Back to Skills

task-decomposition

verified

Break down PRD and Architecture into implementable engineering tasks. Use when user wants to create a task list, plan implementation work, decompose features into stories and tasks, or needs to translate requirements into actionable engineering work. Reads from docs/PRD.md and docs/ARCHITECTURE.md, outputs docs/TASKS.md.

View on GitHub

Marketplace

porkbutts-n-taters

Porkbutts/porkbutts-n-taters

Plugin

vibe-coding

Repository

Porkbutts/porkbutts-n-taters
1stars

skills/task-decomposition/SKILL.md

Last Verified

January 21, 2026

Install Skill

Select agents to install to:

Scope:
npx add-skill https://github.com/Porkbutts/porkbutts-n-taters/blob/main/skills/task-decomposition/SKILL.md -a claude-code --skill task-decomposition

Installation paths:

Claude
.claude/skills/task-decomposition/
Powered by add-skill CLI

Instructions

# Task Decomposition

Transform a PRD and Architecture document into a structured task list with epics, user stories, and atomic engineering tasks. Read `docs/PRD.md` and `docs/ARCHITECTURE.md`, analyze the work required, and produce `docs/TASKS.md` with tasks that are independently implementable and testable.

## Workflow

```
START
  │
  ▼
┌─────────────────────────────┐
│ Read docs/PRD.md            │ ◄── Load requirements
│ Read docs/ARCHITECTURE.md   │ ◄── Load technical design
└──────────┬──────────────────┘
           │
           ▼
┌─────────────────────────────┐
│ Identify Epics              │ ◄── Group by major feature areas
└──────────┬──────────────────┘
           │
           ▼
┌─────────────────────────────┐
│ Break into User Stories     │ ◄── Define user-facing value units
└──────────┬──────────────────┘
           │
           ▼
┌─────────────────────────────┐
│ Decompose into Tasks        │ ◄── Atomic, testable engineering work
└──────────┬──────────────────┘
           │
           ▼
┌─────────────────────────────┐
│ Confirm with user           │ ◄── Review structure, get approval
└──────────┬──────────────────┘
           │
           ▼
┌─────────────────────────────┐
│ Generate TASKS.md           │ ◄── Write to docs/TASKS.md
└─────────────────────────────┘
```

## Step 1: Read Input Documents

Read both documents to understand:
- **From PRD**: User scenarios, acceptance criteria, scope boundaries, roles
- **From Architecture**: Tech stack, directory structure, key abstractions, data flow

If either file is missing, ask the user to provide it or suggest using the appropriate skill first (product-design for PRD, architecture for ARCHITECTURE.md).

## Step 2: Identify Epics

Group work into 3-7 epics based on:
- Major feature areas from the PRD
- Architectural boundaries (frontend, backend, integrations)
- User roles or workflows

Each epic should represent a cohesive chunk of functionality that could be released independently.

## Step 3: Break in

Validation Details

Front Matter
Required Fields
Valid Name Format
Valid Description
Has Sections
Allowed Tools
Instruction Length:
6849 chars