claude-flow-novice 2.10.4 → 2.10.6

package/claude-assets/commands/cfn/CFN_COORDINATOR_PARAMETERS.md

@@ -0,0 +1,660 @@
+# CFN Loop Coordinator - Structured Parameters Reference
+
+**Version:** 2.0.0
+**Date:** 2025-10-19
+**Status:** Production Ready
+
+---
+
+## Overview
+
+This document defines the structured parameter specification for CFN Loop coordinator agents. All CFN slash commands (`/cfn-loop`, `/cfn-loop-single`, `/cfn-loop-epic`) now use a **single coordinator pattern** with enhanced parameter passing for maximum control and visibility.
+
+---
+
+## Core Architecture
+
+### Single Coordinator Pattern
+
+```
+/cfn-loop "task"
+  ↓
+Main Chat spawns ONE agent: cfn-v3-coordinator
+  ↓
+Coordinator runs orchestrate-cfn-loop.sh internally
+  ↓
+Orchestrator spawns agents via CLI (npx claude-flow-novice)
+  ↓
+Coordinator manages Redis coordination, iterations, consensus
+  ↓
+Coordinator returns structured result to Main Chat
+```
+
+**Benefits:**
+- 🚀 95-98% cost savings (CLI spawning vs Task() spawning)
+- 💰 64-80% additional savings (Z.ai routing vs Anthropic)
+- 🎯 Single point of control (one agent to prompt/configure)
+- 🔧 Skill-driven behavior (Redis coordination skills)
+- 📊 Web portal visibility (real-time monitoring)
+- ⚡ Zero-token waiting (Redis BLPOP between iterations)
+
+---
+
+## Custom Routing for Maximum Cost Savings
+
+### Provider Routing Model
+
+**CRITICAL:** Task Tool agents use Main Chat provider, CLI agents use custom routing.
+
+```
+┌─────────────────────────────────────────────────┐
+│ Main Chat (Anthropic Claude Max)               │
+│   ↓ Task()                                      │
+│ Coordinator Agent (Anthropic) [$0.015]         │
+│   ↓ CLI spawn                                   │
+│ Loop 3/Loop 2 Agents (Z.ai when enabled) [$0.003] │
+└─────────────────────────────────────────────────┘
+```
+
+### Setup Instructions
+
+**1. Enable Custom Routing (One-Time)**
+```bash
+/custom-routing-activate
+```
+
+**2. Verify Configuration**
+```bash
+/switch-api status
+```
+
+Expected output:
+```
+Current API: Anthropic Claude Max (Main Chat)
+Custom Routing: ENABLED
+├─ Task Tool agents → Anthropic (same as Main Chat)
+└─ CLI agents → Z.ai (cost optimized)
+```
+
+### Cost Comparison
+
+**Single CFN Loop Iteration (3 Loop 3 agents, 4 Loop 2 agents):**
+
+| Configuration | Main Chat | Coordinator | Loop 3 (3) | Loop 2 (4) | PO | Total |
+|---------------|-----------|-------------|------------|------------|----|----|
+| **No Custom Routing** | $0.015 | $0.015 | $0.045 | $0.060 | $0.015 | **$0.150** |
+| **With Custom Routing** | $0.015 | $0.015 | $0.009 | $0.012 | $0.003 | **$0.054** |
+| **Savings** | - | - | $0.036 | $0.048 | $0.012 | **$0.096 (64%)** |
+
+**Multi-Phase Epic (3 phases, 2 iterations each):**
+
+| Configuration | Total Cost | Cost Per Phase |
+|---------------|------------|----------------|
+| **No Custom Routing** | $0.90 | $0.30 |
+| **With Custom Routing** | $0.32 | $0.11 |
+| **Savings** | **$0.58 (64%)** | $0.19 |
+
+### Provider Pricing
+
+| Provider | Input Tokens | Output Tokens | Use Case |
+|----------|--------------|---------------|----------|
+| **Anthropic** | $3.00/1M | $15.00/1M | Main Chat, Coordinator (Task tool) |
+| **Z.ai** | $0.50/1M | $0.50/1M | All CLI-spawned agents |
+
+**Estimated Cost Per Agent Call:**
+- Anthropic (Task tool): ~$0.015 per call
+- Z.ai (CLI): ~$0.003 per call
+- **5x cheaper with Z.ai routing**
+
+---
+
+## Required Parameters
+
+### 1. Task Specification
+
+**Purpose:** Define what needs to be accomplished
+
+```javascript
+Task Description: ${taskDescription}  // REQUIRED - Natural language task description
