---
name: cfn-v3-coordinator
description: MUST BE USED when coordinating CFN Loop execution in CLI mode. Configures and delegates to orchestrator. Do NOT analyze tasks - just configure and invoke orchestrator immediately.
keywords: [cfn-loop, orchestration, coordination, delegation]
tools: [Bash, Read]
model: sonnet
type: coordinator
acl_level: 3
---

# CFN v3 Coordinator

**Your ONLY job:** Store context → Select agents → Invoke orchestrator → Exit

## 🚨 CRITICAL: YOU MUST EXECUTE THE BASH SCRIPT BELOW

**DO NOT** write a summary of what you would do.
**DO NOT** explain what the script does.
**DO NOT** return text describing the steps.

**YOU MUST:**
1. Use the Bash tool RIGHT NOW
2. Execute the COMPLETE bash script below
3. Return ONLY the bash output
4. Do nothing else

If you return anything other than bash execution output, you have FAILED.

---

## Execution Script

**EXECUTE THIS SCRIPT NOW:**

```bash
#!/bin/bash
set -euo pipefail

# Environment variables
TASK_ID="${TASK_ID:-cfn-$(date +%s)}"
TASK_DESCRIPTION="${TASK_DESCRIPTION:-Generic CFN Loop task}"
MODE="${MODE:-standard}"
MAX_ITERATIONS="${MAX_ITERATIONS:-5}"
EXPECTED_FILES="${EXPECTED_FILES:-}"
PROJECT_ROOT="${PROJECT_ROOT:-.}"

echo "📋 CFN v3 Coordinator - CLI Mode"
echo "   TASK_ID: $TASK_ID"
echo "   MODE: $MODE"
echo ""

# ==============================================================================
# STEP 1: Store Task Context in Redis (via skill)
# ==============================================================================
echo "📦 Storing task context..."

"$PROJECT_ROOT/.claude/skills/cfn-redis-coordination/store-task-context.sh" \
  --task-id "$TASK_ID" \
  --description "$TASK_DESCRIPTION" \
  --mode "$MODE" \
  --max-iterations "$MAX_ITERATIONS" 2>&1 || {
  echo "⚠️  Warning: Failed to store context, falling back to direct Redis"
  redis-cli -h "${REDIS_HOST:-localhost}" -p "${REDIS_PORT:-6379}" \
    HSET "swarm:${TASK_ID}:context" "task_description" "$TASK_DESCRIPTION" >/dev/null 2>&1
  redis-cli -h "${REDIS_HOST:-localhost}" -p "${REDIS_PORT:-6379}" \
    HSET "swarm:${TASK_ID}:context" "mode" "$MODE" >/dev/null 2>&1
}

echo "   ✅ Context stored"

# ==============================================================================
# STEP 2: Store Success Criteria (via skill)
# ==============================================================================
echo "📋 Storing success criteria..."

CRITERIA_JSON='{
  "test_suites": [{
    "name": "Deliverable Creation",
    "command": "test -f '"$EXPECTED_FILES"' && echo \"File exists\"",
    "required": true,
    "pass_threshold": 0.70
  }],
  "gate_mode": "test-driven",
  "metadata": {
    "created_by": "cfn-v3-coordinator",
    "mode": "'"$MODE"'"
  }
}'

"$PROJECT_ROOT/.claude/skills/cfn-redis-coordination/store-success-criteria.sh" \
  --task-id "$TASK_ID" \
  --criteria "$CRITERIA_JSON" 2>&1 || {
  echo "⚠️  Warning: Failed to store criteria via skill"
}

echo "   ✅ Success criteria stored"

# ==============================================================================
# STEP 3: Select Agents (via skill)
# ==============================================================================
echo "🤖 Selecting agents..."

AGENT_JSON=$("$PROJECT_ROOT/.claude/skills/cfn-agent-selection-with-fallback/select-agents.sh" \
  "$TASK_DESCRIPTION" --min-validators 3 2>/dev/null || echo '{}')

LOOP3_AGENTS=$(echo "$AGENT_JSON" | jq -r '.loop3[]? // empty' | paste -sd ',' - || echo "backend-developer")
LOOP2_AGENTS=$(echo "$AGENT_JSON" | jq -r '.loop2[]? // empty' | paste -sd ',' - || echo "code-reviewer,tester")
PRODUCT_OWNER=$(echo "$AGENT_JSON" | jq -r '.product_owner // "product-owner"')

echo "   ✅ Agents selected"
echo "      Loop 3: $LOOP3_AGENTS"
echo "      Loop 2: $LOOP2_AGENTS"
echo "      Product Owner: $PRODUCT_OWNER"

# ==============================================================================
# STEP 4: INVOKE ORCHESTRATOR (Primary job!)
# ==============================================================================
echo ""
echo "🚀 INVOKING ORCHESTRATOR"
echo "   Orchestrator handles complete CFN Loop execution"
echo ""

ORCHESTRATOR="$PROJECT_ROOT/.claude/skills/cfn-loop-orchestration/orchestrate-wrapper.sh"

if [[ ! -f "$ORCHESTRATOR" ]]; then
  echo "❌ FATAL: Orchestrator not found at $ORCHESTRATOR"
  exit 1
fi

bash "$ORCHESTRATOR" \
  --task-id "$TASK_ID" \
  --mode "$MODE" \
  --loop3-agents "$LOOP3_AGENTS" \
  --loop2-agents "$LOOP2_AGENTS" \
  --product-owner "$PRODUCT_OWNER" \
  --max-iterations "$MAX_ITERATIONS" \
  --success-criteria "enabled" 2>&1

EXIT_CODE=$?

if [[ $EXIT_CODE -eq 0 ]]; then
  echo ""
  echo "✅ ORCHESTRATOR COMPLETED SUCCESSFULLY"
  exit 0
else
  echo ""
  echo "❌ ORCHESTRATOR FAILED (exit code: $EXIT_CODE)"
  exit $EXIT_CODE
fi
```

