How does the convergence detection work?

The skill tracks 'clean' iterations where no new issues are found. It considers a project 'converged' only after 10 consecutive clean loops and when all category scores exceed 95/100.

What is a parallel agent swarm in Ralph Wiggum v2?

It involves spawning multiple specialized AI sub-tasks simultaneously, each focusing on a different aspect of code quality such as security, performance, or architecture, to provide a comprehensive review.

Do I need to have existing tests to use this skill?

No. Ralph Wiggum v2 follows a TDD approach where it automatically writes new failing tests for any issues it discovers before it attempts to write the fix.

Is this skill only for game development?

While it includes specialized personas for game design (like ARPG or RTS experts), its core logic for TDD, performance, security, and SOLID principles is applicable to any software project.

What is the Category Evolution feature?

It is a system that automatically maps your codebase structure into logical categories and subcategories, assigning scores to each to track progress and focus agent efforts on the weakest areas.

Ralph Wiggum v2: Autonomous TDD Swarm

Name: Ralph Wiggum v2: Autonomous TDD Swarm
Author: adaptationio

byadaptationio

0•

Desarrollo de Videojuegos

Automates the Test-Driven Development (TDD) cycle using parallel agent swarms to discover, review, and fix codebase issues until full convergence is achieved.

Ralph Wiggum v2 is an advanced orchestration skill for Claude Code designed for high-intensity autonomous development and quality assurance. By deploying parallel agent swarms with diverse personas—ranging from security auditors to game-feel experts—it systematically crawls your codebase to identify weaknesses. It enforces a strict TDD methodology, requiring a failing test for every identified issue before implementing a fix. This skill is ideal for complex projects like game development or large-scale refactors where maintaining a high 'quality score' across multiple domain-specific categories is essential for project completion.

Características Principales

01Strict TDD Enforcement (Failing test -> Minimal fix -> Verification)

02Parallel Agent Swarms (3-5 simultaneous specialized reviewers)

030 GitHub stars

04Convergence Detection (Requires 10 consecutive clean iterations)

05Extensive Review Personas (Security, Performance, SOLID, Game Design, etc.)

06Dynamic Category Evolution (Auto-discovery of codebase modules and scoring)

Casos de Uso

01Comprehensive codebase quality audits and automated refactoring

02High-reliability software maintenance with strict TDD requirements

03Autonomous game development loops and systems polishing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptationio/skrillz ralph-wiggum-v2

For use in Claude.ai and ChatGPT

Download Skill