AI coding tools write fast. They also write the same security vulnerabilities repeatedly, because they learned from code that had those mistakes in it. VibeGuard is a security linter built specifically around the exact patterns GitHub Copilot, Cursor, ChatGPT, Claude and Gemini produce: → SQL injection via f-strings → Hardcoded API keys and passwords → JWT without algorithm verification → Command injection via shell=True → MD5 for password hashing → DEBUG=True shipped to production
No reviews yetBe the first to leave a review for VibeGuard
Maker
📌
Hey Product Hunt! 👋
I am David, security engineer and builder of VibeGuard.
I started building this in January after noticing the
same security vulnerabilities appearing repeatedly in
AI-generated code across multiple projects. SQL injection
via f-strings. Hardcoded credentials. JWT bypass. All
confidently produced by Copilot and Cursor because they
learned these patterns from training data.
This week a developer handed me their production codebase
to scan. 533 files. 3 critical vulnerabilities. All from
AI-generated code. All missed by their existing security
pipeline. All fixed with exact one-line commands.
VibeGuard is completely free, open source, and runs
Locally, no cloud, no tracking, and no cost ever.
Happy to answer any questions about how the rules work,
what patterns I have found most common, or how to
Integrate it into your pipeline.
If you use any AI coding tool, run this before you
ship. It takes 30 seconds.
github.com/zeroFhacker/vibeguard