The body is lean — operating rules, I/O contract, and a copy-paste example with no concept explanations Claude already knows; every line is operational guidance, matching the lean/efficient anchor.

Provides a fully executable `echo '...' | python scripts/rest_request.py` command plus concrete JSON request patterns and an explicit output field list, which is copy-paste ready; not the level below because code is real and complete rather than pseudocode.

As a simple single-action skill (read JSON from stdin, run the script) the action is unambiguous via the operating rules and example; the simple-skill note allows a 3 here, and no destructive/batch validation loop is required.

Under 50 lines with well-organized sections and the single bundle file `scripts/rest_request.py` correctly referenced and present on disk; the under-50-lines/no-external-references note permits a 3 on organization alone.

Lists three concrete operations — 'details', 'publication-linkage', and 'DOI lookups' — matching the anchor for multiple specific concrete actions; not the level below because it goes beyond naming a single domain action.

Explicitly states what it does ('Submit ... API requests for details, publication-linkage, and DOI lookups') and when to use it ('Use when a user wants concise preprint metadata summaries'), satisfying both halves with an explicit trigger.

Natural user-facing terms 'bioRxiv', 'medRxiv', and 'preprint metadata' are exactly what a user would say, giving good coverage; 'publication-linkage' is somewhat technical but the natural vocabulary is present, so it is not the level below.

The bioRxiv/medRxiv preprint niche is narrow with distinct triggers unlikely to collide with other skills; not the level below because the domain is specific rather than broadly overlapping.