# Systems Decompose

> Decompose a feature into system components before implementation

- Skill: `sunnypatneedi/systems-decompose` (Agent Skill)
- Install (CLI): `npx skillmds@latest add sunnypatneedi/systems-decompose`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sunnypatneedi/systems-decompose/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: sunnypatneedi (https://skillmd.com/u/sunnypatneedi)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/sunnypatneedi/systems-decompose

---


# Systems Decomposition

Before implementing, let's properly decompose this feature:

## Context

1. **What are we building?** Describe the feature.
2. **Who triggers it?** User action, system event, scheduled?
3. **What should happen?** The happy path outcome.

## Data Flow Mapping

Map the complete flow:

```
[Trigger] → [Input] → [Validation] → [Business Logic] → [Side Effects] → [Output]
```

For each step:

- What data comes in?
- What transformation happens?
- What data goes out?
- What can go wrong?

## Interface Contracts

Define explicit interfaces:

### Input Schema

- What fields are required?
- What are the types and constraints?
- Validation rules?

### Output Schema

- Success response format?
- Error response format?
- All possible error types?

### Error Enumeration

List ALL possible errors:

- Validation errors
- Not found errors
- Authorization errors
- External service errors
- System errors

## Boundaries

Clarify ownership:

- What does this component OWN?
- What does it NOT own?
- What are the integration points?

## Dependencies

For each external dependency:

- What is it?
- Sync or async?
- What if it fails?
- How do we recover?

This decomposition will guide implementation.