Conciseness

The body is dense and additive — per-study-type required inputs, a subsection-by-subsection output structure, and hard rules with exact placeholder syntax — with no explaining of concepts Claude already knows. Two minor trims keep it below anchor 5: the guideline-to-study-type mapping appears twice (Step 1 'Applicable reporting guideline: CONSORT (RCT), STROBE (observational)...' and again in the Quick Reference table), and the quoted line '> "Method Writing produces Methods section text..."' adds little. Well above anchor 3, which requires unnecessary explanation.

Actionability

Highly concrete for an instruction-only skill: the eight IMRAD subsections each specify their required contents, the hard rule gives exact placeholder syntax '[AUTHOR TO SPECIFY: randomization method]', and Step 5 gives a verbatim deliverable format ('CONSORT items covered: [list]. Items not addressed (need from author): [list]'). Below anchor 5 only because there is no worked example of the target output prose (e.g., one model Methods paragraph demonstrating register, tense, and abbreviation style) — a minor gap per the instruction-only guidance in the rubric's scoring notes.

Workflow Clarity

Five clearly sequenced steps with real checkpoints: 'If the study type is unclear, ask one focused question before proceeding', 'If critical items are missing, ask for them before writing', and a post-draft Step 4 guideline check. Below anchor 5 because the feedback loop is not fully closed: Step 4 says to 'Identify any required item that is missing or incomplete' and flag items, but never explicitly instructs revising the draft to close those gaps before delivery — a minor validation gap versus the anchor-5 'fix and re-validate' pattern. Not a destructive/batch operation, so the cap-at-3 rule does not apply.

Progressive Disclosure

Appropriate split: the ~120-line SKILL.md keeps workflow and rules inline while three substantial detail files (imrad_structure.md, reporting_guidelines.md, writing_principles.md — 615/748/786 lines, verified to exist with no nested references) carry the deep material, one level deep and clearly labeled. Two minor gaps versus anchor 5: the Quick Reference table duplicates content that belongs in reporting_guidelines.md, and the references are grouped in a terminal list rather than linked at point of use (e.g., Step 4 does not point to reporting_guidelines.md when it runs the guideline check).

Specificity

Lists multiple concrete actions covering the domain comprehensively: 'Turns your protocol and analysis workflow into publication-ready Methods text', 'ensuring it complies with reporting guidelines (CONSORT, STROBE, PRISMA, TRIPOD)', 'matches what is in the Results section', 'satisfies journal-specific word limits and declarations'. Anchor 5 fits because capability coverage has no noticeable gaps; anchor 4 ('several specific actions; minor gaps') understates it. Voice is third person ('Turns') — 'your protocol' is a possessive object, not a second-person construction like 'You can use this', so no voice penalty applies.

Completeness

Explicitly answers both questions: what ('Turns your protocol and analysis workflow into publication-ready Methods text... ensuring it complies with reporting guidelines...') and when ('Use when writing or revising the Methods section of a biomedical manuscript', plus literal trigger phrases). This matches the anchor-5 good example structure exactly; the 'when' is not merely implied as in anchor 3.

Trigger Term Quality

Explicitly enumerates natural phrases users would actually say: '"write my methods"', '"revise my methods section"', '"how to report my statistics"', '"make my methods CONSORT-compliant"', alongside guideline acronyms and 'Methods section of a biomedical manuscript'. This is comprehensive synonym coverage per the anchor-5 pattern (no file extensions exist for this domain, so none are expected); anchor 4 ('a few natural terms missing') does not apply.

Distinctiveness Conflict Risk

A clear niche — biomedical manuscript Methods sections with named reporting guidelines (CONSORT, STROBE, PRISMA, TRIPOD) — with distinct triggers ('make my methods CONSORT-compliant') unlikely to fire for generic writing or document skills. Anchor 4 ('minor overlap risk with closely related skills') would require some plausible confusion with adjacent skills, which the explicit Methods-section scoping prevents.