[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M753WSJWBNNGR2H89T5DXNPQ":3,"skill-same:01M753WSJWBNNGR2H89T5DXNPQ":61},{"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":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":32,"install":33,"resources":46},"01M753WSJWBNNGR2H89T5DXNPQ","research",null,"Multi-source dev research with parallel scouts. Scales effort to the question's shape, writes a plan and per-angle notes to a run folder, then synthesizes a cited report with a recommendation first. Every claim traces to a note finding with a URL. Inferred triggers are confirmed before a deep run.","jabworks\u002Fagentic-toolkit","jabworks","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F159094119?v=4","01M753P4JBZBTT7WN1PJBPZJZP","JAB Works","01M753P4K5KDGEWGXWT811BZET",2,"MIT","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: research\ndescription: Multi-source dev research with parallel scouts. Scales effort to the question's shape, writes a plan and per-angle notes to a run folder, then synthesizes a cited report with a recommendation first. Every claim traces to a note finding with a URL. Inferred triggers are confirmed before a deep run.\nwhen_to_use: Comparing libraries, frameworks or approaches; surveying how others solve X; \"research X\"; ecosystem or standards surveys that need several sources. Not for a single API or signature lookup (researcher agent, or answer directly); not for implementation tasks (\u002Fworkflow); not for explaining code (answer directly).\nargument-hint: \"[question to research]\"\n---\n\n# \u002Fresearch\n\nBreak a question into angles, send scouts at them in parallel, and return a report you can trust.\n\n## 1. Clarify (only if it changes the research)\n\nAsk only when the answer would change *what gets researched*: which options to compare, a version or time bound, the deciding criterion. Max 3 questions, all at once. Otherwise state your assumptions in `plan.md` and go.\n\n## 2. Classify the shape\n\n| Shape | Example | Scouts |\n|---|---|---|\n| Single fact or API | \"what does `useId` return\" | **0** — answer directly or hand to `researcher` |\n| Comparison | \"Zustand vs Jotai vs Valtio\" | **2–4** — one per option, plus one for criteria if needed |\n| Survey | \"how do teams handle offline sync in 2026\" | **5–6** — one per distinct angle |\n\nMore than 6 needs the user's yes first. Name the cost when you ask: multi-agent research ran at about 15x the tokens of a chat turn in Anthropic's own measurement.\n\n**Depth gate.** How the skill was reached decides whether to fan out:\n\n| Reached by | What happens |\n|---|---|\n| **Explicit** — the user named the skill (`\u002Fcondux:research`, \"use the research skill\", \"condux:research\") or asked for depth in words (\"deep research\", \"thorough\", \"comprehensive\", \"cited report\", \"dig into\") | Proceed with the classified scout count. Do not ask. |\n| **Inferred** — the skill loaded because a question sounded like research (routing hook or trigger description) | Single fact or API: answer directly, never ask. Comparison or survey: ask ONCE, recommendation marked: \"Quick answer (one pass, about 1x cost) or deep research (N scouts, about 15x cost)?\" N is the classified count. Recommend quick for a comparison of well-known options, deep for a survey or a comparison where the user's constraints decide it. |\n\n**Quick answer** means: the main session answers in one pass, with a few searches and fetches through the host's search and fetch tools, sources cited inline, no run folder and no scouts. Same honesty rules: a version or date on dev claims, and say what could not be verified. If the user picks quick, answer and stop; offer deep research in one line only if the answer is visibly thin.\n\n## 3. Open the run folder\n\n```\n.condux\u002Fresearch\u002F\u003CYYYY-MM-DD>-\u003Cslug>\u002F\n  plan.md\n  notes\u002F\u003Cangle>.md\n  report.md\n```\n\nNever overwrite a run folder. A follow-up opens a new one and names the earlier run as background in its `plan.md`.\n\n**Bootstrap.** `.condux\u002F` is created on demand at the git root. Before the first write, check it is ignored:\n\n```bash\ngit check-ignore -q .condux\u002F || echo \"not ignored\"\n```\n\nIf not, offer once: \"condux keeps its working files in `.condux\u002F` — add it to `.gitignore` so they stay out of your commits?\" On yes, append `.condux\u002F` to `.gitignore`; if the user would rather not touch a tracked file, use `.git\u002Finfo\u002Fexclude`. Never edit either without asking. Not a git repo: fall back to CWD and say so once.\n\n## 4. Write `plan.md`\n\nBefore any dispatch, write: the sub-questions, the scout count with the reason, and **every brief verbatim**. Recording the notes paths here also makes a stray write by a scout visible later.\n\n## 5. Dispatch — all scouts in ONE message\n\nParallel only works if every call goes out together. Dispatch the named `scout` agent and nothing else — never a general-purpose or custom-prompt agent.