a11y-auditlisted

# a11y-audit — accessibility via axe-core, plus the pass axe cannot do Reference: [`examples/medibook/tests/a11y.booking.spec.js`](https://github.com/QAIA-Project/QAIA/blob/main/examples/medibook/tests/a11y.booking.spec.js) (0 serious/critical violations). Tooling is fixed: **axe-core driven through Playwright** — the de-facto standard for automated WCAG checks, and it reuses the browser context the functional tests already run in rather than introducing a second driver stack. **Naming a regulation is an implicit promise about what the run means.** This skill's oracle is **WCAG 2.1 A/AA and nothing else**. EN 301 549, the EAA, RGAA 4.1 and Section 508 are all built on WCAG, which is why the run helps with all of them — and each adds obligations WCAG does not contain, which is why a green run is **never** a compliance verdict for any of them. RGAA in particular is a *test method*, not a synonym: **106 numbered criteria, a mandatory page sample, and a compliance rate this skill's output cannot produce** — different numerator, different denominator, different sample. Load `references/regulations.md` before answering any question that contains the word "compliant", and use the wording it gives. **The number that governs this skill: automated tooling detects roughly a third of WCAG success criteria.** The rest are not "advanced" — they are keyboard access, focus visibility, and alt-text relevance, i.e. the failures that actually stop a disabled user. A skill that ships only t