AI Guard is an open-source ESLint plugin, CLI, and GitHub Action built to catch security and reliability problems commonly introduced by AI-generated JavaScript/TypeScript. It includes 18 focused rules, autofixes for selected issues, SARIF support, GitHub Code Scanning integration, and AI coding-agent context files. Instead of trying to detect whether code was written by AI, AI Guard focuses on the risky patterns that AI-assisted development can introduce.

Hey Product Hunt 👋 We built AI Guard because AI coding tools can generate code incredibly fast, but fast generation also means subtle security and reliability issues can slip through. AI Guard doesn't try to detect “AI-written code.” Instead, it looks for recurring risky patterns in JavaScript/TypeScript and catches them through ESLint, your CI pipeline, or GitHub Code Scanning. We're especially interested in hearing from developers who use Claude Code, Cursor, Copilot, or other AI coding agents: what kinds of bugs or mistakes do you find yourself fixing repeatedly after AI-generated code? Give AI Guard a try and tell us where it catches something useful—or where it gets something wrong!!