mcpField
T
task-decomposition

task-decomposition

github.com/d-o-hubAgent workflows1337 installsUpdated 1d ago
Repository

About task-decomposition

A planning skill for agents that turns complex requests into actionable tasks with dependencies, success criteria, and opportunities for parallel work. It also presents decomposition as the first phase of a GOAP execution plan.

Key features

  1. 01Requirements and constraint analysis
  2. 02Hierarchical task breakdown
  3. 03Sequential and parallel dependencies
  4. 04Measurable task success criteria
  5. 05GOAP planning integration

Use cases

  1. 01Breaking down multi-part requests
  2. 02Coordinating multi-phase projects
  3. 03Finding tasks suitable for parallel work

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/d-o-hub/rust-self-learning-memory --skill task-decomposition

After installing, type/task-decompositionin Claude Code to confirm it loaded.

Details

From the project README

Task Decomposition

Break down complex tasks into atomic, actionable goals with clear dependencies.

When to Use

  • Complex user requests with multiple components
  • Multi-phase projects requiring coordination
  • Tasks that could benefit from parallel execution
  • Planning agent coordination strategies

Decomposition Framework

1. Requirements Analysis
  • Primary objective
  • Implicit requirements (quality, performance)
  • Constraints (time, resources)
  • Success criteria
2. Goal Hierarchy
Main Goal
├─ Sub-goal 1
│  ├─ Task 1.1 (atomic)
│  └─ Task 1.2 (atomic)
├─ Sub-goal 2
└─ Sub-goal 3
3. Dependency Types
TypeSymbolExample
SequentialA → B → CB needs A's output
ParallelA─┐ B─┐ C─┘Independent, concurrent
ConvergingA─┐ B─┼─> DD needs A, B, C
ResourceA, BSequential or pooled
4. Success Criteria

For each task:

  • Input: What data/state is needed
  • Output: What artifacts will be produced
  • Quality: Performance, testing, docs requirements

Decomposition Patterns

PatternUse Case
Layer-BasedArchitectural changes (data, logic, API, test, docs)
Feature-BasedNew features (MVP, error handling, optimization, integration)
Phase-BasedLarge projects (research, foundation, core, integration, polish)
Problem-SolutionDebugging (reproduce, diagnose, design, fix, verify, prevent)

Quality Checklist

✓ Atomic and actionable ✓ Dependencies clearly identified ✓ Success criteria measurable ✓ No task too large (>4 hours) ✓ Parallelization opportunities identified

✗ Tasks too large or vague ✗ Missing dependencies ✗ Unclear success criteria ✗ Missing quality/testing tasks

Integration with GOAP

Task decomposition is Phase 1 of GOAP:

  1. Receive request
  2. Apply decomposition
  3. Create execution plan
  4. Execute with monitoring
  5. Report results

Comments

Markdown supported