claude-code-workflow 7.2.27 → 7.2.28

package/.codex/skills/team-visual-a11y/roles/coordinator/commands/dispatch.md

@@ -0,0 +1,188 @@
+# Command: Dispatch
+
+Create the visual accessibility task chain with correct dependencies and structured task descriptions. Supports audit-only and full pipeline modes.
+
+## Phase 2: Context Loading
+
+| Input | Source | Required |
+|-------|--------|----------|
+| User requirement | From coordinator Phase 1 | Yes |
+| Session folder | From coordinator Phase 2 | Yes |
+| Pipeline mode | From session tasks.json `pipeline_mode` | Yes |
+| Target info | From session tasks.json `target` | Yes |
+
+1. Load user requirement and scope from tasks.json
+2. Load pipeline stage definitions from specs/pipelines.md
+3. Read `pipeline_mode` and `target` from tasks.json
+
+## Phase 3: Task Chain Creation (Mode-Branched)
+
+### Task Description Template
+
+Every task is added to tasks.json with structured format:
+
+```json
+{
+  "<TASK-ID>": {
+    "title": "<task title>",
+    "description": "PURPOSE: <what this task achieves> | Success: <measurable completion criteria>\nTASK:\n  - <step 1: specific action>\n  - <step 2: specific action>\n  - <step 3: specific action>\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <URL or file path>\n  - WCAG Level: <AA|AAA>\n  - Upstream artifacts: <artifact-1>, <artifact-2>\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <deliverable path> + <quality criteria>\nCONSTRAINTS: <scope limits, focus areas>",
+    "role": "<role-name>",
+    "prefix": "<PREFIX>",
+    "deps": ["<dependency-list>"],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+### Mode Router
+
+| Mode | Action |
+|------|--------|
+| `audit-only` | Create 4 tasks: [COLOR-001 + TYPO-001 + FOCUS-001] parallel -> REMED-001 |
+| `full` | Create 7 tasks: [COLOR-001 + TYPO-001 + FOCUS-001] -> REMED-001 -> FIX-001 -> [COLOR-002 + FOCUS-002] |
+
+---
+
+### Audit-Only Pipeline Task Chain
+
+**COLOR-001** (color-auditor):
+```json
+{
+  "COLOR-001": {
+    "title": "OKLCH Color Contrast Audit",
+    "description": "PURPOSE: OKLCH color contrast audit for all color combinations | Success: Complete color audit with WCAG 2.1 + APCA contrast ratios and color blindness assessment\nTASK:\n  - Extract all color values (OKLCH, HSL, RGB, hex) from stylesheets\n  - Calculate WCAG 2.1 contrast ratios for all text/background combinations\n  - Calculate APCA Lc values for body and large text\n  - Assess OKLCH lightness ranges for text and backgrounds\n  - Simulate color blindness (protanopia, deuteranopia, tritanopia)\n  - Check dark mode color parity if applicable\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/audits/color/color-audit-001.md | Pass/fail per color combination\nCONSTRAINTS: Read-only analysis | Use Chrome DevTools for computed styles when available",
+    "role": "color-auditor",
+    "prefix": "COLOR",
+    "deps": [],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**TYPO-001** (typo-auditor):
+```json
+{
+  "TYPO-001": {
+    "title": "Typography Readability Audit",
+    "description": "PURPOSE: Typography readability audit across all viewports | Success: Complete typography audit with size, line-height, reading width at each breakpoint\nTASK:\n  - Audit font sizes at each breakpoint (320px, 768px, 1024px, 1400px)\n  - Validate minimum body text >= 16px on mobile\n  - Check line-height ratios (body 1.5-1.75, headings 1.1-1.3)\n  - Validate clamp() functions for responsive scaling\n  - Measure reading width (45-75 characters per line)\n  - Assess font loading strategy (font-display values)\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/audits/typography/typo-audit-001.md | Breakpoint-by-breakpoint report\nCONSTRAINTS: Read-only analysis | Screenshot at multiple viewports if Chrome DevTools available",
