feature-factory

Decomposes feature specs into slices with tests and parallel subagents.

Updated Mar 9, 2026
One-click install
npx skills add https://github.com/alasdairnc/casedive --skill feature-factory
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: feature-factory
Source: https://github.com/alasdairnc/casedive/tree/main/.claude/skills/feature-factory
Command: npx skills add https://github.com/alasdairnc/casedive --skill feature-factory

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Decompose large feature specs into independent slices, write failing tests first per slice, then dispatch parallel subagents to implement each slice until green. This approach reduces risk, accelerates delivery, and clarifies ownership across a multi-file feature.

Core Features & Use Cases

  • Slice the spec into 3–6 bounded pieces with clear contracts.
  • Write red tests for every slice before implementation to ensure test-driven discipline.
  • Dispatch parallel subagents to implement slices concurrently and iterate to green.
  • Validate integration by running the full test suite once all slices are green.

Quick Start

Decompose a feature specification into 3–6 independent slices, write failing tests for each slice, then dispatch parallel subagents to implement until all slices pass.

Frequently Asked Questions about feature-factory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decompose a feature spec into parallel test-driven slices?

You decompose a feature spec by identifying 3–6 independent slices and establishing clear contracts for each. This reduces risk and clarifies ownership across a multi-file feature before writing tests.

What is the best way to parallelize feature implementation using subagents?

The best way to parallelize implementation is to write failing tests for every slice first, then dispatch parallel subagents to iterate concurrently. Subagents work independently until each slice passes its tests.

How do I prevent cross-slice dependencies when implementing features in parallel?

You prevent cross-slice dependencies by establishing clear contracts with guardrails during spec decomposition. These boundaries keep parallel subagents isolated until the final integration phase.

When do I need to run the full test suite during parallel feature development?

You need to run the full test suite during the integration phase, once all parallel subagents finish and every individual slice is green. This validates that the independent slices work together correctly.

Can I use test-driven development for large multi-file features?

Yes, you can apply test-driven development to large multi-file features by writing red tests for every bounded slice before implementation. This ensures test-driven discipline across concurrent development tracks.

Does parallel subagent development work without writing failing tests first?

No, writing failing tests for every slice before implementation is required. This test-driven discipline ensures each parallel subagent has a clear, verifiable target to iterate until green.