AI coding assistants like Cursor, Windsurf, and Claude Code write fast code — but not always secure code. They hardcode API keys, use MD5 for passwords, build SQL queries with f-strings, and you won't catch it until something breaks in production. VibeLint is an MCP server that sits between the AI and your files, Every time your AI agent generates code, VibeLint intercepts it first and scans for security risks before it even touches your files.
No reviews yetBe the first to leave a review for VibeLint
Maker
📌
Hey Product Hunt 👋
I'm a solo developer who built VibeLint after an AI agent hardcoded Stripe key to a public repo at 2am and i did not have time to review the files myself
The thing that frustrated me wasn't that it happened, it's that every security tool I tried scanned code after the code is written and saved . By then it's already in your git history, already potentially exposed.
So I built VibeLint to intercept AI-generated code before it gets written to disk. It runs as an MCP server inside Cursor, Windsurf, Claude Code, and VS Code — so when your AI agent generates code, VibeLint checks it first.
The free version is fully open source and catches the most common issues. I literally got my first paying customer before I even launched someone found it from a single Reddit post, which told me the problem is real.
I'd love brutal feedback from this community. What detectors are missing? What would make you actually use this on a real production project?
Happy to answer anything in the comments. 🙏