agentsclimarketplace

Accessibility review

Skill Amey-Thakur/AI-SKILLS/skills/accessibility/accessibility-review

Audit a UI for real accessibility, keyboard-first, with concrete failures and fixes rather than checkbox compliance. Use when reviewing or building any user interface.From its SKILL.md

Install
npx -y skills add Amey-Thakur/AI-SKILLS --skill accessibility-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.5 KB, 570 tokens by cl100k_base, as published. Nobody here has run it

Accessibility review

Accessibility is whether real people can use the thing: with a keyboard, a screen reader, low vision, or a tremor: not whether a linter is quiet. Audit by use, report by failure.

Method

  1. Walk the keyboard path first. Unplug the mouse mentally and complete the page's primary task with Tab, Shift+Tab, Enter, Space, Escape, and arrows. Every failure is a finding:
    • Focus that vanishes (no visible indicator) or lands somewhere absurd.
    • Interactive things you cannot reach (div onclick without tabindex and key handling: a native <button> fixes all of it at once).
    • Traps: modals you cannot Escape, menus that swallow Tab.
    • Order that jumps around the visual layout.
  2. Read the page as the accessibility tree. Every control needs an accessible name (visible label, aria-label, or alt); every input a programmatic label (for/id or wrapping); images alt text that says what matters (alt="" for decoration); headings a real hierarchy (one h1, no skipped levels: structure, not styling).
  3. Check state and change announcements. Toggles carry aria-pressed or aria-expanded; errors are associated with their field (aria-describedby), not floating red text; async updates that matter (saved, failed, loading) reach aria-live regions; the modal traps focus in and returns it on close.
  4. Check the visual floor. Text contrast 4.5:1 (3:1 for large text and UI parts); information never carried by color alone (the red/green dot needs a label); layout survives 200% zoom; animation respects prefers-reduced-motion; touch targets are not miniature.
  5. Report by task, ranked by exclusion. "A keyboard user cannot send a message" is a blocker; "the decorative divider is announced" is minor. Each finding: who is excluded, where, the failing behavior, the concrete fix: which is usually use the native element.

Rules

  • Native elements first (button, a, label, select, dialog); ARIA is the patch, not the plan. Wrong ARIA is worse than none.
  • Test the actual behaviors, not the attribute presence: aria-label="" passes a grep and fails a person.
  • Accessibility notes belong in the same review as everything else, at the same severity scale. A blocker is a blocker.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most review quality skills give in 570 tokens

Counted across 1,273 of the 2,403 authors here whose files we hold, read 2026-09-06

  • Ask one question at a timein 63 of 1273, across 62 files
  • Provide a recommended answer for each questionin 47 of 1273, across 45 files
  • Rank findings by severityin 44 of 1273
  • Use parameterized queries for database accessin 38 of 1273, across 20 files
  • Validate all user input with schemasin 33 of 1273, across 15 files
  • Store secrets in environment variablesin 32 of 1273, across 14 files
  • Explore the codebase to answer questionsin 31 of 1273, across 29 files
  • Store tokens in httpOnly cookiesin 30 of 1273, across 12 files
  • Implement rate limiting on API endpointsin 30 of 1273, across 12 files
  • Sanitize user-provided HTMLin 29 of 1273, across 11 files
  • Return generic error messages to usersin 28 of 1273, across 10 files
  • Cite file and line for every findingin 28 of 1273, across 25 files

Said here and by no other author read

  • Navigate the interface using only keyboard controls
  • Verify every control has an accessible name
  • Ensure headings follow a logical hierarchy
  • Check state changes use appropriate ARIA attributes
  • Validate text contrast meets minimum ratios
  • Ensure information is not conveyed by color alone

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.