Content
85%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.
The body is an unusually disciplined instruction-only skill: concrete tool parameters, exact output conventions, honest-sizing rules, and a well-sequenced workflow with real validation checkpoints. Its main structural weakness is that reference material (tool semantics, report format spec) is inlined in a long single file rather than offloaded to reference files.
Suggestions
Move the detailed MCP tool semantics (get_serp_results, get_ranked_keywords conventions) into a references/tools.md file, keeping a one-line summary of each tool in SKILL.md.
Move the full output-format specification and HTML skeleton into a references/report-format.md, inlining only the section list and key writing rules.
Trim the get_serp_results paragraph by moving edge-case conventions (merge behavior, seenIn handling) into the same reference file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with non-obvious domain judgment — tool semantics, evidence conventions, sizing rules — and assumes Claude's competence without padding or basic-concept explanations. A few spots could be trimmed (the long get_serp_results conventions paragraph, the inline canonical gloss), which keeps it at anchor 4 rather than 5. | 4 / 5 |
Actionability | Guidance is fully executable throughout: exact tool calls with parameters (resultTypes: ["organic"], scope: "exact_url", depth 20), exact position notation rules ("#10 (page 1)"), a copy-paste-ready HTML skeleton for the deliverable, and defined columns for the opportunities.md shortlist. This matches anchor 5 — specific examples cover the common cases. | 5 / 5 |
Workflow Clarity | Six clearly sequenced steps (Orient, Investigate, Shortlist, Choose, Size, Review/write) with explicit validation checkpoints: re-run deciding queries before writing, complete missing lookups before ranking, a reviewer pass with a concrete checklist, verification of new factual claims, and crawl-failure investigation before anything else. This matches anchor 5; the destructive/batch cap does not apply since the audit is read-only plus report writing. | 5 / 5 |
Progressive Disclosure | The body has good section structure but is a single ~135-line file with no bundle files, and reference-like material (the MCP tool semantics section and the full output-format specification with HTML skeleton) is inlined rather than split into separate reference files. Anchor 3 fits best: structure is present, but content that could live one level deeper is inline, and the under-50-line simple-skill exception does not apply. | 3 / 5 |
Total | 17 / 20 Passed |