Summary

This repository introduces "PRD Led Context Engineering," a methodology for building AI-powered products by treating shared human-AI memory as maintainable infrastructure. It targets product teams aiming to accelerate development without sacrificing alignment, transforming documentation into a queryable Knowledge Graph. The core benefit is enabling teams to manage complex products that exceed individual cognitive limits by leveraging AI as a collaborative team member.

How It Works

The system operates on a "Memory as Infrastructure" philosophy, where documentation serves as the persistent, queryable Source of Truth (SoT). It utilizes Unique IDs (e.g., BR-xxx, UJ-xxx) to build a Knowledge Graph, enabling "Just-in-Time Context" loading for AI and humans, optimizing "Context Density" and reducing hallucinations. A "Progressive PRD" approach structures development through gated lifecycle stages (v0.1-v1.0), ensuring focused progress. Automated "Readiness Scoring" assesses project health and identifies critical blockers.

Quick Start & Requirements

Primary interaction involves Python scripts, such as python scripts/readiness.py run for computing readiness scores. The methodology is adaptable to different AI agents (e.g., .claude/, .gemini/), implying a Python environment is necessary. Specific hardware or AI model dependencies are not explicitly detailed but are inherent to AI-assisted development. Links to detailed documentation on the readiness protocol are provided within the README.

Highlighted Details

• Progressive PRD Lifecycle: A gated, 10-stage workflow (v0