Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An efficient, highly actionable reference-style body: dense parameter table, copy-paste JSON patterns, and useful anti-patterns with no wasted tokens. Main gaps are the unstated invocation mechanism (tool/endpoint name) and the never-explicitized usage sequence (cache check → keyword query → parameter selection → call).
Suggestions
State the invocation mechanism explicitly — the tool, endpoint, or command name the JSON payloads are sent to — so the Quick Reference examples are fully copy-paste ready.
Add a short numbered usage sequence (check cache → formulate keyword query → select parameters → invoke) making the cache-first rule an explicit step instead of an Anti-Pattern aside.
Move the 24-row Parameters table into a bundled reference file (e.g. references/parameters.md) and keep a short core-parameter summary inline, letting SKILL.md act as a true overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is a lean tabular parameter reference plus nine distinct JSON use-case snippets, with zero padding and no explanation of concepts Claude already knows — "Cap how many top results get full content fetched; lower (e.g. 3) to keep listings cheap" is typical of the terse, earned tokens throughout. Fits the 5 anchor (lean, assumes competence); there is no over-explanation to justify a 4. | 5 / 5 |
Actionability | Quick Reference and Patterns provide copy-paste-ready JSON payloads (e.g. `{ "query": "prisma migrations guide", "include_domains": ["prisma.io"], "max_results": 3 }`) and the Parameters table gives types, defaults, and selection guidance. However, the actual invocation mechanism — tool, endpoint, or command name — is never stated, so guidance is mostly but not fully executable: the 4 anchor, not 5. | 4 / 5 |
Workflow Clarity | As a single-action tool skill it benefits from the simple-skill path, and the 'When to use' column plus Anti-Patterns ("DON'T send natural-language questions; use keyword phrases") make parameter choice clear. But the implied sequence — check cache first ("DON'T search without checking the cache first"), formulate a keyword query, select parameters, invoke — is never laid out as steps, and the invocation itself is implicit, leaving minor gaps at the 4 anchor rather than an unambiguous single action worth 5. No destructive or batch operations, so no cap applies. | 4 / 5 |
Progressive Disclosure | Sections are well organized (Quick Reference, Parameters, Always-Emitted Fields, Patterns, Anti-Patterns, When NOT to use, See Also) and the See Also links are one level deep and clearly signaled. At ~107 lines with no bundle files, the 24-row parameter table is reference material inlined in the overview rather than split out, which is a minor organization gap — the 4 anchor rather than the fully-split 5. | 4 / 5 |
Total | 17 / 20 Passed |