Task Decomposition

OAK Task Decomposition

This skill provides expertise in breaking down OAK implementation plans into actionable, well-structured tasks suitable for the tasks.md format and export to issue trackers.

OAK Task System Overview

Tasks in the OAK Workflow

/oak.plan-create    /oak.plan-research    /oak.plan-tasks      /oak.plan-export
      │                    │                    │                     │
      ↓                    ↓                    ↓                     ↓
  plan.md           research/*.md          tasks.md    ────►   GitHub Issues
                                                               or ADO Work Items
                                               │
                                               ↓
                                       /oak.plan-implement
                                         (execute tasks)

File Structure

oak/plan/<plan-name>/
├── plan.md              # Goals, scope, research topics
├── research/            # Research findings
├── tasks.md             # Generated task breakdown (THIS FILE)
└── .manifest.json       # Tracks task completion state

When to Use This Skill

Use when you need to:

• Generate tasks.md from plan and research (/oak.plan-tasks)
• Structure work for sprint planning
• Define task dependencies and ordering
• Create acceptance criteria aligned with plan goals
• Prepare tasks for export to GitHub/ADO

OAK Task Hierarchy

Level 1: Epic

• Theme-level grouping (maps to plan Goals)
• Multiple sprints of work
• Business objective alignment

Level 2: Story

• User-visible functionality
• Single sprint deliverable
• Clear acceptance criteria

Level 3: Task

• Technical implementation unit
• Hours to a few days of work
• Independently testable

Level 4: Subtask

• Granular work item
• Single session of work
• Part of larger task

OAK tasks.md Format

This is the expected format for oak/plan/<name>/tasks.md:

# Tasks: [Plan Name]

## Summary
[Brief overview of task breakdown approach]

## Epic 1: [Epic Title - from plan Goal]

### Story 1.1: [Story Title]
**Priority**: High/Medium/Low
**Estimate**: S/M/L or points
**Depends on**: [Story IDs if any]

#### Acceptance Criteria
- [ ] [Criterion 1 - verifiable]
- [ ] [Criterion 2 - from plan success criteria]
- [ ] Tests pass per constitution requirements

#### Tasks
- [ ] **Task 1.1.1**: [Action verb] [specific what] (S)
  - File: `path/to/file.py`
  - Details: [Implementation notes from research]

- [ ] **Task 1.1.2**: [Action verb] [specific what] (M)
  - Depends on: 1.1.1
  - File: `path/to/file.py`
  - Pattern: [Reference similar implementation]

### Story 1.2: [Story Title]
**Priority**: Medium
**Estimate**: M

#### Acceptance Criteria
- [ ] [Criterion]

#### Tasks
- [ ] **Task 1.2.1**: [Task description] (S)

## Epic 2: [Epic Title]
...

## Dependencies Graph

1.1.1 → 1.1.2 → 1.2.1 ↓ 1.1.3 → 2.1.1

## Testing Tasks (Constitution-Driven)

Per constitution testing requirements:
- [ ] Unit tests for [component] - coverage target: [%]
- [ ] Integration tests for [workflow]
- [ ] [TDD note if constitution requires test-first]

## Documentation Tasks

Per constitution documentation requirements:
- [ ] Update [specific docs]
- [ ] Add inline comments for complex logic
- [ ] Update API docs (if applicable)

## Notes
- [Important consideration from research]
- [Risk mitigation approach]

Task Generation from Plan + Research

Step 1: Map Goals to Epics

# From plan.md:
## Goals
- Implement user authentication       → Epic 1: Authentication
- Add API rate limiting               → Epic 2: Rate Limiting
- Create admin dashboard              → Epic 3: Admin Dashboard

Step 2: Derive Stories from Scope

# From plan.md:
## Scope
### In Scope
- OAuth2 integration                  → Story 1.1: OAuth2 Setup
- Session management                  → Story 1.2: Session Handling
- JWT token refresh                   → Story 1.3: Token Management

Step 3: Inform Tasks from Research

# From research/auth-patterns.md:
## Recommendation
Use existing AuthService pattern at src/services/auth.py

# Becomes:
- [ ] **Task 1.1.1**: Extend AuthService with OAuth2 provider (M)
  - File: `src/services/auth.py`
  - Pattern: Follow existing provider pattern (lines 45-80)

Step 4: Apply Constitution Requirements

# From oak/constitution.md:
## Testing
- MUST have 80% coverage for new code
- SHOULD use TDD for complex logic

# Add to tasks:
## Testing Tasks (Constitution-Driven)
- [ ] Write tests BEFORE implementation (TDD per constitution)
- [ ] Achieve 80% coverage for auth module

OAK Phased Task Structure

Standard OAK phases (adjust based on constitution):

Phase 1: Setup & Investigation

- [ ] Setup: Review plan context and research findings
- [ ] Setup: Identify affected modules from codebase exploration
- [ ] Setup: Configure dependencies per constitution

Phase 2: Core Implementation (or Phase 3 if TDD)

- [ ] Implement: [Requirement 1]
  - File: [path]
  - Function: [name]
  - Pattern: [reference]

Phase 3: Testing (or Phase 2 if TDD)

- [ ] Test: [Component] unit tests
  - File: `tests/unit/test_[component].py`
  - Coverage: [target %]

Phase 4: Integration

- [ ] Integration: Connect with [related system]
- [ ] Integration: End-to-end workflow verification

Phase 5: Polish & Documentation

- [ ] Documentation: Update [files per constitution]
- [ ] Quality: Run linters per constitution
- [ ] Quality: Final constitution compliance check

Task Quality Checklist

Every task should have:

• Clear title - Action verb + specific what
• File reference - Where work happens
• Size estimate - S/M/L or points
• Dependencies - What must come first
• Acceptance criteria at story level

Estimation Guidelines (OAK Convention)

T-Shirt Sizing

Story Points (Alternative)

Export Considerations

Tasks in tasks.md can export to:

GitHub Issues (/oak.plan-export)

• Epics → Milestones or Labels
• Stories → Issues with acceptance criteria
• Tasks → Checklist items in issue body

Azure DevOps (/oak.plan-export)

• Epics → Epics
• Stories → User Stories
• Tasks → Tasks linked to stories

Best Practices

1. INVEST criteria - Independent, Negotiable, Valuable, Estimatable, Small, Testable
2. Vertical slices - Deliver user value, not horizontal layers
3. Right-size tasks - S/M preferred, L should be rare
4. Front-load risk - Put unknown work early
5. Include testing - Tests are not optional add-ons
6. Reference files - Always include specific paths

Integration with OAK Commands

Quick Reference

• Tasks location: oak/plan/<name>/tasks.md
• State tracking: oak/plan/<name>/.manifest.json
• Input: plan.md goals + research/*.md findings
• Constitution: Check oak/constitution.md for test/doc requirements
• Export targets: GitHub Issues, Azure DevOps Work Items