+    "role": "typo-auditor",
+    "prefix": "TYPO",
+    "deps": [],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**FOCUS-001** (focus-auditor):
+```json
+{
+  "FOCUS-001": {
+    "title": "Focus & Keyboard Accessibility Audit",
+    "description": "PURPOSE: Focus management and keyboard accessibility audit | Success: Complete focus audit with tab order, indicator visibility, ARIA coverage\nTASK:\n  - Audit tab order for logical sequence and no tab traps\n  - Check focus indicator visibility (outline >= 2px, contrast >= 3:1)\n  - Verify :focus-visible usage for keyboard vs mouse distinction\n  - Check skip link presence and functionality\n  - Audit focus traps in modals/dialogs (Tab cycles, Escape dismisses)\n  - Verify ARIA live regions for dynamic content\n  - Check ARIA roles and states on interactive elements\n  - Validate keyboard operability (Enter/Space, Arrow keys)\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/audits/focus/focus-audit-001.md | Element-by-element focus report\nCONSTRAINTS: Read-only analysis | Tab through elements if Chrome DevTools available",
+    "role": "focus-auditor",
+    "prefix": "FOCUS",
+    "deps": [],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**CRITICAL**: COLOR-001, TYPO-001, FOCUS-001 have NO deps -- they run in PARALLEL.
+
+**REMED-001** (remediation-planner):
+```json
+{
+  "REMED-001": {
+    "title": "Prioritized Remediation Plan",
+    "description": "PURPOSE: Synthesize all 3 audit findings into prioritized remediation plan | Success: Complete remediation plan with severity ranking, code-level fix guidance, WCAG criterion mapping\nTASK:\n  - Read color, typography, and focus audit reports\n  - Merge and deduplicate findings\n  - Prioritize by severity (Critical > High > Medium > Low)\n  - Group by file/component for efficient fixing\n  - Provide code-level fix guidance (specific CSS/HTML changes)\n  - Map each fix to WCAG success criterion\n  - Estimate effort per fix\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Upstream artifacts: audits/color/color-audit-001.md, audits/typography/typo-audit-001.md, audits/focus/focus-audit-001.md\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/remediation/remediation-plan.md | All critical/high issues addressed\nCONSTRAINTS: Read-only synthesis | No code modifications",
+    "role": "remediation-planner",
+    "prefix": "REMED",
+    "deps": ["COLOR-001", "TYPO-001", "FOCUS-001"],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+---
+
+### Full Pipeline Task Chain
+
+Same as audit-only (COLOR-001, TYPO-001, FOCUS-001, REMED-001), plus:
+
+**FIX-001** (fix-implementer):
+```json
+{
+  "FIX-001": {
+    "title": "Implement Accessibility Fixes",
+    "description": "PURPOSE: Implement accessibility fixes from remediation plan | Success: All critical and high severity issues fixed with passing contrast ratios and ARIA validation\nTASK:\n  - Read remediation plan for prioritized fix list\n  - Apply OKLCH color corrections for contrast compliance\n  - Add/fix focus styles (outline, outline-offset, :focus-visible)\n  - Add missing ARIA attributes (role, aria-label, aria-live, aria-expanded)\n  - Add reduced-motion media queries where needed\n  - Fix typography (clamp(), line-height, max-width for reading)\n  - Add skip link if missing\n  - Apply fixes in priority order (critical first)\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Upstream artifacts: remediation/remediation-plan.md\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: Modified source files + <session>/fixes/fix-summary-001.md | All critical/high fixes applied\nCONSTRAINTS: Modify only files identified in remediation plan | Preserve existing functionality",
