---

name: cfn-task-config-init description: Initialize structured task configuration for CFN Task Mode with scope boundaries and deliverables dependencies: []

CFN Task Config Initialization

Generates structured task configuration file for CFN Task Mode to ensure product owner has complete scope context.

Purpose

Without explicit scope configuration:

• Product owner cannot distinguish sprint vs epic completion
• No clear in-scope / out-of-scope boundaries
• Missing deliverables list for validation
• No acceptance criteria for quality gates

Config Structure

{
  "taskId": "cfn-phase-[timestamp]",
  "taskDescription": "User's original task description",
  "mode": "mvp|standard|enterprise",
  "spawnMode": "task",
  "scope": {
    "epicGoal": "1-2 sentence high-level objective",
    "inScope": ["Specific achievable objectives"],
    "outOfScope": ["Clear boundaries"],
    "deliverables": ["Exact file paths"],
    "directory": "Target creation path",
    "acceptanceCriteria": ["Measurable requirements"]
  },
  "agents": {
    "loop3": ["implementer-agents"],
    "loop2": ["validator-agents"]
  },
  "thresholds": {
    "gate": 0.75,
    "consensus": 0.90,
    "maxIterations": 10
  },
  "createdAt": "ISO-8601 timestamp"
}

Usage

./.claude/skills/cfn-task-config-init/initialize-config.sh \
  --task-description "Implement JWT authentication" \
  --mode "standard" \
  --task-id "cfn-phase-1730545678"

Returns: Path to created config file

Config Storage

Location: .cfn/task-configs/task-[task-id].json

Lifecycle:

• Created: At CFN Loop Task Mode start
• Read by: Main Chat coordinator, Product Owner (both modes)
• Updated: Never (immutable scope contract)
• Cleaned: After 7 days (configurable TTL)

Scope Extraction

The script analyzes task description to extract:

Epic Goal:

• Summary of high-level objective (1-2 sentences)
• Derived from imperative verbs: "Implement", "Build", "Create"

In-Scope:

• Explicit requirements from task description
• Concrete deliverables mentioned
• Technology stack specified

Out-of-Scope:

• Common scope creep patterns (analytics, admin UI, monitoring)
• Advanced features not mentioned
• Enterprise/scaling concerns (unless mode=enterprise)

Deliverables:

• File paths inferred from task type
• Standard patterns (src/, tests/, docs/)
• Technology-specific conventions (*.ts for TypeScript, *.rs for Rust)

Acceptance Criteria:

• Test coverage thresholds
• Security requirements
• Documentation completeness
• Performance benchmarks (if applicable)

Agent Selection

Based on task analysis:

Keywords → Implementers:

• "API", "backend", "server" → backend-dev
• "React", "Vue", "frontend" → react-frontend-engineer
• "mobile", "iOS", "Android" → mobile-dev
• "infrastructure", "deploy" → devops-engineer
• "package", "npm", "library" → npm-package-specialist

Complexity → Validators:

• Simple (1-2 files): reviewer, tester
• Standard (3-5 files): +architect, +security-specialist
• Complex (>5 files): +code-analyzer, +performance-benchmarker

Integration

In /cfn-loop command:

# Step 1: Initialize config (Task Mode only)
if [[ "$SPAWN_MODE" == "task" ]]; then
  CONFIG_PATH=$(./claude/skills/cfn-task-config-init/initialize-config.sh \
    --task-description "$ARGUMENTS" \
    --mode "$MODE" \
    --task-id "$TASK_ID")
fi

In Main Chat (Task Mode):

// Read config before spawning agents
const config = JSON.parse(Read(configPath));

// Pass scope to all agents
const loop3Context = `
Task: ${config.taskDescription}

Scope:
- Epic Goal: ${config.scope.epicGoal}
- In-Scope: ${config.scope.inScope.join(', ')}
- Out-of-Scope: ${config.scope.outOfScope.join(', ')}
- Deliverables: ${config.scope.deliverables.join(', ')}
- Directory: ${config.scope.directory}
- Acceptance Criteria:
  ${config.scope.acceptanceCriteria.map(c => `  - ${c}`).join('\n')}
`;

In Product Owner:

// Extract context from config
const config = JSON.parse(Read(configPath));

// Make informed decision
if (decision === "PROCEED") {
  // Check if sprint complete vs epic complete
  if (config.scope.sprints && currentSprint < totalSprints) {
    nextAction = `Proceed to Sprint ${currentSprint + 1}`;
  } else {
    nextAction = "Epic complete, all deliverables met";
  }
}

Benefits

Scope Clarity:

• Product owner knows exact boundaries
• Prevents "consensus on vapor" (approving plans without code)
• Clear deliverable verification

Consistent Context:

• All agents receive same scope information
• No context drift between iterations
• Product owner has complete decision context

Sprint Management:

• Distinguishes sprint completion from epic completion
• Tracks progress across multi-sprint epics
• Enables autonomous sprint transitions

Dependencies

• jq for JSON manipulation
• .cfn/task-configs/ directory (auto-created)
• Task Mode only (CLI mode uses Redis context storage)

Error Handling

Missing task description: Exit with error Invalid mode: Default to "standard" Config file exists: Overwrite with warning JSON validation fails: Retry with escaped strings

Output

Success:

✅ Config initialized: .cfn/task-configs/task-cfn-phase-1730545678.json

Error:

❌ Failed to initialize config: [error details]
Exit code: 1