\n\nA vague brief makes scouts repeat each other or wander. Each brief carries:\n\n| Part | Rule |\n|---|---|\n| Objective | one angle, with a clear boundary against its siblings |\n| Key questions | the specific things the notes must answer |\n| Suggested sources | starting points (official docs, changelogs, repos) |\n| Constraints | the user's version and time bounds, carried through unchanged |\n| Notes path | absolute: `\u003Cgit-root>\u002F.condux\u002Fresearch\u002F\u003Crun>\u002Fnotes\u002F\u003Cangle>.md` |\n\nWait for every scout to return before step 6. Their summaries say what was covered; the evidence is in the notes files.\n\n## 6. Gap round (at most one)\n\nRead every notes file. If a *critical* gap remains — one that blocks the recommendation — dispatch a single follow-up round of scouts aimed at only those gaps. Minor gaps go into the report as unknowns. A second round needs the user's consent.\n\n## 7. Synthesize `report.md`\n\nRead the notes from disk, not from scout summaries.\n\n| Part | Contains |\n|---|---|\n| Recommendation \u002F answer | first, decision-ready, with the deciding reason |\n| Comparison table | when the question has options |\n| Evidence sections | claims with inline `([source](url))` |\n| Risks and unknowns | unresolved conflicts and gaps, stated plainly |\n| Sources | the full list, each with its version or date |\n\nWhen notes conflict, present both sides — don't pick one quietly.\n\n## 8. Citation check\n\nWalk the report claim by claim. Each must trace to a note finding that has a URL. Drop any claim that doesn't, or keep it visibly flagged as unverified. No re-fetching — the notes are the evidence.\n\n## 9. Deliver\n\nIn chat: 3–5 lines (the answer, the confidence, the biggest unknown) and the report path.\n\nOffer promotion **once**: \"Does this feed a spec? If so, which one?\" On a named spec:\n\n1. Copy the report to `specs\u002F\u003Cthat-spec>\u002Fresearch-\u003CYYYY-MM-DD>-\u003Ctopic>.md`.\n2. Add a link row to that spec's `index.md`.\n3. The copy keeps external URLs only. It must never cite a `.condux\u002F` path.\n\nNo fitting spec: the report stays in `.condux\u002F`.\n\n## Sequential fallback\n\nIf `scout` cannot be spawned — no agent tool, or no scout agent (Cursor, or Codex before the agent installer is re-run) — say so once: \"no scout agent here — researching sequentially, slower, same output.\"\n\nThen do the work yourself: take each angle in turn, using the host's search and fetch tools under the same rules (10–15 calls per angle, primary sources, version or date on every finding, conflicts recorded). Write the same `notes\u002F\u003Cangle>.md` files in the scout's shape, one block per key question, then continue from step 7:\n\n```\n## \u003Ckey question>\n**Answer**: \u003C1–2 lines>\n**Findings**\n- \u003Cclaim> — \u003Csource URL> — \u003Csource version or date>\n**Conflicts**\n- \u003Csource A says X; source B says Y — both cited>\n**Gaps**\n- \u003Cwhat could not be found, and why>\n```\n\nA finding without a URL goes under Gaps.\n\n## What does NOT happen\n\n- No scouts for a single fact — answer it or use `researcher`\n- No fan-out from an inferred trigger without asking\n- No more than 6 scouts without the user's yes and the cost named\n- No dispatch before `plan.md` holds the briefs\n- No general-purpose or custom-prompt agent standing in for `scout`\n- No overwriting a run folder\n- No second gap round without consent\n- No claim in the report without a URL in the notes\n- No `.condux\u002F` path cited from a promoted copy\n- No promotion without asking which spec\n","# \u002Fresearch\n\nBreak a question into angles, send scouts at them in parallel, and return a report you can trust.\n\n## 1. Clarify (only if it changes the research)\n\nAsk only when the answer would change *what gets researched*: which options to compare, a version or time bound, the deciding criterion. Max 3 questions, all at once. Otherwise state your assumptions in `plan.md` and go.\n\n## 2. Classify the shape\n\n| Shape | Example | Scouts |\n|---|---|---|\n| Single fact or API | \"what does `useId` return\" | **0** — answer directly or hand to `researcher` |\n| Comparison | \"Zustand vs Jotai vs Valtio\" | **2–4** — one per option, plus one for criteria if needed |\n| Survey | \"how do teams handle offline sync in 2026\" | **5–6** — one per distinct angle |\n\nMore than 6 needs the user's yes first. Name the cost when you ask: multi-agent research ran at about 15x the tokens of a chat turn in Anthropic's own measurement.\n\n**Depth gate.