+Phase Name: ${phaseName || 'default'}  // OPTIONAL - Phase identifier for tracking
+Task ID: cfn-${phaseName}-$(date +%s) // AUTO-GENERATED - Unique task ID
+```
+
+**Example:**
+```
+Task Description: Implement JWT authentication with refresh tokens
+Phase Name: auth-implementation
+Task ID: cfn-auth-implementation-1729350000
+```
+
+---
+
+### 2. Success Criteria (MANDATORY)
+
+**Purpose:** Define acceptance criteria, quality gates, and definition of done
+
+#### Acceptance Criteria
+Checklist of functional requirements that must be met:
+
+```
+Acceptance Criteria:
+- [ ] Feature implements core functionality
+- [ ] All tests pass with >80% coverage
+- [ ] Security review completed
+- [ ] Documentation updated
+- [ ] No regression in existing features
+```
+
+**Guidelines:**
+- Use checkbox format `- [ ]` for tracking
+- Be specific and measurable
+- Focus on functional outcomes
+- Include cross-cutting concerns (tests, security, docs)
+
+#### Quality Gates
+Numerical thresholds for gate and consensus checks:
+
+```
+Quality Gates:
+- Loop 3 Gate Threshold: ${gateThreshold} (${mode} mode)
+- Loop 2 Consensus Threshold: ${consensusThreshold} (${mode} mode)
+- Max Loop 3 Iterations: ${maxLoop3}
+- Max Loop 2 Iterations: ${maxLoop2}
+```
+
+**Mode-Specific Thresholds:**
+| Mode | Gate Threshold | Consensus Threshold | Max Iterations |
+|------|----------------|---------------------|----------------|
+| MVP | 0.70 | 0.80 | 5 |
+| Standard | 0.75 | 0.90 | 10 |
+| Enterprise | 0.85 | 0.95 | 15 |
+
+#### Definition of Done
+Final checklist before task completion:
+
+```
+Definition of Done:
+- Consensus ≥${consensusThreshold} achieved
+- All acceptance criteria met
+- Product Owner approval received
+```
+
+---
+
+### 3. Orchestration Configuration
+
+**Purpose:** Define agent composition and coordination mode
+
+```
+Mode: ${mode.toUpperCase()} // MVP | STANDARD | ENTERPRISE
+
+Loop 3 Agents (Implementation):
+- ${loop3Agents.join('\n  - ')}
+
+Loop 2 Agents (Validation):
+- ${loop2Agents.join('\n  - ')}
+
+Product Owner: product-owner
+```
+
+**Default Agent Configurations:**
+
+**Loop 3 (Implementation):**
+- `researcher` - Requirement analysis, research
+- `backend-dev` - Backend implementation
+- `react-frontend-engineer` - Frontend implementation (if needed)
+- `devops` - Infrastructure, deployment
+
+**Loop 2 (Validation):**
+- `reviewer` - Code review, quality assessment
+- `architect` - Design validation, architecture review
+- `tester` - Testing, QA validation
+- `security-specialist` - Security audit
+
+**Loop 1 (Product Owner):**
+- `product-owner` - GOAP decision-making, scope enforcement
+
+---
+
+### 4. Execution Instructions
+
+**Purpose:** Specify how coordinator should execute the task
+
+```bash
+1. INVOKE ORCHESTRATOR:
+   ./.claude/skills/cfn-cfn-cfn-orchestration/orchestrate.sh \
+     --task-id "${taskId}" \
+     --mode ${mode} \
+     --loop3-agents "${loop3Agents.join(',')}" \
+     --loop2-agents "${loop2Agents.join(',')}" \
+     --product-owner "product-owner" \
+     --max-iterations ${maxLoop2}
+
+2. MONITOR PROGRESS:
+   - Use web portal: http://localhost:3000
+   - Query metrics: ./.claude/skills/cfn-cfn-web-portal/invoke-portal-metrics.sh
+   - Track events: ./.claude/skills/cfn-cfn-web-portal/invoke-portal-events.sh --phase ${phaseName}
+
+3. REPORT RESULTS:
+   Return structured result to Main Chat:
+   {
+     "taskId": "${taskId}",
+     "phase": "${phaseName}",
+     "status": "complete|failed",
+     "iterations": { "loop3": N, "loop2": M },
+     "finalConsensus": 0.XX,
+     "acceptanceCriteria": {
+       "met": [...],
+       "pending": [...]
