Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Deep visual accessibility QA: OKLCH-based perceptual color contrast, typography readability at all viewports, focus-visible completeness, WCAG AA/AAA audit at rendered level. Built on team-worker agent architecture -- all worker roles share a single agent definition with role-specific Phase 2-4 loaded from roles/<role>/role.md.
Architecture
Skill(skill="team-visual-a11y", args="task description")
|
SKILL.md (this file) = Router
|
+--------------+--------------+
| |
no --role flag --role <name>
| |
Coordinator Worker
roles/coordinator/role.md roles/<name>/role.md
|
+-- analyze -> dispatch -> spawn workers -> STOP
|
+-------+-------+-------+
v v v |
[3 auditors spawn in PARALLEL] |
color-auditor typo-auditor focus-auditor
| | |
+---+---+---+---+
v
remediation-planner
|
v
fix-implementer (inner_loop)
|
v
[re-audit: color + focus in PARALLEL]
After spawning, use wait_agent({ timeout_ms: 1800000 }) to collect results. If result.timed_out, send STATUS_CHECK via followup_task (wait 3 min), then FINALIZE with interrupt (wait 3 min), then mark timed_out and close agents. Use close_agent({ target }) each worker.
Model Selection Guide
Visual accessibility is a precision pipeline where auditors need thorough analysis and fix-implementer needs careful code changes.
Role
reasoning_effort
Rationale
color-auditor
high
OKLCH calculations, contrast ratio precision
typo-auditor
high
Multi-breakpoint analysis, clamp() validation
focus-auditor
high
ARIA patterns, keyboard navigation completeness
remediation-planner
high
Synthesize 3 audit reports into actionable plan
fix-implementer
medium
Implementation follows defined remediation plan
Audit-to-Remediation Context Flow
All 3 audit findings must reach remediation-planner via coordinator's upstream context:
// After COLOR-001 + TYPO-001 + FOCUS-001 all complete, coordinator sends findings to planner
spawn_agent({
agent_type: "team_worker",
task_name: "REMED-001",
fork_turns: "none",
message: `## Upstream Context
Color audit: <session>/audits/color/color-audit-001.md
Typography audit: <session>/audits/typography/typo-audit-001.md
Focus audit: <session>/audits/focus/focus-audit-001.md`
})
Send audit findings to running remediation-planner
Assign new work from reviewed plan
followup_task
Assign FIX task after remediation plan ready
Check running agents
list_agents
Verify agent health during resume
Agent Health Check
Use list_agents({}) in handleResume and handleComplete:
// Reconcile session state with actual running agents
const running = list_agents({})
// Compare with tasks.json active tasks
// Reset orphaned tasks (in_progress but agent gone) to pending
Named Agent Targeting
Workers are spawned with task_name: "<task-id>" enabling direct addressing:
