What I post.
What you learn.
AI-Powered Pentest
Full attack chains — from recon to exploit — using Claude + Kali. Every step shown. Every command explained. Real findings, not lab scenarios.
Vibe-Code Vulnerabilities
AI-generated apps have predictable security holes. I catalog them systematically. You'll recognize these patterns in the wild immediately.
Breaking AI Agents
Prompt injection, jailbreaks, LLM hijacking, agent manipulation. The attack surface of AI systems is massive and mostly unexplored.