Skip to content
SkillModel-led2,372 tokens in use

Better Accessibility

Reviews and fixes keyboard and focus behavior, ARIA, accessible names, forms, announcements, motion and zoom, citing the WCAG 2.2 criterion each failure breaks.

12,497 stars on the repository MIT Licensejakubkrehel/skills · skills/better-accessibility
npx skills add jakubkrehel/skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Better Accessibility costs

35tokens idle
2,372tokens in use
8,412tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure8/25
  • Examples0/15
  • Output15/15
  • Size15/15
  • Resources13/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity96
  • Maintenance52
  • Community48
  • Maturity24
  • Momentum40

Better Accessibility is an agent skill that reviews a UI in two walks: keyboard-only, where every flow must finish without a mouse, and screen-reader, where every control announces a name, a role and its state. Each failure is reported against the WCAG 2.2 Level A or AA criterion it breaks, and the fix is written in the project’s own styling system.

It favors native elements over custom ones, and removing ARIA over adding it. Contrast measurement, text sizing, spacing, copy and animation recipes are handed to the sibling skills that own them.

What You Get

  • A review that groups findings under the principle each one violates, ordered by severity, with the file and line, the current code, the fix and the WCAG criterion
  • Fixes for focus rings, focus order, dialogs, route changes and roving tabindex, written in the project’s styling system
  • Form guidance covering labels, autocomplete tokens, inputmode, inline errors tied to fields with aria-describedby and focus on the first invalid field
  • Checks for live regions, alt text by purpose, landmarks, skip links, reduced motion, pause controls, 200% text resize and 320px reflow
  • A Block or Approve verdict, with every check it could not run listed as Not verified

Common Use Cases

  • Auditing a new modal, menu or combobox for keyboard support and focus return before it ships
  • Fixing a form whose errors are not announced and whose submit button is disabled until it is valid
  • Checking an icon-only toolbar for accessible names and hit areas
  • Reviewing a page for reduced-motion handling, zoom behavior and viewport meta settings

What Makes It Unique

The skill separates a failure from a preference. Only a WCAG 2.2 A or AA criterion, a task a user cannot finish or a listed trigger can reach HIGH, and heuristics such as 44px targets or the one-h1 rule stay recommendations. It also names the sibling skills that own contrast, type, layout, copy and animation, so each rule has one home.

Under The Hood

The skill is a SKILL.md with seven reference files for focus and keyboard, forms, hit areas, motion and zoom, screen readers, semantics and ARIA, and reporting. It contains no scripts. The agent reads your source and, when a browser is available, tabs through the flow, reads computed names and roles from the accessibility tree and runs an automated audit. A review leaves your files untouched, and edits happen only when you ask for the fixes to be applied.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers