Content
63%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-structured and appropriately brief with clean progressive disclosure to a real reference file, but the in-body guidance itself is high-level: it names what to do without examples, tooling, or concrete invalid/valid attribute pairs. Mild redundancy across Quick Reference, Check, Fix, and Code Review sections dilutes an otherwise clean simple-skill layout.
Suggestions
Inline 2-3 concrete invalid-vs-valid attribute examples (e.g., `aria-labeledby` → `aria-labelledby`) in the Fix section so the skill is actionable without opening the reference.
Name the specific verification tooling once (e.g., axe DevTools or the W3C WAI-ARIA spec attribute list) in the Check or Code Review section instead of the generic "browser accessibility tooling".
Merge the Quick Reference bullets into Check/Fix (or drop them) to remove duplicated content and tighten the token budget.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short, but the opening sentence ("Invalid ARIA attributes are ignored by browsers and screen readers...") explains a concept Claude already knows, and the Quick Reference bullets ("Use only valid, standard-defined ARIA attributes") largely duplicate the Check/Fix sections. Mostly efficient, but it could be tightened by merging these redundant sections. | 3 / 5 |
Actionability | Guidance like "Identify any invalid or misspelled ARIA attributes on HTML elements" and "Replace invalid ARIA attributes with their correct names as defined in the WAI-ARIA specification" is directionally concrete but incomplete — no tooling (axe, WAVE, browser devtools), no invalid-vs-valid attribute examples, and no inline spec lookup guidance, with all specifics deferred to references/rule.md. This lands between minimal high-level hints and executable guidance. | 3 / 5 |
Workflow Clarity | For a simple single-purpose review skill, the Check → Fix → Explain sequence makes the core action (flag invalid attributes, supply correct names) reasonably unambiguous, and the Code Review section adds verification framing ("note how to verify the fix with browser accessibility tooling"). It loses a point because the relationship between Check and Code Review sections is unclear and there is no explicit post-fix re-validation step. | 4 / 5 |
Progressive Disclosure | A concise overview with a single well-signaled, one-level-deep reference: "For full implementation details, code examples, and framework-specific guidance, see `references/rule.md`" — and references/rule.md exists in the bundle and covers exactly that. Content is appropriately split with easy navigation. | 5 / 5 |
Total | 15 / 20 Passed |