slack-auth-1770276413/.claude/commands/breakdown-feature.md
jordan 5812d7a868
All checks were successful
ci/woodpecker/manual/woodpecker Pipeline was successful
ci/woodpecker/push/woodpecker Pipeline was successful
Initialize project from skeleton template
2026-02-05 07:26:54 +00:00

1.8 KiB

description argument-hint allowed-tools
Break a feature into implementation tasks <feature-slug> Bash, Read, Write, Edit, Glob, Grep

Break down feature into tasks: $ARGUMENTS

Instructions

1. Load Feature, Spec, and Design

sdlc feature show $ARGUMENTS --json

Read both:

  • .sdlc/features/$ARGUMENTS/spec.md
  • .sdlc/features/$ARGUMENTS/design.md

Both must exist before creating a task breakdown.

2. Identify Implementation Units

From the design, identify discrete units of work. Each task should:

  • Be completable in a single coding session
  • Have clear inputs and outputs
  • Be independently testable
  • Have minimal overlap with other tasks

3. Create Tasks via CLI

For each task, register it:

sdlc task add $ARGUMENTS "Task title - brief description"

Order tasks by dependency -- foundational work first, integration last.

4. Write Detailed Task Descriptions

Write to .sdlc/features/$ARGUMENTS/tasks.md:

# Tasks: [Feature Title]

## Task Order (dependency sequence)

### T1: [Title]
- **Scope:** [What exactly to implement]
- **Files:** [Expected files to create/modify]
- **Depends on:** [None / T-id]
- **Acceptance criteria:**
  - [ ] [Specific, testable criterion]

### T2: [Title]
...

5. Register the Artifact

sdlc artifact create $ARGUMENTS tasks

6. Report

List all tasks in order, total count, and estimated dependency chain depth (longest path through the task graph).

Critical Rules

  • NEVER create tasks without reading both spec and design
  • ALWAYS order tasks by dependency -- no task should reference work not yet done
  • Each task MUST be completable in a single session
  • ALWAYS include acceptance criteria per task
  • NEVER create monolithic tasks -- split until each is focused