Conciseness

The body is table-heavy and terse, assuming Claude's competence with no background-concept explanations. Minor redundancy keeps it below anchor 5: the gate output format appears in both the Phase 4 block and the Output Examples section, and severity categorization is shown in both the Phase 3 table and the Severity Reference section.

Actionability

Mostly executable guidance: a copy-paste-ready command ('node ${SKILL_DIR}/scripts/static-checks.js ${PROJECT_ROOT}'), a concrete JSON output schema, real template file paths for subagent prompts, and concrete decision/commit-annotation formats. Below anchor 5 because 'Present findings via gate interaction' leaves the interaction mechanism abstract (which tool, what format) and subagent prompts are delegated rather than shown.

Workflow Clarity

Five clearly sequenced phases with an explicit validation checkpoint up front (static checks must run and produce pass/fail/warn before expert review), an error-handling table covering script failure, timeout, and clean states, and a fix-then-await-approval loop in Phase 5. This matches anchor 5: clear sequence, explicit validation, feedback loops for error recovery.

Progressive Disclosure

The body is an orchestration-level overview with well-signaled, one-level-deep references (criteria.md, review-checklists.md, verification.md) that all exist as real files, and the script bundle matches the body's claims about orchestrator/layers. Below anchor 5 because references/co-change.md exists in the bundle but is never referenced from SKILL.md, a minor navigation gap.

Specificity

The description names the domain (epistemic protocol verification) only through trigger verbs like "verify protocols" and "validate definitions", but never states what the skill actually does (static checks, parallel expert-review subagents, severity synthesis). This matches the anchor 'names domain and 1-2 concrete actions, but not comprehensive' — above anchor 2 because the actions are domain-specific rather than generic, below anchor 4 because no list of specific capabilities is given.

Completeness

The 'when' is explicit and strong (trigger phrases plus 'Invoke explicitly with /verify for pre-commit validation'), but the 'what' is only weakly implied through the trigger verbs — no separate statement of the skill's capabilities or mechanism. This sits between anchor 3 ('what' weakly implied) and anchor 4 ('when' could be more explicit), landing at 3 since the what-side is the weaker half.

Trigger Term Quality

Six quoted natural phrases ('verify protocols', 'check consistency before commit', 'validate definitions', 'run pre-commit checks', 'verify soundness') with good synonym variation across verify/check/validate/ensure. Good coverage but not comprehensive — natural terms like 'lint' or 'sanity check' are absent — placing it at anchor 4 rather than 5.

Distinctiveness Conflict Risk

The niche (epistemic protocol consistency before commit) is mostly distinct with clear commit-time triggers, though phrases like 'verify soundness' overlap with generic verification/lint skills. Minor overlap risk with closely related skills — anchor 4, not 5 (not a fully clear niche) and not 3 (triggers are well-differentiated).