[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EFC8BGF19RPHSA0F5VMT7A":3,"skill-same:01M7EFC8BGF19RPHSA0F5VMT7A":89},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":47,"commitSha":48,"skillDirectory":49,"forksCount":50,"repositoryHtmlUrl":51,"repositoryHomepageUrl":6,"ownerHtmlUrl":52,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":53,"install":54,"resources":66},"01M7EFC8BGF19RPHSA0F5VMT7A","self-evolving-rules",null,"Use when the operator manually runs memory or rule evolution — \"add a memory\", \"merge memories\", \"archive\", \"rewrite the rule\", \"move across layers\", \"clean stale memory\", \"sync L1 + rules\", or trims always-on instruction load \u002F CLAUDE.md length \u002F rules dump — and needs the 6-step strip-and-route process plus optional Step 7 extraction and Step 8 context handoff to domain-modeling.","winter6205\u002Farthurpower","winter6205","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F136674824?v=4","01M7EF9EWNYDKHX9MR3AG5QV7A","01M7EF9EXCC1CNFXQNV6PS3M8B",2,"MIT","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\r\nname: self-evolving-rules\r\ndescription: \"Use when the operator manually runs memory or rule evolution — \\\"add a memory\\\", \\\"merge memories\\\", \\\"archive\\\", \\\"rewrite the rule\\\", \\\"move across layers\\\", \\\"clean stale memory\\\", \\\"sync L1 + rules\\\", or trims always-on instruction load \u002F CLAUDE.md length \u002F rules dump — and needs the 6-step strip-and-route process plus optional Step 7 extraction and Step 8 context handoff to domain-modeling.\"\r\nbucket: engineering\r\ndisable-model-invocation: true\r\nversion: 2.2.0\r\ntags: [memory, rules, self-evolving, strip, routing, L1, L2, archive, merge, always-on, instruction-budget, progressive-disclosure]\r\nauthor: arthurpower-team\r\nrelated_skills: [skill-authoring, session-handoff, verification-before-completion, using-agent-skills, domain-modeling]\r\ntype: discipline\r\n---\r\n\r\n# Self-Evolving Rules\r\n\r\n> **User-invoked.** `disable-model-invocation: true`. The model does not auto-fire this skill from description match. The operator types the skill name (or `\u002Fself-evolving-rules`) when memory \u002F rule work starts.\r\n\r\nThis skill turns operator feedback and scattered notes into **one durable rule in the right layer**, after strip and conflict resolution. It does **not** author new skills (that is `arthurpower:skill-authoring`).\r\n\r\n## When to use\r\n\r\nOperator is doing any of these memory \u002F rule jobs:\r\n\r\n| Branch | Typical phrasing |\r\n|--------|------------------|\r\n| Add | \"add a memory\", \"write a new rule\", \"add one rule\" |\r\n| Merge | \"merge memories\", \"dedup fragments\" |\r\n| Archive | \"archive\", \"mark SUPERSEDED\", \"mark ARCHIVED\" |\r\n| Rewrite | \"rewrite the rule\", \"change wording\", \"revise existing\" |\r\n| Migrate | \"move across layers\", \"promote auto memory to rules\" |\r\n| Clean | \"clean stale memory\", \"delete old rule files\" |\r\n| Config sync | \"sync L1 + auto memory + rules\", \"refresh the index\" |\r\n\r\nAlso use after a completed iteration when the operator wants Step 7 knowledge extraction.\r\n\r\n## When not to use\r\n\r\n- Single-line typo with no rule semantics\r\n- Pure command execution with no memory write\r\n- Skill body already strip-clean and only needs structural skill-authoring work → hand off to `arthurpower:skill-authoring` (this skill never authors skill content itself)\r\n- Writing domain glossary \u002F ADR → this skill never writes them directly; Step 8 drives `domain-modeling` (the sole write authority) to persist them\r\n- Session continuity only → `session-handoff`\r\n- Operator said not to touch memory \u002F rules this turn\r\n\r\n### Skill-write preflight (hand-off, not authoring)\r\n\r\nWhen the task will create or edit a **skill**, still start here for strip \u002F locate \u002F conflict \u002F provenance. After Steps 1–5 are clean, **hand off implementation** to `arthurpower:skill-authoring`. Do not skip this preflight and jump straight to skill-authoring when the change carries new durable rules, wrappers to strip, or cross-layer conflicts.\r\n\r\n## Procedure\r\n\r\n### Step 1 — Distill\r\n\r\nCompress the feedback into **one imperative sentence**.\r\n\r\n**Done when:** exactly one sentence exists; if more than one concern, split into separate iterations.