+    "role": "fix-implementer",
+    "prefix": "FIX",
+    "deps": ["REMED-001"],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**COLOR-002** (color-auditor):
+```json
+{
+  "COLOR-002": {
+    "title": "Re-audit Color Contrast After Fixes",
+    "description": "PURPOSE: Re-audit color contrast after fixes applied | Success: All color combinations pass WCAG target level\nTASK:\n  - Re-extract all color values from modified stylesheets\n  - Re-calculate WCAG 2.1 contrast ratios\n  - Re-calculate APCA Lc values\n  - Compare before/after for each fixed combination\n  - Verify no regressions in unfixed combinations\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Upstream artifacts: fixes/fix-summary-001.md, audits/color/color-audit-001.md\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/re-audit/color-audit-002.md | Before/after comparison with pass/fail\nCONSTRAINTS: Read-only verification | Focus on fixed items + regression check",
+    "role": "color-auditor",
+    "prefix": "COLOR",
+    "deps": ["FIX-001"],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**FOCUS-002** (focus-auditor):
+```json
+{
+  "FOCUS-002": {
+    "title": "Re-audit Focus Management After Fixes",
+    "description": "PURPOSE: Re-audit focus management after fixes applied | Success: All focus indicators visible with correct ARIA attributes\nTASK:\n  - Re-audit tab order after DOM changes\n  - Verify focus indicator fixes (outline, contrast)\n  - Check new ARIA attributes are valid\n  - Verify skip link implementation\n  - Test keyboard operability of fixed elements\n  - Compare before/after for each fixed element\nCONTEXT:\n  - Session: <session-folder>\n  - Target: <target>\n  - WCAG Level: <level>\n  - Upstream artifacts: fixes/fix-summary-001.md, audits/focus/focus-audit-001.md\n  - Shared memory: <session>/.msg/meta.json\nEXPECTED: <session>/re-audit/focus-audit-002.md | Before/after comparison with pass/fail\nCONSTRAINTS: Read-only verification | Focus on fixed items + regression check",
+    "role": "focus-auditor",
+    "prefix": "FOCUS",
+    "deps": ["FIX-001"],
+    "status": "pending",
+    "findings": "",
+    "error": ""
+  }
+}
+```
+
+**CRITICAL**: COLOR-002 and FOCUS-002 both blocked only by FIX-001 -- they run in PARALLEL after fixes.
+
+---
+
+## Phase 4: Validation
+
+Verify task chain integrity:
+
+| Check | Method | Expected |
+|-------|--------|----------|
+| Task count correct | tasks.json count | audit-only: 4, full: 7 |
+| Dependencies correct | Trace dependency graph | Acyclic, correct deps |
+| No circular dependencies | Trace dependency graph | Acyclic |
+| 3 audit tasks have NO deps | Pattern check | COLOR-001, TYPO-001, FOCUS-001 are parallel |
+| REMED-001 blocked by all 3 | Check deps | [COLOR-001, TYPO-001, FOCUS-001] |
+| Task IDs use correct prefixes | Pattern check | COLOR/TYPO/FOCUS/REMED/FIX |
+| Structured descriptions complete | Each has PURPOSE/TASK/CONTEXT/EXPECTED/CONSTRAINTS | All present |
+
+If validation fails, fix the specific task and re-validate.

package/.codex/skills/team-visual-a11y/roles/coordinator/commands/monitor.md

@@ -0,0 +1,281 @@
+# Monitor Pipeline
+
+Synchronous pipeline coordination using spawn_agent + wait_agent.
+
+## Constants
+
+- WORKER_AGENT: team_worker
+- MAX_GC_ROUNDS: 2
+
+## Handler Router
+
+| Source | Handler |
+|--------|---------|
+| "capability_gap" | handleAdapt |
+| "check" or "status" | handleCheck |
+| "resume" or "continue" | handleResume |
+| All tasks completed | handleComplete |
+| Default | handleSpawnNext |
+
+## handleCallback
+
+Worker completed (wait_agent returns). Process and advance.
