ralph

Execute PRD-driven multi-iteration workflows with reviewer verification and progress tracking.

1|Updated Mar 27, 2026
One-click install
npx skills add https://github.com/oneforce/oh-my-gemini-cli --skill ralph-oneforce
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/oneforce/oh-my-gemini-cli/tree/main/skills/ralph
Command: npx skills add https://github.com/oneforce/oh-my-gemini-cli --skill ralph-oneforce

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Self-regulated task execution often ends prematurely or without proper verification. Ralph provides a PRD-driven persistence loop that continues iterating until all user stories are passes and reviewer-verified, guaranteeing completion with structured progress and verification.

Core Features & Use Cases

  • PRD-driven iteration: decomposes complex tasks into verifiable user stories and guides multi-turn work until completeness.
  • Parallel execution with persistence: coordinates concurrent tasks while maintaining session state across retries.
  • Mandatory reviewer verification: requires explicit sign-off against acceptance criteria before completion.
  • Progress tracking: maintains progress logs and learnings to improve future iterations for similar projects.
  • Use Case: long-running software builds or research tasks that demand rigorous verification and artifact stability before delivery.

Quick Start

Start Ralph with your initial task prompt to allow it to scaffold a PRD-driven workflow and guide iterative verification 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 enforce mandatory reviewer verification for long-running task execution?

You can enforce mandatory reviewer verification by using a PRD-driven workflow that requires explicit sign-off against acceptance criteria before marking any task complete. This ensures long-running, multi-step projects pass formal completion checks.

What is the best way to track progress and decompose work into verifiable user stories?

The best way to track progress is through a PRD-driven persistence loop that decomposes complex tasks into verifiable user stories. This approach maintains progress logs and learnings to guide multi-turn work until all stories pass and are reviewer-verified.

How do I manage parallel execution with persistent state across task retries?

You can manage parallel execution with persistent state by applying a workflow that coordinates concurrent tasks while maintaining session state across retries. This guarantees structured progress and artifact stability throughout multi-iteration projects.

Can I use PRD-driven iteration for research tasks that demand rigorous verification?

Yes, PRD-driven iteration is suitable for long-running research tasks or software builds that demand rigorous verification. It guides iterative work with structured progress tracking and mandatory reviewer checks before final delivery.

When do I need structured progress tracking with explicit acceptance criteria?

You need structured progress tracking with explicit acceptance criteria when handling long-running, multi-step projects that require persistent state. This prevents premature execution endings and guarantees complete, verified task delivery through formal completion checks.