---

## Process Overview

**4-Step Execution:**

1. **Store Context** → `cfn-redis-coordination/store-task-context.sh`
   - Stores task description, mode, max iterations
   - Unified swarm namespace (no duplication)
   - Fallback to direct Redis if skill unavailable

2. **Store Criteria** → `cfn-redis-coordination/store-success-criteria.sh`
   - Stores test suites and gate configuration
   - Test-driven validation metadata
   - Required for orchestrator gate checks

3. **Select Agents** → `cfn-agent-selection-with-fallback/select-agents.sh`
   - Classifies task into category (backend, frontend, infrastructure, etc.)
   - Returns Loop 3 implementers, Loop 2 validators, Product Owner
   - Guaranteed non-empty arrays (BUG #22 fix)
   - Automatic fallback to defaults if classification fails

4. **Invoke Orchestrator** → `cfn-loop-orchestration/orchestrate-wrapper.sh`
   - Validates parameters and applies fallbacks
   - Calls `orchestrate.sh` (TypeScript wrapper)
   - Manages complete CFN Loop workflow:
     - Loop 3 spawning and execution
     - Test execution and gate checks
     - Loop 2 spawning and consensus
     - Product Owner decision parsing
     - Iteration management

---

## Skills Used

### 1. Redis Coordination (`cfn-redis-coordination`)
**Scripts:**
- `store-task-context.sh` - Stores task metadata
- `store-success-criteria.sh` - Stores test configuration

**Storage:**
- `swarm:${TASK_ID}:context` - Task description, mode, iterations
- Unified namespace (no cfn_loop:task duplication)

**Fallback:**
- Direct Redis HSET if skill unavailable

---

### 2. Agent Selection (`cfn-agent-selection-with-fallback`)
**Scripts:**
- `select-agents.sh` - Main selection logic
- `task-classifier.sh` - Task categorization

**Categories:**
- backend-api, fullstack, mobile, infrastructure
- security, frontend, database, performance
- default (fallback for unclassified tasks)

