ralph

Loop iterative work with parallel delegation and architect verification until completion.

Updated Apr 23, 2026
One-click install
npx skills add https://github.com/wmasfoe/notch-pet --skill ralph-wmasfoe
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/wmasfoe/notch-pet/tree/main/.codex/skills/ralph
Command: npx skills add https://github.com/wmasfoe/notch-pet --skill ralph-wmasfoe

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph addresses the challenge of reliably completing complex tasks that span multiple iterations by enforcing persistent state and a mandatory verification step before declaring completion.

Core Features & Use Cases

  • Persistent iteration loop that retries work until success and requires architect sign-off.
  • Parallel delegation across agents to accelerate progress while maintaining guardrails and traceability.
  • Built-in execution policy and state management for background tasks across iterations.

Quick Start

Start Ralph and provide your task prompt; it will loop, persist progress across iterations, retry on failures, and require architect verification before completion.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I automate task completion for long-running operations that require formal verification?

Automated task completion is handled by looping over work in persistent iterations until an architect-verified state is reached. It coordinates iterative planning and automatic retries to ensure long-running operations are reliably finished.

How does parallel delegation work for complex multi-step tasks?

Parallel delegation accelerates progress by distributing work across agents while maintaining guardrails and traceability. This allows complex tasks to be executed concurrently without losing state or oversight.

Why does my automated workflow fail to persist progress across multiple iterations?

Progress fails to persist without built-in state management for background tasks. Enforcing persistent context across iterations ensures that long-running operations retain their state and can automatically retry upon failures.

Can I use background execution policies for tasks that span multiple iterations?

Background execution policies are explicitly supported for tasks spanning multiple iterations. The system enforces running long operations in the background while maintaining persistent state and mandatory verification before completion.

What is the best way to guarantee a complex task is fully completed before stopping?

Guaranteeing task completion requires a mandatory verification step before declaring success. The system enforces architect sign-off by looping over work, retrying failures, and demanding formal verification of the final state.

Do I need explicit model passing to run iterative delegation workflows?

Explicit model passing is required to run iterative delegation workflows. The system enforces this policy alongside background execution and mandatory verification to ensure reliable, architect-verified task completion.