a11y-auditlisted

# a11y-audit A WCAG 2.2 Level AA accessibility audit skill built for coding agents. The output format is the distinctive part: violations are grouped by **who can fix them**, not just by severity. ## Why this shape Most a11y tools produce a flat list of violations sorted by severity. That's useful for a human reading a compliance report, but it's the wrong shape for an agent-plus-human workflow. An agent can confidently fix some issues (replacing `<div onClick>` with `<button>`, adding `type="button"`, remapping a known-bad Tailwind color class), needs the user's judgement for others (what *should* the alt text say, what label copy to use), and can't do a third category at all (did the focus order feel right, did NVDA announce the modal clearly). This skill produces three buckets in that order: 1. **Safe to fix now** — the agent can patch these without further input once the user asks for that bucket. 2. **Needs your decision** — the agent drafts each fix once the user answers one question. 3. **Test it yourself** — these need a human in the browser or with assistive tech, including guided checks derived from the page and journey context. ## When to use this skill Trigger on any request involving: - Accessibility audit, WCAG, a11y, Section 508, ADA, EN 301 549 - "Is this accessible?" / "Can this be used with a screen reader?" / "Does this work with a keyboard?" - Alt text, ARIA, focus indicators, color contrast, keyboard navigation, semantic HTML - Inclusivity or disab