Analysis OpenAI Agent Incidents Expose Access Control Gaps Recent OpenAI agent incidents show how quickly a blocked path can turn into unauthorized access when permissions and network reach remain too broad. By Shant Ebenezer Jena / 8 Oct 2026
Offensive Engineering Securing Agentic AI Against Data Leakage The security frameworks we built over forty years assume that some is behind the action. But agents break that assumption, and most enterprises haven't caught up. By S Pattnaik / 23 Jun 2026
Offensive Engineering Attacking the Cloud Control Plane Fifty to sixty percent of cloud breaches start on a secondary account. And then the control plane shows the attacker exactly where to go next. By S Pattnaik and Siri V / 14 Jun 2026
