Harness Kit

skill

a11y

Accessibility audit, remediation, and verification. WCAG 2.2 AA compliance. Three-agent protocol: audit (find issues) → remediate (fix them) → critique (verify fixes). Use when: "accessibility audit", "a11y", "WCAG", "screen reader", "keyboard navigation", "contrast check", "aria fix", "accessibility sprint", "audit accessibility", "fix accessibility", "a11y issues", "a11y check". Trigger: /a11y

/a11y

Source: skills/a11y/SKILL.md

What it does

Accessibility audit, remediation, and verification. WCAG 2.2 AA compliance. Three-agent protocol: audit (find issues) → remediate (fix them) → critique (verify fixes). Use when: "accessibility audit", "a11y", "WCAG", "screen reader", "keyboard navigation", "contrast check", "aria fix", "accessibility sprint", "audit accessibility", "fix accessibility", "a11y issues", "a11y check". Trigger: /a11y

Workflow role

Support primitive in the Harness Kit operating loop.

Source contract preview

This generated excerpt gives readers the beginning of the live primitive contract before they jump to GitHub.

Audit, fix, and verify accessibility. WCAG 2.2 AA.

**Target:** $ARGUMENTS

## Execution Stance

You are the executive orchestrator.
- Keep severity decisions, scope tradeoffs, and final PASS/FAIL judgment on the lead model.
- Delegate audit, remediation, and critique to separate focused subagents.
- Prefer parallel exploration for independent checks; keep remediation sequential when fixes interact.

## Delegation Floor

Delegation floor applies: probe the roster first; dispatch two or more
providers for substantive work; direct solo only for mechanical, emergency,
user-forbidden, or fewer-than-two-providers cases. See
`harnesses/shared/AGENTS.md` (Roster).

Local lane guidance: Use specialized lanes for accessibility auditor,
remediation builder, keyboard/screen-reader verifier, and cold WCAG critic.
Prefer different providers for audit and critique when available; native
in-thread subagents may supplement but do not satisfy the roster floor.

## Routing

| Intent | Action |
|--------|--------|
| `/a11y` or `/a11y audit` | Full triad: audit → remediate → critique |
| `/a11y audit <target>` | Audit only — read `references/audit.md` |
| `/a11y fix <target>` | Remediate only — read `references/remediate.md` |
| `/a11y verify` | Critique recent changes — read `references/critique.md` |

If no sub-command, run the full triad below.

## Three-Agent Protocol

### Phase 1: Audit (read-only)

Launch an Explore subagent to find issues. It does NOT fix anything.

**Automated scan:**
- If a dev server is running, scan with Playwright + axe-core
  (`wcag2a`, `wcag2aa`, `wcag22aa` tags)
- If no server, scan component source statically

**Static analysis — grep for anti-patterns:**
- `div onClick` or `span onClick` without `role` + `tabIndex` + `onKeyDown`
- `<img` without `alt=`
- `<table>` without `<caption>` or `aria-label`
...

What to verify