ralph

Coordinates parallel task delegation with verified, retryable completion across iterations.

Updated Feb 11, 2026
One-click install
npx skills add https://github.com/qazz92/oh-my-droid --skill ralph-qazz92
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/qazz92/oh-my-droid/tree/main/oh-my-droid/skills/ralph
Command: npx skills add https://github.com/qazz92/oh-my-droid --skill ralph-qazz92

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph enables long-running tasks to complete reliably by maintaining a persistent execution loop that retries on failure and enforces verification before finalization.

Core Features & Use Cases

  • Persistent iteration across retries to guarantee task completion
  • Parallel delegation of independent subtasks to accelerate progress
  • Verification-driven finalization with fresh evidence before marking complete
  • Use cases: multi-step workflows where tasks span multiple iterations, require retry logic, and must be verified

Quick Start

Initiate Ralph to continue the current task loop, delegate subtasks in parallel, retry on failures, and verify results before completion.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I persist long-running tasks to completion with verification?

Persistent task execution loops maintain state across iterations, retrying failed parallel subtasks and enforcing fresh evidence verification before marking any long-running task complete.

What's the best way to retry failed subtasks during parallel execution?

Parallel delegation distributes independent subtasks, and the persistent execution loop automatically retries failures across multiple iterations to guarantee full task completion.

When do I need verification-driven finalization for task management?

Verification-driven finalization is needed for multi-step workflows that span multiple iterations, requiring fresh evidence and explicit progress review before final task completion.

Can I delegate independent subtasks in parallel and still guarantee completion?

Yes, parallel delegation accelerates progress by distributing independent subtasks while the persistent loop tracks state and retries failures until verification passes.

Why does my multi-step workflow fail to complete without state tracking?

Without persistent state tracking, multi-step workflows lose progress across iterations, causing incomplete execution when subtasks fail and lack retry policies or verification checks.

Does task persistence work for workflows requiring multiple iterations?

Task persistence is designed specifically for multi-iteration workflows, maintaining an execution loop that retries subtasks and applies verifier-based completion using fresh evidence.