Attackers host AntiFox AI agents on Alibaba Cloud to hit AWS

The security firm Arctic Wolf says attackers use their own copies of AntiFox on Alibaba Cloud. AntiFox is an open-source tool that uses AI agents to automate penetration tests, which look for weak points. Arctic Wolf saw it in actual attacks on web services and Amazon Web Services (AWS) accounts. In some attacks, the attackers got production secrets.

What was seen

In one attack, the attacker put a webshell, a hidden remote control page, on an exposed server. It then read settings and database connection strings. Correct AWS credentials let the attacker make hundreds of requests in about 30 minutes. In a second attack, the attacker read, copied and deleted S3 objects.

The numbers

The source gives only some numbers. Most AWS tries were stopped, but some worked. Arctic Wolf gives no count of victims. AntiFox is a fork of ARTEX 0.1.1 and is at version 2.0.

What is not known

The publishers forbid use against live systems. The source says they give no technical control for the ban. It does not name the attackers.

Insight: Open attack tools are hard to close

ARTEX became closed-source after attacks on banks in South Korea. CrowdStrike linked it to a campaign that hit more than 200 organizations. AntiFox, a fork of an early ARTEX version, stays open on GitHub. This seems to show that closing one tool does not stop its copies. Leaders must monitor cloud keys, not tool names.