| name | task-decomposition |
| description | Break down complex tasks into atomic, actionable goals with clear dependencies and success criteria. Use when planning multi-step projects, coordinating agents, or decomposing complex requests. Triggers: "break down task", "decompose", "plan steps", "task breakdown".
|
| category | agent |
| license | MIT |
| metadata | {"author":"d-oit","version":"0.2.10","adapted-from":"d-o-hub/github-template-ai-agents"} |
Task Decomposition
Decompose high-level objectives into manageable, testable sub-tasks.
When to Use
- Complex requests with multiple components
- Multi-phase projects requiring coordination
- Tasks benefiting from parallel execution
Framework
1. Requirements Analysis
Extract: Primary objective, implicit requirements, constraints, success criteria.
2. Goal Hierarchy
Main Goal
├─ Sub-goal 1 → Tasks 1.1, 1.2
├─ Sub-goal 2 → Task 2.1
└─ Sub-goal 3
Atomic Criteria: Single action, defined inputs/outputs, one agent, testable.
3. Dependency Mapping
- Sequential: A → B → C
- Parallel: A, B, C (independent)
- Converging: A, B, C → D
4. Success Criteria
Define: Inputs, outputs, quality standards.
Process
Step 1: Understand Goal
Request: [Original]
Goal: [Main objective]
Type: [Implementation/Debug/Refactor]
Complexity: [Simple/Medium/Complex]
Step 2: Identify Components (3-7)
Step 3: Decompose Components
Step 4: Map Dependencies
Step 5: Assign Priorities
- P0 (Critical), P1 (Important), P2 (Nice-to-have)
Step 6: Estimate Complexity
- Low (<30min), Medium (30min-2hr), High (>2hr)
Patterns
Layer-Based (Rust)
- Data types / schemas
- Business logic
- API / CLI interface
- Integration tests
- Documentation
Feature-Based
- Core (MVP)
- Error handling (
thiserror/anyhow)
- Performance (benches)
- Integration
- Testing
- Docs
Problem-Solution
- Reproduce
- Diagnose
- Design
- Fix
- Verify
- Prevent
Quality Checklist
- Atomic and actionable tasks
- Dependencies identified
- Measurable success criteria
- Appropriate complexity estimates
- No task >4 hours
- Parallelization opportunities found
Rationalizations
| Rationalization | Reality |
|---|
| "I already know what needs to be done, no need to decompose" | Skipping decomposition leads to scope creep, missed dependencies, and rework. |
| "The task is too simple to break down" | Even simple tasks benefit from explicit success criteria and dependency mapping. |
Red Flags