idd-spec-auditlisted

Repo-local, dogfood-only semantic audit of the IDD instruction corpus for leaked session context, cross-file contradictions, fresh-memory completability gaps, and automation blockers. Use only in the kurone-kito/idd-skill source repository, on request, to audit .github/instructions, skills/issue-authoring, CLAUDE.md, or .github/copilot-instructions.md. Read-only — never edits files or mutates issues.
kurone-kito/idd-skill · ★ 4 · AI & Automation · score 70

Install: claude install-skill kurone-kito/idd-skill

# IDD Spec Audit  ## Status **Dogfood-only.** This skill lives at `.claude/skills/idd-spec-audit/` in the `kurone-kito/idd-skill` source repository only. It has no canonical `skills/` source and is not an `audit/sync-manifest.json` `syncPairs`/`fileSets` entry, unlike `.claude/skills/issue-authoring/` (which mirrors `skills/issue-authoring/`) — there is nothing to distribute to adopter repositories yet. See [Promotion checkpoint](#promotion-checkpoint-final-step) below. `scripts/audit-docs.mjs` catches byte-level drift — sync-pair mismatches, budgets, config-vs-instruction agreement — but not semantic drift: prose that contradicts a sibling file, leaked session context, passages that are not completable from a cold read, or wording that stalls an autonomous step. This skill runs that check as N parallel LLM read passes, adapted from `mew-ton/soloscrum`'s `define-pr-lifecycle` audit model. ## Scope - **Audit targets** (findings may cite these): `.github/instructions/**/*.md` (including `lite/`), `skills/issue-authoring/**/*.md`, `CLAUDE.md`, and `.github/copilot-instructions.md`. Audit these files even though `.github/instructions/**` is itself the generated `target` side of an `audit/sync-manifest.json` sync pair (mirrored from `idd-template/`) — this skill audits the corpus a worker session actually reads, not the `idd-template/` source, so being a sync target never exempts a file here. - **Reference-only inputs** (read for