A high-performance 3-tier AI agent methodology designed for autonomous software implementation with rigorous quality control and architectural integrity.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install kj-ralph-loop
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install kj-ralph-loop using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Ralph Loop v2?
Ralph Loop v2 is a sophisticated autonomous development methodology designed to streamline how AI agents handle coding tasks within Openclaw Skills. Inspired by the principles of Geoffrey Huntley and refined for native performance, it orchestrates a hierarchical structure of agents to ensure code is not just written, but verified and architecturally sound. By separating concerns between planning and execution, this skill eliminates common AI pitfalls like context drift and unverified claims.
At its core, Ralph Loop v2 focuses on small, manageable task cycles and evidence-based completion. It is built to handle everything from rapid single-file prototyping to complex multi-file projects in languages like Rust or environments like Godot. By integrating Ralph Loop into your Openclaw Skills repertoire, you enable a development environment where YAGNI principles are enforced and every line of code is backed by functional evidence.
Ralph Loop v2 Use Cases
- Orchestrating autonomous implementation of complex software features from high-level user prompts.
- Managing high-quality code generation for single-file tools and HTML5 games.
- Systematizing the debugging process for persistent architectural errors using a 3-strike rule.
- Enforcing strict specification compliance and YAGNI (You Ain't Gonna Need It) logic in AI workflows.
- Automating the creation of implementation plans and technical specifications folders.
How Ralph Loop v2 Works
- The Main agent receives a development request and spawns a Ralph Mother sub-agent to handle the implementation lifecycle.
- The Mother agent analyzes requirements and populates a specs directory with individual markdown files for each feature.
- An IMPLEMENTATION_PLAN.md is generated, breaking the project into 2-5 minute micro-tasks.
- Mother spawns Worker agents for the highest priority task, providing them with specific specification text to maintain high focus.
- Workers implement the code and must perform actual verification (tests, curl commands, or browser checks) to gather evidence.
- Mother performs a two-stage review of the Worker's output, checking for both specification compliance and technical code quality.
- Upon successful review, the Mother agent updates the plan, commits the changes, and moves to the next task until completion.
Ralph Loop v2 Setup
To deploy Ralph Loop v2, ensure the methodology is active within your Openclaw Skills environment. The framework runs natively and does not require external plugins.
# The system will automatically initialize the following structure on implementation requests:
mkdir -p specs/
touch IMPLEMENTATION_PLAN.md
Simply provide a development prompt to the Main agent to trigger the Ralph Loop mother-worker orchestration.
Ralph Loop v2 Data Schema & Taxonomy
The skill organizes implementation data using a standardized file-based memory system to track state across Openclaw Skills agents:
| Component | File Path | Purpose |
|---|---|---|
| Feature Specs | specs/*.md |
Defined requirements for individual modules or features. |
| Implementation Plan | IMPLEMENTATION_PLAN.md |
Central source of truth for TODO/DONE status and verification evidence. |
| Architecture Refs | references/workflow.md |
Detailed guidance for complex project structures and TDD patterns. |
| Worker State | Disk-based | Workers record state to disk to allow for context isolation and recovery. |
Ralph Loop v2 Advanced Features
- 3-tier Agent Orchestration: Separates user interaction, strategic planning, and tactical execution for maximum efficiency.
- Evidence Before Claims: Rejects completion reports that do not include actual execution logs, test outputs, or visual captures.
- Systematic 3-Strike Debugging: Forces an architectural re-evaluation if an error persists through three modification attempts, preventing infinite loops.
- Context-Isolated Workers: Optimizes performance by passing only relevant specifications to sub-agents, keeping them in the smart-zone.
- Native Openclaw Skills Integration: Operates without the overhead of external plugins or complex environment dependencies.
SKILL.md
Loading
Related Openclaw Skills
Parallel Agent Dispatch
A sophisticated workflow designed to dispatch multiple independent subagents to handle unrelated tasks concurrently, maximizing throughput and reducing execution time.
EvoWeb.ai Website Builder
Transform natural language descriptions into fully functional, responsive websites and landing pages in under five minutes.
Godot Engine Game Development
A comprehensive development framework for building 2D and 3D games using Godot 4.x and GDScript.
GitHub Pro
An advanced automation skill for managing GitHub repositories, CI/CD pipelines, and pull requests via the GitHub CLI.
UI/UX Pro Max
An advanced design intelligence skill for building, auditing, and implementing polished user interfaces and design systems.
Web Deploy GitHub Pages
Automate the entire lifecycle of creating, building, and deploying professional static websites to GitHub Pages.
Featured*
