Agents / Agents

Accessibility Auditor

Agent

Audits your screens against accessibility standards and fixes what it finds — keyboard traps, missing labels, broken focus, contrast failures — so the product works for everyone and passes the audits that matter.

Best for

  • Products that have never had an accessibility pass
  • Meeting WCAG requirements for a contract or a law
  • Building accessibility in before retrofit gets expensive

What you give it

  • The screens or flows to audit, and the standard you must meet
  • Decisions where a fix changes the design

What you get back

  • Findings by severity with the affected users named — who exactly is locked out
  • Fixes applied where the pattern is clear; options where design must change
  • A re-test after fixing, plus guidance your team can apply to new screens

How it works

  1. Walks each flow keyboard-only, then with assistive-technology semantics in mind.
  2. Checks structure (headings, landmarks, labels), behaviour (focus, announcements) and presentation (contrast, zoom, motion).
  3. Ranks findings by who is excluded and how completely — not by how easy the fix is.
  4. Fixes code-level issues directly, using your component patterns.
  5. Returns design-level conflicts (brand colour vs contrast) as decisions with options.

Example

You: Audit the signup and checkout flows against WCAG 2.1 AA.

Result: Nine findings: two blockers (checkout button unreachable by keyboard; error messages announced to nobody), four serious (labels, contrast, focus order), three minor. Blockers and serious fixed and re-tested; one contrast fix returned as a design option because the brand colour itself fails.

Limits — please read

  • Automated and code-level checks catch a lot, but real assistive-technology user testing remains the gold standard — it will tell you when that is needed.
  • Design-system-wide fixes are proposed once at the system level, not patched screen by screen.
  • Standards versions differ; it audits against the one you name.