[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79AFG5AM8BNP37788BZQ59H":3,"skill-same:01M79AFG5AM8BNP37788BZQ59H":47},{"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":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M79AFG5AM8BNP37788BZQ59H","zeugnis-cycle",null,"Use when the user asks for a bounded self-optimization loop that re-runs zeugnis's audits until convergence or a pass cap, wants to 'run the zeugnis cycle', 'keep fixing findings until done', or asks for propose-only vs fix-mode passes. Runs at most max_passes_per_invocation (default 5) passes, applies fixes only in fixable domains, and stops early on convergence — every bound is enforced by scripts\u002Fcycle_engine.py, not by this file's instructions.","Anselmoo\u002Fwerkstoff","Anselmoo","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F13209783?v=4","01M799YTB3D7RZ4P3B8CSR2X85","Anselm Hahn","01M799YTBRAM6719D9V0NRGB4R",1,"MIT","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Fplugins\u002Fzeugnis\u002Fskills\u002Fzeugnis-cycle\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: zeugnis-cycle\ndescription: \"Use when the user asks for a bounded self-optimization loop that re-runs zeugnis's audits until convergence or a pass cap, wants to 'run the zeugnis cycle', 'keep fixing findings until done', or asks for propose-only vs fix-mode passes. Runs at most max_passes_per_invocation (default 5) passes, applies fixes only in fixable domains, and stops early on convergence — every bound is enforced by scripts\u002Fcycle_engine.py, not by this file's instructions.\"\n---\n\nRun zeugnis's bounded self-optimization loop: repeatedly find the\n\"constraint\" domain (the one with the most urgent open findings), audit\nit, and — in fix mode — apply or draft fixes, until the loop converges or\nhits its pass cap. The pass cap, the reopen thrash-guard, the fixable-\ndomain gating, and the convergence check are all enforced by\n`scripts\u002Fcycle_engine.py` and `scripts\u002Flib\u002Fledger.py`; this file only\ndescribes how to drive that engine. If your own count of passes ever\ndisagrees with what the engine reports, trust the engine — it is the\nsource of truth, not your running tally.\n\n## Setup (once per invocation)\n\n1. Determine `repo_root`, `mode` (`\"fix\"` or `\"propose\"` — default\n   `\"propose\"` unless the user explicitly asked to apply fixes),\n   `max_passes` (default 5, never pass a value above 20 to the engine —\n   it will refuse), and `max_reopens` (default 3, never above 10).\n2. Generate one `invocation_id` for this whole run (e.g.\n   `python3 -c \"import uuid; print(uuid.uuid4())\"`) and reuse it for every\n   symbol-index build this invocation needs — never build a fresh one per\n   pass or per domain. This is what makes the symbol-index snapshot\n   shared rather than rebuilt (rule: symbol-index-shared-per-invocation).\n   Resolve it now:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fsymbol_index_cli.py\" resolve \u003Crepo_root> --invocation-id \u003Cinvocation_id>\n   ```\n   This is the \"CLI wrapper the cycle skill points you at\" that\n   `zeugnis-assertion-audit` and `zeugnis-contract-drift`'s \"Shared symbol\n   index\" sections already reference — tell those domain skills this same\n   `invocation_id` when dispatching them so they hit the cache instead of\n   rebuilding.\n\n## Per-pass loop\n\nRepeat the following until the engine tells you to stop:\n\n3. Ask the engine what to do next:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fcycle_engine.py\" plan-next-pass \u003Crepo_root> \\\n       --max-passes \u003Cmax_passes> --mode \u003Cmode>\n   ```\n   If this exits 2 with a stderr JSON object naming `\"maxPasses\"`\n   (`{\"error\": \"...\", \"maxPasses\": N}`), the pass cap is already\n   reached — stop the loop immediately and go to \"Wrap up\" below. Any\n   other non-zero exit is a different failure (a bad `repo_root`, a\n   corrupted ledger, an invalid `--mode`) surfacing as an uncaught\n   crash rather than this deliberate refusal — surface that error to\n   the user and stop; do not treat it as \"cap reached\" or proceed to\n   \"Wrap up\" as if the loop finished normally. Otherwise it prints\n   `{\"passNumber\": N, \"domain\": \"...