Hey everyone, Shantanu here 👋 Over the last few months, I’ve been building a lot with AI tools — and one thing became obvious: We’re shipping apps faster than ever… but we have no idea how secure they are. Most vibe-coded / AI-generated apps aren’t tested like real production systems. Traditional tools either don’t work well with AI flows or require heavy setup and expertise. So I built JAKU. It uses multiple autonomous agents to behave like attackers — crawling your app, probing vulnerabilities, and chaining them into real exploit paths (XSS, SQLi, prompt injection, etc.). The goal was simple: → No security expertise required → No manual babysitting → Just “tell me how my app can be broken” Would love your feedback — especially from folks building with AI or shipping fast 🚀

Jaku.app

Your AI app is vulnerable. JAKU finds how to break it.

Your AI app is vulnerable. JAKU finds how to break it.