Work Breakdown Skill

Break down implementation plans into bd-compatible issues. This skill produces epics, tasks, dependencies, and estimates that can be directly executed via bd commands.

When to Invoke

• User provides an implementation plan and asks to break it down

Workflow

Step 1: Read the Implementation Plan

If the user references a file, read it first. Extract:

• Objective: What is the end goal?
• Phases: Are there distinct implementation phases?
• Deliverables: What outputs are expected?
• Dependencies: What depends on what?

Step 2: Identify Epics (Work Packages)

Group work into 3-7 epics. Common patterns:

• Phase 0: Setup/Infrastructure
• Phase 1-N: Feature implementation phases
• Testing & Quality
• Documentation & Deployment

Step 3: Decompose into Tasks

For each epic, create tasks that are:

• 30-480 minutes (0.5-8 hours) of focused work
• Deliverable-oriented: produces a testable output
• Single-responsibility: one executor can complete it
• Ensure your tasks are aligned with the development guidelines in @docs/MAINTENANCE.md

Step 4: Map Dependencies

Identify blocking relationships:

• Task B cannot start until Task A completes
• Epic 2 depends on Epic 1 completion
• Testing depends on implementation

Step 5: Assign Estimates and Priorities

Use bd-compatible values:

• Priority: 0 (critical) to 4 (backlog)
• Estimate: minutes (60=1hr, 240=4hrs)
• Type: epic, task, feature, bug, chore

Step 6: Present WBS for Approval

Show the user:

1. Summary table with all tasks
2. Dependency graph (text representation)
3. Parallel opportunities
4. Total effort estimate

Ask: "Ready to create these issues in beads?"

Step 7: Execute bd Commands

After user approval, execute commands in this order:

# 1. Create epics first (capture IDs from --json output)
bd create "Epic Name" -t epic -p [priority] -e [total_minutes] -d "Description" --json

# 2. Create tasks as children of epics
bd create "Task Name" -t task -p [priority] -e [minutes] --parent [epic_id] -d "Description" --json

# 3. Add dependencies between tasks
bd dep add [dependent_id] [blocker_id]

# 4. Show ready work
bd ready

Output Format

Present the WBS as a markdown table:

## Work Breakdown: [Project Name]

### Summary
- Total: N tasks across M epics
- Effort: X-Y hours
- Critical path: Task1 -> Task2 -> Task3

### Epic 1: [Name] (P1, ~Xhr)

| ID | Task | Type | Pri | Est | Depends | Risk |
|----|------|------|-----|-----|---------|------|
| 1.1 | [Title] | task | P2 | 120m | - | Low |
| 1.2 | [Title] | task | P2 | 180m | 1.1 | Med |

### Epic 2: [Name] (P2, ~Xhr)
...

bd Quick Reference

Field	Flag	Values
type	-t	epic, task, feature, bug, chore
priority	-p	0-4 (0=critical)
estimate	-e	minutes
parent	--parent	epic ID
description	-d	text
labels	-l	comma-separated

Dependencies: bd dep add [child] [blocker] - child depends on blocker

Key Rules

1. Always get approval before creating issues
2. Create epics first to get their IDs for --parent
3. Use --json flag to capture created issue IDs
4. Add dependencies after all tasks exist
5. Run bd ready at the end to show actionable work