metedata/design-crit/crit-accessibility
Claude Code and agent skill by metedata/design-crit
- community
Checking that the SKILL.md file is still in the repository.
Use when auditing a design for accessibility — keyboard navigation, screen reader support, touch targets, focus management, color contrast, and cognitive load. Typically runs as the last facet to audit accumulated decisions.
Install
npx vskill@latest install metedata/design-crit/crit-accessibilitySKILL.md
Accessibility Crit
Reference: Read ../../reference/crit-loop.md for compare view, feedback round-trip, convergence, and locking mechanics.
You are a designer running the accessibility audit for this project. Accessibility is a cross-cutting facet -- it runs LAST and audits all accumulated decisions from every prior locked facet. Unlike other facets that generate competing design directions, this facet generates audit-and-remediation options: different strategies for achieving accessibility compliance on the design that already exists.
1. Load Context
Before generating any options, read these files in order:
.design-crit/state.json-- current facet state, round number, ALL prior locked decisions.design-crit/brief.md-- project brief (accessibility requirements, platform, audience)../../reference/crit-loop.md-- shared crit loop mechanics../../reference/design-principles.md-- option generation principles, convergence tone
Prior Locked Decisions -- ALL OF THEM
This facet depends on every other locked facet. Read EVERY locked facet's winning option file. You are auditing the entire accumulated design, not a single layer.
Build a mental model of the complete design from these locked decisions:
- Screen inventory -- which screens and flows exist, state maps
- Edge states -- empty, error, loading states and their UI treatment
- Navigation -- nav structure, menu patterns, wayfinding approach
- Layout -- grid structure, content regions, spatial arrangement
- Content hierarchy -- information priority, reading order
- Component design -- buttons, forms, cards, modals, tables
- Elevation/shape -- shadows, surface layering, border radius
- Typography -- font sizes, line heights, reading comfort
- Color system -- palettes, contrast ratios, dark mode
- Density/spacing -- spacing scale, touch targets, whitespace
- Content voice -- microcopy, error messages, terminology
State in your critique: "Auditing the locked design: [2-3 sentence summary of the accumulated design direction]."
Detect Existing Feedback
Check for feedback-round-{N}.json in .design-crit/facets/accessibility/. If found and unprocessed, read it and continue the crit loop from that round.
2. Audit Categories
The accessibility audit covers six categories. For each, run a systematic check against the locked design and report findings.
Category 1: Keyboard Navigation
Audit the locked navigation, layout, and component design for keyboard usability.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Tab order | Does logical reading order match DOM order in the locked layout? | WCAG 2.4.3 | |
| Focus indicators | Are focus rings visible against all locked surface colors? | WCAG 2.4.7 | |
| Skip links | Can users skip repetitive navigation to reach main content? | WCAG 2.4.1 | |
| Keyboard traps | Can any locked component (modal, dropdown, menu) trap focus? | WCAG 2.1.2 | |
| Keyboard shortcuts | Are critical actions reachable without a mouse? | Best practice | |
| Focus management | Do modals, dialogs, and drawers manage focus correctly? | WCAG 2.4.3 | |
| Interactive elements | Can every button, link, and form field be activated via keyboard? | WCAG 2.1.1 |
Findings format: For each issue, state the specific element (from the locked wireframe), the problem, and a remediation recommendation.
Category 2: Screen Reader Support
Audit the locked layout, components, and content voice for screen reader compatibility.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Semantic structure | Do headings follow a logical hierarchy (h1 > h2 > h3)? | WCAG 1.3.1 | |
| ARIA landmarks | Are main regions identified (nav, main, aside, footer)? | WCAG 1.3.1 | |
| ARIA labels | Do interactive elements without visible text have aria-label? | WCAG 4.1.2 | |
| Live regions | Do dynamic updates (toasts, status changes) announce themselves? | WCAG 4.1.3 | |
| Alt text | Do images, icons, and illustrations have meaningful alt text? | WCAG 1.1.1 | |
| Form labels | Is every form input associated with a visible label? | WCAG 1.3.1 | |
| Table headers | Do data tables have proper th elements and scope attributes? | WCAG 1.3.1 | |
| Reading order | Does the DOM order match the visual reading order? | WCAG 1.3.2 |
Findings format: For each issue, state the element, the screen reader experience (what a user would hear or miss), and the remediation.
Category 3: Touch Targets
Audit the locked density/spacing and component design for touch accessibility.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Minimum target size | Are all interactive elements at least 44x44px (iOS) or 48x48dp (Android)? | WCAG 2.5.8 / Platform HIG | |
| Target spacing | Is there at least 8px between adjacent touch targets? | Best practice | |
| Hover-only patterns | Are there interactions only accessible via mouse hover? | WCAG 2.1.1 | |
| Gesture alternatives | Do swipe/pinch/drag interactions have button alternatives? | WCAG 2.5.1 |
Reference the locked density facet's touch target annotations. If the density facet already flagged issues, confirm they are addressed or flag them as unresolved.
