Agent skill
task-decomposition
Use when the Architect is breaking down change requests into implementable tasks, defining acceptance criteria, estimating task size, mapping dependencies, or creating technical sub-tasks for Developer and Integrator.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/task-decomposition-dsivov-ai-development-team
SKILL.md
Task Decomposition Expertise
When This Applies
Apply this guidance when:
- Breaking a CR into implementation tasks
- Creating technical sub-tasks from architecture
- Writing acceptance criteria for tasks
- Mapping task dependencies
- Estimating effort and sizing
Decomposition Process
Step 1: Identify Work Streams
From the CR and architecture, identify:
- Data layer tasks — Models, migrations, storage
- Logic layer tasks — Business logic, services, validation
- Interface layer tasks — APIs, UI, integrations
- Test tasks — Unit tests, e2e tests, test data
- Infrastructure tasks — Config, deployment, CI/CD
Step 2: Size Each Task
A good task is:
- Completable in one session — If it takes more, break it down further
- Independently testable — Can be verified without other tasks
- Clearly scoped — Developer knows exactly what files to touch
- Has clear done criteria — Unambiguous definition of complete
Size Guidelines
| Size | Lines of Code | Files | Complexity |
|---|---|---|---|
| Small | < 50 | 1-2 | Straightforward |
| Medium | 50-200 | 2-5 | Some decisions required |
| Large | 200+ | 5+ | Break down further |
Step 3: Define Dependencies
Map which tasks must complete before others can start:
TASK-001 (data model) ─┬─▶ TASK-003 (API endpoint)
TASK-002 (auth logic) ─┘ │
▼
TASK-004 (integration tests)
Mark dependent tasks in their descriptions. Assign upstream tasks first.
Step 4: Write Acceptance Criteria
Each task needs:
- Given — Starting conditions and context
- When — The action or change being made
- Then — Expected outcome, verifiable result
Example:
Given: User model exists with email field
When: Login endpoint receives valid credentials
Then: Returns JWT token with 24h expiry and user profile
Task Assignment Guidelines
| Task Type | Assign To |
|---|---|
| Feature implementation | Developer |
| Unit/e2e tests | Integrator |
| API design review | Self (Architect) |
| Database migrations | Developer (implement) → Integrator (commit) |
| CI/CD changes | DevOp |
| Deployment | DevOp → Production Engineer |
Common Decomposition Patterns
New Feature
- Define data model → Developer
- Implement business logic → Developer
- Create API endpoint → Developer
- Write unit tests → Integrator
- Write e2e tests → Integrator
- Update documentation → Architect
Bug Fix
- Reproduce and document → Developer
- Implement fix → Developer
- Add regression test → Integrator
- Verify fix → Architect (review)
Refactoring
- Document current behavior with tests → Integrator
- Refactor implementation → Developer
- Verify tests still pass → Integrator
- Review architecture impact → Architect
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?