\", \"mode\": \"...\"}`.\n4. Run that domain's audit skill's Find (and, per that domain's own\n   rules, Verify) phase, scoped efficiently since you likely already have\n   recent findings from a prior invocation — reuse `analysis\u002Fzeugnis\u002F\u003Cdomain>_summary.json`\n   if it's fresh, otherwise re-run the audit as that skill's own SKILL.md\n   describes. \"Fresh\" means the same check `scripts\u002Fstatus_dashboard.py`\n   already implements: the summary file's mtime is not older than\n   `git -C \u003Crepo_root> log -1 --format=%ct` (its\n   `staleRelativeToLatestCommit` field) — not a recorded commit or\n   fingerprint field, which no domain `*_summary.json` stores. Do not weaken that domain's own mandatory-verification rules\n   just because you're inside a cycle pass — `assertion_audit` still runs\n   its Verify phase unconditionally, `contract_drift` and\n   `agentic_reliability` still run theirs unless explicitly skipped.\n5. For each open finding in the constraint domain, decide its handling:\n   - **mode is `\"propose\"`**: never apply anything. For a fixable finding\n     (see below), still just describe the proposed fix in your report —\n     do not dispatch `zeugnis-remediator` at all in propose mode. Record\n     that finding's outcome (step 6) as `\"blocked\"` with a `blockReason`\n     of `\"not applied: cycle is in propose mode\"` — there is no separate\n     outcome for \"described but not fixed\"; `\"blocked\"` is the existing\n     value that keeps the finding open without claiming a fix happened.\n   - **mode is `\"fix\"` and the finding IS in zeugnis's fixable set**\n     (`dependency_audit` any category, `contract_drift` any category,\n     `agentic_reliability` category `excessive-tool-grant` only — check\n     with the same table `scripts\u002Flib\u002Fconstants.py FIXABLE_DOMAINS`\n     encodes, don't just guess): open a remediation scope, dispatch the\n     remediator, then close the scope:\n     ```\n     python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fremediation_scope_cli.py\" open \u003Crepo_root> \\\n         --finding-id \u003Cid> --domain \u003Cdomain> --category \u003Ccategory> \\\n         --target-file \u003Crepo-relative file>\n     ```\n     If this exits 3, the finding is NOT fixable — the script refused to\n     even open a scope for it. Treat it as advisory and do not dispatch\n     `zeugnis-remediator`; this can happen even in fix mode if a finding's\n     category was miscategorized upstream, and the script is the\n     authority here, not your own read of the finding.\n     If it exits 0, dispatch `zeugnis-remediator` with exactly this one\n     finding. A `PreToolUse` hook enforces that its one `Edit` call must\n     target the locked file and that no second edit is possible in this\n     scope — you do not need to police that yourself. After the\n     remediator returns, run:\n     ```\n     python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fremediation_scope_cli.py\" close \u003Crepo_root>\n     ```\n   - **the finding is in `assertion_audit`, or `agentic_reliability`\n     outside `excessive-tool-grant`**: this is draft-only or advisory —\n     never open a remediation scope for it, never dispatch\n     `zeugnis-remediator`. For `assertion_audit` specifically, you may ask\n     the `assertion-auditor` agent to run in **Suggest mode** to draft a\n     replacement assertion, but that draft is reported, never applied —\n     no `Edit` call happens for it, ever.\n6. Build this pass's outcome record and tell the engine:\n   ```json\n   {\"passNumber\": N, \"domain\": \"...\", \"mode\": \"fix\"|\"propose\",\n    \"outcomes\": [{\"findingId\": \"...\", \"category\": \"...\", \"evidence\": \"...\",\n                   \"severity\": \"...