ralph

Orchestrate iterative task completion with parallel agent delegation and architect verification.

20|6|Updated Feb 13, 2026
One-click install
npx skills add https://github.com/sigridjineth/oh-my-codex --skill ralph-sigridjineth
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/sigridjineth/oh-my-codex/tree/main/skills/ralph
Command: npx skills add https://github.com/sigridjineth/oh-my-codex --skill ralph-sigridjineth

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill ensures complex tasks are fully completed and verified, preventing silent failures and partial implementations.

Core Features & Use Cases

  • Persistence Loop: Continuously works on a task until it's genuinely done.
  • Parallel Execution: Delegates work to specialized agents simultaneously for efficiency.
  • Mandatory Verification: Requires fresh evidence (tests, builds) and architect sign-off before completion.
  • Use Case: Debugging a complex race condition in a multi-module application that requires extensive testing and architectural review.

Quick Start

Use the ralph skill to finish the current task, ensuring all tests pass and an architect approves the changes.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I guarantee task completion and prevent silent failures during complex software engineering projects?

Guaranteed task completion is achieved through a persistent iterative loop that requires fresh evidence, such as passing tests and successful builds, alongside mandatory architect verification before finalizing any work.

How does parallel execution work when delegating tasks to specialized agents?

Parallel execution delegates work to specialized agents simultaneously, allowing multiple task components to progress concurrently while managing background execution for long-running operations to maximize efficiency.

What's the best way to verify complex code changes before considering a task fully done?

Mandatory verification requires fresh evidence like passing tests and successful builds, combined with architect sign-off, to validate task finality and prevent partial implementations from being accepted.

Can I use structured project breakdown for complex multi-module applications?

Optional PRD mode supports structured project breakdown, enabling detailed task decomposition for complex multi-module applications that require extensive testing and architectural review.

Why do I need architect review for task verification?

Architect review ensures evidence-based validation by requiring fresh test and build evidence before sign-off, preventing silent failures and ensuring complex tasks meet architectural standards before completion.

Does this approach work for debugging complex race conditions requiring extensive testing?

Debugging complex race conditions benefits from the persistence loop that continuously works until resolution, parallel agent delegation, and mandatory verification through fresh evidence and architect approval.