ralph

Iterate on tasks until all prd.json user stories pass reviewer verification.

Updated Apr 20, 2026
One-click install
npx skills add https://github.com/jimmi2051/oh-my-copilot --skill ralph-jimmi2051
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/jimmi2051/oh-my-copilot/tree/main/plugins/omc-copilot/skills/ralph
Command: npx skills add https://github.com/jimmi2051/oh-my-copilot --skill ralph-jimmi2051

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph provides a persistence loop that keeps working on a task until all user stories defined in prd.json pass and are reviewer-verified, preventing silent or partial completions.

Core Features & Use Cases

  • PRD-driven iteration with explicit acceptance criteria and reviewer verification
  • Automatic progress tracking across iterations and retries
  • Flexible reviewer selection via --critic and configurable deslop options

Quick Start

Start Ralph to initialize a PRD-driven iteration loop that persists across retries until all stories pass.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I automate task verification until all PRD user stories pass?

PRD-driven iteration loops persist across retries until every user story passes reviewer verification, preventing silent or partial task completions. Progress is automatically tracked across iterations against formal acceptance criteria.

What is a PRD-driven iteration loop for complex task completion?

A PRD-driven iteration loop is a persistence mechanism that solves complex tasks by repeatedly running verification checks until all user stories defined in prd.json pass and receive reviewer sign-off. It supports initialization and optional deslop cleanup passes.

How do I configure reviewer checks for automated workflow verification?

Reviewer checks are configured using the --critic flag to select flexible reviewer runtimes. This allows you to apply optional configurable critics to verify formal acceptance criteria during the iteration loop.

Can I use PRD acceptance criteria to prevent partial task completions?

Yes, defining explicit acceptance criteria in prd.json prevents partial task completions. The persistence loop keeps working until all formal user stories pass reviewer verification, ensuring no silent or incomplete outputs.

What are the limitations of automated iteration loops for workflow review?

Automated iteration loops require formal acceptance criteria defined in prd.json to function properly. Without explicit user stories and reviewer verification configured, the persistence mechanism cannot determine when a task is fully complete.