PRD Builder

A Claude Code skill that turns data models into buildable specs. It reads the outputs of /data-modeler or /migrate-filemaker and produces a prioritized Product Requirements Document (PRD) with features, user stories, acceptance criteria, and a dependency-ordered task checklist.

What it does

• Reads your data-modeler outputs (interview.md, schema.sql, erd.mermaid) or migrate-filemaker outputs (app summary, discovery answers, recommendations, schema, API design, UI spec)
• Infers product decisions from the source material, only asks what's missing
• Recommends a tech stack if you don't have preferences
• Generates a full PRD with features, user stories, acceptance criteria, and a dependency-ordered task checklist

The pipeline

/data-modeler              →  /prd-builder  →  "build from prd.md"
Interview + Schema            PRD               Any dev or AI agent
./data-modeler/               ./prd-builder/prd.md

— or —

/migrate-filemaker         →  /prd-builder  →  "build from prd.md"
Migration Artifacts           PRD               Any dev or AI agent
./migration/                  ./prd-builder/prd.md

Install

Copy the prd-builder folder into your Claude Code skills directory:

~/.claude/skills/prd-builder/

Then after running /data-modeler or /migrate-filemaker, say /prd-builder or "create a PRD from my data model."

Files

• SKILL.md — the skill definition (source detection, discovery, PRD generation)
• stack-recommendations.md — tech stack decision matrix and deployment options

Requirements

• Either /data-modeler or /migrate-filemaker must have already run and produced outputs in ./data-modeler/ or ./migration/
• Claude Code with file system access