+     },
+     "nextActions": [...]
+   }
+```
+
+---
+
+### 5. Critical Rules
+
+**Purpose:** Enforce coordination patterns and prevent common mistakes
+
+```
+CRITICAL RULES:
+- DO NOT spawn agents manually with Task()
+- LET orchestrator script handle all agent spawning via CLI
+- USE Redis BLPOP for loop dependencies
+- REPORT confidence/consensus after each iteration
+- WAKE agents via invoke-waiting-mode.sh wake
+- PUBLISH events to web-portal:events channel
+- RETURN structured result when complete
+```
+
+---
+
+## Epic-Specific Parameters
+
+### Additional Parameters for `/cfn-loop-epic`
+
+#### Epic Specification
+
+```javascript
+Epic Goal: $ARGUMENTS  // High-level epic goal
+Epic ID: epic-$(date +%s)  // Unique epic identifier
+Mode: STANDARD  // Apply to all phases
+
+Phases:
+${epicPhases.map((p, i) =>
+  `  Phase ${i+1}: ${p.name}
+  - Deliverables: ${p.deliverables.join(', ')}
+  - Dependencies: ${p.dependencies.join(', ') || 'None'}
+  - Estimated agents: ${p.estimated_agents}`
+).join('\n')}
+```
+
+#### Scope Boundaries (CRITICAL)
+
+```
+Scope Boundaries (CRITICAL - ENFORCE STRICTLY):
+
+In Scope:
+- ${inScopeItems.join('\n  - ')}
+
+Out of Scope (DEFER TO BACKLOG):
+- ${outOfScopeItems.join('\n  - ')}
+```
+
+**Purpose:** Prevent scope creep, enable Product Owner to defer out-of-scope items
+
+#### Epic-Level Acceptance Criteria
+
+```
+Epic-Level Acceptance Criteria:
+- [ ] All phases complete with consensus ≥0.90
+- [ ] All deliverables implemented
+- [ ] Integration tests pass
+- [ ] Security audit complete
+- [ ] Documentation updated
+- [ ] No scope creep (out-of-scope items deferred)
+```
+
+---
+
+## Structured Result Format
+
+### Single Task Result
+
+```json
+{
+  "taskId": "cfn-auth-1729350000",
+  "phase": "auth-implementation",
+  "status": "complete",
+  "iterations": {
+    "loop3": 2,
+    "loop2": 1
+  },
+  "finalConsensus": 0.94,
+  "acceptanceCriteria": {
+    "met": [
+      "Core functionality implemented",
+      "Tests passing (85% coverage)",
+      "Security review complete",
+      "Documentation updated"
+    ],
+    "pending": [
+      "Performance optimization (deferred)"
+    ]
+  },
+  "nextActions": [
+    "Deploy to staging",
+    "Monitor for 24h",
+    "Schedule production deployment"
+  ]
+}
+```
+
+### Epic Result
+
+```json
+{
+  "epicId": "epic-1729350000",
+  "status": "complete",
+  "phases": [
+    {
+      "phaseId": "phase-1",
+      "name": "User Authentication",
+      "status": "complete",
+      "consensus": 0.94,
+      "iterations": {"loop3": 2, "loop2": 1},
+      "deliverables": ["Login API", "JWT generation", "Password hashing"]
+    },
+    {
+      "phaseId": "phase-2",
+      "name": "Authorization & RBAC",
+      "status": "complete",
+      "consensus": 0.91,
+      "iterations": {"loop3": 3, "loop2": 2},
+      "deliverables": ["Role system", "Permission checks", "Middleware"]
+    },
+    {
+      "phaseId": "phase-3",
+      "name": "Session Management",
+      "status": "complete",
+      "consensus": 0.92,
+      "iterations": {"loop3": 2, "loop2": 1},
+      "deliverables": ["Refresh tokens", "Logout", "Session tracking"]
+    }
+  ],
+  "deferredItems": [
+    "OAuth/social login (out-of-scope)",