\", \"outcome\": \"fixed\"|\"drafted\"|\"blocked\"|\"reopened\",\n                   \"blockReason\": \"... (required when outcome is \\\"blocked\\\")\"}]}\n   ```\n   Use `\"fixed\"` only for a finding `zeugnis-remediator` returned\n   `status: \"applied\"` for. Use `\"blocked\"` for one it returned\n   `status: \"blocked\"` for, or one you determined is not fixable — and\n   always copy its `reason` into `blockReason` here; the engine rejects\n   (exit 2) any `\"blocked\"` outcome with an empty `blockReason` rather\n   than accepting a blocked finding with no explanation on record. Use\n   `\"drafted\"` for an assertion-audit Suggest-mode draft. Use\n   `\"reopened\"` for a finding that was closed in a prior pass\u002Finvocation\n   but is open again this pass. Write this to a scratch JSON file and run:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fcycle_engine.py\" record-pass-result \u003Crepo_root> \\\n       --pass-json \u003Cpath> --max-reopens \u003Cmax_reopens> --max-passes \u003Cmax_passes>\n   ```\n   This is the ONLY place ledger state changes — do not hand-edit\n   `analysis\u002Fzeugnis\u002Fledger.json` yourself. The script rejects (exit 2) any\n   `\"fixed\"` outcome for a non-fixable domain\u002Fcategory or for a\n   `\"propose\"`-mode pass; if you see that rejection, you made a mistake\n   in step 5 — fix your outcome record, don't retry with a different\n   claim to force it through.\n   It prints `{\"converged\": bool, \"shouldContinue\": bool, ...}`. If\n   `shouldContinue` is `false`, stop the loop — either the pass cap was\n   reached or the pass converged (closed zero findings and produced zero\n   fix\u002Fdraft outcomes). Do not start another `plan-next-pass` call after\n   `shouldContinue: false`. In `\"propose\"` mode, `\"blocked\"` outcomes never\n   count toward fix\u002Fdraft outcomes, so a propose-mode pass can report\n   `\"converged\": true` on its very first pass regardless of how many\n   findings were described — do not read that as \"nothing needs\n   attention\" when summarizing for the user; report the open findings\n   themselves. `converged` is also forced `false` while any verified\n   domain's sidecar has `unverifiedCount > 0` (or an unreadable sidecar):\n   the output then lists them under `unverifiedBlockers`. That is an\n   abstain, not a pass: re-run that domain's Verify phase rather than\n   starting more passes (`shouldContinue` stays `false`).\n\n## Wrap up\n\n7. `analysis\u002Fzeugnis\u002Freports\u002FZEUGNIS_CYCLE.md` already has one line appended per\n   pass by the engine. Summarize it for the user: total passes run,\n   which of the three stop reasons applied — convergence, the pass cap, or\n   **blocked by unverified findings** (`converged: false` with\n   `unverifiedBlockers`; name each blocked domain and say its Verify phase\n   must be re-run — that is not convergence and not a clean result) —\n   findings closed, findings escalated (reopened past `max_reopens`), and\n   anything left open.\n\n## What NOT to do\n\n- Do not run more than `max_passes` passes in one invocation, even if the\n  user asks you to \"just keep going\" — a fresh invocation (a new call to\n  this skill) is how they get more passes; `plan-next-pass` will refuse\n  once the cap is hit regardless of what you're asked.\n- Do not dispatch `zeugnis-remediator` for an `assertion_audit` finding or\n  a non-`excessive-tool-grant` `agentic_reliability` finding under any\n  framing — the remediation-scope script refuses to open a scope for\n  these, and the `PreToolUse` hook refuses the edit even if you tried to\n  route around the scope step.\n- Do not treat a `\"blocked\"` remediator outcome as a failure to retry\n  immediately — record it as `\"blocked\"` and let it surface as an open\n  finding for the next pass or the user's manual attention.\n","Run zeugnis's bounded self-optimization loop: repeatedly find the\n\"constraint\" domain (the one with the most urgent open findings), audit\nit, and — in fix mode — apply or draft fixes, until the loop converges or\nhits its pass cap. The pass cap, the reopen thrash-guard, the fixable-\ndomain gating, and the convergence check are all enforced by\n`scripts\u002Fcycle_engine.py` and `scripts\u002Flib\u002Fledger.py`; this file only\ndescribes how to drive that engine. If your own count of passes ever\ndisagrees with what the engine reports, trust the engine — it is the\nsource of truth, not your running tally.\n\n## Setup (once per invocation)\n\n1. Determine `repo_root`, `mode` (`\"fix\"` or `\"propose\"` — default\n   `\"propose\"` unless the user explicitly asked to apply fixes),\n   `max_passes` (default 5, never pass a value above 20 to the engine —\n   it will refuse), and `max_reopens` (default 3, never above 10).