+
+1. Determine role from completed task prefix:
+
+| Task Prefix | Role |
+|-------------|------|
+| `COLOR-*` | color-auditor |
+| `TYPO-*` | typo-auditor |
+| `FOCUS-*` | focus-auditor |
+| `REMED-*` | remediation-planner |
+| `FIX-*` | fix-implementer |
+
+2. Mark task completed in tasks.json: `state.tasks[taskId].status = 'completed'`
+3. Record completion in session state
+
+4. Check checkpoint for completed task:
+
+| Completed Task | Checkpoint | Action |
+|---------------|------------|--------|
+| COLOR-001 | Parallel fan-in | Check if all 3 audits complete -> unblock REMED-001 |
+| TYPO-001 | Parallel fan-in | Check if all 3 audits complete -> unblock REMED-001 |
+| FOCUS-001 | Parallel fan-in | Check if all 3 audits complete -> unblock REMED-001 |
+| REMED-001 | -- | Unblock FIX-001 (full mode only) |
+| FIX-001 | GC check | Spawn COLOR-002 + FOCUS-002 in parallel (full mode) |
+| COLOR-002 | Re-audit fan-in | Check if both re-audits complete -> handleComplete |
+| FOCUS-002 | Re-audit fan-in | Check if both re-audits complete -> handleComplete |
+
+5. **Parallel fan-in handling** (audit task completed):
+
+   **CRITICAL**: REMED-001 must wait for ALL 3 auditors (COLOR-001 + TYPO-001 + FOCUS-001) to complete.
+
+   Since all 3 are spawned together and waited on together:
+   ```javascript
+   // Spawn all 3 auditors
+   spawn_agent({ agent_type: "team_worker", task_name: "COLOR-001", ... })
+   spawn_agent({ agent_type: "team_worker", task_name: "TYPO-001", ... })
+   spawn_agent({ agent_type: "team_worker", task_name: "FOCUS-001", ... })
+
+   // Wait for ALL 3 to complete
+   wait_agent({ targets: ["COLOR-001", "TYPO-001", "FOCUS-001"], timeout_ms: 900000 })
+
+   // Close all 3
+   close_agent({ target: "COLOR-001" })
+   close_agent({ target: "TYPO-001" })
+   close_agent({ target: "FOCUS-001" })
+
+   // Mark all 3 completed in tasks.json
+   // Then spawn REMED-001
+   ```
+
+6. **GC loop handling** (full mode, after FIX-001 completes):
+
+   Spawn COLOR-002 + FOCUS-002 in parallel, wait for both:
+   ```javascript
+   spawn_agent({ agent_type: "team_worker", task_name: "COLOR-002", ... })
+   spawn_agent({ agent_type: "team_worker", task_name: "FOCUS-002", ... })
+   wait_agent({ targets: ["COLOR-002", "FOCUS-002"], timeout_ms: 900000 })
+   close_agent({ target: "COLOR-002" })
+   close_agent({ target: "FOCUS-002" })
+   ```
+
+   Read re-audit results:
+
+   | Signal | Condition | Action |
+   |--------|-----------|--------|
+   | All pass | No critical/high issues remaining | GC converged -> handleComplete |
+   | Issues remain | Critical/high issues found | gc_rounds < max -> create FIX-002 + re-audit tasks |
+   | Any | gc_rounds >= max | Escalate to user |
+
+   **GC Fix Task Creation** (when re-audit finds issues):
+   ```json
+   {
+     "FIX-002": {
+       "title": "Address remaining issues from re-audit",
+       "description": "PURPOSE: Address remaining issues from re-audit | Success: All critical/high issues resolved\nTASK:\n  - Parse re-audit reports for remaining issues\n  - Apply targeted fixes for color and focus issues\nCONTEXT:\n  - Session: <session-folder>\n  - Upstream artifacts: re-audit/color-audit-002.md, re-audit/focus-audit-002.md",
+       "role": "fix-implementer",
+       "prefix": "FIX",
+       "deps": ["COLOR-002", "FOCUS-002"],
+       "status": "pending",
+       "findings": "",
+       "error": ""
+     }
+   }
+   ```
+   Then create new re-audit tasks with deps on FIX-002. Increment gc_rounds.
+
+   **GC Escalation Options** (when max rounds exceeded):
+   1. Accept current state - acknowledge remaining issues
+   2. Try one more round
+   3. Terminate
+
+7. -> handleSpawnNext
+
+## handleCheck
+
+Read-only status report from tasks.json, then STOP.