\r\n\r\n### Step 2 — Locate\r\n\r\nGrep across three layers for the same topic:\r\n\r\n1. Project `.\u002FCLAUDE.md` and `.\u002F.claude\u002Frules\u002F`\r\n2. User-level CLAUDE.md and user-level rules topic files\r\n3. Auto memory index + topic files for this project\r\n\r\n**Done when:** every same-topic hit is listed with path + line, or operator confirms \"this is new\".\r\n\r\n### Step 3 — Strip\r\n\r\nRemove wrappers; keep content + real provenance.\r\n\r\nFull strip categories and placeholders: [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\nRun the genericization grep suite from the `no-meta-specific` rule (or skill validator private-prefix checks). Hits outside necessary exemptions must be removed before write.\r\n\r\n**Done when:** strip checklist clean; provenance present; no self-certifying wrapper remains.\r\n\r\n### Step 4 — Resolve conflicts\r\n\r\nApply the priority ladder in [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\n**Done when:** at most one surviving rule per topic; losers deleted or marked SUPERSEDED with pointer.\r\n\r\n### Step 5 — Polish (optional)\r\n\r\nTighten wording. Prefer leading words already in the operator's vocabulary. No new constraints without provenance.\r\n\r\n**Done when:** wording is shorter or equal length and meaning unchanged.\r\n\r\n**Instruction budget check (always-on rules only).** When the routed layer is global L1, project `CLAUDE.md`, or any `@import`-ed rules file: each new imperative unit increases the always-on instruction load. Confirm universal applicability (must fire on >50% of sessions for the same project), prefer `file:line` or markdown link over `@import` when the new rule is task-shaped, and reject duplicate rewording of existing bullets. Done when the touched file still has fewer imperative units than before, or the additions are universally-applicable.\r\n\r\n### Step 6 — Route and write\r\n\r\nWrite only the target layer from the routing table in [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\nRe-read the written file. If the project has `.evals\u002F`, run post-iteration regression (below) before claiming done.\r\n\r\n**Done when:** file exists at the routed path; re-read matches intent; no secret values written.\r\n\r\n### Step 7 — Extract knowledge (conditional)\r\n\r\nAfter Step 6, optionally spawn a **read-only** subagent. Details: [`references\u002Fstep-7-extraction.md`](references\u002Fstep-7-extraction.md).\r\n\r\n**Done when:** skip conditions applied, or candidates filtered and counts reported. Subagent wrote zero files.\r\n\r\n### Step 8 — Hand new context to domain-modeling (conditional)\r\n\r\nIf this iteration surfaced a new project term, an overloaded \u002F ambiguous term, or a hard-to-reverse decision, invoke `domain-modeling` **via the Skill tool** (`arthurpower:domain-modeling`) to persist it — `docs\u002FCONTEXT.md` (or `CONTEXT-MAP.md`) for vocabulary, `docs\u002Fadr\u002F` for a decision that meets the ADR three-condition gate. The Skill-tool invocation is required, not a formality: `pre-tool-skill-routing-sync.cjs` refreshes `~\u002F.claude\u002F.routing-state.json` only when a skill is invoked through the Skill tool, and `pre-context-write-guard.cjs` permits these writes only while that state names `domain-modeling`. Merely reading domain-modeling's SKILL.md leaves the state on `self-evolving-rules` and the write is blocked. This skill is a **driver**, never the writer: `domain-modeling` is the sole write authority for those artifacts.\r\n\r\n**Skip when:** the iteration touched only memory \u002F rules wording with no new vocabulary or decision.\r\n\r\n**Done when:** no new term \u002F decision surfaced (skipped), or `domain-modeling` was invoked via the Skill tool and confirmed the write landed. This skill wrote zero `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` bytes itself.\r\n\r\n## Post-iteration regression\r\n\r\nWhen rules changed and the project has `.evals\u002F`:\r\n\r\n```bash\r\nbash .evals\u002Frun.sh\r\n```\r\n\r\nBaseline must stay N\u002FN pass. Fail → roll back the rule write and restart from Step 3.\r\n\r\nNo `.evals\u002F` → skip, note the gap.\r\n\r\n## Engineering changes ride-along\r\n\r\nIf this iteration ships with hook \u002F skill \u002F rules \u002F settings changes (not memory-only), follow the project boundary-testing protocol. Subagent smoke alone is not acceptance.