ralph-loop

Autonomous iterative development is driven by generating a PRD, converting it into a structured prd.json format, and executing user stories sequentially with integrated quality checks. This loop manages the entire software lifecycle from coding to committing with minimal human intervention.

A PRD-driven autonomous development loop is a process where an AI agent breaks down a Product Requirements Document into machine-readable JSON, then iteratively implements user stories, runs quality gates, and tracks progress until the feature is complete.

Yes, you can manually execute development steps one by one. The system supports both full autonomous loops via a shell script driver and manual step-by-step execution, ensuring each user story passes quality checks before proceeding.

You can convert a PRD into machine-readable JSON by asking the AI agent to generate the requirements document and then format it into a structured prd.json file, which serves as the input for sequential user story execution.

No external dependencies are required to run the autonomous development loop. The system utilizes internal scripts, references, and assets to manage the shell script driver and execute the iterative PRD-driven software engineering process.

The best way to kickstart a new feature is to ask the AI to create a PRD for the specific feature, convert it to prd.json, and run the autonomous development loop to handle coding, testing, and progress tracking iteratively.