visual-responsive-accessibility-audit

Audits web interfaces for visual and responsive accessibility barriers mapped to WCAG 2.2.

Updated Jul 7, 2026
One-click install
npx skills add https://github.com/vmitsaras/Agent-Skills --skill visual-responsive-accessibility-audit-vmitsaras
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: visual-responsive-accessibility-audit
Source: https://github.com/vmitsaras/Agent-Skills/tree/main/skills/frontend-a11y/visual-responsive-accessibility-audit
Command: npx skills add https://github.com/vmitsaras/Agent-Skills --skill visual-responsive-accessibility-audit-vmitsaras

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Visual and responsive accessibility issues like low contrast, clipped text at zoom, hidden focus indicators, and undersized touch targets are hard to review consistently, and screenshot-only reviews often lead to unsupported WCAG failure claims. This Skill provides an evidence-aware audit workflow that separates what screenshots can prove from what requires runtime testing or measurement. ## Core Features & Use Cases - Evidence-classified auditing: Classifies every observation as screenshot, source, runtime, measurement, or coverage-gap evidence, and labels findings as confirmed, provisional, or needs-validation. - Comprehensive WCAG 2.2 coverage: Tests contrast, color-only communication, text resize, zoom, reflow, orientation, text spacing, hover/focus content, focus visibility and obscuring, target size, and forced-colors modes with criterion-specific mapping guidance. - Structured finding contract: Produces findings with stable IDs, severity, confidence, framework mappings, remediation, and verification steps, plus an evidence ledger and coverage-gap report. - Use Case: Given screenshots of a responsive checkout flow at multiple breakpoints, the audit confirms visible clipping issues, flags apparent contrast risks as provisional pending measurement, and lists the runtime tests needed to validate zoom and reflow behavior. ## Quick Start Audit the attached page screenshots and component source files for visual and responsive accessibility barriers, mapping confirmed findings to WCAG 2.2 criteria.

Frequently Asked Questions about visual-responsive-accessibility-audit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I audit a website for WCAG 2.2 visual accessibility?▼

Test text and non-text contrast against 4.5:1 and 3:1 thresholds, verify color is not the only means of conveying information, and exercise text resize, browser zoom, reflow at 320 CSS pixels, text spacing overrides, focus visibility, and 24 by 24 CSS-pixel target sizes. Record measured ratios, viewport dimensions, and states for each finding.

How to test responsive reflow and zoom accessibility?▼

Test text resize to 200% for WCAG 1.4.4 and browser zoom up to a 320 CSS-pixel equivalent width for reflow under 1.4.10, checking for clipped, overlapping, or hidden content and horizontal scrolling. Record both zoom percentage and resulting CSS viewport dimensions, since zoom is a test method rather than a standalone criterion.

Can screenshots alone prove a WCAG contrast failure?▼

No. A screenshot can confirm visible symptoms like clipping or overlap in its captured state, but contrast failures require measured rendered foreground and background colors. Apparent low-contrast pairings should be labeled provisional until measured with a named method.

Does this audit cover keyboard navigation and screen readers?▼

Only partially. It includes focus visibility and focus obscuring because they are visual and layout-dependent, but keyboard reachability, focus order, screen-reader output, and semantic structure are routed to specialist audits such as keyboard-focus-accessibility-audit.

What are the limitations of a visual accessibility audit?▼

The result is a bounded technical audit, not legal advice, certification, or a WCAG conformance claim. Untested viewports, states, themes, and platform modes are reported as coverage gaps, and untested or not-reproduced areas are never reported as passes.