ralph

Coordinate persistent multi-step execution loops with verification and evidence collection.

11|3|Updated Feb 18, 2026
One-click install
npx skills add https://github.com/kangig94/coral --skill ralph-kangig94
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/kangig94/coral/tree/main/skills/ralph
Command: npx skills add https://github.com/kangig94/coral --skill ralph-kangig94

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Persistent execution loop with verification to ensure tasks are completed with traceable evidence, preventing partial or falsely claimed outcomes.

Core Features & Use Cases

  • End-to-end task persistence: Orchestrates multi-step work with stepwise verification and evidence collection.
  • Adaptive routing: Supports --codex and --red modes for delegated planning and adversarial testing.
  • Audit-ready logs: Maintains verifiable outputs and decision records suitable for audits and reviews.

Quick Start

Initiate a persistent task sequence, monitor progress with verification, and collect evidence upon completion.

Frequently Asked Questions about ralph

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

FAQPage Schema
How do I maintain persistent task execution with verification for multi-step agents?

Persistent task execution is maintained by orchestrating a continuous execution loop with stepwise verification gates, ensuring multi-step agents produce complete and evidenced outcomes. This approach prevents partial or falsely claimed results through auditable logs.

What is a persistent execution loop and how does it ensure task completion?

A persistent execution loop is a task orchestration mechanism that processes multi-step plans with built-in verification gates and evidence collection. It ensures complete task completion by validating auditable decision records and verifiable outputs.

How do I set up automated task orchestration with verification gates and audit logs?

Automated task orchestration with verification gates is set up by initiating a persistent task sequence that monitors progress and collects evidence. The system maintains audit-ready logs and decision records to satisfy explicit routing and verification requirements.

Can I use adaptive routing for delegated planning and adversarial testing in task management?

Adaptive routing for delegated planning and adversarial testing is supported through dedicated routing modes. You can use the codex mode for delegated planning and the red mode for adversarial testing to validate multi-step task execution across diverse environments.

What is the best way to prevent partial or falsely claimed outcomes in complex plan execution?

Preventing partial or falsely claimed outcomes in complex plan execution is best achieved through a persistent execution loop with built-in verification. This approach enforces stepwise evidence collection and maintains verifiable outputs suitable for end-to-end completion audits.

Why does automated task management fail to provide traceable evidence for completed work?

Automated task management fails to provide traceable evidence when it lacks a persistent execution loop with integrated verification gates. Implementing stepwise evidence collection and auditable decision records ensures complete, evidenced results for end-to-end task completion.