[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6CN7B3ZJYT51DWACFX3YMX3":3,"skill-same:01M6CN7B3ZJYT51DWACFX3YMX3":58},{"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":30,"linkOnly":31,"markdownContent":32,"markdownBody":33,"frontMatter":34,"refName":36,"commitSha":37,"skillDirectory":38,"forksCount":39,"repositoryHtmlUrl":40,"repositoryHomepageUrl":6,"ownerHtmlUrl":41,"ownerBlogUrl":42,"ownerDescription":6,"updatedAt":43,"install":44,"resources":57},"01M6CN7B3ZJYT51DWACFX3YMX3","strategy",null,"Hold the long arc as an auditable record of bets and beliefs — append-only markdown cards under workspace\u002Fstrategy\u002F for theses, assumptions, decisions, and signals, cross-linked by [[slug]] so a thesis resolves to the assumptions holding it up and the signals pressing on it. TRIGGER: a strategic bet stated or revised, an assumption named, a competitor or market signal noticed, a consequential decision made, or a request to review the strategy book. SKIP: operational commitments with an owner and a due date (that is ops) and day-to-day choices with no horizon.","joshuamschultz\u002FArc","joshuamschultz","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F2312349?v=4","01M6CN5XFME1WP2XMKCZBH4MMG","Josh","01M6CN5XMCNJTZPB9HE6N2MD2R",3,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fblueprints\u002Fstrategic-ceo\u002Fskills\u002Fstrategy\u002FSKILL.md",[],[5,19,20,21,22,23,24,25,26,27,28,29],"review","markdown","decision","workspace","decisions","append","ops","only","up","Market","record","2026-11-10 10:12:30",false,"---\nname: strategy\ndescription: \"Hold the long arc as an auditable record of bets and beliefs — append-only markdown cards under workspace\u002Fstrategy\u002F for theses, assumptions, decisions, and signals, cross-linked by [[slug]] so a thesis resolves to the assumptions holding it up and the signals pressing on it. TRIGGER: a strategic bet stated or revised, an assumption named, a competitor or market signal noticed, a consequential decision made, or a request to review the strategy book. SKIP: operational commitments with an owner and a due date (that is ops) and day-to-day choices with no horizon.\"\nversion: 2.0.0\n---\n\n## Files\n\n- `strategy\u002Ftheses\u002F\u003Cslug>.md` — one card per strategic bet.\n- `strategy\u002Fassumptions\u002F\u003Cslug>.md` — one card per assumption a thesis rests on.\n- `strategy\u002Fdecisions\u002F\u003Cslug>.md` — one card per consequential decision.\n- `strategy\u002Fsignals\u002Fsignals.md` — the single running signal log.\n\nAll paths are relative to the agent's own workspace, written with the ordinary\nfile tools (`read`, `write`, `ls`, `grep`). There is no strategy tool to call and\nnone is needed — this is the agent's own workspace.\n\n## Contract\n\n1. Every thesis, assumption, decision, and signal lands on its card at\n   `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md`.\n2. Cards are **appended to, never rewritten**: read, add the entry at the end,\n   write the whole file back. The superseded belief is the record's whole value.\n3. A new card starts with exactly this header and nothing else:\n\n   ```\n   # \u003CTitle>\n\n   - slug: \u003Cslug>\n   - type: \u003Ckind>\n   ```\n\n4. Each entry is a UTC timestamp heading followed by `- key: value` lines,\n   omitting every field with no value.\n5. References between cards are written `[[slug]]`, never a bare name.\n6. **Every assumption records what would falsify it.** An assumption with no\n   `falsified_by` is an opinion, and is reported back as such.\n7. A challenge appends to the assumption's own card with `status: challenged` —\n   it never edits the original entry away.\n\n## Knowledge\n\n- **Thesis and assumption are separated for falsifiability.** A thesis is what is\n  believed; an assumption is the specific claim that, if false, takes the thesis\n  with it. Merged, nothing is ever checkable and a strategy dies of forgetting\n  rather than of being wrong.