Content
53%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 comprehensive, largely actionable WCAG reference with strong code examples, but it front-loads the entire standard into SKILL.md. It would score much higher by keeping a concise overview with the audit workflow and moving the per-principle checklists into reference files, and by cutting concepts (POUR, conformance levels) that Claude already knows.
Suggestions
Split the four per-principle audit checklists (Perceivable, Operable, Understandable, Robust) into separate reference files (e.g., references/perceivable.md) and keep only a summarized checklist with clear one-level-deep links in SKILL.md.
Add an explicit audit workflow with validation checkpoints, e.g.: 1) run automated scan (axe CLI), 2) triage violations by impact, 3) verify each with manual checks, 4) apply remediation patterns, 5) re-run the scan to confirm zero violations.
Remove or drastically condense content Claude already knows (POUR principles, A/AA/AAA table, standard contrast ratio values) and keep only the operational checklists, tooling, and fix patterns that add value.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~550-line body restates substantial WCAG knowledge Claude already has: the POUR principles block, the A/AA/AAA conformance table, per-criterion checklists, and standard contrast ratios. Not a 2 because the content is mostly structured checklists and code rather than padded prose; not a 4 because the full enumeration of every success criterion is unnecessary explanation that could be trimmed or delegated to the WCAG spec link. | 3 / 5 |
Actionability | Concrete, mostly executable guidance throughout: before/after HTML fixes for labels and link text, CSS for focus outlines and reduced motion, a keyboard-accessible custom element class, an axe-core Playwright snippet, and copy-paste CLI commands ('npx @axe-core/cli https://example.com'). Not a 5 because the axe-core snippet assumes unstated setup (page/test harness imports) and some checklist items give no verification method. | 4 / 5 |
Workflow Clarity | Sections imply a rough flow (Core Concepts → Audit Checklist → Automated Testing → Remediation Patterns → Best Practices), but no explicit audit sequence exists and there are no validation checkpoints (e.g., how to reconcile automated results with manual checks, or when an audit is complete). Not a 2 because the section ordering and per-criterion checklists give real structure; not a 4 because the ordering is implicit and no stepwise process is stated. | 3 / 5 |
Progressive Disclosure | A monolithic 550-line SKILL.md with no bundle files at all; the per-principle checklists (Perceivable/Operable/Understandable/Robust) clearly belong in separate reference files loaded only when needed. Not a 1 because section headers and a Resources section provide navigable structure. | 2 / 5 |
Total | 12 / 20 Passed |