MiniMax Code Review

AI-powered GitHub Pull Request code review using MiniMax models. Automatic PR comments, bug detection, and improvement suggestions via GitHub Actions.

Features

• 🚀 Detect bugs
• 🔍 Suggest improvements
• 🧠 AI-driven PR feedback
• ⚡ Works with GitHub Actions

Quickstart

Add this to your .github/workflows/code-review.yml:

name: AI Code Review with MiniMax

on:
  pull_request:
    types: [opened, synchronize]

permissions:
  pull-requests: write

jobs:
  review:
    name: Review
    runs-on: ubuntu-latest
    steps:
      - name: Code Review
        uses: tarmojussila/minimax-code-review@v0.4.0
        with:
          MINIMAX_API_KEY: ${{ secrets.MINIMAX_API_KEY }}

Inputs

Input	Required	Default	Description
MINIMAX_API_KEY	Yes	—	Your MiniMax API key
MINIMAX_MODEL	No	MiniMax-M2.5	MiniMax model to use for review
MINIMAX_SYSTEM_PROMPT	No	See below	Custom system prompt for the AI reviewer
MINIMAX_REVIEWER_NAME	No	MiniMax Code Review	Name shown in the review comment header
EXCLUDE_PATTERNS	No	*.lock,package-lock.json,yarn.lock,pnpm-lock.yaml	Comma-separated file patterns to exclude from review
MAX_DIFF_CHARS	No	0 (unlimited)	Maximum total characters for the diff sent to the API

The default system prompt is:

You are an expert code reviewer. Review the provided code changes and give clear, actionable feedback.

You can override it to focus on specific concerns, enforce coding standards, or adjust the review tone, e.g.:

You are a security-focused code reviewer. Identify vulnerabilities, unsafe patterns, and authentication issues. Skip style comments.

Configuration

To use this action, you must add your MiniMax API key as a GitHub secret.

1️⃣ Get your MiniMax API key

Generate an API key from your MiniMax dashboard.

2️⃣ Add the API key to your repository

1. Go to your GitHub repository

2. Click Settings

3. Navigate to Secrets and variables → Actions

4. Click New repository secret and add:

   • Name: MINIMAX_API_KEY — Value: your MiniMax API key

Advanced configuration

Instead of using default values for MINIMAX_MODEL, MINIMAX_SYSTEM_PROMPT, and MINIMAX_REVIEWER_NAME, you can override them, and manage them as GitHub Actions variables. This lets you update the model, review prompt, or reviewer name without touching the workflow file.

1️⃣ Add the variables to your repository

1. Go to your GitHub repository

2. Click Settings

3. Navigate to Secrets and variables → Actions

4. Click the Variables tab

5. Click New repository variable and add:

   • Name: MINIMAX_MODEL — Value: e.g. MiniMax-M2.5
   • Name: MINIMAX_SYSTEM_PROMPT — Value: your custom system prompt
   • Name: MINIMAX_REVIEWER_NAME — Value: e.g. AI Code Review

2️⃣ Reference them in your workflow

      - name: Code Review
        uses: tarmojussila/minimax-code-review@v0.4.0
        with:
          MINIMAX_API_KEY: ${{ secrets.MINIMAX_API_KEY }}
          MINIMAX_MODEL: ${{ vars.MINIMAX_MODEL }}
          MINIMAX_SYSTEM_PROMPT: ${{ vars.MINIMAX_SYSTEM_PROMPT }}
          MINIMAX_REVIEWER_NAME: ${{ vars.MINIMAX_REVIEWER_NAME }}

Contributing

Contributions are welcome. See the CONTRIBUTING file for more information.

License

This project is licensed under the MIT License. See the LICENSE file for more information.