against a quality bar with parallel specialized reviewers","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Freview\u002FSKILL.md",[],[127,132,133,134,83,135,136,137,25,28,113],"specialized","benchmarking","code","parallel","quality","strategy",{"skillNo":139,"name":140,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":142,"categories":143,"keywords":144,"publishedAt":30},"01M6GNH1MYFXC7BP9MW004W6HG","refine","Use when improving a draft until it meets a review target — the improve step after Check","https:\u002F\u002Fgithub.com\u002Funclejobs-ai\u002Fsecond-claude-code\u002Fblob\u002Fmain\u002Fskills\u002Frefine\u002FSKILL.md",[],[140,145,146,25,147,28,148,127],"improve","check","draft","it"]