Which skills can this loop review by default?

It is pre-configured to review the ralph-prompt suite, including the builder, single-task, multi-task, project, and research prompt generators.

How does this skill track progress through iterations?

The skill maintains a RALPH_REVIEW_STATE.json file and a detailed iteration log to monitor improvements, track consecutive clean reviews, and ensure state persistence across loop cycles.

What is the Ralph Wiggum technique in Claude Code?

It is a self-correcting prompt engineering method where the AI iteratively reviews its own output against strict research-backed criteria until specific quality goals are met.

When does the review loop officially terminate?

The loop terminates only after the targeted skills pass two consecutive comprehensive reviews with zero recommendations for improvement, signifying a 'perfected' state.

Ralph Skill Review Loop

Name: Ralph Skill Review Loop
Author: adaptationio

byadaptationio

0•

Productivity & Workflow

Automates the continuous improvement and validation of Claude Code skills using the Ralph Wiggum self-review technique.

This meta-skill implements a sophisticated self-improving review loop designed to audit and enhance the Ralph Wiggum prompt generator suite. By executing a recursive process that tracks state across iterations, it evaluates skills against 16 distinct quality criteria—including YAML frontmatter accuracy, technique alignment, and structural integrity. The skill provides a deterministic path to perfection by continuing the cycle until two consecutive clean reviews are achieved, ensuring your custom Claude behaviors meet production-grade standards for reliability and performance.

Key Features

01Automated Git integration for tracking incremental skill improvements

02Strict termination logic requiring two consecutive error-free validation passes

03Continuous self-improvement loop with JSON-based state persistence

040 GitHub stars

05Comprehensive evaluation against 16 specific Ralph Wiggum technique criteria

06Standardized quality reporting via iteration logs and findings documents

Use Cases

01Automating the QA process for AI-driven developer workflows and skill creation

02Refining complex prompt templates to meet specific performance benchmarks

03Validating a suite of Claude Code skills before production deployment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptationio/skrillz ralph-skill-review-loop

For use in Claude.ai and ChatGPT

Download Skill