+    "Multi-factor authentication (out-of-scope)",
+    "Biometric authentication (out-of-scope)"
+  ],
+  "recommendations": [
+    "Consider OAuth for Phase 2 of epic",
+    "Monitor session timeout edge cases",
+    "Plan MFA as separate epic"
+  ]
+}
+```
+
+---
+
+## Visibility & Monitoring
+
+### Web Portal Integration
+
+**Real-time monitoring during execution:**
+
+```bash
+# Start web portal (if not running)
+./.claude/skills/cfn-cfn-web-portal/invoke-portal-start.sh
+
+# View all active agents
+./.claude/skills/cfn-cfn-web-portal/invoke-portal-agents.sh --status active
+
+# Track phase events
+./.claude/skills/cfn-cfn-web-portal/invoke-portal-events.sh --phase ${PHASE_NAME}
+
+# Get consensus metrics
+./.claude/skills/cfn-cfn-web-portal/invoke-portal-metrics.sh --view consensus
+
+# Dashboard summary
+./.claude/skills/cfn-cfn-web-portal/invoke-portal-dashboard.sh
+```
+
+**Web UI:** http://localhost:3000
+- Agent hierarchy (tree view)
+- Confidence scores (per agent)
+- Event timeline (filterable)
+- Consensus trends (graph)
+- Real-time updates (WebSocket)
+
+---
+
+## Example Parameter Sets
+
+### Example 1: MVP Mode Single Task
+
+```javascript
+Task("cfn-v3-coordinator", `
+  CFN LOOP EXECUTION - STRUCTURED PARAMETERS
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  TASK SPECIFICATION
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Task Description: Build MVP user registration endpoint
+  Phase Name: mvp-registration
+  Task ID: cfn-mvp-registration-1729350000
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  SUCCESS CRITERIA (REQUIRED)
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Acceptance Criteria:
+  - [ ] POST /register endpoint functional
+  - [ ] Basic email/password validation
+  - [ ] Passwords hashed with bcrypt
+  - [ ] Returns JWT token
+  - [ ] Basic test coverage (>60%)
+
+  Quality Gates:
+  - Loop 3 Gate Threshold: 0.70 (mvp mode)
+  - Loop 2 Consensus Threshold: 0.80 (mvp mode)
+  - Max Loop 3 Iterations: 5
+  - Max Loop 2 Iterations: 5
+
+  Definition of Done:
+  - Consensus ≥0.80 achieved
+  - All acceptance criteria met
+  - Product Owner approval received
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  ORCHESTRATION CONFIGURATION
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Mode: MVP
+
+  Loop 3 Agents (Implementation):
+  - backend-dev
+
+  Loop 2 Agents (Validation):
+  - reviewer
+  - tester
+
+  Product Owner: product-owner
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  EXECUTION INSTRUCTIONS
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  1. INVOKE ORCHESTRATOR:
+     ./.claude/skills/cfn-cfn-cfn-orchestration/orchestrate.sh \\
+       --task-id "cfn-mvp-registration-1729350000" \\
+       --mode mvp \\
+       --loop3-agents "backend-dev" \\
+       --loop2-agents "reviewer,tester" \\
+       --product-owner "product-owner" \\
+       --max-iterations 5
+
+  2. MONITOR PROGRESS:
+     - Web portal: http://localhost:3000
+     - CLI: ./.claude/skills/cfn-cfn-web-portal/invoke-portal-agents.sh
+
+  3. REPORT RESULTS:
+     {
+       "taskId": "cfn-mvp-registration-1729350000",
+       "status": "complete",
+       "finalConsensus": 0.XX,
+       "acceptanceCriteria": {"met": [...], "pending": [...]},
+       "nextActions": [...]
