ralph

Automate iterative task loops with architect verification until completion.

Updated Feb 1, 2026
One-click install
npx skills add https://github.com/suhwan/claude-registry --skill ralph-suhwan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/suhwan/claude-registry/tree/main/skills/orchestration/ralph
Command: npx skills add https://github.com/suhwan/claude-registry --skill ralph-suhwan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ralph provides an automated, self-referential loop that repeats task execution until a defined completion is achieved, with built-in architect verification to ensure quality and alignment with requirements.

Core Features & Use Cases

  • Iterative refinement: Repeats task cycles to progressively improve output until acceptance criteria are met.
  • PRD-driven planning: Optional PRD initialization to structure work before execution.
  • Ultrawork integration: Parallel task execution across specialized agents to speed delivery.
  • Use Case: Teams needing end-to-end task verification and architecture oversight for complex projects.

Quick Start

Activate Ralph to begin iterative task execution until a verified completion is produced.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I automate iterative task refinement until completion criteria are met?

You can automate iterative task refinement by running an automated loop that repeats execution until a defined completion promise is achieved. This enforces progressive output improvement until acceptance criteria are fully satisfied.

What is architect verification in an automated workflow loop?

Architect verification in an automated workflow loop is a strict quality check that ensures task outputs align with requirements. It enforces architecture oversight to guarantee end-to-end reliability before the loop exits.

How do I structure complex tasks using PRD-driven planning before execution?

You can structure complex tasks by initializing an optional PRD to define work structure before execution begins. This drives planning and progress tracking throughout the iterative task loop.

Can I use parallel execution across multiple subagents for complex task automation?

Yes, you can use parallel execution across multiple subagents through ultrawork integration. This speeds up delivery by distributing specialized tasks while maintaining architect verification.

What are the limitations of self-referential task loops for workflow automation?

Self-referential task loops require clearly defined acceptance criteria to exit successfully. Without strict completion promises and architect verification steps, the loop may fail to achieve end-to-end reliability for complex projects.