Conciseness

Nearly every line is a directive, checklist prompt, output template, or rule — 'Score each of Nielsen's 10 heuristics 0–4', 'tag with **P0–P3 severity**' — with no concept explanations. Minor trimmable instances exist: tone coaching is repeated ('Be brutally honest', the six-bullet '**Remember**' block restating style directives) and the suggested-command vocabulary is listed in full twice (Priority Issues and Phase 4 rules). Not 5 because of those redundancies; not 3 because there is no padding or explanation of things Claude already knows.

Actionability

Concrete, executable guidance throughout: a full heuristic scoring table with 'Total ??/40' format, worked persona output examples ('Alex (Power User): No keyboard shortcuts detected. Form requires 8 clicks for primary action.'), sample questions, an explicit command vocabulary, and conditional rules ('If the user chose a limited scope, only include items within that scope'). The gap: workflow-critical specifics are deferred to referenced files — the '8-item cognitive load checklist', 'P0–P3 severity definitions', and the persona 'selection table' — none of which exist in the bundle, so those steps cannot be fully executed as shipped. Not 5 due to those unavailable details; not 3 since the inline guidance itself is specific and executable apart from the missing references.

Workflow Clarity

Four numbered phases with explicit entry/exit gating ('**After presenting findings**, use targeted questions', '**After receiving the user's answers**'), a mandatory precondition ('if no design context exists yet, you MUST run /teach-impeccable first'), conditional branches ('If findings are straightforward (e.g., only 1–2 clear issues), skip questions and go directly to Phase 4'), and per-phase rules lists. Not 4: sequencing, gating, and conditionals are explicit throughout with no checkpoint gaps relevant to this read-only evaluation task (the destructive/batch cap does not apply).

Progressive Disclosure

The body uses a consistent, clearly signaled one-level-deep citation pattern ('> *Consult [cognitive-load](reference/cognitive-load.md) for the working memory rule and 8-item checklist*'), but none of the three referenced files (cognitive-load.md, heuristics-scoring.md, personas.md) exist in the bundle — the promised details (checklist, severity definitions, persona catalog) don't resolve. Additionally, content that arguably belongs in those references (the full 10-row heuristic table) is inline. Not 4 because missing referenced files is more than a minor organization gap; not 2 because the body itself is well-structured and references are prominently signaled rather than buried.

Specificity

Lists multiple specific concrete actions with comprehensive coverage: 'assessing visual hierarchy, information architecture, emotional resonance, cognitive load, and overall quality with quantitative scoring, persona-based testing, and actionable feedback' — five named assessment dimensions plus three output modalities that map directly onto the skill's phases. Not 4: the enumerated coverage has no notable gaps relative to what the skill does.

Completeness

Explicitly answers both: what — 'Evaluate design from a UX perspective, assessing visual hierarchy, information architecture, emotional resonance, cognitive load, and overall quality' — and when, with concrete trigger phrases — 'Use when the user asks to review, critique, evaluate, or give feedback on a design or component.' Matches the anchor-5 pattern exactly. Not 4: the 'when' clause is fully explicit with concrete triggers rather than weakly implied.

Trigger Term Quality

Good natural-term coverage via four verb synonyms — 'review, critique, evaluate, or give feedback' — but object coverage is limited to 'a design or component', missing common natural terms users would say such as 'UI', 'interface', 'page', 'mockup', 'wireframe', or 'usability'. Not 5 because those natural synonyms are missing; not 3 because the verb coverage is broad and phrased the way users actually ask.

Distinctiveness Conflict Risk

A clear niche ('design from a UX perspective') with domain-specific trigger phrasing; mostly distinct from generic skills. Minor overlap risk remains with closely related sibling design skills — the body itself lists /audit and /polish in the same ecosystem, and 'review... a design' could plausibly trigger a design-audit command instead. Not 5 due to that minor overlap with closely related commands.