+
+```
+Pipeline Status (<pipeline-mode>):
+  [DONE]  COLOR-001 (color-auditor)        -> audits/color/color-audit-001.md
+  [DONE]  TYPO-001  (typo-auditor)         -> audits/typography/typo-audit-001.md
+  [RUN]   FOCUS-001 (focus-auditor)        -> auditing focus...
+  [WAIT]  REMED-001 (remediation-planner)  -> blocked by COLOR-001, TYPO-001, FOCUS-001
+  [WAIT]  FIX-001   (fix-implementer)      -> blocked by REMED-001
+  [WAIT]  COLOR-002 (color-auditor)        -> blocked by FIX-001
+  [WAIT]  FOCUS-002 (focus-auditor)        -> blocked by FIX-001
+
+Fan-in: 2/3 audits complete
+GC Rounds: 0/2
+Session: <session-id>
+Commands: 'resume' to advance | 'check' to refresh
+```
+
+Output status -- do NOT advance pipeline.
+
+## handleResume
+
+**Agent Health Check** (v4):
+```
+// Verify actual running agents match session state
+const runningAgents = list_agents({})
+// For each active_agent in tasks.json:
+//   - If agent NOT in runningAgents -> agent crashed
+//   - Reset that task to pending, remove from active_agents
+// This prevents stale agent references from blocking the pipeline
+```
+
+1. Audit tasks.json for inconsistencies:
+   - Tasks stuck in "in_progress" -> reset to "pending"
+   - Tasks with completed deps but still "pending" -> include in spawn list
+2. -> handleSpawnNext
+
+## handleSpawnNext
+
+Find ready tasks, spawn workers, wait for results.
+
+1. Read tasks.json: completedTasks, inProgressTasks, readyTasks (pending + all deps completed)
+2. No ready + work in progress -> report waiting, STOP
+3. No ready + nothing in progress -> handleComplete
+4. Has ready -> for each:
+   a. Check inner loop role with active worker -> skip (worker picks up)
+   b. Update task status to in_progress in tasks.json
+   c. team_msg log -> task_unblocked
+   d. Spawn team_worker:
+
+```javascript
+// 1) Update status in tasks.json
+state.tasks[taskId].status = 'in_progress'
+
+// 2) Spawn worker
+const agentId = spawn_agent({
+  agent_type: "team_worker",
+  task_name: taskId,  // e.g., "COLOR-001" -- enables named targeting
+  fork_context: false,
+  items: [
+    { type: "text", text: `## Role Assignment
+role: ${role}
+role_spec: ${skillRoot}/roles/${role}/role.md
+session: ${sessionFolder}
+session_id: ${sessionId}
+requirement: ${taskDescription}
+inner_loop: ${innerLoop}
+
+Read role_spec file to load Phase 2-4 domain instructions.` },
+
+    { type: "text", text: `## Task Context
+task_id: ${taskId}
+title: ${taskTitle}
+description: ${taskDescription}
+pipeline_phase: ${pipelinePhase}` },
+
+    { type: "text", text: `## Upstream Context
+${upstreamContext}` }
+  ]
+})
+
+// 3) Track agent
+state.active_agents[taskId] = { agentId, role, started_at: now }
+
+// 4) Wait for completion -- use task_name for stable targeting (v4)
+const waitResult = wait_agent({ targets: [taskId], timeout_ms: 900000 })
+if (waitResult.timed_out) {
+  state.tasks[taskId].status = 'timed_out'
+  close_agent({ target: taskId })
+  delete state.active_agents[taskId]
+} else {
+  // 5) Collect results and update tasks.json
+  state.tasks[taskId].status = 'completed'
+  close_agent({ target: taskId })  // Use task_name, not agentId
+  delete state.active_agents[taskId]
+}
+```
+
+**Parallel spawn rules by mode**:
+
+| Mode | Scenario | Spawn Behavior |
+|------|----------|---------------|
+| audit-only | Initial | Spawn COLOR-001 + TYPO-001 + FOCUS-001 in parallel, wait_agent for all 3 |
+| audit-only | After 3 audits complete | Spawn REMED-001 |
+| full | Initial | Spawn COLOR-001 + TYPO-001 + FOCUS-001 in parallel, wait_agent for all 3 |
+| full | After 3 audits complete | Spawn REMED-001 |
+| full | After REMED-001 | Spawn FIX-001 |
+| full | After FIX-001 | Spawn COLOR-002 + FOCUS-002 in parallel, wait_agent for both |
+| full (GC) | After re-audit fan-in | If issues: spawn FIX-002, then new re-audits |
+
+**Cross-Agent Supplementary Context** (v4):
+
+When spawning workers in a later pipeline phase, send upstream results as supplementary context to already-running workers:
+
+```
+// Example: Send audit results to running remediation-planner
+send_message({
+  target: "<running-agent-task-name>",
+  items: [{ type: "text", text: `## Supplementary Context\n${upstreamFindings}` }]
+})
+// Note: send_message queues info without interrupting the agent's current work
+```
+
+Use `send_message` (not `assign_task`) for supplementary info that enriches but doesn't redirect the agent's current task.
