Security
AI agents are going beyond instructions. The past month showed where the real problem lies
Claude reached real systems, an agent tried to manipulate a developer, and hundreds of OpenAI agents coordinated outside the intended scenario. What these incidents reveal about…
Read more → SecurityBots are getting better at attacking login and checkout. Cloudflare is launching self-learning protection on its Enterprise plan
Cloudflare has launched a bot-protection mechanism that continuously learns from current traffic. We explain why that matters for login, customer accounts and checkout.
Read more → SecurityWeekly radar: WordPress has vulnerabilities. Custom sites do too. Only WordPress shows you an “Update” button
More WordPress vulnerabilities? That is not proof WordPress is bad. Custom solutions also need updates, maintenance and security checks.
Read more → SecurityAn AI agent was only supposed to book a gym class. Along the way, it started hacking the system
An AI agent was only supposed to book someone into a class. It found flaws, bypassed the rules and removed someone from the waiting list. A…
Read more →