Claude-Code-Workflow team-visual-a11y

Unified team skill for visual accessibility QA. OKLCH color contrast, typography readability, focus management, WCAG AA/AAA audit at rendered level. Uses team-worker agent architecture. Triggers on "team visual a11y", "accessibility audit", "visual a11y".

install
source · Clone the upstream repo
git clone https://github.com/catlog22/Claude-Code-Workflow
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/catlog22/Claude-Code-Workflow "$T" && mkdir -p ~/.claude/skills && cp -r "$T/.codex/skills/team-visual-a11y" ~/.claude/skills/catlog22-claude-code-workflow-team-visual-a11y-d1f175 && rm -rf "$T"
manifest: .codex/skills/team-visual-a11y/SKILL.md
source content

Team Visual Accessibility

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]

Role Registry

RolePathPrefixInner Loop
coordinatorroles/coordinator/role.md----
color-auditorroles/color-auditor/role.mdCOLOR-*false
typo-auditorroles/typo-auditor/role.mdTYPO-*false
focus-auditorroles/focus-auditor/role.mdFOCUS-*false
remediation-plannerroles/remediation-planner/role.mdREMED-*false
fix-implementerroles/fix-implementer/role.mdFIX-*true

Role Router

Parse

$ARGUMENTS
:

  • Has
    --role <name>
    -> Read
    roles/<name>/role.md
    , execute Phase 2-4
  • No
    --role
    ->
    roles/coordinator/role.md
    , execute entry router

Delegation Lock

Coordinator is a PURE ORCHESTRATOR. It coordinates, it does NOT do.

Before calling ANY tool, apply this check:

Tool CallVerdictReason
spawn_agent
,
wait_agent
,
close_agent
,
send_message
,
followup_task
ALLOWEDOrchestration
list_agents
ALLOWEDAgent health check
request_user_input
ALLOWEDUser interaction
mcp__ccw-tools__team_msg
ALLOWEDMessage bus
Read/Write
on
.workflow/.team/
files
ALLOWEDSession state
Read
on
roles/
,
commands/
,
specs/
ALLOWEDLoading own instructions
Read/Grep/Glob
on project source code
BLOCKEDDelegate to worker
Edit
on any file outside
.workflow/
BLOCKEDDelegate to worker
Bash("ccw cli ...")
BLOCKEDOnly workers call CLI
Bash
running build/test/lint commands
BLOCKEDDelegate to worker

If a tool call is BLOCKED: STOP. Create a task, spawn a worker.

No exceptions for "simple" tasks. Even a single-file read-and-report MUST go through spawn_agent.


Shared Constants

  • Session prefix:
    VA
  • Session path:
    .workflow/.team/VA-<date>-<slug>/
  • team_name:
    visual-a11y
  • CLI tools:
    ccw cli --mode analysis
    (read-only),
    ccw cli --mode write
    (modifications)
  • Message bus:
    mcp__ccw-tools__team_msg(session_id=<session-id>, ...)
  • Max GC rounds: 2

Worker Spawn Template

Coordinator spawns workers using this template:

spawn_agent({
  agent_type: "team_worker",
  task_name: "<task-id>",
  fork_turns: "none",
  message: `## Role Assignment
role: <role>
role_spec: <skill_root>/roles/<role>/role.md
session: <session-folder>
session_id: <session-id>
requirement: <task-description>
inner_loop: <true|false>

Read role_spec file (<skill_root>/roles/<role>/role.md) to load Phase 2-4 domain instructions.

## Task Context
task_id: <task-id>
title: <task-title>
description: <task-description>
pipeline_phase: <pipeline-phase>

## Upstream Context
<prev_context>`
})

Parallel Fan-in Spawn (3 Auditors)

The 3 auditors run in parallel. Spawn all 3, then wait for all 3:

// Spawn 3 auditors in parallel
spawn_agent({
  agent_type: "team_worker",
  task_name: "COLOR-001",
  fork_turns: "none",
  message: `## Role Assignment
role: color-auditor
role_spec: ${skillRoot}/roles/color-auditor/role.md
session: ${sessionFolder}
session_id: ${sessionId}
requirement: ${colorTaskDescription}
inner_loop: false

Read role_spec file to load Phase 2-4 domain instructions.

