Truth-drift auditor — find & fix comment/doc/CI/version/dependency rot that contradicts the current code, with root + transitive verification (correct beats delete). Use when the user wants to remove stale comments, fix doc drift (CLAUDE.md/README/CHANGELOG), modernize CI publish auth, sync package versions, or question why a sub-library is used instead of a parent/successor package. Triggers on "stale comments", "derot", "doc rot", "comments are out of date", "docs don't match the code", "clean up the docs", "why this dependency", "überarbeitung".
Run risky changes through 8 disciplined phases that map, build, fix, and verify — each unlocks more power only if work stays honest and resumable. Use when a spec or audit needs parallel discovery, lane execution, checkpoints surviving context loss, and a ship-or-die finale.
This skill should be used when the user asks to clean up dead code, remove suppressions/warnings, fix duplication or imports, verify the build is clean with zero suppressions, run a frontend design audit (--goggles), or make a public-API brutal break (--guillotine). Smart infra: deletion permit, audit ledger, break manifest. 4 teammates per phase. Use when cleanup/scope/intensity, goggles, or guillotine is requested.
Analyze, lint, and check a local Claude Code add-on or capability, explain the result, and rank the top fixes first. Use when the user says "audit this", "why did it score that way", "what should I fix first", "validate the manifest", "lint the hooks", or "benchmark the real token usage".
Analyze and check a local Claude Code SKILL.md folder: review frontmatter, size, broken links, trigger words, and token budget, then rank the top fixes. Use when the user says "audit this", "why did it score that way", "this is too big", "it is not triggering", or "measure the real token usage".
Convert cc-plugin-eval findings into a concrete rewrite brief, then hand off to skill-creator for the edit pass. Use when the user already ran an evaluation and now wants a fix plan, after asking "what should I fix first" or "rewrite this using the findings".
Audit a local Claude Code bundle that has .claude-plugin/plugin.json: validate the manifest, lint hooks, check mcp/lsp servers, monitors, and agents, then rank fixes. Use when the user says "audit this", "validate the manifest", "lint the hooks", "why did it score that way", or "benchmark this".
Build a local rubric extension so a team can add custom checks that emit schema-compatible findings, metrics, and artifacts on top of the deterministic engine. Use when the user wants their own scoring rules, domain-specific audits, or a custom report wired into analyze runs.