ralph

Execute persistent task loops with parallel agent calls and architect verification.

Updated May 25, 2026
One-click install
npx skills add https://github.com/BetaS/swm-gis-workshop --skill ralph-betas
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/BetaS/swm-gis-workshop/tree/main/.codex/skills/ralph
Command: npx skills add https://github.com/BetaS/swm-gis-workshop --skill ralph-betas

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Ralph ensures that complex tasks are completed to a high standard, with every step verified by an architect before final completion.

Core Features & Use Cases

  • Persistence Loop: Continuously works on a task until it is fully complete and architect-verified.
  • Parallel Execution: Executes independent agent calls simultaneously for efficiency.
  • Session Persistence: Retains context and progress across multiple iterations.
  • Verification: Requires fresh evidence and tiered architect review for completion.
  • Use Case: Ideal for tasks that require high assurance of completion, such as security-sensitive updates or complex system integrations.

Quick Start

Run the ralph skill to initiate a persistent task loop for guaranteed completion and verification.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I ensure complex task completion with mandatory architect verification?

To ensure complex task completion with architect verification, use a persistent execution loop that continuously processes tasks until an architect reviews and approves the final output. This guarantees high-assurance completion for multi-step workflows.

What is a persistence loop for multi-step execution and when do I need it?

A persistence loop for multi-step execution is a continuous processing cycle that retains session context across iterations until a task is fully complete. It is needed for high-assurance tasks like security-sensitive updates and complex system integrations.

Can I execute parallel agent calls for independent tasks during the verification process?

Yes, you can execute parallel agent calls for independent tasks during the verification process. The system supports simultaneous execution of independent agent calls to maximize efficiency while maintaining mandatory architect verification for final completion.

Does session persistence retain context and progress across multiple task iterations?

Yes, session persistence retains context and progress across multiple task iterations. This ensures the persistent loop maintains continuous workflow awareness, preventing data loss during multi-step execution until the architect verification is complete.

What is the best way to handle security-sensitive updates requiring high-assurance completion?

The best way to handle security-sensitive updates requiring high-assurance completion is using a tiered architect review process with fresh evidence. This persistent loop approach ensures every step is verified before final completion is approved.

Why does task verification require fresh evidence and tiered architect review?

Task verification requires fresh evidence and tiered architect review to guarantee the highest standard of completion. This prevents unverified assumptions from passing, ensuring complex system integrations and security-sensitive updates are fully validated.