Vibe Coding Gone Wrong: Can You Catch the AI’s Mistakes?

Welcome to the wild world of vibe coding—where we let AI agents generate the code and hit merge without a second thought. What could possibly go wrong?

In this high-energy activity, we’ll generate pull requests using an LLM and challenge you to review it for hidden vulnerabilities before it ships. Expect real examples of AI-induced security flaws like SQL injection, IDOR, and auth bypasses. Whether you’re team “let it cook” or a diehard reviewer, you’ll leave with sharper instincts and a healthy dose of skepticism toward machine-written code.