home / skills / dasien / claudemultiagenttemplate / work-breakdown

This skill breaks down features into actionable tasks with scope, dependencies, and clear acceptance criteria for implementable work.

npx playbooks add skill dasien/claudemultiagenttemplate --skill work-breakdown

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
3.5 KB
---
name: "Work Breakdown"
description: "Break down features and requirements into actionable, implementable tasks with clear scope and acceptance criteria"
category: "analysis"
required_tools: ["Read", "Write", "Grep"]
---

# Work Breakdown

## Purpose
Break down large features, enhancements, or requirements into small, actionable tasks that can be implemented independently with clear success criteria.

## When to Use
- Planning implementation of new features
- Breaking down enhancement specifications
- Creating task lists for workflows
- Organizing complex work into manageable pieces
- Planning sprints or iterations

## Key Capabilities
1. **Task Identification** - Identify discrete, implementable work units
2. **Dependency Analysis** - Determine task order and relationships
3. **Scope Definition** - Create clear, bounded task descriptions

## Approach
1. **Read Requirements** - Understand the full scope of work
2. **Identify Components** - What parts of the system are affected?
3. **Break into Phases** - Architecture → Implementation → Testing → Documentation
4. **Define Tasks** - Each should be:
   - Completable in 2-8 hours
   - Independently testable
   - Has clear acceptance criteria
   - Assigned to appropriate agent
5. **Order by Dependencies** - What must happen first?
6. **Validate Completeness** - Do tasks cover all requirements?

## Example
**Context**: Feature - "Add user authentication with OAuth"

**Work Breakdown**:
```markdown
## Phase 1: Architecture & Design
**Task 1.1**: Design authentication flow
- Agent: architect
- Duration: 2-3 hours
- Output: Authentication sequence diagram
- Acceptance: Covers login, logout, token refresh

**Task 1.2**: Design database schema for users/sessions
- Agent: architect
- Duration: 2 hours
- Output: Schema definition with indexes
- Acceptance: Normalized, handles OAuth tokens

## Phase 2: Implementation
**Task 2.1**: Implement OAuth integration
- Agent: implementer
- Duration: 4-6 hours
- Dependencies: Task 1.1 complete
- Output: OAuth client code
- Acceptance: Can authenticate with Google/GitHub

**Task 2.2**: Create user session management
- Agent: implementer
- Duration: 3-4 hours
- Dependencies: Task 1.2, 2.1 complete
- Output: Session middleware
- Acceptance: Sessions persist, timeout after 24hrs

**Task 2.3**: Add authentication middleware
- Agent: implementer
- Duration: 2-3 hours
- Dependencies: Task 2.2 complete
- Output: Route protection middleware
- Acceptance: Protected routes require auth

## Phase 3: Testing
**Task 3.1**: Write authentication tests
- Agent: tester
- Duration: 3-4 hours
- Dependencies: Phase 2 complete
- Output: Test suite
- Acceptance: >90% coverage, tests auth flows

## Phase 4: Documentation
**Task 4.1**: Document OAuth setup
- Agent: documenter
- Duration: 2 hours
- Dependencies: Phase 2 complete
- Output: Setup guide
- Acceptance: Includes provider config steps
```

**Why This Breakdown Works**:
- Each task is 2-6 hours (right-sized)
- Clear dependencies prevent blocking
- Phases group related work
- Each has specific agent and acceptance criteria
- Can be implemented/tested independently

## Best Practices
- ✅ Tasks should be 2-8 hours of work (not days)
- ✅ Each task has clear acceptance criteria
- ✅ Assign appropriate agent to each task
- ✅ Group related tasks into phases
- ✅ Identify dependencies explicitly
- ✅ Ensure tasks cover all requirements
- ❌ Avoid: Tasks that are too large (>1 day)
- ❌ Avoid: Vague acceptance criteria ("make it work")
- ❌ Avoid: Circular dependencies between tasks

Overview

This skill breaks down features, enhancements, and requirements into small, actionable tasks with clear scope and acceptance criteria. It produces ordered, right-sized work items, identifies dependencies, and assigns responsibility so teams or agents can implement independently. The output focuses on implementability and testability.

How this skill works

I read the full requirement or feature description and identify affected system components. I split work into phases (architecture, implementation, testing, documentation), create discrete tasks that are completable in 2–8 hours, and attach clear acceptance criteria. I then order tasks by dependencies and validate coverage so nothing essential is left unaddressed.

When to use it

  • Planning a new feature implementation
  • Breaking an enhancement into execution-ready work
  • Creating sprint or iteration task lists
  • Organizing multi-agent or cross-team workflows
  • Preparing handoffs between architects, implementers, and testers

Best practices

  • Keep tasks 2–8 hours each; avoid multi-day work items
  • Give each task explicit acceptance criteria and expected output
  • Assign an appropriate owner or agent for each task
  • Group related tasks into phases (architecture → implementation → testing → docs)
  • Explicitly list dependencies and ordering to prevent blocking
  • Validate that task set covers all requirements and edge cases

Example use cases

  • Add OAuth-based user authentication split into design, implementation, tests, and docs
  • Turn a high-level performance improvement requirement into measurable sub-tasks with benchmarks
  • Create a rollout plan for a new API version with migration tasks and compatibility tests
  • Prepare a sprint backlog from a product spec, with clear owners and acceptance for each item
  • Decompose cross-service feature into service-level tasks, integration tests, and deployment steps

FAQ

How granular should each task be?

Aim for tasks that a single person or agent can finish in 2–8 hours and independently test. If it takes longer, split further.

What counts as good acceptance criteria?

Concrete, testable statements such as expected outputs, pass/fail conditions, performance targets, or required configuration steps.