We know that security is often the weak point in vibe coding.

To help the community catch common issues before deployment, I built SafeVibe.

It is a free, non-commercial, and collaborative "observatory" for tracking security vulnerabilities specifically found in AI-generated applications.

What it does:

-Vulnerability list: a database of common security issues observed in vibe-coded apps (seeded with 36 issues gathered from recent discussions). -Fixes: clear guidance on how to patch or mitigate each specific vulnerability. -LLM Audit Prompts: you can select specific vulnerabilities and the site generates a prompt you can paste into your LLM/agent to audit your code for those exact problems.

This is a community project and completely open to contributions. I’d love your feedback on the current data and the audit prompts.