ralph

Automate end-to-end task execution with PRD-driven iteration and reviewer verification.

16|3|Updated Mar 12, 2026
One-click install
npx skills add https://github.com/sehoon787/my-claude --skill ralph-sehoon787
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/sehoon787/my-claude/tree/main/skills/omc/ralph
Command: npx skills add https://github.com/sehoon787/my-claude --skill ralph-sehoon787

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph enables end-to-end task execution with persistent state and mandatory reviewer verification, preventing silent incompletes and ensuring accountability across iterations.

Core Features & Use Cases

  • PRD-driven iteration that tracks progress across stories
  • Parallel delegation to specialized agents based on complexity
  • Mandatory verification and progress persistence before completion
  • Use cases include long-running tasks that require multi-round refinement and formal sign-off

Quick Start

Provide the task and let Ralph iterate across iterations until every user story passes verification.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I automate long-running multi-agent tasks with mandatory verification?

PRD-driven iteration automates long-running multi-agent tasks by tracking progress across user stories and enforcing mandatory reviewer sign-off before completion. Ralph applies persistent state management to prevent silent incompletes across retries and ensure accountability.

What is PRD-based story tracking for multi-round task refinement?

PRD-based story tracking monitors task progress across defined user stories over multiple iterations. It pairs per-story acceptance criteria with persistent state management to ensure long-running tasks meet formal sign-off requirements before completion.

Can I delegate tasks to specialized agents in parallel for complex iterations?

You can delegate tasks to specialized agents in parallel based on complexity. Ralph assigns work across agent tiers during PRD-driven iteration, applying persistent state management so every story passes reviewer verification before formal sign-off.

Does persistent state management work for tasks that require retries across iterations?

Persistent state management works across retries by maintaining task progress data throughout PRD-driven iterations. Ralph uses this mechanism to prevent silent incompletes, ensuring tasks resume correctly and pass mandatory reviewer verification before completion.

When do I need formal sign-off and reviewer verification for automated task execution?

Formal sign-off and reviewer verification are needed for long-running, multi-story tasks that require multi-round refinement. Ralph enforces mandatory verification with persistent state management to ensure accountability before any task is marked complete.