Conciseness

The body is lean and assumes Claude's intelligence — it does not explain what a linter or cron is beyond a load-bearing one-line contrast, and every line earns its place with no padded concept explanations.

Actionability

As an instruction-only skill it provides concrete, specific guidance — 'Evidence is a cited file and line' and the four-part finding template (documented intent / implemented reality / attacker and victim / concrete fix) — which the rubric treats as equivalent to executable code when guidance is actionable.

Workflow Clarity

A clear 5-step sequence (Establish intent → Gather evidence → Compare → Classify → Avoid hand-wavy) with explicit validation gates such as 'If you cannot cite both sides of the gap, it is a question to investigate, not a finding.'

Progressive Disclosure

Under 50 lines with no need for external references, the body is well-organized into clear sections (Purpose, Context, Method, What counts, Notes), which the rubric permits to score 5 for short self-contained skills.

Specificity

The description lists several concrete actions — 'Defines what counts as documented intent, what counts as implementation evidence, which mismatches matter, and how to avoid hand-wavy findings' — giving comprehensive coverage of the method rather than minor gaps.

Completeness

It explicitly answers both what ('The method for finding the gap...') and when ('Use when auditing AI-built code, reviewing access control...') with concrete trigger phrases, matching the anchor-5 pattern.

Trigger Term Quality

'Use when auditing AI-built code, reviewing access control against documented permissions, or checking whether a codebase matches its own documentation' provides good natural keyword coverage, though it lacks common synonyms/variations for a 5.

Distinctiveness Conflict Risk

It carves a clear niche — the intent-vs-implementation gap that 'generic scanners miss' — with distinct triggers and minimal overlap risk with other skills.