| name | breakdown |
| description | Execution breakdown workflow. Transforms approved designs into detailed task breakdowns for build agents. Use after /ar approval. |
| disable-model-invocation | true |
| allowed-tools | Read, Glob, Grep, Write |
| argument-hint | [path/to/design-doc.md] |
Opus Execution Planning Protocol
You transform approved designs into execution breakdowns that Sonnet build agents can implement independently. Your breakdowns must be so complete that Sonnet build agents never need to ask for clarification.
Phase Tracking
Before any work, create ALL tasks in full detail using TaskCreate. Pass the subject, activeForm, and description from each task below verbatim. Then progress through tasks sequentially — mark in_progress before starting, completed after finishing. Do not begin a task until the prior task is completed.
Task 1: Load and analyze design document
- activeForm: Analyzing design document
- description: Read the design document provided as the
/breakdown argument. Understand the architecture decisions, data flow, and technical approach. Identify the scope, modules involved, complexity, and dependencies between components.
Task 2: Group tasks and define agent sections
-
activeForm: Grouping tasks
-
description: Split the work into groups that:
- Contain ~5 tasks each (flexible, based on logical splits)
- Have no file conflicts between groups (enables parallel execution)
- Keep tightly-coupled tasks together (same module, shared state)
Common groupings: backend core → backend API → frontend, or by independent modules.
If some tasks are dependent on others, be explicit about it and list them in order.
Task 3: Write execution documents
-
activeForm: Writing execution documents
-
description: Use the template at .claude/resources/claude-execution-template.md. Create a directory at PROCESS_DOCS/breakdowns/<feature-name>/ with one self-contained document per build agent.
Directory structure:
PROCESS_DOCS/breakdowns/<feature-name>/
├── 01-<scope>.md # Build Agent 1
├── 02-<scope>.md # Build Agent 2
└── 03-<scope>.md # Build Agent 3
Each doc is the build agent's entire context. It must contain everything that agent needs to implement its tasks without reading anything else:
- Title -
# Build Agent N: [Scope/Module]
- Dependencies - Which other agent docs must complete first (by filename), or "None (parallel)"
- Overview - Objective, scope (includes/excludes), dependencies, complexity — scoped to this agent's work
- Technical Approach - Architecture decisions, module placement, integration points, data flow — only what this agent touches
- Task Breakdown - ~5 detailed tasks with:
- Acceptance criteria (specific, measurable)
- Exact file paths
- Dependencies between tasks within this agent
- Code examples showing the pattern
- Test cases — named tests with setup, assertion, and file path. The builder implements them, not invents them.
- Testing Strategy - Framework, structure, coverage targets for this agent's scope (per-task test cases go in the task breakdown, not here)
- Risk Mitigation - 3-5 risks with probability, impact, mitigation, fallback, detection
- Success Criteria - Functional and non-functional requirements for this agent
- Implementation Notes - Gotchas, helpful commands, critical configuration
Use kebab-case for directory and file names (e.g., plugin-restructure/, 01-build-script.md).
Key principles:
- Be specific: Show exact configuration, not "configure Redis"
- Show, don't tell: Provide code examples, not just descriptions
- Measurable criteria: "3 connections per IP" not "reasonable limit"
- Exact paths:
backend/app/security/rate_limiter.py not "in security module"
Task 4: Run quality checklist
Common Planning Mistakes to Avoid
Too Vague
BAD: "Implement WebSocket connection"
Specific
GOOD: "Implement WebSocket endpoint at /ws/voice with:
- Rate limiting (3 connections per IP)
- Session initialization with UUID
- JSON message validation using orjson
- Close code 1008 for policy violations"
Missing Test Details
BAD: "Write tests for connection manager"
BAD: "Tests written and passing" (as acceptance criterion without specifying which tests)
Test Cases Defined at Plan Time
GOOD:
- **Test File**: `tests_managers/test_connection_manager.py`
- **Test Cases**:
- `test_connection_limit_per_ip`: Connect 3 clients from same IP → 4th rejected with 1008
- `test_memory_monitoring`: Mock memory at 86% → triggers cleanup, verify oldest connection dropped
- `test_lru_eviction`: Connect A, B, C in order → evict → A removed, B and C remain
- `test_concurrent_connections`: 10 simultaneous connect attempts → exactly MAX_TOTAL accepted
- **Framework**: pytest-asyncio with mock WebSocket fixtures
- **Setup**: `conftest.py` fixture providing `ConnectionManager` with `MAX_CONNECTIONS_PER_IP=3`
The builder implements these exact tests. If the planner can't name the test cases, the acceptance criteria aren't specific enough.
Ambiguous Paths
BAD: "Create rate limiter in security module"
Exact Paths
GOOD: "Create rate limiter at backend/app/security/rate_limiter.py"
Success Metrics
Your plan is successful when:
- Sonnet never asks "where should this go?"
- Sonnet never asks "what framework should I use?"
- Sonnet never asks "how should I handle this error?"
- All acceptance criteria are met on first implementation
- Tests pass without modification
- Code follows all style guidelines
Remember
- Over-specify rather than under-specify - Sonnet can ignore extra detail, but can't guess missing detail
- Show, don't tell - Provide code examples, not just descriptions
- Think like Sonnet - What would you need to know to implement this without any context?
- Test your plan - Read it as if you knew nothing about the project