name	a11y-audit
description	Audit a web page/component against WCAG 2.2 AA — semantics, keyboard, contrast, labels, focus, alt text. Use when reviewing UI for accessibility or before shipping a page.

a11y-audit

Check against WCAG 2.2 AA. Report findings as must-fix vs. should-fix.

Semantics: real elements — <button> not a clickable <div>; <nav>/<main>/<header> landmarks; lists are lists.
Keyboard: every interactive element reachable and operable by Tab/Enter/Space; logical order; no traps.
Focus: visible focus indicator (don't outline: none without a replacement).
Labels: every input has a <label>/aria-label; icon-only buttons have accessible names.
Images: meaningful alt; decorative images alt="".
Contrast: text ≥ 4.5:1 (≥ 3:1 for large text). State the measured ratio, don't eyeball it.
Forms: errors announced (aria-live), associated with their field, not color-only.
Motion: respect prefers-reduced-motion.

Findings grouped must-fix / should-fix, each with the exact element and the fix.
Don't claim "accessible" from reading code alone — say what was checked statically vs. needs a real screen-reader/keyboard pass.

name	a11y-audit
description	Audit a web page/component against WCAG 2.2 AA — semantics, keyboard, contrast, labels, focus, alt text. Use when reviewing UI for accessibility or before shipping a page.

Check against WCAG 2.2 AA. Report findings as must-fix vs. should-fix.

Semantics: real elements — <button> not a clickable <div>; <nav>/<main>/<header> landmarks; lists are lists.
Keyboard: every interactive element reachable and operable by Tab/Enter/Space; logical order; no traps.
Focus: visible focus indicator (don't outline: none without a replacement).
Labels: every input has a <label>/aria-label; icon-only buttons have accessible names.
Images: meaningful alt; decorative images alt="".
Contrast: text ≥ 4.5:1 (≥ 3:1 for large text). State the measured ratio, don't eyeball it.
Forms: errors announced (aria-live), associated with their field, not color-only.
Motion: respect prefers-reduced-motion.

Findings grouped must-fix / should-fix, each with the exact element and the fix.
Don't claim "accessible" from reading code alone — say what was checked statically vs. needs a real screen-reader/keyboard pass.

Mais deste repositório