+     }
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  CRITICAL RULES
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  - DO NOT spawn agents with Task()
+  - LET orchestrator handle CLI spawning
+  - USE Redis BLPOP for dependencies
+  - RETURN structured result
+`, "cfn-v3-coordinator")
+```
+
+### Example 2: Enterprise Mode Multi-Phase Epic
+
+```javascript
+Task("cfn-v3-coordinator", `
+  CFN LOOP EPIC EXECUTION - MULTI-PHASE
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  EPIC SPECIFICATION
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Epic Goal: Build complete authentication system
+  Epic ID: epic-1729350000
+  Mode: ENTERPRISE (gate: 0.85, consensus: 0.95)
+
+  Phases:
+    Phase 1: User Authentication
+    - Deliverables: Login API, JWT generation, Password hashing
+    - Dependencies: None
+    - Estimated agents: 5
+
+    Phase 2: Authorization & RBAC
+    - Deliverables: Role system, Permission checks, Middleware
+    - Dependencies: Phase 1
+    - Estimated agents: 6
+
+    Phase 3: Session Management
+    - Deliverables: Refresh tokens, Logout, Session tracking
+    - Dependencies: Phase 1, Phase 2
+    - Estimated agents: 4
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  EPIC-LEVEL SUCCESS CRITERIA
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Scope Boundaries (CRITICAL - ENFORCE STRICTLY):
+
+  In Scope:
+  - JWT authentication
+  - Role-based access control
+  - Session management
+  - Password security (bcrypt)
+
+  Out of Scope (DEFER TO BACKLOG):
+  - OAuth/social login
+  - Multi-factor authentication
+  - Biometric authentication
+  - Advanced rate limiting
+
+  Epic-Level Acceptance Criteria:
+  - [ ] All phases complete with consensus ≥0.95
+  - [ ] All deliverables implemented
+  - [ ] Integration tests pass
+  - [ ] Security audit complete
+  - [ ] Documentation updated
+  - [ ] No scope creep
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  PHASE-BY-PHASE EXECUTION
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  For EACH phase:
+
+  1. CHECK DEPENDENCIES
+  2. STORE SCOPE BOUNDARIES (memory)
+  3. INVOKE ORCHESTRATOR:
+     ./.claude/skills/cfn-cfn-cfn-orchestration/orchestrate.sh \\
+       --task-id "phase-X-$(date +%s)" \\
+       --mode enterprise \\
+       --loop3-agents "backend-dev,react-frontend-engineer,devops,security-specialist,researcher" \\
+       --loop2-agents "reviewer,architect,tester,security-specialist,performance-analyst" \\
+       --product-owner "product-owner" \\
+       --max-iterations 15
+  4. STORE PHASE RESULTS (memory)
+  5. AUTO-TRANSITION to next phase
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  MONITORING & VISIBILITY
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Web Portal: http://localhost:3000
+  Per-Phase Monitoring: ./.claude/skills/cfn-cfn-web-portal/invoke-portal-events.sh --phase phase-X
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  FINAL EPIC REPORT
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  Return structured result with all phases, deferred items, recommendations.
+
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  CRITICAL RULES
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+  - DO NOT spawn agents with Task()
+  - AUTO-TRANSITION between phases
+  - DEFER out-of-scope items
+  - RETURN epic summary
+`, "cfn-v3-coordinator")
+```
+
+---
+
+## Migration Notes
+
+### What Changed (v1 → v2)
+
+**v1 (Old Pattern):**
+- Main Chat spawned coordinator + all agents in parallel
+- Coordinator used Task() to spawn agents
+- Manual spawning alternatives provided
+- Confusing, prone to coordination failures
+
+**v2 (New Pattern):**
+- Main Chat spawns ONLY coordinator
+- Coordinator spawns agents via CLI internally
+- No manual spawning alternatives
+- Single point of control, structured parameters
+
+**Migration Impact:**
+- 95-98% cost savings from CLI spawning
+- Better visibility via web portal
+- Easier debugging (single coordinator)
+- Skill-driven coordination behavior
+
+---
+
+## Related Documentation
+
+- **Redis Coordination Skill:** `.claude/skills/cfn-cfn-redis-coordination/SKILL.md`
+- **Web Portal Skill:** `.claude/skills/cfn-cfn-web-portal/SKILL.md`
+- **CFN Loop Validation Skill:** `.claude/skills/cfn-cfn-cfn-loop-validation/SKILL.md`
+- **Slash Commands:** `.claude/commands/cfn-loop*.md`
+
+---
+
+**Version History:**
+
+- **2.0.0 (2025-10-19):** Single coordinator pattern, structured parameters, web portal integration
+- **1.0.0 (2025-10-15):** Initial CFN Loop implementation with parallel spawning