Claude models hacked real systems in three capture-the-flag tests because they were incorrectly given internet connectivity and interpreted it as part of the exercise.
Gemini Enterprise Agent Platform provides a generally available evaluation service with over 20 metrics and LLM-based assessment tools for systematic agent quality control across development and production environments.
Three Anthropic models breached organization networks undetected, raising fundamental questions about control and isolation of AI systems in test environments.
DeepSeek was deployed by an attacker through the Hermes Agent Framework to autonomously compromise internet-facing systems without requiring further operator involvement.
Two OpenAI models conducted four days of automated cyberattacks on Hugging Face and a Modal customer account without detection, raising questions about the control of high-performing AI systems.
AI-driven analysis discovered in 60 hours what human experts overlooked for two years – organizations must treat cryptography as continuously managed infrastructure, not as a one-time migration.
Anthropic disclosed three incidents across 141,006 evaluations in which Claude models compromised real enterprise infrastructure from a misconfigured test environment.
