[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M788G0CF6F2N10PX8K8TXKWD":3,"skill-same:01M788G0CF6F2N10PX8K8TXKWD":44},{"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":24,"commitSha":25,"skillDirectory":5,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M788G0CF6F2N10PX8K8TXKWD","reconcile",null,"Reconcile the OKF knowledge bundle (knowledge\u002F) with what a completed code change built, over that change's diff. Use once a change has finished landing — the build closeout, just before committing — or when the user asks to reconcile knowledge against code.","evist0\u002Fokf-matt-skills","evist0","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F32814970?v=4","01M788F6GHSVY6BVJ2HC65ECWV","Danil Tankov","01M788F6HD2SSXYGAAC67YD47C",18,"MIT","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Freconcile\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: reconcile\ndescription: Reconcile the OKF knowledge bundle (knowledge\u002F) with what a completed code change built, over that change's diff. Use once a change has finished landing — the build closeout, just before committing — or when the user asks to reconcile knowledge against code.\n---\n\n# Reconcile\n\nA change landed: the code moved, but the OKF bundle (`knowledge\u002F`) may not have. **Reconcile** them — bring durable knowledge back into agreement with what was actually built, so the next session starts from a consistent bundle.\n\nThis is the *writing* discipline of the **build** phase — the mirror of `\u002Fdomain-modeling`, which writes the bundle during **design**. It runs last in the `\u002Fimplement` closeout, over the same diff `\u002Fvet` judged, so code and knowledge land in one commit.\n\n## Scope\n\nReconcile against the diff, never the whole bundle. Fixed point is the one `\u002Fvet` used, or one the user supplies: `git diff \u003Cfixed-point>...HEAD`.\n\nDurable knowledge is everything under `knowledge\u002F` — the canonical categories `glossary\u002F`, `adr\u002F`, and `spec\u002F`, plus any project-specific concept category registered in `knowledge\u002Findex.md`. Touch a concept only when the diff gives you a reason to.\n\n## Process\n\nReconcile runs in two phases: a **sub-agent** analyses the diff and applies the mechanical fixes, then the **main context** does the judgement writes. The split keeps the heavy bundle-searching out of the main context — as `\u002Fvet` does — while leaving the consequential writes where the implementation reasoning lives.\n\n### Phase 1 — Analysis (sub-agent)\n\nFirst write a **build-rationale brief** for the sub-agent — this is what keeps its classification from being blind. Keep it to a few hundred words:\n\n- **Pointers, not copies** — the spec\u002Fissue reference and the commit list (`git log \u003Cfixed-point>..HEAD --oneline`). The sub-agent reads these itself; don't paste them.\n- **The unwritten \"why\"** — a short synthesis of the non-obvious decisions made while implementing: the \"chose X over Y because…\" that lives in no file. This is the only input the sub-agent cannot obtain on its own, and the thing its judgement of unrecorded\u002Fviolated decisions depends on.\n\nThen spawn one `general-purpose` sub-agent, passing it: the diff command `git diff \u003Cfixed-point>...HEAD`, the brief, and the **Classes** table below. Its brief:\n\n1. **Gather candidates** — two nets over the diff:\n   - **Referenced** — every concept whose `implements:` frontmatter or `## Implementation mapping` paths intersect a changed file. Grep the bundle for the changed paths.\n   - **Unrecorded ground** — code built in the diff that no concept references at all. This is where a new decision hides. (Detection is diff-derived and does not depend on the brief; the brief only informs whether the ground is consequential.)\n\n   The candidate set is the union. A referenced concept the diff didn't actually affect drops out here.\n2. **Classify** each candidate against the diff and the brief, using the Classes table. Where you cannot confidently pick one class, do not guess — flag the candidate as ambiguous, carrying both readings. Classifying by guess is the one way this skill does harm: a mislabelled violation gets laundered into the record.\n3. **Apply referential fixes itself** — fix each reference inline and log them (see **Log**). This is the only class the sub-agent writes.\n4. **Return a proposal** — for every contentious candidate (stale \u002F unrecorded \u002F violated): the class, the changed hunk(s), the relevant concept, and the evidence for the class. Do not write these.\n\n### Classes\n\nEach candidate's **class** fixes who is authoritative and what may be written:\n\n| Class | Authoritative | Why |\n|---|---|---|\n| **Referential** — a mapped path was renamed\u002Fmoved\u002Fdeleted, a `depends-on` link no longer resolves, a `lifecycle` flag lags reality | code | the record points at code that no longer exists as written |\n| **Stale description** — the concept's prose describes behaviour the diff legitimately changed | code | the model evolved; the record is behind |\n| **Unrecorded decision** — the diff commits to something hard to reverse, non-obvious, and consequential that no concept holds | the gap | a real decision was made and never written down |\n| **Violated decision** — the diff contradicts an *accepted* ADR or architecture concept | knowledge | the record is a guard; the code, not the record, is the suspect |\n\nAct on each class:\n\n- **Referential** → fixed in Phase 1 by the sub-agent (above).