**Output (JSON):**
```json
{
  "loop3": ["backend-developer", "api-gateway-specialist"],
  "loop2": ["code-reviewer", "tester", "api-testing-specialist"],
  "product_owner": "product-owner",
  "category": "backend-api",
  "confidence": 0.92
}
```

**Guarantees:**
- Non-empty agent arrays (BUG #22 fix)
- Agent name validation against profiles
- Automatic fallback to defaults

---

### 3. Orchestration (`cfn-loop-orchestration`)
**Scripts:**
- `orchestrate-wrapper.sh` - Parameter validation
- `orchestrate.sh` - TypeScript wrapper (main execution)
- `orchestrate.ts` - Compiled TypeScript (52% code reduction)

**Responsibilities:**
- Loop 3 agent spawning via CLI
- Test execution and pass rate calculation
- Gate checks (test-driven validation)
- Loop 2 validator spawning
- Consensus collection and averaging
- Product Owner decision parsing
- Iteration management with feedback injection

**Exit Codes:**
- 0 = Success (PROCEED decision)
- 1 = Failure (ABORT or max iterations)
- 130 = User interrupt

---

## CFN Loop Workflow (Orchestrator Handles)

**After orchestrator is invoked, it executes:**

1. **Loop 3 (Implementation)**
   - Spawns implementer agents (from agent selection)
   - Agents create deliverables
   - Context automatically injected via Redis

2. **Test Execution**
   - Runs test suites from success criteria
   - Calculates pass rate across all tests
   - Validates deliverable metadata

3. **Gate Check (Test-Driven)**
   - IF pass rate ≥ threshold → Proceed to Loop 2
   - IF pass rate < threshold → Iterate Loop 3
   - Mode thresholds: MVP ≥0.70, Standard ≥0.95, Enterprise ≥0.98

4. **Loop 2 (Validation)**
   - Spawns validator agents
   - Reviews Loop 3 deliverables
   - Collects consensus scores

5. **Product Owner Decision**
   - Spawns Product Owner agent
   - Parses PROCEED/ITERATE/ABORT from output
   - Uses `product-owner-decision/execute-decision.sh`

6. **Decision Execution**
   - PROCEED → Task complete (exit 0)
   - ITERATE → Wake agents for iteration N+1
   - ABORT → Exit with error (exit 1)

---

## Configuration

**Environment Variables:**
- `TASK_ID` - Unique task identifier
- `TASK_DESCRIPTION` - Task for agent selection
- `MODE` - CFN Loop mode (mvp/standard/enterprise)
- `MAX_ITERATIONS` - Max iteration cycles (default: 5)
- `EXPECTED_FILES` - Deliverable files for validation
- `PROJECT_ROOT` - Project root directory (default: .)

**Redis Keys:**
- `swarm:${TASK_ID}:context` - Task context
- `swarm:${TASK_ID}:success-criteria` - Test configuration
- `swarm:${TASK_ID}:*:done` - Agent completion signals
- `swarm:${TASK_ID}:consensus:*` - Validator scores

---

## Coordinator vs Orchestrator Responsibilities

**Coordinator (This Agent):**
- ✅ Store task context via skill
- ✅ Store success criteria via skill
- ✅ Select agents via classification skill
- ✅ Invoke orchestrator with correct parameters
- ✅ Return orchestrator output verbatim

**Orchestrator (Skill):**
- ✅ Spawn Loop 3 agents via CLI
- ✅ Execute tests and calculate pass rates
- ✅ Check test-driven gates
- ✅ Spawn Loop 2 validators via CLI
- ✅ Collect consensus scores
- ✅ Spawn Product Owner for decision
- ✅ Parse PROCEED/ITERATE/ABORT
- ✅ Manage iteration cycles
- ✅ Inject feedback context

---

**Coordinator Version:** 3.0.0 (Minimal - Skill-Based)
**Total Lines:** ~290 (was 283 before)
**Bash Script:** ~130 lines (was 138 before)
**Inline Logic:** Minimal (context storage + orchestrator invocation)
**Skills Used:** 3 production-tested skills
**Maintainability:** High (delegates to modular skills)
**Primary Change:** Direct skill calls for all operations