What problem does it solve? Web interfaces often pass automated lint checks while still blocking keyboard users, hiding focus states, or failing contrast requirements. This Skill provides a structured audit workflow that combines automated scanning with real keyboard, visual, and semantic inspection so accessibility issues are found and fixed by severity. ## Core Features & Use Cases - Multi-layer inspection: Runs automated scans (axe-core or equivalent), then verifies keyboard traversal, visual contrast, focus visibility, and DOM/accessibility-tree semantics. - Severity-based fixing: Classifies findings from P0 (blocks task completion) to P3 (polish) based on actual task impact rather than WCAG criterion numbers alone. - Re-test gate: Re-runs the same conditions after fixes, including 200% zoom, 320px reflow, reduced-motion, and mobile viewport checks. - Use Case: Before shipping a new checkout flow, audit the rendered pages to confirm keyboard-only users can complete purchase, focus is visible, and form errors are announced, then fix P0/P1 issues and re-verify. ## Quick Start Audit the accessibility of the checkout page using automated scan, keyboard navigation, and visual inspection, then report findings by severity.