Lean and assumption-respecting: it never explains what FAIR, a DOI, or an accession number is, and every bullet in 'Default stance' and every workflow step earns its place without padding.

Gives concrete executable guidance — an explicit access-route classification enum, a literal ready-to-paste output-format block, and a CN-to-EN terminology conversion table — which is actionable despite being instruction-only (code absence is not penalized per scoring notes).

An 8-step sequenced workflow includes an explicit validation checkpoint ('Run the FAIR and metadata audit before finalizing') and returns unresolved-field flags, matching the clear-sequence-with-validation anchor; no destructive/batch cap applies.

SKILL.md is a concise overview with a 'Related files' table that signals when to open each of six references, all verified one level deep with no nested .md links, giving easy navigation.

Lists multiple concrete actions ('Prepare, audit, or revise') across four distinct deliverables (Data Availability statements, repository plans, dataset citations, FAIR metadata checklists), matching the 'multiple specific concrete actions' anchor.

Explicitly answers both what (prepare/audit/revise the four deliverables) and when via a clear 'Use when the user asks about...' clause with many triggers, so it is not capped at 2.

Covers natural user phrasing broadly — 'Nature data availability, research data sharing, repository selection, accession numbers, restricted or sensitive data, source data, supplementary datasets' — terms a user would actually say.

A clear Nature-family data-availability niche with distinct triggers ('Nature data availability', 'FAIR metadata for academic publication', 'Chinese-to-English data availability wording') makes overlap with other skills unlikely.