\r\n\r\n## Acceptance Criteria\r\n\r\n- [ ] Invoked for memory \u002F rule evolution, or as strip\u002Froute preflight before handing skill implementation to skill-authoring\r\n- [ ] Step 1 produced one imperative sentence\r\n- [ ] Step 2 grepped all three layers (or operator confirmed new)\r\n- [ ] Step 3 strip clean under genericization checks\r\n- [ ] Step 4 left a single rule per topic\r\n- [ ] Step 6 wrote only the routed layer; re-read OK\r\n- [ ] No API key \u002F token \u002F password values in any written file\r\n- [ ] Step 7 skipped for cause, or candidates filtered with counts\r\n- [ ] Step 8 skipped (no new term \u002F decision), or `domain-modeling` invoked and confirmed the write; this skill wrote zero `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` bytes\r\n- [ ] `.evals\u002Frun.sh` run when present; fail rolled back\r\n\r\n## Verification\r\n\r\n```\r\nSkill type: discipline\r\nBar level: discipline\r\n```\r\n\r\n```bash\r\n# After write: genericization must be clean on the touched file\r\n# (use the no-meta-specific grep suite; for skills, hand off to skill-authoring\r\n#  which runs its own validator from the plugin skills root)\r\n\r\n# If project evals exist:\r\nbash .evals\u002Frun.sh; echo EXIT=$?\r\n```\r\n\r\n### Rationalization Table\r\n\r\n| Excuse | Why it sounds right | Reality | Required action |\r\n|--------|---------------------|---------|-----------------|\r\n| \"The new rule is obviously correct; skip grep\" | Saves time | Misses same-topic rules → silent conflict | Always grep all three layers first |\r\n| \"I'll write memory directly; skill is overhead\" | One-line fix feels trivial | Bypasses strip + routing + provenance | Invoke this skill before any L1\u002FL2 write |\r\n| \"Keep both rules; they mostly agree\" | Feels safer than deleting | Downstream agents pick at random | Merge to one rule after priority ladder |\r\n| \"Promote my inference to a hard constraint\" | Confidence feels like evidence | Invented authority without provenance | Tag `[agent inference · unconfirmed]` or drop |\r\n| \"Run Step 7 on every tiny edit\" | More extraction feels thorough | Noise candidates flood the queue | Honor skip conditions |\r\n| \"Accept every high-looking candidate\" | Subagent sounded confident | Project-specific junk becomes L1 | Filter by scope + rating |\r\n| \"Regression tests can wait\" | Eval suite is slow | Rule change can break routing | Run `.evals\u002Frun.sh` when present |\r\n| \"Docs-only; skip boundary testing\" | Hook text is not product code | Silent bugs ship as discipline | Follow boundary-testing when shipping with code |\r\n| \"Always-on: add one more IMPORTANT, it's harmless\" | Single line feels negligible | Decay is uniform, not additive; perimeter bias amplifies tails | Confirm universal applicability + count imperative units before adding |\r\n\r\n### Red Flags — STOP\r\n\r\n- Editing CLAUDE.md \u002F rules \u002F auto memory without this skill\r\n- Skipping cross-layer grep before write\r\n- Leaving two conflicting rules after a merge\r\n- Shipping self-certifying wrappers instead of strip\r\n- Step 7 subagent allowed to write files\r\n- Accepting all Step 7 candidates without rating filter\r\n- Claiming done without post-iteration eval when `.evals\u002F` exists\r\n- Skipping strip\u002Froute preflight and jumping straight to skill-authoring when the skill change carries new durable rules or wrappers\r\n- Authoring skill content inside this skill instead of handing off to skill-authoring\r\n- Adding imperative units to always-on layers without universal-applicability check or instruction-budget review\r\n- Writing `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` directly from this skill instead of driving `domain-modeling` (write-authority violation)\r\n- Reading `domain-modeling`'s SKILL.md instead of invoking it via the Skill tool — routing-state stays on `self-evolving-rules`, so `pre-context-write-guard.cjs` blocks the write\r\n- Skipping Step 8 when a new term or hard-to-reverse decision surfaced during the iteration\r\n\r\n## References\r\n\r\n- Step 7 details: [`references\u002Fstep-7-extraction.md`](references\u002Fstep-7-extraction.md)\r\n- Strip + routing + priority: [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md)\r\n- Always-on \u002F instruction budget hygiene: [`references\u002Falways-on-hygiene.md`](references\u002Falways-on-hygiene.md)\r\n- Extra rationalization copy: [`references\u002Frationalization-red-flags.md`](references\u002Frationalization-red-flags.md)\r\n- Genericization rule: project \u002F user `no-meta-specific` rule file\r\n- New skill creation: `arthurpower:skill-authoring`\r\n- Project decision SSOT when always-on layout is locked: `docs\u002Fadr\u002F0008-always-on-progressive-disclosure.md` (if present)\r\n","# Self-Evolving Rules\r\n\r\n> **User-invoked.