Content
71%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 a highly actionable, copy-paste-ready report template with a clear input-to-output workflow, and the rubric's instruction-only provisions mean the absence of code is not penalized. Its weaknesses are the ~400-line monolithic structure with zero progressive disclosure and padding from a glossary of concepts Claude already knows plus verbose framing examples.
Suggestions
Split the report template, glossary, and tone/dollar-framing guidance into reference files (e.g. references/report-template.md, references/client-language.md), keeping SKILL.md as a concise overview with clearly signaled one-level-deep links.
Remove or drastically trim the glossary entries for concepts Claude already knows (JSON-LD, SERP, LCP, INP, CLS, E-E-A-T), keeping only genuinely domain-specific terms like GEO, llms.txt, and sameAs conventions.
Add validation checkpoints to the workflow — e.g. verify all required input report files exist before generating, confirm all 12 sections are present and the composite score matches the formula before writing the final GEO-CLIENT-REPORT.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The template tables, score formula, and tone guidance earn their tokens, but the glossary explains concepts Claude already knows (JSON-LD, SERP, LCP, INP, CLS, E-E-A-T) and the repeated 'Translate for the client' and dollar-value framing blocks pad the document. It is mostly efficient with some genuinely unnecessary explanation that could be tightened. | 3 / 5 |
Actionability | The guidance is fully executable: exact 12-section report structure, the composite score formula with component weights and rounding rules, complete table templates with placeholders, per-tier action criteria (e.g. 'Quick Win criteria: < 4 hours by one person'), and explicit tone rules — everything needed to produce the deliverable directly. | 5 / 5 |
Workflow Clarity | 'How to Use This Skill' gives a clear 5-step sequence with input dependencies and conditional branches (competitor section 'if competitor URLs provided', SSR callout), but there are no validation checkpoints such as verifying all input report files exist or confirming the generated report meets the stated completeness/length criteria. Not a destructive or batch operation, so the cap of 3 does not apply. | 4 / 5 |
Progressive Disclosure | The document is well-sectioned and navigable, but it is a ~400-line monolith with no bundle files at all — the glossary, methodology/data-source appendix, and tone/dollar-framing guidance could live in reference files while SKILL.md stays an overview. Some structure exists, but content that could be separate is fully inlined with no references. | 3 / 5 |
Total | 15 / 20 Passed |