ESLint AI Guard

Version updated for https://github.com/YashJadhav21/eslint-plugin-ai-guard to version v1.3.0.

  • This action is used across all versions by 2 repositories.

Action Type

This is a Node action using Node version 20.

Go to the GitHub Marketplace to find the latest changes.

Action Summary

Purpose and Functionality: The GitHub Action ai-guard is designed to prevent AI-generated code from introducing reliability and security bugs into your project by identifying and flagging patterns that are consistently incorrect.

Problems Solved: It helps catch errors such as missing error handling, asynchronous callback errors, insecure API usage, and dead scaffolding in generated code.

Key Capabilities:

  • Inline PR Annotations: Highlights issues directly within GitHub pull requests.
  • SARIF Reports for Code Scanning: Provides detailed reports for GitHub Code Scanning to support automated security audits.
  • Blocking Merges: Prevents the merge of potentially risky changes until issues are addressed.

Installation and Quick Start: No additional configuration is required; it automatically integrates with GitHub workflows.

What’s Changed

AI Guard v1.3.0

This release focuses on improving reliability, developer experience, and GitHub integration.

What’s new

  • Reduced false positives after validation against multiple real-world repositories.
  • Improved ai-guard init --yes for one-command project setup.
  • Better GitHub Actions integration with SARIF output and PR annotations.
  • Added FAQs, troubleshooting guides, and benchmark documentation.
  • Improved self-validation and workflow reliability.

GitHub Action

Run AI Guard directly in your CI:

- uses: YashJadhav21/eslint-plugin-ai-guard@v1
  with:
    preset: recommended

The Action can:

  • Review changed files in pull requests
  • Generate SARIF reports
  • Upload results to GitHub Code Scanning
  • Add inline PR annotations
  • Fail builds based on severity thresholds

Package

npm install --save-dev eslint-plugin-ai-guard

or

npx ai-guard init --yes

Resources

Demo

ezgif com-speed

Thanks to everyone who has tried AI Guard, opened discussions, shared feedback, or adopted it in their projects. Every suggestion has helped improve the rules and reduce false positives.