EU Regulatory Changes

A new academic paper titled "SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents" has been published on arXiv, proposing a framework for systematically generating security ...

This paper, published on arXiv, introduces AgentRedBench, a new framework for testing the security of large language model (LLM) agents that are integrated with third-party software-as-a-service (S...

This publication introduces PyFEX, a new technical framework designed to detect evasive Python-based cyber threats by systematically exploring all possible execution paths in malicious code. While ...

This paper, published on arXiv under the AI Safety framework, introduces a formal mathematical model called "Unicity" for verifying the safety of atomic swaps in decentralized systems. Atomic swaps...

This paper, published on arXiv, introduces a novel monitoring framework called Stateful Online Monitoring designed to detect coordinated attacks by multiple AI agents operating in distributed envir...