Gli agenti AI automatizzano l’intera catena di attacco, dalla patch all’exploit

AI agents automate the entire attack chain, from patch to exploit

Google, Amazon, and the FBI have documented the operational use of autonomous frameworks such as ARTEX, CAIRN, CyberStrikeAI, and PentAGI to conduct multi-stage intrusions at speeds far exceeding human capabilities. Google identified exploit_agent, a pipeline that automatically compares code before and after a patch, identifies the vulnerability, generates an exploit, and executes it without human intervention. China-linked actors also reportedly used CAIRN to compromise ten organizations in the defense, advanced manufacturing, civil rights, and media sectors. Amazon observed ARTEX agents searching for credentials in exposed AI applications, including LiteLLM and Dify, to reuse the stolen keys and support further operations. The FBI linked ARTEX to attacks against the finance, healthcare, public administration, and technology sectors, with some cases achieving privileged access in under three hours. The strategic consequence is a drastic reduction in the time available to install patches, detect reconnaissance, and contain an intrusion, making behavioral defenses and near-automated response processes essential.

Source: Security Affairs

Leave a Reply

Your email address will not be published. Required fields are marked *


Post Comment