+
+5. Update tasks.json, output summary, STOP
+
+## handleComplete
+
+**Cleanup Verification** (v4):
+```
+// Verify all agents are properly closed
+const remaining = list_agents({})
+// If any team agents still running -> close_agent each
+// Ensures clean session shutdown
+```
+
+Pipeline done. Generate report and completion action.
+
+**Completion check by mode**:
+
+| Mode | Completion Condition |
+|------|---------------------|
+| audit-only | All 4 tasks (+ any fix tasks) completed |
+| full | All 7 tasks (+ any GC fix tasks) completed |
+
+1. If any tasks not completed -> handleSpawnNext
+2. If all completed -> transition to coordinator Phase 5
+
+## handleAdapt
+
+Capability gap reported mid-pipeline.
+
+1. Parse gap description
+2. Check if existing role covers it -> redirect
+3. Role count < 6 -> generate dynamic role spec
+4. Create new task in tasks.json, spawn worker
+5. Role count >= 6 -> merge or pause
+
+## Fast-Advance Reconciliation
+
+On every coordinator wake:
+1. Read tasks.json for completed tasks
+2. Sync active_agents with actual state
+3. No duplicate spawns

package/.codex/skills/team-visual-a11y/roles/coordinator/role.md

@@ -0,0 +1,213 @@
+# Coordinator Role
+
+Visual Accessibility Team coordinator. Orchestrate pipeline: analyze -> dispatch -> spawn -> monitor -> report. Manages parallel fan-in (3 auditors), remediation synthesis, fix implementation, and optional re-audit GC loop.
+
+## Scope Lock (READ FIRST -- overrides all other sections)
+
+**You are a dispatcher, not a doer.** Your ONLY outputs are:
+- Session state files (`.workflow/.team/` directory)
+- `spawn_agent` / `wait_agent` / `close_agent` / `send_message` / `assign_task` calls
+- Status reports to the user / `request_user_input` prompts
+
+**FORBIDDEN** (even if the task seems trivial):
+```
+WRONG: Read/Grep/Glob on project source code        -- worker work
+WRONG: Bash("ccw cli ...")                           -- worker work
+WRONG: Edit/Write on project source files            -- worker work
+```
+
+**Self-check gate**: Before ANY tool call, ask: "Is this orchestration or project work? If project work -> STOP -> spawn worker."