** `disable-model-invocation: true`. The model does not auto-fire this skill from description match. The operator types the skill name (or `\u002Fself-evolving-rules`) when memory \u002F rule work starts.\r\n\r\nThis skill turns operator feedback and scattered notes into **one durable rule in the right layer**, after strip and conflict resolution. It does **not** author new skills (that is `arthurpower:skill-authoring`).\r\n\r\n## When to use\r\n\r\nOperator is doing any of these memory \u002F rule jobs:\r\n\r\n| Branch | Typical phrasing |\r\n|--------|------------------|\r\n| Add | \"add a memory\", \"write a new rule\", \"add one rule\" |\r\n| Merge | \"merge memories\", \"dedup fragments\" |\r\n| Archive | \"archive\", \"mark SUPERSEDED\", \"mark ARCHIVED\" |\r\n| Rewrite | \"rewrite the rule\", \"change wording\", \"revise existing\" |\r\n| Migrate | \"move across layers\", \"promote auto memory to rules\" |\r\n| Clean | \"clean stale memory\", \"delete old rule files\" |\r\n| Config sync | \"sync L1 + auto memory + rules\", \"refresh the index\" |\r\n\r\nAlso use after a completed iteration when the operator wants Step 7 knowledge extraction.\r\n\r\n## When not to use\r\n\r\n- Single-line typo with no rule semantics\r\n- Pure command execution with no memory write\r\n- Skill body already strip-clean and only needs structural skill-authoring work → hand off to `arthurpower:skill-authoring` (this skill never authors skill content itself)\r\n- Writing domain glossary \u002F ADR → this skill never writes them directly; Step 8 drives `domain-modeling` (the sole write authority) to persist them\r\n- Session continuity only → `session-handoff`\r\n- Operator said not to touch memory \u002F rules this turn\r\n\r\n### Skill-write preflight (hand-off, not authoring)\r\n\r\nWhen the task will create or edit a **skill**, still start here for strip \u002F locate \u002F conflict \u002F provenance. After Steps 1–5 are clean, **hand off implementation** to `arthurpower:skill-authoring`. Do not skip this preflight and jump straight to skill-authoring when the change carries new durable rules, wrappers to strip, or cross-layer conflicts.\r\n\r\n## Procedure\r\n\r\n### Step 1 — Distill\r\n\r\nCompress the feedback into **one imperative sentence**.\r\n\r\n**Done when:** exactly one sentence exists; if more than one concern, split into separate iterations.\r\n\r\n### Step 2 — Locate\r\n\r\nGrep across three layers for the same topic:\r\n\r\n1. Project `.\u002FCLAUDE.md` and `.\u002F.claude\u002Frules\u002F`\r\n2. User-level CLAUDE.md and user-level rules topic files\r\n3. Auto memory index + topic files for this project\r\n\r\n**Done when:** every same-topic hit is listed with path + line, or operator confirms \"this is new\".\r\n\r\n### Step 3 — Strip\r\n\r\nRemove wrappers; keep content + real provenance.\r\n\r\nFull strip categories and placeholders: [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\nRun the genericization grep suite from the `no-meta-specific` rule (or skill validator private-prefix checks). Hits outside necessary exemptions must be removed before write.\r\n\r\n**Done when:** strip checklist clean; provenance present; no self-certifying wrapper remains.\r\n\r\n### Step 4 — Resolve conflicts\r\n\r\nApply the priority ladder in [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\n**Done when:** at most one surviving rule per topic; losers deleted or marked SUPERSEDED with pointer.\r\n\r\n### Step 5 — Polish (optional)\r\n\r\nTighten wording. Prefer leading words already in the operator's vocabulary. No new constraints without provenance.\r\n\r\n**Done when:** wording is shorter or equal length and meaning unchanged.\r\n\r\n**Instruction budget check (always-on rules only).