Conciseness

The body is dense and assumes Claude's competence (no concept explanations), but several sections restate the same rules: dry-run-before-mutation appears in Philosophy, Current Project Contracts, Workflow, Constraints, and Execution Boundaries; the "init --update is not the routine upgrade lane" warning appears in Current Project Contracts, Gotchas, and Anti-Patterns; the environment.toml ownership rule and the ci-migrate/manual-deletion warning each appear three times. This matches 'mostly efficient but includes some unnecessary explanation or could be tightened' rather than the score-4 anchor, since whole duplicate sections could be consolidated.

Actionability

Commands are copy-paste ready and cover the common cases with explicit variants: "harness next --json", "pnpm exec tsx src/cli.ts next --json" for the source repo, "harness ci-migrate prepare --provider circleci --dry-run", and gate commands like "pnpm check" / "pnpm test:deep". This matches the fully-executable anchor with common cases covered, rather than score 4 where concrete commands have minor gaps.

Workflow Clarity

The 7-step Workflow is clearly sequenced with preview-before-mutation (step 4), applied-lane execution (step 5), and gates before reporting (step 6), plus an explicit feedback loop in Validation: "stop at first blocking gate, fix it, and rerun the exact failed check". Because validation and rollback (snapshot-backed ci-migrate abort) are present for the destructive/CI-migration lane, the destructive-cap rule does not apply and the top anchor fits better than the minor-gaps anchor.

Progressive Disclosure

The Progressive Disclosure section signals real one-level-deep bundle files with purpose ("references/contract.yaml for the machine-readable contract", "references/evals.yaml for benchmark and quality gates", "references/task-profile.json for evaluator thresholds" — all verified present). It stays at score 4 rather than 5 because contract detail like the CI-provider ownership matrix and per-repo learnings is inlined in the body instead of a reference, and cross-skill "Infrastructure/references/..." paths cannot be verified within this bundle, leaving minor organization gaps.

Specificity

The description enumerates multiple concrete actions on a named domain — "install, bootstrap, upgrade, audit, diagnose, or explain @brainwav/coding-harness" — and extends coverage with "harness init/upgrade, CI migration, governance gates, command discovery, and Codex environment action sync". This matches the anchor for multiple specific concrete actions with comprehensive coverage; score 4 ('minor gaps in coverage') fits worse since the when-clause itself enumerates the sub-areas.

Completeness

It explicitly answers both: what ("install, bootstrap, upgrade, audit, diagnose, or explain @brainwav/coding-harness") and when ("Use when users need to ... in a repository") with concrete trigger phrases. This matches the anchor for clearly and explicitly answering both; the score-4 anchor ('when could be more explicit') applies less since the trigger clause is fully explicit and domain-scoped.

Trigger Term Quality

Natural terms users would say are present ("install", "upgrade", "CI migration", "explain", "diagnose"), giving good coverage, but "governance gates", "action sync", and "bootstrap" lean technical and natural variants like "set up" or "repair" are missing, so it falls just below the comprehensive-synonyms anchor.

Distinctiveness Conflict Risk

Tying triggers to the specific package "@brainwav/coding-harness" with its own vocabulary (ci-migrate, command discovery, environment action sync) creates a clear niche with distinct triggers and minimal conflict risk, matching the top anchor rather than the score-4 'minor overlap risk with closely related skills' anchor.