Conciseness

The body is dense and operational — no concept-teaching padding (it never explains what power or an endpoint is), and tables/commands carry most of the load. However, the opening paragraph restates the frontmatter description, and the Purpose section's three tool bullets duplicate the Scripts table's descriptions of the same three scripts — trimmable redundancy that fits the score-4 anchor ('minor instances of over-explanation that could be trimmed') rather than the fully lean score-5 anchor.

Actionability

Commands are copy-paste ready and backed by real bundle files: 'python3 scripts/sample_size_estimator.py --design proportions --p1 0.30 --p2 0.45 --dropout 0.15', 'endpoint_selector.py --input endpoints.json --profile {drug|device|biologic|diagnostic|digital-therapeutic}', onboarding commands, and the /ar:setup example. '--sample' invocations cover common cases and every referenced script and template exists in the bundle — fully matching the score-5 anchor.

Workflow Clarity

The 5-step Workflow is clearly sequenced with real checkpoints: 'If >1 primary, plan multiplicity control', 'Trace the effect/difference/HR to a published or anchor-based source', 'inflate for dropout', and the final sign-off gate ('a biostatistician, medical monitor, and regulatory owner sign'). It falls short of the score-5 anchor because there is no feedback loop for adverse verdicts — nothing says what to do when the scorer returns REDESIGN or NO-GO (e.g., address blockers and re-score).

Progressive Disclosure

Structure is good against the actual bundle: three topical reference files are listed with one-line content summaries ('references/study_design_canon.md — ICH E8(R1)... '), all one level deep, and scripts plus the asset template are correctly cited. But content that belongs in a reference is inlined: the ~25-line forcing-question library (walked only by '/cs:grill-research-ops') and the opt-in autoresearch section could be split out, matching the score-4 anchor ('most content is appropriately placed; minor organization gaps') rather than the score-5 anchor ('content appropriately split').

Specificity

The description lists multiple concrete actions covering all three of the skill's tools: 'selecting and classifying endpoints (primary / key-secondary / exploratory, with surrogate-endpoint flagging)', 'estimating sample size and power for two-arm designs (means / proportions / survival)', and 'scoring a study plan for feasibility and a GO / GO-WITH-CONDITIONS / REDESIGN / NO-GO phase-gate decision'. Coverage is comprehensive with no identifiable gaps, matching the score-5 anchor rather than the score-4 anchor ('minor gaps in coverage').

Completeness

Both questions are explicitly answered: the 'what' is the three concrete actions (endpoint selection/classification, sample-size/power estimation, feasibility scoring), and the 'when' opens with 'Use when designing a prospective clinical study before submission' plus two additional trigger scenarios. This matches the score-5 anchor ('clearly and explicitly answers both what AND when with concrete trigger phrases'), not the score-4 anchor where the 'when' 'could be more explicit or specific'.

Trigger Term Quality

Good natural keyword coverage: 'designing a prospective clinical study', 'endpoints', 'sample size', 'power', 'feasibility', 'phase-gate'. However, the most common user phrasing 'clinical trial' is absent (only 'clinical study' appears), as are 'power analysis' / 'protocol design', so a few natural synonyms are missing — the score-4 anchor rather than the comprehensive score-5 anchor.

Distinctiveness Conflict Risk

The description carves out a clear niche (prospective study design pre-submission) and proactively disambiguates: 'Distinct from ra-qm-team, which handles the regulatory/QM submission (ISO 13485, EU MDR, FDA 510(k)/PMA/QSR), not the study design'. Triggers (endpoint selection, sample size, phase-gate) are distinct, matching the score-5 anchor ('clear niche with distinct triggers; minimal conflict risk').