Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is highly actionable — concrete curl commands, precise thresholds, and a ready output template — with a clear sequenced audit workflow, but the monolithic body explains concepts Claude already knows and would benefit from being split into reference files to improve token efficiency and progressive disclosure.
Suggestions
Move the long output template and per-category scoring rubrics into reference files (e.g. AUDIT_TEMPLATE.md, SCORING.md) and link to them from SKILL.md, reducing the inline token load.
Cut explanations of concepts Claude already knows (what LCP/INP/CLS measure, what HTTPS is, what a viewport meta tag does) and keep only the audit-specific thresholds and checks.
Resolve the duplicate per-category scoring schemes (e.g. Category 7's 'Scoring Detail' vs 'Category Scoring' table) into a single, consistent point breakdown to remove redundancy.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~525-line body is mostly organized reference material but explains concepts Claude already knows (what LCP/INP/CLS measure, what HTTPS does) and carries duplicate per-category scoring schemes, so it could be tightened rather than being lean end-to-end. | 2 / 3 |
Actionability | Provides fully executable, copy-paste-ready commands (e.g. `curl -o /dev/null -s -w 'TTFB: %{time_starttransfer}s\n' [URL]`, `curl -s [URL]` for SSR detection), concrete thresholds (TTFB < 800ms, 48x48 CSS pixels, max-age=31536000), and a complete output template. | 3 / 3 |
Workflow Clarity | A clear numbered 'How to Use This Skill' sequence (collect URLs, fetch, run 8 categories, score, generate report) with each category structured into ordered checks; the task is a read-only audit so the destructive/batch validation-loop cap does not apply. | 3 / 3 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are empty) and the entire audit lives in a single ~525-line monolithic file that could be split (e.g. scoring rubric, AI-crawler table, output template as separate references); sections are well-organized but content that should be separate is inline. | 2 / 3 |
Total | 10 / 12 Passed |