Content
35%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 well-organized as a checklist and provides a useful severity scheme and output template, but it is verbose with web-quality knowledge Claude already has and lacks executable audit guidance — notably, it never references the bundled scripts/analyze.sh. Tightening the inline explanations and wiring in the script would materially improve it.
Suggestions
Trim the per-check explanations that restate common knowledge (LCP/INP/CLS thresholds, WCAG ratios, meta-description lengths); keep them as a terse checklist and delegate depth to the referenced sibling skills.
Reference and invoke scripts/analyze.sh from the 'How it works' section so Claude has an executable first step instead of only 'Analyze the provided code/project'.
Move the bulk of the 150+ inline category checks into a reference file (e.g. references/checks.md) and keep SKILL.md as a concise overview pointing to it, improving progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body extensively restates well-known Lighthouse/WCAG knowledge Claude already has (e.g. 'LCP < 2.5s ... Optimize images, fonts, and server response time', WCAG 4.5:1 contrast, meta description 150-160 chars), matching 'noticeably verbose; several unnecessary explanations or padded sections'. It is not a 3 because the padding is pervasive rather than a few stray over-explanations, and not a 1 because the material is organized rather than a formless wall. | 2 / 5 |
Actionability | The 'How it works' section gives only high-level hints ('Analyze the provided code/project for quality issues') with no executable commands, and the bundled scripts/analyze.sh is never referenced or invoked from the body. This matches 'minimal concrete guidance; high-level hints but missing the specific steps to execute'. It is not a 3 because there is no executable code or command sequence to run the audit, and not a 1 because the output-format template and severity table offer some concrete structure. | 2 / 5 |
Workflow Clarity | A four-step sequence is listed (Analyze, Categorize by severity, Provide recommendations, Include code examples) but with no validation or verification checkpoints, matching 'steps listed but validation gaps; sequence present but checkpoints missing or implicit'. It is not a 4 because no checkpoints or feedback loops are present, and not a 2 because the steps are coherent and in a sensible order. | 3 / 5 |
Progressive Disclosure | The References section clearly signals one-level-deep sibling-skill links, but the 150+ inlined checks remain in SKILL.md rather than being split into reference files, and the bundled scripts/analyze.sh is never linked from the body. This matches 'some structure but could be better organized; references present but not clearly signaled; content that should be separate is inline'. It is not a 4 because of the unreferenced bundle script and the inlined bulk checklist, and not a 2 because structure and clear sibling references do exist. | 3 / 5 |
Total | 10 / 20 Passed |