** How the skill was reached decides whether to fan out:\n\n| Reached by | What happens |\n|---|---|\n| **Explicit** — the user named the skill (`\u002Fcondux:research`, \"use the research skill\", \"condux:research\") or asked for depth in words (\"deep research\", \"thorough\", \"comprehensive\", \"cited report\", \"dig into\") | Proceed with the classified scout count. Do not ask. |\n| **Inferred** — the skill loaded because a question sounded like research (routing hook or trigger description) | Single fact or API: answer directly, never ask. Comparison or survey: ask ONCE, recommendation marked: \"Quick answer (one pass, about 1x cost) or deep research (N scouts, about 15x cost)?\" N is the classified count. Recommend quick for a comparison of well-known options, deep for a survey or a comparison where the user's constraints decide it. |\n\n**Quick answer** means: the main session answers in one pass, with a few searches and fetches through the host's search and fetch tools, sources cited inline, no run folder and no scouts. Same honesty rules: a version or date on dev claims, and say what could not be verified. If the user picks quick, answer and stop; offer deep research in one line only if the answer is visibly thin.\n\n## 3. Open the run folder\n\n```\n.condux\u002Fresearch\u002F\u003CYYYY-MM-DD>-\u003Cslug>\u002F\n  plan.md\n  notes\u002F\u003Cangle>.md\n  report.md\n```\n\nNever overwrite a run folder. A follow-up opens a new one and names the earlier run as background in its `plan.md`.\n\n**Bootstrap.** `.condux\u002F` is created on demand at the git root. Before the first write, check it is ignored:\n\n```bash\ngit check-ignore -q .condux\u002F || echo \"not ignored\"\n```\n\nIf not, offer once: \"condux keeps its working files in `.condux\u002F` — add it to `.gitignore` so they stay out of your commits?\" On yes, append `.condux\u002F` to `.gitignore`; if the user would rather not touch a tracked file, use `.git\u002Finfo\u002Fexclude`. Never edit either without asking. Not a git repo: fall back to CWD and say so once.\n\n## 4. Write `plan.md`\n\nBefore any dispatch, write: the sub-questions, the scout count with the reason, and **every brief verbatim**. Recording the notes paths here also makes a stray write by a scout visible later.\n\n## 5. Dispatch — all scouts in ONE message\n\nParallel only works if every call goes out together. Dispatch the named `scout` agent and nothing else — never a general-purpose or custom-prompt agent.\n\nA vague brief makes scouts repeat each other or wander. Each brief carries:\n\n| Part | Rule |\n|---|---|\n| Objective | one angle, with a clear boundary against its siblings |\n| Key questions | the specific things the notes must answer |\n| Suggested sources | starting points (official docs, changelogs, repos) |\n| Constraints | the user's version and time bounds, carried through unchanged |\n| Notes path | absolute: `\u003Cgit-root>\u002F.condux\u002Fresearch\u002F\u003Crun>\u002Fnotes\u002F\u003Cangle>.md` |\n\nWait for every scout to return before step 6. Their summaries say what was covered; the evidence is in the notes files.\n\n## 6. Gap round (at most one)\n\nRead every notes file. If a *critical* gap remains — one that blocks the recommendation — dispatch a single follow-up round of scouts aimed at only those gaps. Minor gaps go into the report as unknowns. A second round needs the user's consent.\n\n## 7. Synthesize `report.md`\n\nRead the notes from disk, not from scout summaries.\n\n| Part | Contains |\n|---|---|\n| Recommendation \u002F answer | first, decision-ready, with the deciding reason |\n| Comparison table | when the question has options |\n| Evidence sections | claims with inline `([source](url))` |\n| Risks and unknowns | unresolved conflicts and gaps, stated plainly |\n| Sources | the full list, each with its version or date |\n\nWhen notes conflict, present both sides — don't pick one quietly.\n\n## 8. Citation check\n\nWalk the report claim by claim. Each must trace to a note finding that has a URL. Drop any claim that doesn't, or keep it visibly flagged as unverified. No re-fetching — the notes are the evidence.\n\n## 9. Deliver\n\nIn chat: 3–5 lines (the answer, the confidence, the biggest unknown) and the report path.\n\nOffer promotion **once**: \"Does this feed a spec? If so, which one?\" On a named spec:\n\n1. Copy the report to `specs\u002F\u003Cthat-spec>\u002Fresearch-\u003CYYYY-MM-DD>-\u003Ctopic>.md`.\n2. Add a link row to that spec's `index.md`.\n3. The copy keeps external URLs only. It must never cite a `.condux\u002F` path.\n\nNo fitting spec: the report stays in `.condux\u002F`.\n\n## Sequential fallback\n\nIf `scout` cannot be spawned — no agent tool, or no scout agent (Cursor, or Codex before the agent installer is re-run) — say so once: \"no scout agent here — researching sequentially, slower, same output.\"\n\nThen do the work yourself: take each angle in turn, using the host's search and fetch tools under the same rules (10–15 calls per angle, primary sources, version or date on every finding, conflicts recorded). Write the same `notes\u002F\u003Cangle>.md` files in the scout's shape, one block per key question, then continue from step 7:\n\n```\n## \u003Ckey question>\n**Answer**: \u003C1–2 lines>\n**Findings**\n- \u003Cclaim> — \u003Csource URL> — \u003Csource version or date>\n**Conflicts**\n- \u003Csource A says X; source B says Y — both cited>\n**Gaps**\n- \u003Cwhat could not be found, and why>\n```\n\nA finding without a URL goes under Gaps.