parallel-decompose

PRD Decomposition Workflow

Decompose a PRD and Tech Spec into parallel-executable tasks with contracts, dependency graphs, and agent prompts.

Workflow Overview

flowchart TB
    Start[("PRD + Tech Spec")]
    Dir["1. Determine Output Dir"]
    Analyze["2. Analyze Tech Spec"]
    Manifest["3. Create manifest.json"]
    Context["4. Create context.md"]
    Arch["5. Create architecture.md"]
    Contracts["6. Create contracts/"]
    Tasks["7. Generate task files"]
    Graph["8. Create task-graph.md"]
    Prompts["9. Generate prompts/"]
    Update["10. Update manifest"]
    Report["11. Report Results"]

    Start --> Dir
    Dir --> Analyze
    Analyze --> Manifest
    Manifest --> Context
    Context --> Arch
    Arch --> Contracts
    Contracts --> Tasks
    Tasks --> Graph
    Graph --> Prompts
    Prompts --> Update
    Update --> Report
    Report --> Done(["Ready for cpo run"])

Step 1: Determine Output Directory

PARALLEL_DIR="parallel/TS-0042-inventory-system"

• If Tech Spec provided: Extract ID and slug from Tech Spec file
• If --name provided: Use parallel/{slug}/

Step 2: Tech Spec Analysis

If a Tech Spec (TS-XXXX) is provided:

1. Read and validate Tech Spec:

   • Verify status is APPROVED or REFERENCE
   • Warn if DRAFT (not ready for decomposition)

2. Extract from Tech Spec:

   • Design Overview -> Use for architecture.md
   • Data Model -> Use for contracts/types.py
   • API Specification -> Use for contracts/api-schema.yaml
   • Component boundaries -> Use for task ownership

3. Skip redundant steps:

   • If Tech Spec has complete Data Model, skip generating types
   • If Tech Spec has complete API Spec, skip generating api-schema.yaml

If no Tech Spec provided: Generate contracts from PRD and display warning:

"Consider creating a Tech Spec first for better contract definitions"

Step 3: Create manifest.json

Use the cpo format:

{
  "tech_spec_id": "TS-0042",
  "name": "inventory-system",
  "technology": "python",
  "python_version": "3.11",
  "waves": [
    {
      "number": 1,
      "tasks": [
        { "id": "task-001", "agent": "python-experts:django-expert" },
        { "id": "task-002", "agent": "python-experts:django-expert" }
      ],
      "validation": "from apps.users.models import User; print('Wave 1 OK')"
    }
  ],
  "metadata": {
    "tech_spec": "tech-specs/approved/TS-0042-inventory.md",
    "generated_at": "2025-01-15T10:00:00Z",
    "total_tasks": 3,
    "max_parallel": 2,
    "critical_path": ["task-001", "task-003"]
  }
}