Conciseness

The ~75-line body is lean and imperative throughout ("Read the index first. Do not grep the entire wiki on every query", "Every claim cites a page. No uncited assertions.") and assumes Claude's competence — no priming of concepts Claude already knows. Not 4: lines like "The wiki has done the hard work" are working guidance about trusting pre-synthesized pages, not trimmable padding.

Actionability

Mostly executable: exact script invocations ("python <plugin>/scripts/wiki_search.py --vault . --query \"<terms>\" --limit 5", the append_log.py command with args), a verbatim offer-to-file blockquote, and named frontmatter fields. Not 5: the `<plugin>` placeholder is never resolved to a concrete path, and "Update wiki/index.md (inline or via script)" leaves the script unnamed. Not 3: the commands given are real and copy-paste ready, not pseudocode.

Workflow Clarity

The 6-step workflow is clearly sequenced, with a fallback search loop (step 4) and a red-flags checklist giving explicit recovery actions ("Answering without reading the index → go back", "Citing only one source for a multi-source question → broaden"). Not 5: there is no validation checkpoint after the write path — nothing confirms the filed page landed in the index or passes lint. The destructive/batch cap at 3 does not apply since the write operations (create page, append log) are additive, not destructive.

Progressive Disclosure

Good structure: the body summarizes the workflow and delegates the full version and page templates to clearly signaled references ("Follow engineering/llm-wiki/skills/llm-wiki/references/query-workflow.md", page-formats.md) one level deep. Not 5: the referenced files live in another skill's bundle (this skill ships no references/, scripts/, or assets/ of its own), so the paths are long, repeated, and unverifiable from this skill's directory — a minor navigation gap.

Specificity

The description lists multiple concrete actions with specifics — "Reads index.md first, drills into 3-10 relevant pages across categories, synthesizes an answer with inline [[wikilink]] citations, and offers to file the answer back into the wiki as a new comparison or synthesis page" — covering the full read-select-synthesize-cite-file pipeline. Not 4: there is no noticeable coverage gap in the actions named.

Completeness

Both halves are explicit: the "what" is the concrete action chain, and the "when" is "Spawn when the user asks a substantive question the wiki might answer, says 'what does the wiki say about X'..." with concrete trigger phrases. This matches the anchor-5 exemplar structure exactly.

Trigger Term Quality

Natural user phrasings are quoted directly — "what does the wiki say about X", "compare A and B across my sources", "explore a topic" — alongside the synonym "substantive question the wiki might answer". Not 4: the quoted variants cover the query, compare, and explore phrasings users would actually say, not just a few keywords.

Distinctiveness Conflict Risk

The niche — answering queries against an LLM Wiki vault — is unmistakable, and the quoted triggers are wiki-specific, giving minimal conflict risk with other skills. Not 4: the only mildly broad clause ("or wants to explore a topic") stays anchored to "the wiki might answer", so overlap risk is minimal rather than minor.