Content
78%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.
A well-structured, largely actionable skill body with excellent progressive disclosure — on-demand reference pointers, a command quick-reference, tier detection, and a required cost gate on the paid batch workflow. It falls short of top marks on conciseness (a hedged time-sensitive 2026 context block), actionability (a few steps defer all specifics to references), and workflow consistency (Tier 2 omitted from the output spec).
Suggestions
Tighten the 'AI & 2026 context' blockquote to the operationally relevant facts (which GBP API fields are new, what to confirm before promising agentic booking) and drop the reporting-caveat scaffolding; it is time-sensitive detail that costs tokens.
Name the specific DataForSEO endpoints/task POSTs inline for the review-intelligence and GBP-audit workflows (as is already done for the Maps SERP `location_coordinate` param) so the common cases are executable without opening a reference file.
Fix the Output section to say 'Tier 0, 1, or 2' so the deliverable spec matches the three-tier detection section.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational (SoLV formula, exact grid-file JSON schema, copy-paste cost-warning block) and assumes Claude's competence, but the 'AI & 2026 context' blockquote carries hedged time-sensitive details (dates, I/O 2026 claims with sourcing caveats) that could be trimmed, and the Security section restates general policy. Efficient with minor instances of over-explanation fits the 4 anchor rather than the every-token-earns-its-place 5 anchor. | 4 / 5 |
Actionability | Mostly executable guidance: exact cost-confirmation template, concrete Bing WebFetch URL, named MCP tool for tier detection, 2/1/0 GBP scoring rules, and exact output-file schema. However several steps stay high-level ('Fetch Google reviews via DataForSEO Reviews API', Apple Maps is a 'manual check') with specifics deferred to reference files, so it sits at the 4 anchor rather than the fully-executable 5 anchor. | 4 / 5 |
Workflow Clarity | Each command has a numbered, sequenced workflow; the paid batch operation (geo-grid) is explicitly gated by a required cost-estimate confirmation, and a seven-scenario error-handling table supplies recovery actions. It stays at 4 because some workflows (review intelligence, schema generation) lack checkpoints and the Output section says 'Tier 0 or Tier 1' while Tier 2 is defined in the tier-detection section — minor validation/consistency gaps consistent with the 4 anchor, not the 5 anchor's explicit validation throughout. | 4 / 5 |
Progressive Disclosure | Clear overview structure: a Quick Reference command table, per-section 'Load:' pointers, and a consolidated Reference Files list stating each file's purpose with 'Load on-demand as needed (do NOT load all at startup)'. References are one level deep and well signaled, matching the 5 anchor; no bundle files ship alongside this SKILL.md, so scoring reflects the in-file referencing structure. | 5 / 5 |
Total | 17 / 20 Passed |