accessibility-auditor-design

Workflow

1. Define audit scope: Identify the product surfaces, user flows, and platforms to audit. Collect the applicable compliance level (WCAG 2.1 AA minimum). Deliverable: scoped audit brief listing screens, flows, and success criteria.
2. Inventory existing state: Pull current accessibility annotations, design tokens (colour contrast ratios, focus states, touch targets), and any prior audit reports. Deliverable: baseline inventory document.
3. Execute automated checks: Run automated tooling (axe-core, Lighthouse accessibility, colour-contrast analyzers) across all scoped surfaces. Record every violation with severity, WCAG criterion, and screenshot. Deliverable: raw automated findings log.
4. Perform manual testing: Test keyboard navigation, screen reader compatibility (VoiceOver, NVDA, TalkBack), magnification, reduced motion, and high-contrast mode. Document interaction sequences that fail. Deliverable: manual testing results matrix.
5. Classify and prioritize findings: Score each issue by severity (critical, major, minor), affected user population, and remediation effort. Apply the scoring rubric at references/scoring-rubric.md. Group by WCAG principle (Perceivable, Operable, Understandable, Robust). Deliverable: prioritized issue register.
6. Draft remediation recommendations: For each finding, specify the design change required -- updated colour tokens, focus ring additions, ARIA label corrections, reflow adjustments. Assign to design or engineering. Deliverable: remediation plan with owners and effort estimates.
7. Compile audit report: Assemble findings, screenshots, compliance scores, and remediation plan into a structured report. Include executive summary with pass/fail verdict. Deliverable: accessibility audit report.
8. Present and hand off: Walk stakeholders through critical findings and remediation timeline. File issues in the backlog with audit references. Deliverable: backlog tickets linked to audit findings.

Anti-Patterns

• Automated-only auditing: Relying solely on automated tools and skipping manual assistive technology testing. Why: automated scanners catch roughly 30% of accessibility issues; keyboard traps, screen reader announcement order, and cognitive load problems require human evaluation.
• Severity inflation: Marking every finding as critical to force immediate action. Why: teams deprioritize everything when everything is urgent, and real critical issues (keyboard traps, missing alt text on primary actions) get buried.
• Retrofitting bias: Recommending overlay widgets or bolt-on fixes instead of design-level corrections. Why: overlays introduce new accessibility failures, create maintenance debt, and do not address root cause design decisions.
• Scope creep into redesign: Expanding the audit into a full redesign initiative without explicit approval. Why: the audit's value is diagnostic; mixing diagnosis with solution work delays the report and blurs accountability.

Output

On success: Produces an accessibility audit report containing executive summary, compliance scorecard (per WCAG principle), prioritized issue register with screenshots, and a remediation plan with ownership assignments. Delivered as a structured document shared with design, engineering, and product leadership.

On failure: Report the audit scope that could not be completed, the blocking reason (missing platform access, unavailable assistive technology, incomplete design tokens), and recommended next steps to unblock. Every gap must specify what is needed to resume.

Related Skills

• accessibility-checker-design -- Pre-handoff accessibility checks feed into post-launch audit baselines.
• design-review-runner -- Design reviews may surface accessibility concerns that trigger a full audit.
• session-analysis -- Session recordings reveal real-user accessibility friction that scopes audit focus areas.