\n- **Strategy fails less from bad thinking than from forgetting** — the assumption\n  nobody revisited, the competitor move noticed and dropped, the bet that quietly\n  stopped being true. Appending is what makes revisiting possible.\n- **A challenge is evidence, not a verdict.** Record it with its source and leave\n  the verdict `under review` unless one was actually reached.\n- **The slug is the identity.** Lowercase, non-alphanumeric runs → `-`, strip ends.\n- **Get the timestamp from the clock**: run `date -u +'%Y-%m-%d %H:%M'` and use\n  exactly what it returns. The prompt carries the date but not the time.\n- Never invent a horizon, a confidence, or a source. An empty field is honest.\n\n### Fields by kind\n\n| Kind | Fields |\n|---|---|\n| `theses` | `name`, `claim`, `rationale`, `horizon`, `owner`, `status` (default `active`) |\n| `assumptions` | `name`, `thesis` (link), `statement`, `falsified_by`, `confidence`, `status: holding` |\n| `assumptions` (challenge) | `challenge`, `source`, `verdict` (default `under review`), `status: challenged` |\n| `decisions` | `name`, `choice`, `tradeoff`, `reversible`, `thesis` (link), `revisit_on` |\n| `signals` | `signal`, `actor`, `thesis` (link), `implication`, `source` |\n\nFree-text context goes on a final `- note:` line.\n\n## Steps\n\n**To log a thesis, assumption, or decision:**\n\n1. Run `date -u +'%Y-%m-%d %H:%M'`.\n2. Derive the slug from the name.\n3. `read` `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md`; if absent, start from the header block.\n4. Append a blank line, the `## \u003Ctimestamp>` heading, and one `- key: value` line\n   per non-empty field, then `- note:` if there is context.\n5. `write` the whole file back.\n6. Reply in one line. For an assumption with no `falsified_by`, add\n   `⚠ nothing would falsify this — it is an opinion, not an assumption.`\n\n**To challenge an assumption:**\n\n1. Append a new entry to `strategy\u002Fassumptions\u002F\u003Cassumption-slug>.md` carrying\n   `challenge`, `source`, `verdict`, and `status: challenged`.\n2. `grep` the theses for that assumption's slug and name every thesis resting on\n   it in the reply — the point of the challenge is what it puts at risk.\n\n**To log a signal:**\n\nAppend to the shared `strategy\u002Fsignals\u002Fsignals.md`, linking the thesis it bears on.\n\n**To review the strategy book:**\n\n1. `ls` `strategy\u002Ftheses\u002F`, `read` each, take the **last** value of each field.\n2. For each thesis, `grep` the assumptions for its slug.\n3. Report one block per thesis: its latest `status` and `claim`, then its\n   assumptions, marking any whose latest `status` is `challenged`.\n4. A thesis whose assumptions are all challenged is called out as a bet whose\n   footing has gone.\n\n## Red Flags & Rationalizations\n\n| Thought | Reality |\n|---|---|\n| \"The thesis and the assumption are the same claim.\" | Then nothing is falsifiable. Split them: the assumption is what can be checked. |\n| \"The old belief is wrong now, I'll replace it.\" | The superseded belief is the record's value. Append the new one. |\n| \"This assumption is obviously true.\" | Then `falsified_by` is easy to write. If it isn't, it's an opinion. |\n| \"The challenge settles it.\" | Evidence is not a verdict. Leave `under review` unless one was reached. |\n| \"I know roughly what time it is.\" | You know the date, not the time. Run `date -u` or the entry lands at 00:00. |\n| \"The review should show the first status I find.\" | Cards are append-only; the first is the oldest. Take the last. |\n\n## Validation\n\nA log is correct when:\n\n- [ ] The card exists at `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md` with the exact header block.\n- [ ] Every prior entry is present, unmodified.\n- [ ] The `## YYYY-MM-DD HH:MM` heading came from `date -u`, not `00:00`.\n- [ ] Every assumption has a `falsified_by`, or was flagged as an opinion.\n- [ ] A challenge appended to the assumption card with `status: challenged`.\n- [ ] Every cross-reference is `[[slug]]`, not a bare name.