Eye Security defends against ARTEX LLM pentesting attack

Security researchers describe defending a client against an autonomous, LLM-driven penetration testing framework known as ARTEX. The attack unfolded at machine speed, utilizing parallel agents that rapidly discovered secrets, bypassed rate limiters, modified infrastructure, and attempted container breakouts. The incident response team ultimately contained the breach by employing their own auditable LLM agents alongside strict network isolation and credential rotation. Open article →