build-stories

Break plan documents into ordered markdown story files with agent assignments.

Updated Mar 5, 2026
One-click install
npx skills add https://github.com/jaisonerick/spec-plugin --skill build-stories
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: build-stories
Source: https://github.com/jaisonerick/spec-plugin/tree/main/plugins/spec-plugin/skills/build-stories
Command: npx skills add https://github.com/jaisonerick/spec-plugin --skill build-stories

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill breaks a plan and its supporting architecture into a sequence of self-contained, agent-sized story files so an orchestrator can reliably assign work without manual story authoring. It removes the friction of converting high-level specs into executable increments that fit an AI agent's context window and includes testing guidance and acceptance criteria so each increment is verifiable.

Core Features & Use Cases

  • Plan decomposition: Reads the plan spec, plan-specific architecture, overall architecture, and roadmap, then decomposes them into ordered markdown stories.
  • Agent assignment: Marks each story with the appropriate agent type (engineer or designer) and enforces the design→engineer pairing pattern where applicable.
  • Deliverable guidance: Each story includes deliverables, implementation guidance, exact file paths, acceptance criteria, and testing guidance sized for a single AI session.
  • Indexing and ordering: Produces a stories.md index with ordering, status, and a dependency graph to drive incremental execution.

Quick Start

Create ordered stories for plan-0-infrastructure by reading specs/plan-0-infrastructure.md, specs/plan-0-infrastructure/architecture.md, specs/architecture.md, and specs/roadmap.md and scanning the codebase for existing stories and artifacts.

Frequently Asked Questions about build-stories

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

FAQPage Schema
How do I break down a software development plan into AI-sized stories for agent execution?

Each generated story file includes implementation guidance, exact file paths, acceptance criteria, and testing guidance. It also marks the appropriate agent assignment, enforcing a design-to-engineer pairing pattern where applicable for reliable incremental execution.

Do I need existing architecture and roadmap documents to generate ordered story files?

The Skill generates numbered markdown story files and a stories.md index. This index tracks story ordering, status, and a dependency graph to drive incremental execution by orchestrator agents.

What is the best way to assign engineer and designer agents to decomposed plan increments?

This approach removes the friction of manual story authoring by converting high-level specs into executable increments that fit an AI agent's context window, ensuring each increment is verifiable with acceptance criteria.

Can I use plan decomposition for software projects without a pre-existing codebase to scan?

Without codebase scanning, the generated story files may lack exact file paths and existing artifact references, potentially reducing the precision of implementation guidance for the assigned agents.