\n2. Generate one `invocation_id` for this whole run (e.g.\n   `python3 -c \"import uuid; print(uuid.uuid4())\"`) and reuse it for every\n   symbol-index build this invocation needs — never build a fresh one per\n   pass or per domain. This is what makes the symbol-index snapshot\n   shared rather than rebuilt (rule: symbol-index-shared-per-invocation).\n   Resolve it now:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fsymbol_index_cli.py\" resolve \u003Crepo_root> --invocation-id \u003Cinvocation_id>\n   ```\n   This is the \"CLI wrapper the cycle skill points you at\" that\n   `zeugnis-assertion-audit` and `zeugnis-contract-drift`'s \"Shared symbol\n   index\" sections already reference — tell those domain skills this same\n   `invocation_id` when dispatching them so they hit the cache instead of\n   rebuilding.\n\n## Per-pass loop\n\nRepeat the following until the engine tells you to stop:\n\n3. Ask the engine what to do next:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fcycle_engine.py\" plan-next-pass \u003Crepo_root> \\\n       --max-passes \u003Cmax_passes> --mode \u003Cmode>\n   ```\n   If this exits 2 with a stderr JSON object naming `\"maxPasses\"`\n   (`{\"error\": \"...\", \"maxPasses\": N}`), the pass cap is already\n   reached — stop the loop immediately and go to \"Wrap up\" below. Any\n   other non-zero exit is a different failure (a bad `repo_root`, a\n   corrupted ledger, an invalid `--mode`) surfacing as an uncaught\n   crash rather than this deliberate refusal — surface that error to\n   the user and stop; do not treat it as \"cap reached\" or proceed to\n   \"Wrap up\" as if the loop finished normally. Otherwise it prints\n   `{\"passNumber\": N, \"domain\": \"...\", \"mode\": \"...\"}`.\n4. Run that domain's audit skill's Find (and, per that domain's own\n   rules, Verify) phase, scoped efficiently since you likely already have\n   recent findings from a prior invocation — reuse `analysis\u002Fzeugnis\u002F\u003Cdomain>_summary.json`\n   if it's fresh, otherwise re-run the audit as that skill's own SKILL.md\n   describes. \"Fresh\" means the same check `scripts\u002Fstatus_dashboard.py`\n   already implements: the summary file's mtime is not older than\n   `git -C \u003Crepo_root> log -1 --format=%ct` (its\n   `staleRelativeToLatestCommit` field) — not a recorded commit or\n   fingerprint field, which no domain `*_summary.json` stores. Do not weaken that domain's own mandatory-verification rules\n   just because you're inside a cycle pass — `assertion_audit` still runs\n   its Verify phase unconditionally, `contract_drift` and\n   `agentic_reliability` still run theirs unless explicitly skipped.\n5. For each open finding in the constraint domain, decide its handling:\n   - **mode is `\"propose\"`**: never apply anything. For a fixable finding\n     (see below), still just describe the proposed fix in your report —\n     do not dispatch `zeugnis-remediator` at all in propose mode. Record\n     that finding's outcome (step 6) as `\"blocked\"` with a `blockReason`\n     of `\"not applied: cycle is in propose mode\"` — there is no separate\n     outcome for \"described but not fixed\"; `\"blocked\"` is the existing\n     value that keeps the finding open without claiming a fix happened.\n   - **mode is `\"fix\"` and the finding IS in zeugnis's fixable set**\n     (`dependency_audit` any category, `contract_drift` any category,\n     `agentic_reliability` category `excessive-tool-grant` only — check\n     with the same table `scripts\u002Flib\u002Fconstants.py FIXABLE_DOMAINS`\n     encodes, don't just guess): open a remediation scope, dispatch the\n     remediator, then close the scope:\n     ```\n     python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fremediation_scope_cli.py\" open \u003Crepo_root> \\\n         --finding-id \u003Cid> --domain \u003Cdomain> --category \u003Ccategory> \\\n         --target-file \u003Crepo-relative file>\n     ```\n     If this exits 3, the finding is NOT fixable — the script refused to\n     even open a scope for it. Treat it as advisory and do not dispatch\n     `zeugnis-remediator`; this can happen even in fix mode if a finding's\n     category was miscategorized upstream, and the script is the\n     authority here, not your own read of the finding.