mcpField
D
decomposing-tasks

decomposing-tasks

github.com/dralgorhythmDevelopment12865 installsUpdated 14w ago
Repository

About decomposing-tasks

A task-decomposition skill for development work. It guides teams through scoping, identifying components, sequencing dependencies, and checking whether tasks can be demonstrated or tested independently.

Key features

  1. 01Clarifies scope and major components
  2. 02Breaks work into one- or two-day tasks
  3. 03Orders dependencies and execution
  4. 04Uses vertical, horizontal, or SPIDR slicing
  5. 05Defines acceptance and completion criteria

Use cases

  1. 01Split a large feature into testable tasks
  2. 02Plan implementation order and dependencies
  3. 03Write acceptance criteria for tasks

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/dralgorhythm/claude-agentic-framework --skill decomposing-tasks

After installing, type/decomposing-tasksin Claude Code to confirm it loaded.

Details

From the project README

Decomposing Tasks

Workflows

  • Understand Scope: What is the full scope of work?
  • Identify Components: What are the major pieces?
  • Break Down: Split into 1-2 day tasks
  • Order: Define dependencies and sequence
  • Validate: Can each task be demo'd/tested?

Task Size Guidelines

Too Big
  • "Build the authentication system"
  • Cannot be completed in 1-2 days
  • Cannot be tested independently
Just Right
  • "Create login API endpoint"
  • "Add password validation"
  • "Implement JWT token generation"
Too Small
  • "Add semicolon"
  • "Rename variable"
  • Not worth tracking separately

Decomposition Strategies

Vertical Slicing

Cut through all layers for one feature.

Feature: User Login
├── API endpoint (POST /login)
├── Input validation
├── Database query
├── JWT generation
├── Error handling
└── Integration test
Horizontal Slicing

Complete one layer at a time.

Database Layer
├── Users table
├── Sessions table
└── Indexes

API Layer
├── Login endpoint
├── Logout endpoint
└── Refresh endpoint
SPIDR Method
  • Spike: Research/investigation task
  • Path: Main flow implementation
  • Interface: API/UI contract
  • Data: Storage and schemas
  • Rules: Business logic and validation

Task Template

## Task: [Title]

**Description**: [What needs to be done]

**Acceptance Criteria**:
- [ ] [Criterion 1]
- [ ] [Criterion 2]

**Dependencies**: [What must be done first]

**Estimate**: [X hours/points]

Definition of Done

Every task should have:

  • Code complete and reviewed
  • Tests written and passing
  • Documentation updated
  • Deployed to staging

Comments

Markdown supported