\n- **Stale description** → Phase 2: show the rewritten prose for that concept, one concept at a time; write on the user's confirmation.\n- **Unrecorded decision** → Phase 2: propose an ADR or concept, *sparingly* — only when the decision clears all three of hard-to-reverse, non-obvious, consequential. Write on confirmation. Most changes clear none and record nothing. If the unrecorded ground is a new *kind* of entity that fits no existing category (`glossary`, `adr`, `spec`, or a registered one), the extensibility rule applies: propose a new category, get the user's sanction, and register it in `knowledge\u002Findex.md` before writing. Default to the existing categories.\n- **Violated decision** → Phase 2: leave the recorded decision exactly as written and surface it as a finding. A live contradiction with an accepted decision is a signal to fix the code or reopen the ADR deliberately — never to quietly edit the record to match the breach.\n\n### Phase 2 — Confirm & write (main context)\n\nTake the sub-agent's proposal and act on each contentious candidate by its class, as the bullets above specify.\n\nFor any candidate the sub-agent flagged as ambiguous, **put it to the user**: state both readings, and what each costs if you pick wrong. Leave nothing unclassified.\n\n### Log\n\nEvery write to the bundle appends a dated section to `knowledge\u002Flog.md` (newest first, ISO date heading) and refreshes the touched concept's `timestamp`. Two writers append here: the sub-agent's auto-applied referential fixes enter as **one summary block**, not a line each; the main context's Phase 2 writes go in **after** the sub-agent returns, newest on top.\n\n## Completion\n\nDone when:\n\n- every **referenced** candidate is classified and actioned,\n- every changed file has been checked for **unrecorded ground**,\n- every bundle write is mirrored in `knowledge\u002Flog.md`, and\n- every divergence you could not classify has been put to the user.\n\nA candidate left unclassified, or a write missing from the log, means not done.\n","# Reconcile\n\nA change landed: the code moved, but the OKF bundle (`knowledge\u002F`) may not have. **Reconcile** them — bring durable knowledge back into agreement with what was actually built, so the next session starts from a consistent bundle.\n\nThis is the *writing* discipline of the **build** phase — the mirror of `\u002Fdomain-modeling`, which writes the bundle during **design**. It runs last in the `\u002Fimplement` closeout, over the same diff `\u002Fvet` judged, so code and knowledge land in one commit.\n\n## Scope\n\nReconcile against the diff, never the whole bundle. Fixed point is the one `\u002Fvet` used, or one the user supplies: `git diff \u003Cfixed-point>...HEAD`.\n\nDurable knowledge is everything under `knowledge\u002F` — the canonical categories `glossary\u002F`, `adr\u002F`, and `spec\u002F`, plus any project-specific concept category registered in `knowledge\u002Findex.md`. Touch a concept only when the diff gives you a reason to.\n\n## Process\n\nReconcile runs in two phases: a **sub-agent** analyses the diff and applies the mechanical fixes, then the **main context** does the judgement writes. The split keeps the heavy bundle-searching out of the main context — as `\u002Fvet` does — while leaving the consequential writes where the implementation reasoning lives.\n\n### Phase 1 — Analysis (sub-agent)\n\nFirst write a **build-rationale brief** for the sub-agent — this is what keeps its classification from being blind. Keep it to a few hundred words:\n\n- **Pointers, not copies** — the spec\u002Fissue reference and the commit list (`git log \u003Cfixed-point>..HEAD --oneline`). The sub-agent reads these itself; don't paste them.\n- **The unwritten \"why\"** — a short synthesis of the non-obvious decisions made while implementing: the \"chose X over Y because…\" that lives in no file. This is the only input the sub-agent cannot obtain on its own, and the thing its judgement of unrecorded\u002Fviolated decisions depends on.\n\nThen spawn one `general-purpose` sub-agent, passing it: the diff command `git diff \u003Cfixed-point>...HEAD`, the brief, and the **Classes** table below. Its brief:\n\n1. **Gather candidates** — two nets over the diff:\n   - **Referenced** — every concept whose `implements:` frontmatter or `## Implementation mapping` paths intersect a changed file. Grep the bundle for the changed paths.\n   - **Unrecorded ground** — code built in the diff that no concept references at all. This is where a new decision hides. (Detection is diff-derived and does not depend on the brief; the brief only informs whether the ground is consequential.)\n\n   The candidate set is the union. A referenced concept the diff didn't actually affect drops out here.\n2. **Classify** each candidate against the diff and the brief, using the Classes table. Where you cannot confidently pick one class, do not guess — flag the candidate as ambiguous, carrying both readings. Classifying by guess is the one way this skill does harm: a mislabelled violation gets laundered into the record.\n3. **Apply referential fixes itself** — fix each reference inline and log them (see **Log**). This is the only class the sub-agent writes.\n4. **Return a proposal** — for every contentious candidate (stale \u002F unrecorded \u002F violated): the class, the changed hunk(s), the relevant concept, and the evidence for the class. Do not write these.