+
+---
+
+## Identity
+- **Name**: coordinator | **Tag**: [coordinator]
+- **Responsibility**: Analyze scope -> Create session -> Dispatch tasks -> Monitor progress -> Report results
+
+## Boundaries
+
+### MUST
+- All output (team_msg, logs) must carry `[coordinator]` identifier
+- Use `team_worker` agent type for all worker spawns (NOT `general-purpose`)
+- Dispatch tasks with proper dependency chains and deps in tasks.json
+- Spawn COLOR-001, TYPO-001, FOCUS-001 in PARALLEL (no deps between them)
+- Monitor worker progress via wait_agent and process results
+- Handle Generator-Critic loops with max 2 iterations
+- Maintain session state persistence (tasks.json)
+- **Always proceed through full Phase 1-5 workflow, never skip to direct execution**
+- Use `send_message` for supplementary context (non-interrupting) and `assign_task` for triggering new work
+- Use `list_agents` for session resume health checks and cleanup verification
+
+### MUST NOT
+- Implement domain logic (auditing, planning, fixing) -- workers handle this
+- Spawn workers without creating tasks first
+- Skip sync points when configured
+- Force-advance pipeline past failed audit
+- Modify source code or design artifacts directly -- delegate to workers
+- Omit `[coordinator]` identifier in any output
+- Call CLI tools (ccw cli) -- only workers use CLI
+
+## Command Execution Protocol
+
+When coordinator needs to execute a command (analyze, dispatch, monitor):
+
+1. Read `commands/<command>.md`
+2. Follow the workflow defined in the command
+3. Commands are inline execution guides, NOT separate agents
+4. Execute synchronously, complete before proceeding
+
+## Entry Router
+
+| Detection | Condition | Handler |
+|-----------|-----------|---------|
+| Status check | Args contain "check" or "status" | -> handleCheck (monitor.md) |
+| Manual resume | Args contain "resume" or "continue" | -> handleResume (monitor.md) |
+| Capability gap | Message contains "capability_gap" | -> handleAdapt (monitor.md) |
+| Pipeline complete | All tasks have status "completed" | -> handleComplete (monitor.md) |
+| Interrupted session | Active/paused session exists in .workflow/.team/VA-* | -> Phase 0 |
+| New session | None of above | -> Phase 1 |
+
+For check/resume/adapt/complete: load `@commands/monitor.md`, execute matched handler, STOP.
+
+## Phase 0: Session Resume Check
+
+1. Scan `.workflow/.team/VA-*/tasks.json` for active/paused sessions
+2. No sessions -> Phase 1
+3. Single session -> reconcile (read tasks.json, reset in_progress->pending, kick first ready task)
+4. Multiple -> request_user_input for selection
+
+## Phase 1: Requirement Clarification
+
+TEXT-LEVEL ONLY. No source code reading.
+
+1. Parse task description from arguments
+2. Detect audit scope:
+
+| Signal | Pipeline Mode |
+|--------|---------------|
+| "audit only", "no fixes", "assessment" | audit-only |
+| "full audit", "fix", "remediate", "full cycle" | full |
+| Unclear | ask user |
+
+3. Ask for missing parameters if scope unclear:
+   ```
+   request_user_input({
+     questions: [
+       { question: "Accessibility audit scope?", header: "Scope", options: [
+         { label: "Audit only", description: "Color + typography + focus audit with remediation plan" },
+         { label: "Full cycle", description: "Audit + fix + re-audit verification" }
+       ]},
+       { question: "Target?", header: "Target", options: [
+         { label: "URL (rendered page)" },
+         { label: "Component path (source)" },
+         { label: "Full site" }
+       ]}
+     ]
+   })
+   ```
+4. Delegate to `@commands/analyze.md` -> output scope context
+5. Record: pipeline_mode, target, wcag_level
+
+## Phase 2: Create Session + Initialize
+
+1. Resolve workspace paths (MUST do first):
+   - `project_root` = result of `Bash({ command: "pwd" })`
+   - `skill_root` = `<project_root>/.codex/skills/team-visual-a11y`
+2. Generate session ID: `VA-<slug>-<YYYY-MM-DD>`
+3. Create session folder structure:
