Accessibility Testing logo

Accessibility Testing

Community
seb1n
accessibility-testing

Audit web interfaces against WCAG 2.1 AA/AAA standards, identify violations, and produce actionable remediation reports with code fixes. Use when the user requests accessibility testing or provides relevant inputs for this workflow.

Overview

Publisherseb1n
Repositoryawesome-ai-agent-skills
Skill nameaccessibility-testing
Stars
188
Forks
35
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by seb1n on GitHub. Read the source before you install it.

Installation

Install the Accessibility Testing AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/seb1n/awesome-ai-agent-skills.git /tmp/awesome-ai-agent-skills
mkdir -p .claude/skills
cp -r /tmp/awesome-ai-agent-skills/design-and-ui-ux/accessibility-testing .claude/skills/accessibility-testing
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Accessibility Testing in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Accessibility Testing on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Accessibility Testing is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Accessibility Testing

This skill enables the agent to perform thorough accessibility audits of web pages and components against the Web Content Accessibility Guidelines (WCAG) 2.1 at AA and AAA conformance levels. The agent identifies violations across four principles — Perceivable, Operable, Understandable, Robust — and generates structured compliance reports with specific code fixes. It covers automated checks (color contrast, missing alt text, ARIA misuse), semi-automated checks (keyboard navigation flows, focus management), and manual check guidance (screen reader announcements, cognitive load).

Workflow

  1. Define Audit Scope and Conformance Target: Determine which pages, components, or user flows to audit, and whether the target is WCAG 2.1 AA (most common legal requirement) or AAA (highest conformance). Identify the assistive technologies to consider: screen readers (NVDA on Windows, VoiceOver on macOS/iOS, TalkBack on Android), keyboard-only navigation, and magnification tools.

  2. Run Automated Scans: Execute automated accessibility analysis using axe-core or Lighthouse. These tools detect approximately 30-40% of WCAG violations automatically, including missing alt attributes, insufficient color contrast ratios, missing form labels, duplicate IDs, invalid ARIA roles, and missing document language. Capture each violation with its WCAG criterion reference (e.g., 1.4.3 Contrast Minimum), severity (critical, serious, moderate, minor), the affected HTML element, and the CSS selector path.

  3. Perform Keyboard Navigation Testing: Manually trace every interactive flow using only Tab, Shift+Tab, Enter, Space, Escape, and arrow keys. Verify that focus order matches visual order, that focus indicators are visible (minimum 2px outline with 3:1 contrast), that modal dialogs trap focus correctly, and that no keyboard traps exist. Document any flow where a keyboard user cannot complete a task that a mouse user can.

  4. Test with Screen Readers: Navigate the page with at least one screen reader to verify that all content is announced correctly. Confirm that images have meaningful alt text (or alt="" for decorative images), that headings form a logical hierarchy (no skipped levels), that landmarks (<nav>, <main>, <aside>) are present, and that dynamic content updates are announced via aria-live regions.

  5. Generate the Compliance Report: Produce a structured report organized by WCAG principle and criterion. For each violation, include: the criterion number and name, the severity level, the affected element's HTML snippet, an explanation of why it fails, and the specific code fix. Include a summary table with pass/fail counts per principle and an overall conformance statement.

  6. Provide Remediation Code: For every violation, supply a concrete before-and-after code snippet showing the exact fix. Group fixes by component or page region so developers can address them systematically rather than one element at a time.

Usage

Provide the agent with a URL, a block of HTML, or a React/Vue component. Specify the target conformance level (AA or AAA) and any assistive technology requirements. The agent will audit the content and return a structured report with code fixes.

Examples

Example 1: WCAG Compliance Audit Report

User Request: "Audit this login form for WCAG 2.1 AA compliance."

html
<div class="login">
  <img src="logo.png">
  <input type="text" placeholder="Username">
  <input type="password" placeholder="Password">
  <div class="btn" onclick="login()">Sign In</div>
  <p style="color: #aaa; font-size: 12px;">Forgot your password?</p>
</div>

Agent Response — Accessibility Audit Report:

#CriterionSeverityElementIssue
11.1.1 Non-text ContentCritical<img src="logo.png">Image has no alt attribute. Screen readers cannot describe it.
21.3.1 Info and RelationshipsCritical<input placeholder="Username">Input has no associated <label>. Placeholder text is not a substitute for a label.
31.3.1 Info and RelationshipsCritical<input placeholder="Password">Same label issue as above.
41.4.3 Contrast MinimumSerious<p style="color: #aaa">Text color #aaa on white background has a contrast ratio of 2.32:1. AA requires 4.5:1 for text this size.
52.1.1 KeyboardCritical<div class="btn" onclick>A <div> is not focusable via keyboard. Users cannot Tab to or activate this button.
64.1.2 Name, Role, ValueCritical<div class="btn" onclick>Element has no semantic role. Screen readers announce it as generic text, not as a button.

