ralph

Automating PRD-driven task completion requires a persistence loop that iteratively validates each user story until all pass verification and receive reviewer approval. Ralph advances tasks only when acceptance criteria are fully met, ensuring structured progression across multi-iteration projects.

A PRD-driven persistence loop continuously drives task completion by validating user stories against acceptance criteria until all pass and are reviewer-verified. It executes PRD setup, story selection, implementation, verification, cleanup, regression checks, and mandatory reviewer verification in sequence.

Yes, you can skip PRD generation or post-cleanup passes using optional flags within the persistence loop. This allows you to bypass initial PRD setup or final cleanup phases if your multi-agent workflow already has those steps covered externally.

Verifying acceptance criteria across user stories is handled automatically by the persistence loop, which iteratively checks each story until passes returns true. It then performs regression checks and requires mandatory reviewer verification before marking the task complete.

Ensuring release-readiness across multiple user stories requires post-review cleanup and regression checks after all stories pass verification. Ralph enforces mandatory reviewer verification to guarantee quality and completeness before any task is considered fully finished.