\n\n### Classes\n\nEach candidate's **class** fixes who is authoritative and what may be written:\n\n| Class | Authoritative | Why |\n|---|---|---|\n| **Referential** — a mapped path was renamed\u002Fmoved\u002Fdeleted, a `depends-on` link no longer resolves, a `lifecycle` flag lags reality | code | the record points at code that no longer exists as written |\n| **Stale description** — the concept's prose describes behaviour the diff legitimately changed | code | the model evolved; the record is behind |\n| **Unrecorded decision** — the diff commits to something hard to reverse, non-obvious, and consequential that no concept holds | the gap | a real decision was made and never written down |\n| **Violated decision** — the diff contradicts an *accepted* ADR or architecture concept | knowledge | the record is a guard; the code, not the record, is the suspect |\n\nAct on each class:\n\n- **Referential** → fixed in Phase 1 by the sub-agent (above).\n- **Stale description** → Phase 2: show the rewritten prose for that concept, one concept at a time; write on the user's confirmation.\n- **Unrecorded decision** → Phase 2: propose an ADR or concept, *sparingly* — only when the decision clears all three of hard-to-reverse, non-obvious, consequential. Write on confirmation. Most changes clear none and record nothing. If the unrecorded ground is a new *kind* of entity that fits no existing category (`glossary`, `adr`, `spec`, or a registered one), the extensibility rule applies: propose a new category, get the user's sanction, and register it in `knowledge\u002Findex.md` before writing. Default to the existing categories.\n- **Violated decision** → Phase 2: leave the recorded decision exactly as written and surface it as a finding. A live contradiction with an accepted decision is a signal to fix the code or reopen the ADR deliberately — never to quietly edit the record to match the breach.\n\n### Phase 2 — Confirm & write (main context)\n\nTake the sub-agent's proposal and act on each contentious candidate by its class, as the bullets above specify.\n\nFor any candidate the sub-agent flagged as ambiguous, **put it to the user**: state both readings, and what each costs if you pick wrong. Leave nothing unclassified.\n\n### Log\n\nEvery write to the bundle appends a dated section to `knowledge\u002Flog.md` (newest first, ISO date heading) and refreshes the touched concept's `timestamp`. Two writers append here: the sub-agent's auto-applied referential fixes enter as **one summary block**, not a line each; the main context's Phase 2 writes go in **after** the sub-agent returns, newest on top.\n\n## Completion\n\nDone when:\n\n- every **referenced** candidate is classified and actioned,\n- every changed file has been checked for **unrecorded ground**,\n- every bundle write is mirrored in `knowledge\u002Flog.md`, and\n- every divergence you could not classify has been put to the user.\n\nA candidate left unclassified, or a write missing from the log, means not done.\n",{"name":5,"description":7},"main","f3ea5610b95df91d4f50fb07d76b3fe1548e05f9",4,"https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills","https:\u002F\u002Fgithub.com\u002Fevist0","2026-07-09 03:00:22",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add evist0\u002Fokf-matt-skills --skill reconcile",{"tool":36,"command":37},"bunx","bunx skills add evist0\u002Fokf-matt-skills --skill reconcile",{"tool":39,"command":40},"pnpm","pnpm dlx skills add evist0\u002Fokf-matt-skills --skill reconcile","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Freconcile\u002FSKILL.md",true,[],[45,52,59,66,73,80],{"skillNo":46,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":49,"categories":50,"keywords":51,"publishedAt":19},"01M788H00W6M0Y2S2VDGXN6XPY","write-skill","Reference for writing and editing skills well — the vocabulary and principles that make a skill predictable.","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Fwrite-skill\u002FSKILL.md",[],[],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M788GY0ZW3K1Y1NWH0ES6S3F","which-skill","Ask which skill or flow fits your situation. A router over the skills in this repo.","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Fwhich-skill\u002FSKILL.md",[],[],{"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":19},"01M788GV67Q09BEARBX3PCKZB9","wayfinder","Plan a huge chunk of work — more than one agent session can hold — as a shared map of investigation tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Fwayfinder\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M788GPMNP9KQ02GC0AEVN22M","vet","Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue\u002FPRD asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to \"review since X\".","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Fvet\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M788GM23Z4V9DZH7WX5D1XHJ","triage","Move issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Ftriage\u002FSKILL.md",[],[],{"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":86,"publishedAt":19},"01M788GGNFSFAM4WNHNDQXRHD9","to-tickets","Break a plan, spec, or the current conversation into a set of tracer-bullet tickets, each declaring its blocking edges, published to the configured tracker — edges as text in a local file, or native blocking links on a real tracker.","https:\u002F\u002Fgithub.com\u002Fevist0\u002Fokf-matt-skills\u002Fblob\u002Fmain\u002Fto-tickets\u002FSKILL.md",[],[]]