\n\n## What does NOT happen\n\n- No scouts for a single fact — answer it or use `researcher`\n- No fan-out from an inferred trigger without asking\n- No more than 6 scouts without the user's yes and the cost named\n- No dispatch before `plan.md` holds the briefs\n- No general-purpose or custom-prompt agent standing in for `scout`\n- No overwriting a run folder\n- No second gap round without consent\n- No claim in the report without a URL in the notes\n- No `.condux\u002F` path cited from a promoted copy\n- No promotion without asking which spec\n",{"name":5,"description":7,"when_to_use":24,"argument-hint":25},"Comparing libraries, frameworks or approaches; surveying how others solve X; \"research X\"; ecosystem or standards surveys that need several sources. Not for a single API or signature lookup (researcher agent, or answer directly); not for implementation tasks (\u002Fworkflow); not for explaining code (answer directly).","[question to research]","main","5277649b071c1c0cdf4f15eee861b1608fc1bbcb","skills\u002Fresearch",0,"https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit","https:\u002F\u002Fgithub.com\u002Fjabworks","2026-11-03 15:03:51",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add jabworks\u002Fagentic-toolkit --skill research",{"tool":39,"command":40},"bunx","bunx skills add jabworks\u002Fagentic-toolkit --skill research",{"tool":42,"command":43},"pnpm","pnpm dlx skills add jabworks\u002Fagentic-toolkit --skill research","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002FSKILL.md",true,[47,53,57],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"evals\u002Fgate_eval.json","REFERENCE","参考文档",626,"https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Fevals\u002Fgate_eval.json",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"evals\u002Fquality_eval.json",2234,"https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Fevals\u002Fquality_eval.json",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"evals\u002Ftrigger_eval.json",2657,"https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fresearch\u002Fevals\u002Ftrigger_eval.json",[62,68,74,81,88,95],{"skillNo":63,"name":64,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7540BPNTFZH7V41P5E2S692","orphan","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcomposition\u002Fbad\u002Fskills\u002Forphan\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7540AG2KFTQ8DJMNV6806MV","no-blurb","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcomposition\u002Fbad\u002Fskills\u002Fno-blurb\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M75404N306PY1JFPQ0GGNTYB","workflow","Routes any dev task into the right execution tier (Small\u002FMedium\u002FLarge) and carries condux's operating rules — proportional effort, lazy skill loading, soft gates, implement-yourself-by-default.","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M753ZZ6TTA06GRJAG679BQ2H","usage-trace","Checks and manages durable, local usage capture — ingest, health status, and the opt-in limit snapshot — for Claude Code and Codex. Capture itself runs automatically at session start.","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fusage-trace\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M753ZSKJSF8WFS7AXR2N65SE","toolkit-skill-standards","Use when writing or reviewing a SKILL.md for jabworks\u002Fagentic-toolkit — frontmatter budgets, the trigger contract (\"Use when…\" description or when_to_use field), progressive disclosure into references\u002F, and keeping a new description out of existing skills' trigger space. Triggers include \"review this skill description\", \"write the frontmatter\", \"does this collide with an existing skill\", \"where should this content live\". Not for scaffold\u002Fsync mechanics; use toolkit-foundry.","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Ftoolkit-skill-standards\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M753ZKX3KTVYWX15A9A0WZN0","toolkit-research-frontier","Use when planning improvement work on jabworks\u002Fagentic-toolkit itself — open problems (trigger-collision automation, YAML-strict frontmatter checks, docs-catalog enforcement), existing assets, the next three concrete steps, and the four-front library-health campaign with commands and success criteria. Triggers include \"let's improve our toolkit\", \"let's improve this repo\", \"what should we improve next\", \"toolkit roadmap\", \"open problems\", \"run the health campaign\".","https:\u002F\u002Fgithub.com\u002Fjabworks\u002Fagentic-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Ftoolkit-research-frontier\u002FSKILL.md",[],[]]