ralph

Iterate tasks persistently with retries, parallel execution, and architect verification.

3|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/Tienching/oh-my-codebuddy --skill ralph-tienching
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/Tienching/oh-my-codebuddy/tree/main/.codex/skills/ralph
Command: npx skills add https://github.com/Tienching/oh-my-codebuddy --skill ralph-tienching

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph provides a persistence loop that keeps working on a task until it is fully complete and architect-verified, preventing premature completion.

Core Features & Use Cases

  • Persistent iteration across retries for long-running tasks with explicit verification before completion.
  • Parallel execution orchestration with session persistence and automatic retry on failure.
  • Suitable for workflows requiring structured state management and auditable completion evidence.

Quick Start

Initiate the Ralph loop with a defined task and allow it to persist across iterations until completion is verified.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I prevent premature completion in long-running task workflows?

To prevent premature completion in long-running task workflows, use a persistent execution loop that enforces mandatory architect verification before reporting completion. This ensures tasks are fully complete through structured execution policies and progress checkpoints.

What is architect verification in state management for parallel execution?

Architect verification in state management is a formal verification cycle that validates parallel execution results before completion. It enforces progress checkpoints and rollback safety, ensuring long-running workflows meet structured execution policies before final reporting.

How do I set up a persistent task loop with automatic retry on failure?

Set up a persistent task loop by defining a target task and initiating the orchestration with session persistence. The loop will automatically retry on failure and iterate across execution states until the task is fully complete and architect-verified.

Does parallel execution orchestration support rollback safety for long-running tasks?

Yes, parallel execution orchestration supports rollback safety for long-running tasks through structured state tracking and escalation rules. It enforces progress checkpoints to ensure data integrity throughout the task iteration process.

Why does my task orchestration report completion before verifying the final state?

Task orchestration reports completion early when it lacks a mandatory verification cycle. Enforcing a formal architect verification with progress checkpoints prevents premature completion by requiring auditable evidence before the task is marked complete.

Can I use task persistence to maintain state across retries for complex workflows?

Yes, you can use task persistence to maintain state across retries for complex workflows. It provides session persistence and state tracking, ensuring long-running workflows have auditable completion evidence and structured execution policies.