Conciseness

The body is dense and prescriptive (prose-limit table, fenced-block table, arrow-style table) with no padding that explains concepts Claude already knows; the manual-vs-lesson sentences define the skill's own terminology. It stops short of 5 because of trimmable redundancy: the "TOTAL 0" invariant is stated four times (intro, Method step 12, Build and audit, checklist) and "Size commits by judgment" appears as both Method step 13 and again in Commit rules.

Actionability

Fully concrete, executable guidance: exact build commands ("node --test site/test_manuals.js", "node scripts/audit_manuals.js --manual <id>"), an exact commit message format ("feat(manual/<id>): <section slug>"), quantified limits ("at most 25 words", "viewBox=\"0 0 640 H\" with H at most 760, text at least 11 units"), and required keys per fence kind. For an instruction-only skill this is copy-paste-ready and matches anchor 5.

Workflow Clarity

A 13-step numbered Method in dependency order with an explicit validation feedback loop ("Fix every finding until the audit prints `TOTAL 0`", "Fix the text, never the limit"), a capture drift check, and a 10-item "Check before you finish" checklist. Sequence, checkpoints, feedback loop, and checklist are all present, matching anchor 5.

Progressive Disclosure

All six referenced files (research-and-sources.md, capture-kit.md, manual-json.md, section-template.md, prose-rules.md, audit.md) exist in references/, are linked at the point of use, are one level deep, and are indexed in a "Files" section; bulky detail (audit rules, banned words, schema fields) is correctly split out. This matches anchor 5's clear-overview pattern.

Specificity

The description states a concrete what ("Writes a long technical manual that explains one subject at one pinned version, from its purpose to each request and response") plus concrete invariants ("Every listing comes from an offline capture, every claim traces to a ranked source, and an audit must report TOTAL 0"). These are facets of one main action rather than a list of multiple distinct actions, matching anchor 4; it is well above 3 in concreteness but short of 5's multi-action comprehensiveness.

Completeness

It explicitly answers both what ("Writes a long technical manual that explains one subject at one pinned version...") and when ("Use it for...") with a list of concrete trigger phrases. This matches the anchor-5 example pattern exactly; 4 would require the when to be less explicit, which it is not.

Trigger Term Quality

Seven natural quoted triggers ("write a manual", "write a 101 manual for <subject>", "manual.json", "add a manual section", "build a capture kit", "audit the manual", "pin a spec version and document it") give good coverage including the manifest filename. A few natural synonyms a user might say, such as "write documentation for X" or "spec reference/guide", are missing, so it fits anchor 4 rather than 5's comprehensive synonym coverage.

Distinctiveness Conflict Risk

The niche is clear (pinned-version 101 manuals with offline capture kits, ranked sources, and a TOTAL 0 audit) and the triggers ("101 manual", "manual.json", "capture kit", "pin a spec version") are phrases unlikely to fire any other skill. Minimal overlap risk with generic documentation skills, matching anchor 5.