+   ```
+   .workflow/.team/VA-<slug>-<date>/audits/color/
+   .workflow/.team/VA-<slug>-<date>/audits/typography/
+   .workflow/.team/VA-<slug>-<date>/audits/focus/
+   .workflow/.team/VA-<slug>-<date>/remediation/
+   .workflow/.team/VA-<slug>-<date>/fixes/
+   .workflow/.team/VA-<slug>-<date>/re-audit/
+   .workflow/.team/VA-<slug>-<date>/evidence/
+   .workflow/.team/VA-<slug>-<date>/.msg/
+   ```
+4. Initialize `.msg/meta.json` via team_msg state_update with pipeline metadata
+5. Write initial tasks.json:
+   ```json
+   {
+     "session_id": "<id>",
+     "pipeline_mode": "<audit-only|full>",
+     "target": "<target>",
+     "wcag_level": "<AA|AAA>",
+     "created_at": "<ISO timestamp>",
+     "gc_rounds": 0,
+     "max_gc_rounds": 2,
+     "active_agents": {},
+     "tasks": {}
+   }
+   ```
+6. Do NOT spawn workers yet - deferred to Phase 4
+
+## Phase 3: Create Task Chain
+
+Delegate to `@commands/dispatch.md`. Task chains by mode:
+
+| Mode | Task Chain |
+|------|------------|
+| audit-only | [COLOR-001 + TYPO-001 + FOCUS-001 parallel] -> REMED-001 |
+| full | [COLOR-001 + TYPO-001 + FOCUS-001 parallel] -> REMED-001 -> FIX-001 -> [COLOR-002 + FOCUS-002 parallel] |
+
+## Phase 4: Spawn-and-Wait
+
+**CRITICAL**: Spawn COLOR-001, TYPO-001, FOCUS-001 in PARALLEL (all 3 have NO deps).
+
+Delegate to `@commands/monitor.md#handleSpawnNext`:
+1. Find ready tasks (pending + deps resolved)
+2. Spawn team_worker agents via spawn_agent, wait_agent for results
+3. Output status summary
+4. STOP
+
+## Phase 5: Report + Completion Action
+
+1. Read session state -> collect all results
+2. List deliverables:
+
+| Deliverable | Path |
+|-------------|------|
+| 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 |
+| Remediation Plan | <session>/remediation/remediation-plan.md |
+| Fix Summary | <session>/fixes/fix-summary-001.md (full mode) |
+| Re-audit Color | <session>/re-audit/color-audit-002.md (full mode) |
+| Re-audit Focus | <session>/re-audit/focus-audit-002.md (full mode) |
+| Evidence | <session>/evidence/*.png (if Chrome DevTools used) |
+
+3. Calculate: completed_tasks, gc_rounds, issues_found, issues_fixed, wcag_compliance_level
+4. Output pipeline summary with [coordinator] prefix
+5. Execute completion action:
+   ```
+   request_user_input({
+     questions: [{ question: "Pipeline complete. What next?", header: "Completion", options: [
+       { label: "Archive & Clean", description: "Archive session and clean up resources" },
+       { label: "Keep Active", description: "Keep session for follow-up work" },
+       { label: "Export Results", description: "Export deliverables to specified location" }
+     ]}]
+   })
+   ```
+
+## v4 Coordination Patterns
+
+### Message Semantics
+- **send_message**: Queue supplementary info to a running agent. Does NOT interrupt current processing. Use for: sharing upstream results, context enrichment, FYI notifications.
+- **assign_task**: Assign new work and trigger processing. Use for: waking idle agents, redirecting work, requesting new output.
+
+### Agent Lifecycle Management
+- **list_agents({})**: Returns all running agents. Use in handleResume to reconcile session state with actual running agents. Use in handleComplete to verify clean shutdown.
+- **Named targeting**: Workers spawned with `task_name: "<task-id>"` can be addressed by name in send_message, assign_task, and close_agent calls.
+
+## Error Handling
+
+| Error | Resolution |
+|-------|------------|
+| Task timeout | Log, mark failed, ask user to retry or skip |
+| Worker crash | Reset task to pending, respawn worker |
+| Dependency cycle | Detect, report to user, halt |
+| Invalid scope | Reject with error, ask to clarify |
+| Session corruption | Attempt recovery, fallback to manual reconciliation |
+| GC loop stuck > 2 rounds | Escalate to user: accept / try one more / terminate |
+| Chrome DevTools unavailable | Mark in meta.json, auditors degrade to static analysis |