## Task Context
task_id: COLOR-001
title: OKLCH Color Contrast Audit
description: ${colorTaskDescription}
pipeline_phase: audit`
})

spawn_agent({
  agent_type: "team_worker",
  task_name: "TYPO-001",
  fork_turns: "none",
  message: `## Role Assignment
role: typo-auditor
role_spec: ${skillRoot}/roles/typo-auditor/role.md
session: ${sessionFolder}
session_id: ${sessionId}
requirement: ${typoTaskDescription}
inner_loop: false

Read role_spec file to load Phase 2-4 domain instructions.

## Task Context
task_id: TYPO-001
title: Typography Readability Audit
description: ${typoTaskDescription}
pipeline_phase: audit`
})

spawn_agent({
  agent_type: "team_worker",
  task_name: "FOCUS-001",
  fork_turns: "none",
  message: `## Role Assignment
role: focus-auditor
role_spec: ${skillRoot}/roles/focus-auditor/role.md
session: ${sessionFolder}
session_id: ${sessionId}
requirement: ${focusTaskDescription}
inner_loop: false

Read role_spec file to load Phase 2-4 domain instructions.

## Task Context
task_id: FOCUS-001
title: Focus & Keyboard Accessibility Audit
description: ${focusTaskDescription}
pipeline_phase: audit`
})

// Wait for ALL 3 auditors to complete
wait_agent({ timeout_ms: 1800000 })  // 30 min — apply timeout cascade if timed_out

// Close all 3
close_agent({ target: "COLOR-001" })
close_agent({ target: "TYPO-001" })
close_agent({ target: "FOCUS-001" })

// Then spawn remediation-planner with all 3 audit results as upstream context

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.

Rolereasoning_effortRationale
color-auditorhighOKLCH calculations, contrast ratio precision
typo-auditorhighMulti-breakpoint analysis, clamp() validation
focus-auditorhighARIA patterns, keyboard navigation completeness
remediation-plannerhighSynthesize 3 audit reports into actionable plan
fix-implementermediumImplementation 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`
})

User Commands

CommandAction
check
/
status
View execution status graph
resume
/
continue
Advance to next step

Specs Reference

Session Directory

.workflow/.team/VA-<date>-<slug>/
+-- .msg/
|   +-- messages.jsonl         # Team message bus
|   +-- meta.json              # Pipeline config + GC state
+-- audits/
|   +-- color/                 # Color auditor output
|   |   +-- color-audit-001.md
|   +-- typography/            # Typography auditor output
|   |   +-- typo-audit-001.md
|   +-- focus/                 # Focus auditor output
|       +-- focus-audit-001.md
+-- remediation/               # Remediation planner output
|   +-- remediation-plan.md
+-- fixes/                     # Fix implementer output
|   +-- fix-summary-001.md
+-- re-audit/                  # Re-audit output (GC loop)
|   +-- color-audit-002.md
|   +-- focus-audit-002.md
+-- evidence/                  # Screenshots, traces

v4 Agent Coordination

Message Semantics

IntentAPIExample
Queue supplementary info (don't interrupt)
send_message
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:

  • send_message({ target: "REMED-001", message: "..." })
    -- send additional audit findings to remediation-planner
  • followup_task({ target: "FIX-001", message: "..." })
    -- assign implementation from remediation plan
  • close_agent({ target: "COLOR-001" })
    -- cleanup after color audit

Error Handling

ScenarioResolution
Unknown commandError with available command list
Role not foundError with role registry
Session corruptionAttempt recovery, fallback to manual
Fast-advance conflictCoordinator reconciles on next callback
Completion action failsDefault to Keep Active
GC loop stuck > 2 roundsEscalate to user: accept / retry / terminate
Chrome DevTools unavailableDegrade to static analysis only