** When the routed layer is global L1, project `CLAUDE.md`, or any `@import`-ed rules file: each new imperative unit increases the always-on instruction load. Confirm universal applicability (must fire on >50% of sessions for the same project), prefer `file:line` or markdown link over `@import` when the new rule is task-shaped, and reject duplicate rewording of existing bullets. Done when the touched file still has fewer imperative units than before, or the additions are universally-applicable.\r\n\r\n### Step 6 — Route and write\r\n\r\nWrite only the target layer from the routing table in [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md).\r\n\r\nRe-read the written file. If the project has `.evals\u002F`, run post-iteration regression (below) before claiming done.\r\n\r\n**Done when:** file exists at the routed path; re-read matches intent; no secret values written.\r\n\r\n### Step 7 — Extract knowledge (conditional)\r\n\r\nAfter Step 6, optionally spawn a **read-only** subagent. Details: [`references\u002Fstep-7-extraction.md`](references\u002Fstep-7-extraction.md).\r\n\r\n**Done when:** skip conditions applied, or candidates filtered and counts reported. Subagent wrote zero files.\r\n\r\n### Step 8 — Hand new context to domain-modeling (conditional)\r\n\r\nIf this iteration surfaced a new project term, an overloaded \u002F ambiguous term, or a hard-to-reverse decision, invoke `domain-modeling` **via the Skill tool** (`arthurpower:domain-modeling`) to persist it — `docs\u002FCONTEXT.md` (or `CONTEXT-MAP.md`) for vocabulary, `docs\u002Fadr\u002F` for a decision that meets the ADR three-condition gate. The Skill-tool invocation is required, not a formality: `pre-tool-skill-routing-sync.cjs` refreshes `~\u002F.claude\u002F.routing-state.json` only when a skill is invoked through the Skill tool, and `pre-context-write-guard.cjs` permits these writes only while that state names `domain-modeling`. Merely reading domain-modeling's SKILL.md leaves the state on `self-evolving-rules` and the write is blocked. This skill is a **driver**, never the writer: `domain-modeling` is the sole write authority for those artifacts.\r\n\r\n**Skip when:** the iteration touched only memory \u002F rules wording with no new vocabulary or decision.\r\n\r\n**Done when:** no new term \u002F decision surfaced (skipped), or `domain-modeling` was invoked via the Skill tool and confirmed the write landed. This skill wrote zero `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` bytes itself.\r\n\r\n## Post-iteration regression\r\n\r\nWhen rules changed and the project has `.evals\u002F`:\r\n\r\n```bash\r\nbash .evals\u002Frun.sh\r\n```\r\n\r\nBaseline must stay N\u002FN pass. Fail → roll back the rule write and restart from Step 3.\r\n\r\nNo `.evals\u002F` → skip, note the gap.\r\n\r\n## Engineering changes ride-along\r\n\r\nIf this iteration ships with hook \u002F skill \u002F rules \u002F settings changes (not memory-only), follow the project boundary-testing protocol. Subagent smoke alone is not acceptance.\r\n\r\n## Acceptance Criteria\r\n\r\n- [ ] Invoked for memory \u002F rule evolution, or as strip\u002Froute preflight before handing skill implementation to skill-authoring\r\n- [ ] Step 1 produced one imperative sentence\r\n- [ ] Step 2 grepped all three layers (or operator confirmed new)\r\n- [ ] Step 3 strip clean under genericization checks\r\n- [ ] Step 4 left a single rule per topic\r\n- [ ] Step 6 wrote only the routed layer; re-read OK\r\n- [ ] No API key \u002F token \u002F password values in any written file\r\n- [ ] Step 7 skipped for cause, or candidates filtered with counts\r\n- [ ] Step 8 skipped (no new term \u002F decision), or `domain-modeling` invoked and confirmed the write; this skill wrote zero `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` bytes\r\n- [ ] `.evals\u002Frun.sh` run when present; fail rolled back\r\n\r\n## Verification\r\n\r\n```\r\nSkill type: discipline\r\nBar level: discipline\r\n```\r\n\r\n```bash\r\n# After write: genericization must be clean on the touched file\r\n# (use the no-meta-specific grep suite; for skills, hand off to skill-authoring\r\n#  which runs its own validator from the plugin skills root)\r\n\r\n# If project evals exist:\r\nbash .evals\u002Frun.sh; echo EXIT=$?