Application securityMicrosoft finds backdoor using OpenAI Assistants API for C2 communicationsLaura FrenchNovember 5, 2025Rather than leveraging AI capabilities, the attacker stores commands in Assistant messages.
Application securityResearchers use ChatGPT to help reverse engineer XLoader malwareLaura FrenchNovember 4, 2025The AI assistant helped researchers save time, but not without manual intervention.
Application securityResearchers demonstrate Agent2Agent prompt injection riskLaura FrenchNovember 3, 2025A malicious AI agent could cause another agent to perform harmful actions in a multi-stage attack.
Application securityChatGPT Atlas address bar a new avenue for prompt injection, researchers sayLaura FrenchOctober 27, 2025A prompt disguised as a URL could be copied and pasted by an unsuspecting user.
Vulnerability ManagementRobo Bees, side, AI Risk, Red Tiger, SessionReaper, Bad Bots, Willow, Josh Marpet – SWN #523October 24, 2025Venomous Robo Bees and Rabid Cocaine Weasels, sidebar spoofing, AI Risk, Red Tiger, SessionReaper, Bad Bots, Willow, Josh Marpet, and More on this edition of the Security Weekly News.
AI/MLMany companies say they haven’t assessed agentic AI risksSteve ZurierOctober 23, 2025While nearly 60% of respondents plan to use agentic AI, more than half have not assessed the risks.
Application securityAI agents gone rogue: Why businesses need Security Centers of ExcellenceOWASP GenAI Security Project Team October 21, 2025As AI agents run unchecked, experts call for Centers of Excellence to secure and govern autonomous systems.