\n\nA review is correct when every thesis appears once with its latest status and\nevery challenged assumption beneath it is marked.\n\n## Examples\n\n**An assumption with a falsifier:**\n\n```\n\n## 2026-08-14 17:20\n- name: Mid-market buys on compliance\n- thesis: [[federal-first-wedge]]\n- statement: Mid-market buyers will pay a premium for FedRAMP-grade controls\n- falsified_by: Two consecutive quarters where compliance is absent from won-deal reasons\n- confidence: medium\n- status: holding\n```\n\nReply: `Logged assumption 'Mid-market buys on compliance' under [[federal-first-wedge]]`\n\n**An assumption with nothing that would falsify it:**\n\nReply: `Logged assumption 'AI adoption keeps growing' ⚠ nothing would falsify this — it is an opinion, not an assumption.`\n\n**A review:**\n\n```\nfederal-first-wedge — [active] Win federal first, then descend into mid-market\n  - mid-market-buys-on-compliance: holding\n  - procurement-cycles-shorten: ⚠ challenged (verdict: under review)\n```\n","## Files\n\n- `strategy\u002Ftheses\u002F\u003Cslug>.md` — one card per strategic bet.\n- `strategy\u002Fassumptions\u002F\u003Cslug>.md` — one card per assumption a thesis rests on.\n- `strategy\u002Fdecisions\u002F\u003Cslug>.md` — one card per consequential decision.\n- `strategy\u002Fsignals\u002Fsignals.md` — the single running signal log.\n\nAll paths are relative to the agent's own workspace, written with the ordinary\nfile tools (`read`, `write`, `ls`, `grep`). There is no strategy tool to call and\nnone is needed — this is the agent's own workspace.\n\n## Contract\n\n1. Every thesis, assumption, decision, and signal lands on its card at\n   `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md`.\n2. Cards are **appended to, never rewritten**: read, add the entry at the end,\n   write the whole file back. The superseded belief is the record's whole value.\n3. A new card starts with exactly this header and nothing else:\n\n   ```\n   # \u003CTitle>\n\n   - slug: \u003Cslug>\n   - type: \u003Ckind>\n   ```\n\n4. Each entry is a UTC timestamp heading followed by `- key: value` lines,\n   omitting every field with no value.\n5. References between cards are written `[[slug]]`, never a bare name.\n6. **Every assumption records what would falsify it.** An assumption with no\n   `falsified_by` is an opinion, and is reported back as such.\n7. A challenge appends to the assumption's own card with `status: challenged` —\n   it never edits the original entry away.\n\n## Knowledge\n\n- **Thesis and assumption are separated for falsifiability.** A thesis is what is\n  believed; an assumption is the specific claim that, if false, takes the thesis\n  with it. Merged, nothing is ever checkable and a strategy dies of forgetting\n  rather than of being wrong.\n- **Strategy fails less from bad thinking than from forgetting** — the assumption\n  nobody revisited, the competitor move noticed and dropped, the bet that quietly\n  stopped being true. Appending is what makes revisiting possible.\n- **A challenge is evidence, not a verdict.** Record it with its source and leave\n  the verdict `under review` unless one was actually reached.\n- **The slug is the identity.** Lowercase, non-alphanumeric runs → `-`, strip ends.\n- **Get the timestamp from the clock**: run `date -u +'%Y-%m-%d %H:%M'` and use\n  exactly what it returns. The prompt carries the date but not the time.\n- Never invent a horizon, a confidence, or a source. An empty field is honest.\n\n### Fields by kind\n\n| Kind | Fields |\n|---|---|\n| `theses` | `name`, `claim`, `rationale`, `horizon`, `owner`, `status` (default `active`) |\n| `assumptions` | `name`, `thesis` (link), `statement`, `falsified_by`, `confidence`, `status: holding` |\n| `assumptions` (challenge) | `challenge`, `source`, `verdict` (default `under review`), `status: challenged` |\n| `decisions` | `name`, `choice`, `tradeoff`, `reversible`, `thesis` (link), `revisit_on` |\n| `signals` | `signal`, `actor`, `thesis` (link), `implication`, `source` |\n\nFree-text context goes on a final `- note:` line.