Security issues in vibe-coded web applications
Another interesting article I've read today. The author has been looking at an analysis of 20,000+ AI-generated applications. The conclusion is that models are actually getting better at avoiding things like SQL injection or XSS. What shows up instead is repetition — each model has its own set of common secrets it reuses across different generated apps. The same JWT signing secrets, the same placeholder passwords like password123 and admin123, appearing in app after app, along with the same endpoints. This isn’t really a coding mistake. It comes from how the model generates. Which shifts the risk. Vulnerabilities are no longer local and discovered — they become shared and predictable. Once you recognize a pattern, it is likely to appear elsewhere, as systems are generated from the same source. https://www.invicti.com/blog/security-labs/security-issues-in-vibe-coded-web-apps-analyzed?utm_source=substack&utm_medium=email