\r\n```\r\n\r\n### Rationalization Table\r\n\r\n| Excuse | Why it sounds right | Reality | Required action |\r\n|--------|---------------------|---------|-----------------|\r\n| \"The new rule is obviously correct; skip grep\" | Saves time | Misses same-topic rules → silent conflict | Always grep all three layers first |\r\n| \"I'll write memory directly; skill is overhead\" | One-line fix feels trivial | Bypasses strip + routing + provenance | Invoke this skill before any L1\u002FL2 write |\r\n| \"Keep both rules; they mostly agree\" | Feels safer than deleting | Downstream agents pick at random | Merge to one rule after priority ladder |\r\n| \"Promote my inference to a hard constraint\" | Confidence feels like evidence | Invented authority without provenance | Tag `[agent inference · unconfirmed]` or drop |\r\n| \"Run Step 7 on every tiny edit\" | More extraction feels thorough | Noise candidates flood the queue | Honor skip conditions |\r\n| \"Accept every high-looking candidate\" | Subagent sounded confident | Project-specific junk becomes L1 | Filter by scope + rating |\r\n| \"Regression tests can wait\" | Eval suite is slow | Rule change can break routing | Run `.evals\u002Frun.sh` when present |\r\n| \"Docs-only; skip boundary testing\" | Hook text is not product code | Silent bugs ship as discipline | Follow boundary-testing when shipping with code |\r\n| \"Always-on: add one more IMPORTANT, it's harmless\" | Single line feels negligible | Decay is uniform, not additive; perimeter bias amplifies tails | Confirm universal applicability + count imperative units before adding |\r\n\r\n### Red Flags — STOP\r\n\r\n- Editing CLAUDE.md \u002F rules \u002F auto memory without this skill\r\n- Skipping cross-layer grep before write\r\n- Leaving two conflicting rules after a merge\r\n- Shipping self-certifying wrappers instead of strip\r\n- Step 7 subagent allowed to write files\r\n- Accepting all Step 7 candidates without rating filter\r\n- Claiming done without post-iteration eval when `.evals\u002F` exists\r\n- Skipping strip\u002Froute preflight and jumping straight to skill-authoring when the skill change carries new durable rules or wrappers\r\n- Authoring skill content inside this skill instead of handing off to skill-authoring\r\n- Adding imperative units to always-on layers without universal-applicability check or instruction-budget review\r\n- Writing `docs\u002FCONTEXT.md` \u002F `docs\u002Fadr\u002F` directly from this skill instead of driving `domain-modeling` (write-authority violation)\r\n- Reading `domain-modeling`'s SKILL.md instead of invoking it via the Skill tool — routing-state stays on `self-evolving-rules`, so `pre-context-write-guard.cjs` blocks the write\r\n- Skipping Step 8 when a new term or hard-to-reverse decision surfaced during the iteration\r\n\r\n## References\r\n\r\n- Step 7 details: [`references\u002Fstep-7-extraction.md`](references\u002Fstep-7-extraction.md)\r\n- Strip + routing + priority: [`references\u002Fstrip-and-routing.md`](references\u002Fstrip-and-routing.md)\r\n- Always-on \u002F instruction budget hygiene: [`references\u002Falways-on-hygiene.md`](references\u002Falways-on-hygiene.md)\r\n- Extra rationalization copy: [`references\u002Frationalization-red-flags.md`](references\u002Frationalization-red-flags.md)\r\n- Genericization rule: project \u002F user `no-meta-specific` rule file\r\n- New skill creation: `arthurpower:skill-authoring`\r\n- Project decision SSOT when always-on layout is locked: `docs\u002Fadr\u002F0008-always-on-progressive-disclosure.md` (if present)\r\n",{"name":5,"tags":23,"type":36,"author":37,"bucket":38,"version":39,"description":7,"related_skills":40,"disable-model-invocation":46},[24,25,26,27,28,29,30,31,32,33,34,35],"memory","rules","self-evolving","strip","routing","L1","L2","archive","merge","always-on","instruction-budget","progressive-disclosure","discipline","arthurpower-team","engineering","2.2.0",[41,42,43,44,45],"skill-authoring","session-handoff","verification-before-completion","using-agent-skills","domain-modeling",true,"master","1a067ffc9024631d22ca8bf2b47ad81369f28056","arthurpower\u002Fskills\u002Fself-evolving-rules",0,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower","https:\u002F\u002Fgithub.com\u002Fwinter6205","2026-10-28 23:46:38",{"source":8,"skillName":5,"commands":55,"prompt":65,"zipAvailable":46},[56,59,62],{"tool":57,"command":58},"npx","npx skills add winter6205\u002Farthurpower --skill self-evolving-rules",{"tool":60,"command":61},"bunx","bunx skills add winter6205\u002Farthurpower --skill self-evolving-rules",{"tool":63,"command":64},"pnpm","pnpm dlx skills add