Can I use MAGI Test for security audits?

Yes, the CASPER-3 engine specifically focuses on security coverage, negative testing, and boundary conditions to find vulnerabilities.

Do I need extra tools to run this?

The skill automatically uses the ToolSearch function to load the necessary MCP tools for Gemini and Codex integration during execution.

Which AI models are used in MAGI Test?

It utilizes Claude Sonnet (MELCHIOR-1), Gemini Flash (BALTHASAR-2), and Codex/GPT (CASPER-3) to ensure a diverse range of insights.

How does the voting mechanism work?

Each model analyzes the task independently based on its specific focus area, and the results are aggregated into a final report to highlight agreed-upon improvements.

What is the MAGI Test skill?

It is a Claude Code skill that runs three different AI models in parallel to analyze your code for test coverage, quality, and security.

MAGI Test Parallel Review

Name: MAGI Test Parallel Review
Author: weiting-tw

byweiting-tw

•

Security & Testing

Executes a multi-engine parallel analysis to evaluate test coverage, boundary conditions, and error handling across distinct AI models.

MAGI Test leverages a triple-engine consensus architecture inspired by the MAGI supercomputer to provide rigorous code reviews and test assessments. By simultaneously tasking Claude (Strategy), Gemini (Quality/Isolation), and Codex (Security/Edge Cases), it synthesizes a comprehensive consensus report that identifies gaps in test coverage and potential vulnerabilities that a single model might overlook. This skill ensures that your features are evaluated from three distinct perspectives—strategic, qualitative, and defensive—before reaching a final implementation consensus.

Key Features

01Consensus-based reporting and voting mechanism

02Boundary condition and negative testing identification

03Triple-engine parallel analysis (Claude, Gemini, Codex)

042 GitHub stars

05Seamless integration with Claude Code via MCP tools

06Automated test coverage and strategy assessment

Use Cases

01Validating test coverage for critical payment services or business logic

02Identifying edge cases and security vulnerabilities in new feature PRs

03Gaining diverse AI perspectives on code quality and mock isolation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add weiting-tw/magi-system magi-test

For use in Claude.ai and ChatGPT