\n     If it exits 0, dispatch `zeugnis-remediator` with exactly this one\n     finding. A `PreToolUse` hook enforces that its one `Edit` call must\n     target the locked file and that no second edit is possible in this\n     scope — you do not need to police that yourself. After the\n     remediator returns, run:\n     ```\n     python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fremediation_scope_cli.py\" close \u003Crepo_root>\n     ```\n   - **the finding is in `assertion_audit`, or `agentic_reliability`\n     outside `excessive-tool-grant`**: this is draft-only or advisory —\n     never open a remediation scope for it, never dispatch\n     `zeugnis-remediator`. For `assertion_audit` specifically, you may ask\n     the `assertion-auditor` agent to run in **Suggest mode** to draft a\n     replacement assertion, but that draft is reported, never applied —\n     no `Edit` call happens for it, ever.\n6. Build this pass's outcome record and tell the engine:\n   ```json\n   {\"passNumber\": N, \"domain\": \"...\", \"mode\": \"fix\"|\"propose\",\n    \"outcomes\": [{\"findingId\": \"...\", \"category\": \"...\", \"evidence\": \"...\",\n                   \"severity\": \"...\", \"outcome\": \"fixed\"|\"drafted\"|\"blocked\"|\"reopened\",\n                   \"blockReason\": \"... (required when outcome is \\\"blocked\\\")\"}]}\n   ```\n   Use `\"fixed\"` only for a finding `zeugnis-remediator` returned\n   `status: \"applied\"` for. Use `\"blocked\"` for one it returned\n   `status: \"blocked\"` for, or one you determined is not fixable — and\n   always copy its `reason` into `blockReason` here; the engine rejects\n   (exit 2) any `\"blocked\"` outcome with an empty `blockReason` rather\n   than accepting a blocked finding with no explanation on record. Use\n   `\"drafted\"` for an assertion-audit Suggest-mode draft. Use\n   `\"reopened\"` for a finding that was closed in a prior pass\u002Finvocation\n   but is open again this pass. Write this to a scratch JSON file and run:\n   ```\n   python3 -B \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Fcycle_engine.py\" record-pass-result \u003Crepo_root> \\\n       --pass-json \u003Cpath> --max-reopens \u003Cmax_reopens> --max-passes \u003Cmax_passes>\n   ```\n   This is the ONLY place ledger state changes — do not hand-edit\n   `analysis\u002Fzeugnis\u002Fledger.json` yourself. The script rejects (exit 2) any\n   `\"fixed\"` outcome for a non-fixable domain\u002Fcategory or for a\n   `\"propose\"`-mode pass; if you see that rejection, you made a mistake\n   in step 5 — fix your outcome record, don't retry with a different\n   claim to force it through.\n   It prints `{\"converged\": bool, \"shouldContinue\": bool, ...}`. If\n   `shouldContinue` is `false`, stop the loop — either the pass cap was\n   reached or the pass converged (closed zero findings and produced zero\n   fix\u002Fdraft outcomes). Do not start another `plan-next-pass` call after\n   `shouldContinue: false`. In `\"propose\"` mode, `\"blocked\"` outcomes never\n   count toward fix\u002Fdraft outcomes, so a propose-mode pass can report\n   `\"converged\": true` on its very first pass regardless of how many\n   findings were described — do not read that as \"nothing needs\n   attention\" when summarizing for the user; report the open findings\n   themselves. `converged` is also forced `false` while any verified\n   domain's sidecar has `unverifiedCount > 0` (or an unreadable sidecar):\n   the output then lists them under `unverifiedBlockers`. That is an\n   abstain, not a pass: re-run that domain's Verify phase rather than\n   starting more passes (`shouldContinue` stays `false`).\n\n## Wrap up\n\n7. `analysis\u002Fzeugnis\u002Freports\u002FZEUGNIS_CYCLE.md` already has one line appended per\n   pass by the engine. Summarize it for the user: total passes run,\n   which of the three stop reasons applied — convergence, the pass cap, or\n   **blocked by unverified findings** (`converged: false` with\n   `unverifiedBlockers`; name each blocked domain and say its Verify phase\n   must be re-run — that is not convergence and not a clean result) —\n   findings closed, findings escalated (reopened past `max_reopens`), and\n   anything left open.