- Home
- Skills
- UX & UI Design
- Accessibility Audit
More screenshots
Works with the AI Agents you already use
Accessibility Audit
WCAG 2.2 AA audit for any website. You get one shareable HTML report with every issue screenshotted, ranked by who it blocks (Critical, Serious, Polish), with a fix for each.
$15
Accessibility Audit
Example session with this skill installed
/accessibility-audit https://lumen.example.com, check the home page
- Read your context and instructions
- Compiled the accessibility audit
- Generated the document
Audit done. The report is open in your browser: audit-report.html (every screenshot built in, ready to send).
Summary
Critical: 3 · Serious: 4 · Polish: 2 against WCAG 2.2 A/AA
| Severity | Issue | WCAG |
|---|---|---|
| Critical | Focus outline removed everywhere, keyboard users can't see where they are | 2.4.7 |
| Critical | Pricing plans don't stack at phone width, Pro and Scale are off-screen | 1.4.10 |
| Critical | Search and account icon buttons have no name | 4.1.2 |
| Serious | "Watch the tour" button text at about 1.3:1 contrast | 1.4.3 |
| Serious | Hero subtitle and badge below 4.5:1 | 1.4.3 |
| Serious | Email field uses placeholder instead of a label | 3.3.2 |
| Serious | Footer social links are 16px, 4px apart, unlabelled | 2.5.8, 4.1.2 |
| Polish | Third feature card's text sits 6px further in | Visual |
| Polish | Feature tags don't line up across cards | Visual |
Fix first
Bring back a visible focus indicator. Delete the global outline: none or add :focus-visible { outline: 2px solid #7c5cff; outline-offset: 2px; }
2.
Let the pricing plans stack on phones. Use grid-template-columns: repeat(auto-fit, minmax(260px, 1fr)) and remove the min-width: 948px
3. Label the icon buttons. aria-label="Search" and aria-label="Your account"
Note: automated checks passed the email field because they count the placeholder as a name. A person wouldn't, so it's flagged here.
Not checked
- Screen reader behaviour (needs a real screen reader)
- 200% zoom, text spacing and orientation showed no failures in screenshots but need a human to confirm
Every finding in the report has a red-outlined screenshot, a verdict (Does Not Support / Partially Supports) and the exact fix.
audit-report.html
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Accessibility audits that read like a designer wrote them.
Point it at any URL and Claude runs a full WCAG 2.2 A/AA pass, then hands you one HTML report you can send to a client or developer as is. Every issue is screenshotted with the problem outlined in red, ranked by real user impact, and comes with a concrete fix.
What it checks
- Automated rules via axe-core (contrast, names, labels, alt text, structure)
- 200% zoom and reflow at 320px
- Text spacing (1.4.12) and orientation (1.3.4)
- Keyboard focus: visible, logical order, not hidden
- Target size using the real 2.5.8 spacing rule, not just "under 24px"
- Visual polish: alignment, spacing, inconsistent states
How it ranks
- Critical: someone is blocked
- Serious: usable but failing or painful
- Polish: works, but looks off
Plus standard VPAT verdicts (Supports, Partially Supports, Does Not Support).
Good for
- Client audits for freelancers and agencies
- Checking a site before launch or a VPAT
- Catching accessibility regressions on staging
- Giving developers screenshots of exactly what to fix
Honest by design It never marks a criterion as passing on its own. A human confirms. Screen reader testing is out of scope and flagged in every report.
Setup Drag the zip into Claude Code on your computer and say "install this skill into ~/.claude/skills and run its setup commands". Manual steps are in the README.
Needs Node.js 18+. The first run installs Playwright and axe-core and downloads a test browser (about 200MB). The tests run on your own machine.
Known limitations Some large retail and banking sites block automated browsers. Pages behind a login need a logged-in or staging URL.
Includes a sample report so you can see the output before you run it.
UX Skills Lab is independent and not affiliated with Anthropic.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for UX & UI Design.
Bot protection may block automated scans on some sites. Requires Node.js and Playwright. Visual checks for login-walled pages require a shareable URL or manual screenshot assistance.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 11 days ago
- Accounts
Anthropic
- Software
- Playwright
Node.js
Permissions
Allowed Hosts
Network: loads the website you're auditing in a headless browser. File write: saves screenshots and the HTML report to an output folder you choose. Shell: runs node and npm to install dependencies and run the audit scripts. No credentials are requested or stored.