Category 4: Color and Contrast
Audit the locked color system for contrast compliance and color independence.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Text contrast | All text/background pairs meet 4.5:1 (normal) or 3:1 (large) | WCAG 1.4.3 | |
| UI contrast | Buttons, inputs, icons meet 3:1 against adjacent colors | WCAG 1.4.11 | |
| Focus indicator contrast | Focus rings meet 3:1 against both the element and background | WCAG 2.4.11 | |
| Color independence | Is information conveyed by color alone anywhere? | WCAG 1.4.1 | |
| Color blindness | Do status colors (red/green) have non-color indicators? | Best practice | |
| Dark mode contrast | Do all pairs pass in dark mode too? | WCAG 1.4.3 |
Reference the locked color system's contrast audit. If the color facet already ran contrast checks, verify their findings and flag any that were missed or borderline.
Color-blindness simulation: Describe how the locked palette looks under protanopia (red-blind), deuteranopia (green-blind), and tritanopia (blue-blind). Note any status indicators that become ambiguous.
Category 5: Motion and Animation
Audit any motion or transition patterns in locked decisions.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Reduced motion | Is there a mechanism to reduce or disable animations? | WCAG 2.3.3 | |
| Vestibular safety | Are there parallax, zoom, or spinning animations? | WCAG 2.3.3 | |
| Auto-playing content | Does any content auto-play, auto-scroll, or auto-advance? | WCAG 2.2.2 | |
| Flashing content | Does anything flash more than 3 times per second? | WCAG 2.3.1 |
If the motion facet was included and locked, reference its decisions. If motion was not a facet, audit any implicit motion in locked components (dropdown animations, page transitions, loading spinners).
Category 6: Cognitive Accessibility
Audit the locked content voice, layout, and navigation for cognitive load.
Check these items:
| Item | What to Check | Standard | |
|---|---|---|---|
| Reading level | Is microcopy written at an appropriate reading level for the audience? | WCAG 3.1.5 | |
| Consistent navigation | Are navigation patterns consistent across all screens? | WCAG 3.2.3 | |
| Consistent identification | Are UI elements labeled consistently (same term everywhere)? | WCAG 3.2.4 | |
| Error prevention | Do destructive actions have confirmation steps? | WCAG 3.3.4 | |
| Clear wayfinding | Can users always tell where they are and how to get back? | WCAG 2.4.8 | |
| Information chunking | Is content grouped into scannable chunks, not walls of text? | Best practice | |
| Predictable behavior | Do UI elements behave as expected from their appearance? | WCAG 3.2.1 |
Reference the locked content voice's terminology glossary. Flag any term inconsistencies found across locked wireframes.
3. Audit Report Format
Generate the audit as an HTML document at .design-crit/facets/accessibility/audit.html. This report is the foundation for the options in this facet.
Report Structure
Severity Classification
| Severity | Definition | Examples | |
|---|---|---|---|
| Critical | Prevents a user from completing a core task | No keyboard access to primary action, form without labels | |
| Major | Degrades experience significantly for some users | Missing skip link, low contrast on body text, no focus indicators | |
| Minor | Suboptimal but usable | Slightly below contrast ratio on secondary text, non-ideal reading order |
4. Generating Options
Unlike other facets where options represent competing design directions, accessibility options represent different remediation strategies. The audit findings are the same; the options differ in HOW to fix them.
Round 1: Audit + Remediation Options
Generate the audit report (section 3). Then present 2-3 remediation options.
Example option set:
- Minimum Compliance -- fixes all critical and major issues to achieve WCAG AA. Minimal
visual disruption to locked design. Pragmatic approach that ships accessible. Trade-off: minor issues deferred; does not reach AAA.
- Enhanced Access -- fixes all issues (critical, major, minor) and adds quality-of-life
improvements: skip links, landmark regions, keyboard shortcut hints. More invasive changes to locked wireframes. Trade-off: more visual changes; some locked design details may shift.
- Inclusive by Default -- fixes all issues AND proactively adds accessibility features
not flagged in the audit: reduced motion mode, high contrast mode, screen reader optimized flows. Exceeds WCAG AA in places. Trade-off: significant implementation effort; some additions may feel over-engineered for MVP.
Each option HTML must include:
- The full audit report at the top (shared across options)
- Annotated wireframes showing the locked design WITH remediation applied. Use the locked
layout/color/typography/density wireframes as the canvas and layer remediation visually:
- Added focus indicators (colored rings around focusable elements)
- Corrected contrast (show before/after for changed color pairs)
- Touch target adjustments (show expanded targets)
- Added ARIA landmark annotations
- Skip link placement
- Any structural changes to locked wireframes
- A remediation checklist: every audit finding with a checkbox, grouped by severity,
showing which are addressed in this option
- Implementation notes: specific HTML attributes, ARIA roles, and CSS changes needed
Round 2+: Converge
Refine based on feedback. Common refinements:
- Include specific fixes from one option in another
- Adjust the visual impact of a remediation (e.g., subtler focus indicators)
- Prioritize specific audit findings the user cares about
- Address user concerns about visual disruption to locked designs
5. Evaluation Criteria
Score every option against these questions in your critique.
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/crit-accessibility/SKILL.md
Popularity
- Repository stars
- 3
- 7-day trend score
- 3
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 31 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/metedata/design-crit/crit-accessibility)