winter6205\u002Farthurpower --skill self-evolving-rules","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002FSKILL.md",[67,73,77,81,85],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"references\u002Falways-on-hygiene.md","REFERENCE","参考文档",4756,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002Freferences\u002Falways-on-hygiene.md",{"path":74,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":75,"sourceHtmlUrl":76},"references\u002Fgenericization.md",6503,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002Freferences\u002Fgenericization.md",{"path":78,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":79,"sourceHtmlUrl":80},"references\u002Frationalization-red-flags.md",1989,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002Freferences\u002Frationalization-red-flags.md",{"path":82,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":83,"sourceHtmlUrl":84},"references\u002Fstep-7-extraction.md",1906,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002Freferences\u002Fstep-7-extraction.md",{"path":86,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Fstrip-and-routing.md",3018,"https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fself-evolving-rules\u002Freferences\u002Fstrip-and-routing.md",[90,97,104,110,116,123],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":18},"01M7EFDX3VT33HSHDA17T9N9JJ","writing-plans","Use when turning a buildable spec into ordered tracer bullets after architecture-change-reviewer. Covers vertical slicing, dependency order, and headroom for the implementer. Triggers on 写计划, 拆任务, 任务分解, 制定方案, 跨文件改动, 计划.","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fwriting-plans\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":18},"01M7EFDQV6SGJWDK634DWPFR35","wayfinder","Use when a chunk of work is more than one agent session can hold — plan it as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fwayfinder\u002FSKILL.md",[],[],{"skillNo":105,"name":43,"nameZh":6,"description":106,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":107,"categories":108,"keywords":109,"publishedAt":18},"01M7EFDM9Y55GKYFGY96088BYB","This skill should be used when claiming \"完成\" \u002F \"done\" \u002F \"fixed\" \u002F \"passing\" \u002F \"ready to ship\", before trusting a subagent success report, or when checking \"对照计划\" \u002F \"对照 spec\" — whether this run's delivery matches its basis.","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fverification-before-completion\u002FSKILL.md",[],[],{"skillNo":111,"name":44,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":18},"01M7EFDFCY4V0JMQHSEE34A2JS","Use when starting a new session, switching tasks, or unsure which arthurpower skill applies. Routes incoming work to the matching skill via flowchart before invoking. Activates on phrases like \"which skill should I use\", \"skill check\", \"pre-task skill check\", \"how to start\", \"router\", \"skill routing\", or when multiple competing skills could apply. Does not auto-invoke user-invoked authoring skills (skill-authoring, self-evolving-rules).","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fusing-agent-skills\u002FSKILL.md",[],[],{"skillNo":117,"name":118,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":120,"categories":121,"keywords":122,"publishedAt":18},"01M7EFDB70ATNQASXP4BKDYGA6","test-driven-development","Use when implementing logic, fixing a bug, or changing behavior — write the failing test first (RED), then minimal code (GREEN), then refactor. Triggers on \"write the failing test first\", \"Prove-It\", or a bug that needs a reproduction test before the fix.","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[],{"skillNo":124,"name":125,"nameZh":6,"description":126,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":18},"01M7EFD6NPWFG5WSXT0DF07XH1","systematic-debugging","Use when debugging a test failure, build break, unexpected runtime error, flaky test, or hanging process. Triggers on phrases like \"fix this bug\", \"getting an error\", \"failing test\", \"flaky\", \"hanging\", \"halfway through debugging\", \"this used to work\", \"intermittent failure\", \"stack trace shows X\", or any case where behavior does not match expectations. Prevents shotgun debugging.","https:\u002F\u002Fgithub.com\u002Fwinter6205\u002Farthurpower\u002Fblob\u002Fmaster\u002Farthurpower\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[]]