\n\n## What NOT to do\n\n- Do not run more than `max_passes` passes in one invocation, even if the\n  user asks you to \"just keep going\" — a fresh invocation (a new call to\n  this skill) is how they get more passes; `plan-next-pass` will refuse\n  once the cap is hit regardless of what you're asked.\n- Do not dispatch `zeugnis-remediator` for an `assertion_audit` finding or\n  a non-`excessive-tool-grant` `agentic_reliability` finding under any\n  framing — the remediation-scope script refuses to open a scope for\n  these, and the `PreToolUse` hook refuses the edit even if you tried to\n  route around the scope step.\n- Do not treat a `\"blocked\"` remediator outcome as a failure to retry\n  immediately — record it as `\"blocked\"` and let it surface as an open\n  finding for the next pass or the user's manual attention.\n",{"name":5,"description":7},"main","6158c1e4e112c06af2c7ac01b13b2ed7214c0fcf","plugins\u002Fzeugnis\u002Fskills\u002Fzeugnis-cycle",0,"https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff","https:\u002F\u002Fanselmoo.github.io\u002Fwerkstoff\u002F","https:\u002F\u002Fgithub.com\u002FAnselmoo","Likes to Develop","2026-11-02 04:20:59",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Anselmoo\u002Fwerkstoff --skill zeugnis-cycle",{"tool":39,"command":40},"bunx","bunx skills add Anselmoo\u002Fwerkstoff --skill zeugnis-cycle",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Anselmoo\u002Fwerkstoff --skill zeugnis-cycle","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Fplugins\u002Fzeugnis\u002Fskills\u002Fzeugnis-cycle\u002FSKILL.md",true,[],[48,55,62,67,72,77],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M79AJTRBNBX03WWKYJ8ENM45","toy-check","A skill that checks things in the repository and reports what it finds.","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Ftest\u002Fworkflows\u002Ffixtures\u002Ftoy-plugin\u002Fskills\u002Ftoy-check\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M79AJS292F3HTDRXWZVRSN5B","demo-check","Reports which generated files changed since the last generation run. Use when the user asks what drifted. Read-only; never edits a file.","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Ftest\u002Fplugins\u002Ffixtures\u002Fnacharbeit-uncompleted-run\u002Fplugins\u002Fdemo\u002Fskills\u002Fdemo-check\u002FSKILL.md",[],[],{"skillNo":63,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M79AJP4TFQXKT1AAK1WC1JYK","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Ftest\u002Fplugins\u002Ffixtures\u002Fnacharbeit-stale-lock\u002Fplugins\u002Fdemo\u002Fskills\u002Fdemo-check\u002FSKILL.md",[],[],{"skillNo":68,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M79AJM6Y50WP3G2BR4H0M60Z","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Ftest\u002Fplugins\u002Ffixtures\u002Fnacharbeit-preflight-mixed\u002Fplugins\u002Fdemo\u002Fskills\u002Fdemo-check\u002FSKILL.md",[],[],{"skillNo":73,"name":57,"nameZh":6,"description":58,"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":76,"publishedAt":19},"01M79AJGR6DMT8KCCK071V78XM","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Ftest\u002Fplugins\u002Ffixtures\u002Fnacharbeit-hooks-shape\u002Fplugins\u002Fdemo\u002Fskills\u002Fdemo-check\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M79AJBJ4CHC4PZVWKERWF3HF","zirkel-verify-assumptions","Checks exactly ONE named assumption against real evidence in at most 3 steps and reports one outcome: confidence raised (>=90), reading changed, or still unresolved. Use to resolve a single load-bearing uncertainty — typically a blocking entry surfaced by zirkel-clarify-scope: \"verify this one assumption\", \"is it actually true that X\", \"check this before we rely on it\". For multiple uncertainties, invoke once per uncertainty — the step budget is never shared.","https:\u002F\u002Fgithub.com\u002FAnselmoo\u002Fwerkstoff\u002Fblob\u002Fmain\u002Fplugins\u002Fzirkel\u002Fskills\u002Fzirkel-verify-assumptions\u002FSKILL.md",[],[]]