Conciseness

The body is mostly direct commands, but the fact that the tools are available in the headless loop is stated three times ('The same tool is available in the headless pnpm agent loop and every built-in app agent', then again 'The headless pnpm agent loop and built-in app agent also expose read-only framework-search, docs-search, and source-search tools. Use the unified tool for discovery, then the focused tools for full page or file reads'), and the closing 'Authoring visual docs blocks' section is tangential to a docs-lookup skill. Mostly efficient but with noticeable redundancy that could be tightened — matching the score-3 anchor; the duplicated paragraphs push it below 4.

Actionability

The body is dominated by copy-paste-ready commands with concrete patterns and paths: 'pnpm action framework-search --pattern "defineAction"', 'cp node_modules/@agent-native/core-corpus/corpus/templates/mail/actions/archive-email.ts actions/archive-item.ts', and 'rg -n "defineAction" node_modules/@agent-native/core/dist/action.js'. Specific examples cover the common cases (search, focused read, corpus install fallback), matching the score-5 anchor.

Workflow Clarity

There is a clear decision sequence — unified 'framework-search' first, narrow with 'scope: docs' or 'scope: source', then focused readers, with fallbacks if the action runner is unavailable ('If the action runner is unavailable, search the package directly') and if the corpus is missing (install '@agent-native/core-corpus@<installed-core-version>'). It is a read-only lookup skill with no destructive or batch operations, so the validation cap does not apply, and the caution 'refine the pattern or path instead of treating a truncated result as exhaustive' serves as a feedback hint. It sits at 4 rather than 5 because the sequence is spread across prose paragraphs rather than presented as an ordered workflow.

Progressive Disclosure

No bundle files exist, and the skill is appropriately self-contained: well-labeled sections (Rule, Why, How, Reuse Proven Patterns, Useful Slugs, Don't) with the slug table inline where it belongs. It sits at 4 rather than 5 because 'customizing-agent-native' is referenced three times ('read customizing-agent-native before taking ownership of a shared component', 'follow customizing-agent-native's configure/compose/eject ladder') without ever stating where to find it — it is not in the Useful Slugs table and no path or lookup command is given.

Specificity

The description names its domain ('find version-matched Agent-Native framework docs and source bundled in node_modules') with one concrete action, but does not enumerate several specific capabilities — the actual search tooling and corpus mechanics live only in the body. It names the domain plus 1-2 actions without comprehensive coverage, matching the score-3 anchor; it falls short of 4 because no list of several distinct actions is present.

Completeness

Explicitly answers both questions: the what is 'How to find version-matched Agent-Native framework docs and source bundled in node_modules' and the when is the explicit trigger clause 'Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features'. Both are present with concrete trigger phrases, matching the score-5 anchor exactly.

Trigger Term Quality

Strong natural triggers a user of this framework would actually say: '@agent-native/core', 'APIs', 'generated apps', 'workspaces', 'templates'. Good keyword coverage matching the score-4 anchor; not 5 because a few natural variations are missing (e.g. 'toolkit', 'docs lookup', 'source search', 'corpus').

Distinctiveness Conflict Risk

The description is tied to a named internal framework ('@agent-native/core', 'generated apps', 'workspaces'), giving it a clear niche with distinct triggers and minimal conflict risk with generic skills. It matches the score-5 anchor; the framework-specific terms make wrong-skill triggering very unlikely.