\n\n## Steps\n\n**To log a thesis, assumption, or decision:**\n\n1. Run `date -u +'%Y-%m-%d %H:%M'`.\n2. Derive the slug from the name.\n3. `read` `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md`; if absent, start from the header block.\n4. Append a blank line, the `## \u003Ctimestamp>` heading, and one `- key: value` line\n   per non-empty field, then `- note:` if there is context.\n5. `write` the whole file back.\n6. Reply in one line. For an assumption with no `falsified_by`, add\n   `⚠ nothing would falsify this — it is an opinion, not an assumption.`\n\n**To challenge an assumption:**\n\n1. Append a new entry to `strategy\u002Fassumptions\u002F\u003Cassumption-slug>.md` carrying\n   `challenge`, `source`, `verdict`, and `status: challenged`.\n2. `grep` the theses for that assumption's slug and name every thesis resting on\n   it in the reply — the point of the challenge is what it puts at risk.\n\n**To log a signal:**\n\nAppend to the shared `strategy\u002Fsignals\u002Fsignals.md`, linking the thesis it bears on.\n\n**To review the strategy book:**\n\n1. `ls` `strategy\u002Ftheses\u002F`, `read` each, take the **last** value of each field.\n2. For each thesis, `grep` the assumptions for its slug.\n3. Report one block per thesis: its latest `status` and `claim`, then its\n   assumptions, marking any whose latest `status` is `challenged`.\n4. A thesis whose assumptions are all challenged is called out as a bet whose\n   footing has gone.\n\n## Red Flags & Rationalizations\n\n| Thought | Reality |\n|---|---|\n| \"The thesis and the assumption are the same claim.\" | Then nothing is falsifiable. Split them: the assumption is what can be checked. |\n| \"The old belief is wrong now, I'll replace it.\" | The superseded belief is the record's value. Append the new one. |\n| \"This assumption is obviously true.\" | Then `falsified_by` is easy to write. If it isn't, it's an opinion. |\n| \"The challenge settles it.\" | Evidence is not a verdict. Leave `under review` unless one was reached. |\n| \"I know roughly what time it is.\" | You know the date, not the time. Run `date -u` or the entry lands at 00:00. |\n| \"The review should show the first status I find.\" | Cards are append-only; the first is the oldest. Take the last. |\n\n## Validation\n\nA log is correct when:\n\n- [ ] The card exists at `strategy\u002F\u003Ckind>\u002F\u003Cslug>.md` with the exact header block.\n- [ ] Every prior entry is present, unmodified.\n- [ ] The `## YYYY-MM-DD HH:MM` heading came from `date -u`, not `00:00`.\n- [ ] Every assumption has a `falsified_by`, or was flagged as an opinion.\n- [ ] A challenge appended to the assumption card with `status: challenged`.\n- [ ] Every cross-reference is `[[slug]]`, not a bare name.\n\nA review is correct when every thesis appears once with its latest status and\nevery challenged assumption beneath it is marked.\n\n## Examples\n\n**An assumption with a falsifier:**\n\n```\n\n## 2026-08-14 17:20\n- name: Mid-market buys on compliance\n- thesis: [[federal-first-wedge]]\n- statement: Mid-market buyers will pay a premium for FedRAMP-grade controls\n- falsified_by: Two consecutive quarters where compliance is absent from won-deal reasons\n- confidence: medium\n- status: holding\n```\n\nReply: `Logged assumption 'Mid-market buys on compliance' under [[federal-first-wedge]]`\n\n**An assumption with nothing that would falsify it:**\n\nReply: `Logged assumption 'AI adoption keeps growing' ⚠ nothing would falsify this — it is an opinion, not an assumption.