ralph-loop

Select and process the highest-priority failing backlog item per invocation.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/synqing/BT-HIL --skill ralph-loop-synqing
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph-loop
Source: https://github.com/synqing/BT-HIL/tree/main/.claude/skills/ralph-loop
Command: npx skills add https://github.com/synqing/BT-HIL --skill ralph-loop-synqing

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Ralph Loop automates autonomous WORKER mode execution by selecting the highest-priority FAILING item that has ralph_loop.enabled=true, while tracking iteration state and recording attempts for traceability.

Core Features & Use Cases

  • Executes exactly ONE backlog item per invocation to prevent runaway automation
  • Reads PRD context when available and enforces project constraints
  • Manages iterations, records attempts, and commits or reverts changes based on success or failure

Quick Start

Run Ralph Loop to automatically process the next eligible backlog item.

Frequently Asked Questions about ralph-loop

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

FAQPage Schema
How do I automate autonomous backlog execution for failing PRD items?

Autonomous backlog execution is automated by selecting the highest-priority failing item with ralph_loop.enabled=true and tracking attempts. The process reads PRD context, commits on success, and reverts on failure to ensure traceable iteration state.

Why does my autonomous worker mode process only handle one backlog item per run?

Handling one backlog item per invocation prevents runaway automation. This behavior is strictly enforced to ensure deterministic development and verification, requiring you to run the process again for the next eligible failing item.

What happens to code changes when an automated backlog iteration fails?

When an automated backlog iteration fails, the process automatically reverts all code changes. If the iteration succeeds, it commits the changes and records the attempt for full iteration tracking and traceability.

How does iteration tracking work with PRD context and project constraints?

Iteration tracking reads available PRD context to enforce project constraints like CENTER_ORIGIN and NO_RAINBOWS during execution. It records each attempt and relies on sub-skills to maintain deterministic development and verification standards.

Can I use autonomous worker mode without defining project constraints like NO_RAINBOWS?

Autonomous worker mode relies on defined constraints such as CENTER_ORIGIN and NO_RAINBOWS to guide deterministic development. Without applying these constraints, the automated execution and sub-skill verification may not function as intended.