Tag
MOLE is a benchmark for evaluating defenses that detect harmful actions by AI agents operating under limited review budgets. It introduces an open benchmark with 150 AI-operated accounts and compares various monitors across different scenarios.
This post shares a GitHub repository, featuring a curated set of security vulnerability samples and benchmarks, to evaluate the capabilities of static analysis tools and large models in vulnerability mining and secure code generation, covering multiple languages and the latest research findings.
Microsoft launched its first cybersecurity-specialized AI model, MAI-Cyber-1-Flash, and a new agentic cybersecurity platform called Perception, aiming to compete with Anthropic, Google, and OpenAI in AI-powered security.
A security study reveals that most AI agents in production are vulnerable to simple system prompt extraction attacks, leaking sensitive configuration and credentials. The article details common attack techniques and effective defenses.