跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
设计与 UI

dev-accessibility-audit

Use to audit web/app accessibility, including semantics, keyboard flow, focus, labels, ARIA, contrast, and responsive behavior.

BlueSkyXN · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add BlueSkyXN/Codex-is-all-you-need --skill dev-accessibility-audit

技能说明


name: dev-accessibility-audit description: Use to audit web/app accessibility, including semantics, keyboard flow, focus, labels, ARIA, contrast, and responsive behavior. metadata: version: "0.3" updated: "2026-06-12"

Accessibility audit workflow

Use this workflow when a frontend change or existing UI needs accessibility review.

Steps

  1. Define the scope.

    • Pages, routes, components, dialogs, forms, tables, navigation, or flows.
    • Target platforms and any stated compliance level.
    • Known user complaints or previous violations.
  2. Collect evidence.

    • Source components and templates.
    • Existing tests, Storybook stories, screenshots, or browser states.
    • Design tokens for color, spacing, focus, and typography.
  3. Check core interaction requirements.

    • Semantic HTML and landmark structure.
    • Keyboard-only navigation and visible focus.
    • Logical tab order and focus trapping for modals.
    • Labels, descriptions, errors, and required field announcements.
    • Button, link, menu, tab, combobox, table, and dialog roles.
  4. Check perception and responsiveness.

    • Color contrast and non-color cues.
    • Text wrapping and zoom behavior.
    • Hit target size for touch interfaces.
    • Motion, animation, reduced-motion support, and loading states.
  5. Validate with available tools.

    • Run lint, tests, browser inspection, accessibility tree snapshots, axe-like checks, or manual keyboard smoke tests when available.
    • Record what was not tested when assistive technology is unavailable.
  6. Prioritize remediation.

    • Blocking issues that prevent task completion.
    • High-risk WCAG or keyboard failures.
    • Usability improvements and regression tests.

Output

Return:

  1. Scope reviewed
  2. Blocking accessibility findings
  3. Important non-blocking findings
  4. Recommended fixes
  5. Validation performed
  6. Residual risk and untested assistive technology

Do not

  • Do not claim formal WCAG compliance without the required test evidence.
  • Do not rely only on automated checks.
  • Do not add ARIA when semantic HTML already solves the problem.
  • Do not ignore keyboard and focus behavior.