`\n\n**A review:**\n\n```\nfederal-first-wedge — [active] Win federal first, then descend into mid-market\n  - mid-market-buys-on-compliance: holding\n  - procurement-cycles-shorten: ⚠ challenged (verdict: under review)\n```\n",{"name":5,"version":35,"description":7},"2.0.0","main","e2fb5cfd858fb29298ab468e61a273f1e372b0e7","blueprints\u002Fstrategic-ceo\u002Fskills\u002Fstrategy",0,"https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz","https:\u002F\u002Fjoshuaschultz.com","2026-10-22 08:51:07",{"source":8,"skillName":5,"commands":45,"prompt":55,"zipAvailable":56},[46,49,52],{"tool":47,"command":48},"npx","npx skills add joshuamschultz\u002FArc --skill strategy",{"tool":50,"command":51},"bunx","bunx skills add joshuamschultz\u002FArc --skill strategy",{"tool":53,"command":54},"pnpm","pnpm dlx skills add joshuamschultz\u002FArc --skill strategy","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fblueprints\u002Fstrategic-ceo\u002Fskills\u002Fstrategy\u002FSKILL.md",true,[],[59,70,101,120,129,142],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":30},"01M6CN91Q5XSKA66KDKFV9BRTQ","notes","Read a note back with the notes_echo tool.","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Ftests\u002Ffixtures\u002Fnotes_connector\u002Fskills\u002Fnotes\u002FSKILL.md",[],[61,66,67,68,69],"back","tool","read","Echo",{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":81,"publishedAt":30},"01M6CN8Z0BXKCQEJGJG21DFPA7","workflow-builder","Turn a repeatable business process described in conversation into a named, validated ArcFlow workflow draft.","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Fsrc\u002Farcagent\u002Fbuiltins\u002Fcapabilities\u002Fskills\u002Fworkflow-builder\u002FSKILL.md",[76],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100],"builder","workflow builder","workflow","create","run","automate this","SET","automate","node","edit","list","business","process","channel","remove","files","build","conversation","do",{"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":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":30},"01M6CN8VZT2YXB8EFJRHSCZ735","update-tool","Update an existing @tool source file in workspace\u002Fcapabilities\u002F, bumping its semver per major\u002Fminor\u002Fpatch.","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Fsrc\u002Farcagent\u002Fbuiltins\u002Fcapabilities\u002Fskills\u002Fupdate-tool\u002FSKILL.md",[],[108,109,67,110,111,112,113,114,115,68,116,117,118,119,22],"update tool","update","bump","bug","behavior","patch","version","fix","change","semver","capabilities","file",{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":127,"publishedAt":30},"01M6CN8SJRCCV05HMF91ZNAK8Z","update-skill","Update an existing skill's SKILL.md body in workspace\u002Fcapabilities\u002Fskills\u002F, bumping its frontmatter semver.","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Fsrc\u002Farcagent\u002Fbuiltins\u002Fcapabilities\u002Fskills\u002Fupdate-skill\u002FSKILL.md",[126],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},[109,128,117,118,22,68,115,114,110],"update skill",{"skillNo":130,"name":131,"nameZh":6,"description":132,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":133,"categories":134,"keywords":135,"publishedAt":30},"01M6CN8PRNV3S8D4BZN9EYHWVZ","create-tool","Author a new @tool-decorated Python file in the agent workspace, validate it, and register it via reload().","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Fsrc\u002Farcagent\u002Fbuiltins\u002Fcapabilities\u002Fskills\u002Fcreate-tool\u002FSKILL.md",[],[136,67,85,137,138,139,119,140,22,68,141,98],"create tool","agent workspace","python","validate","Bash","write",{"skillNo":143,"name":144,"nameZh":6,"description":145,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":146,"categories":147,"keywords":149,"publishedAt":30},"01M6CN8M1G6VJPG9KQN9CA04BZ","create-skill","Scaffold a new skill folder under workspace\u002Fcapabilities\u002Fskills\u002F with frontmatter and the seven required sections.","https:\u002F\u002Fgithub.com\u002Fjoshuamschultz\u002FArc\u002Fblob\u002Fmain\u002Fpackages\u002Farcagent\u002Fsrc\u002Farcagent\u002Fbuiltins\u002Fcapabilities\u002Fskills\u002Fcreate-skill\u002FSKILL.md",[148],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},[85,150,84,151,118,152,22,153,68,141,91,154,155],"create skill","scaffold","write a skill","capture","document","new skill"]