Remediation — Fixed Code:

html
<form class="login" onsubmit="login(event)">
  <img src="logo.png" alt="Acme Corp logo">

  <label for="username">Username</label>
  <input id="username" type="text" autocomplete="username" required>

  <label for="password">Password</label>
  <input id="password" type="password" autocomplete="current-password" required>

  <button type="submit" class="btn">Sign In</button>

  <a href="/forgot-password" style="color: #595959; font-size: 14px;">
    Forgot your password?
  </a>
</form>

Changes made: added alt to the image, replaced placeholder-only inputs with explicit <label> elements, replaced the <div> with a semantic <button>, changed the link color from #aaa (2.32:1) to #595959 (7.0:1), wrapped the form in a <form> element for proper submit handling, and added autocomplete attributes.

Example 2: Fixing a React Component

User Request: "Fix accessibility issues in this React notification component."

Before (inaccessible):

jsx
function Notification({ message, onClose }) {
  return (
    <div className="notification">
      <span>{message}</span>
      <span className="close-x" onClick={onClose}></span>
    </div>
  );
}

After (accessible):

jsx
function Notification({ message, onClose }) {
  return (
    <div role="alert" aria-live="assertive" className="notification">
      <p>{message}</p>
      <button
        type="button"
        onClick={onClose}
        aria-label="Dismiss notification"
        className="close-btn"
      >
      </button>
    </div>
  );
}

Fixes applied: (1) Added role="alert" and aria-live="assertive" so screen readers announce the notification immediately when it appears. (2) Replaced the <span onClick> with a <button> so it is keyboard-focusable and announced as an interactive control. (3) Added aria-label="Dismiss notification" because the "✕" character alone does not convey the button's purpose to screen reader users. (4) Changed the inner <span> to a <p> for proper text semantics.

Best Practices

  • Run automated tools first, but never rely on them alone: axe-core catches roughly 30-40% of WCAG issues. The remaining 60-70% require manual keyboard testing, screen reader verification, and cognitive review.
  • Test with real screen readers, not just ARIA validators: An element may have correct ARIA markup but still produce confusing announcements. VoiceOver and NVDA sometimes interpret the same markup differently.
  • Fix critical and serious issues before moderate ones: Prioritize violations that completely block access (missing keyboard operability, no alt text on functional images) over cosmetic issues (minor contrast shortfalls on decorative elements).
  • Use semantic HTML before reaching for ARIA: A <button> needs no role="button". A <nav> needs no role="navigation". ARIA is a repair tool for situations where semantic HTML is not sufficient, not a replacement for it.
  • Include accessibility checks in CI pipelines: Run axe-core or pa11y in automated tests so new violations are caught before they reach production. Fail the build on critical violations.
  • Document accessibility decisions: When a component intentionally deviates from a guideline (e.g., a custom combobox pattern), document the rationale and the alternative approach used to maintain equivalent access.

Edge Cases

  • Single-page applications with client-side routing: Page navigation does not trigger a browser page load, so screen readers are not notified of the new content. Use an aria-live="polite" region to announce route changes, or programmatically move focus to the new page's <h1>.
  • Dynamic content loaded after initial render: Content injected via JavaScript after page load is invisible to screen readers unless wrapped in an aria-live region or focus is explicitly managed. For toast notifications use aria-live="assertive"; for feed updates use aria-live="polite".
  • Complex data tables: Tables with merged cells, nested headers, or sortable columns require explicit scope, headers, and aria-sort attributes. Test that a screen reader user can understand which header applies to each data cell.
  • Custom interactive widgets (sliders, date pickers, comboboxes): These have no native HTML equivalent. Follow the WAI-ARIA Authoring Practices 1.2 patterns exactly, implementing the full keyboard interaction model specified for each widget type.
  • Third-party embedded content (iframes, widgets): You cannot fix accessibility inside a third-party iframe. Document the issue, add a descriptive title attribute to the <iframe>, and provide an accessible alternative when the embedded content is critical to the user flow.

Frequently asked questions

What does the Accessibility Testing AI skill do?

Audit web interfaces against WCAG 2.1 AA/AAA standards, identify violations, and produce actionable remediation reports with code fixes. Use when the user requests accessibility testing or provides relevant inputs for this workflow.

Why use Accessibility Testing on TypingMind?

Because you install it once and use it with any model. Accessibility Testing is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Accessibility Testing in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/seb1n/awesome-ai-agent-skills/tree/main/design-and-ui-ux/accessibility-testing. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Accessibility Testing?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Accessibility Testing?

As many as you like. As long as a model supports skills, you can use Accessibility Testing with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Accessibility Testing AI skill free?

Yes. It is published on GitHub by seb1n under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