cheese-factory

Parallelize approved multi-part specifications into file-disjoint curds with wiring DAG integration.

15|1|Updated May 4, 2026
One-click install
npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheese-factory
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cheese-factory
Source: https://github.com/paulnsorensen/easy-cheese/tree/main/skills/cheese-factory
Command: npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheese-factory

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Repetitively implementing a multi-step product spec serially slows delivery and makes it harder to review changes safely when the work can be cleanly split into independent behavioral chunks.

Core Features & Use Cases

  • Spec decomposition into curds, seed, and wiring: turns an approved spec into 5+ file-disjoint behavioral curds plus integration wiring tasks.
  • Parallel execution with guardrails: runs per-curd work as independent sub-agent pipelines and validates with focused test targets and project quality gates.
  • Fan-out → fan-in integration with conflict handling: merges curds in order, applies wiring sequentially via a DAG, and halts on wiring/merge conflicts rather than guessing.
  • Post-merge review and PR planning: performs press/age/cure integration review in fresh context and produces a PR layout (single, orthogonal, stacked, or diamond) for publishing 1–N reviewable PRs.

Quick Start

Ask your agent to run the pipeline for an approved spec by invoking the command: /cheese-factory .cheese/specs/your-spec-slug.md

Frequently Asked Questions about cheese-factory

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

FAQPage Schema
How do I parallelize a multi-step spec into independent pull requests?

Parallelizing a spec into independent pull requests requires decomposing the approved specification into five or more file-disjoint behavioral units, executing them as independent sub-agent pipelines, and integrating the results through a wiring DAG to produce reviewable PRs.

What is the best way to split a large development spec for parallel execution?

Splitting a large development spec involves breaking it down into file-disjoint behavioral curds mapped to acceptance criteria, allowing each unit to be developed and validated independently before being merged in order through a conflict-aware integration DAG.

How does spec decomposition handle merge conflicts during parallel development?

Spec decomposition handles merge conflicts during parallel development by merging curds sequentially and applying wiring through a DAG, halting execution immediately upon encountering wiring or merge conflicts rather than guessing the resolution.

Can I use parallel development for specs with fewer than five acceptance criteria?

Parallel development using this approach applies to scenarios where the spec can be decomposed into five or more acceptance-criterion-mapped units, ensuring enough independent file-disjoint behavioral chunks exist to justify fan-out execution.

Why does parallel spec execution produce multiple PRs instead of a monolithic change?

Parallel spec execution produces multiple PRs instead of a monolithic change because it performs post-merge integration review in fresh context and generates a PR layout—single, orthogonal, stacked, or diamond—designed for publishing safely reviewable grouped commits.

What quality gates are applied during parallel spec implementation?

Quality gates applied during parallel spec implementation include phase-gated execution with focused test targets and project quality checks, validating each behavioral curd through independent sub-agent pipelines before fan-in integration.