The original headline is "6 Security Holes We Keep Finding in Vibe-Coded Apps"

Short summary

AI-generated codebases repeatedly expose six security vulnerabilities: hardcoded API keys, missing server-side validation, frontend-only auth, overly permissive CORS, missing authorization checks, and exposed secrets in git history. These aren't exotic—junior devs would catch them in code review. The post provides actionable checks for each issue.

• •Six repeating security issues found in 80% of AI-generated codebases
• •Most common: hardcoded API keys and secrets exposed in client code and git history
• •Root cause is skipped code review; article provides specific security checks for each vulnerability

Generated with AI, which can make mistakes.