ARTEX Goes Closed-Source After Bank Hacks
The developer of ARTEX is closing the source of its AI penetration-testing agent after the tool was linked to attacks on banks.
Autonomous security agents have a clear dual-use problem: the same capabilities that help defenders can also lower the barrier for attackers.
What happened
The developer of ARTEX said the penetration-testing agent would become closed-source and future public releases would stop after researchers linked the tool to attacks on South Korean banks.
The tool was reportedly used alongside other AI coding systems during the incidents.
Why it matters
Automated penetration testing can help security teams find weaknesses faster.
But once those tools can autonomously probe systems, chain exploits and operate across live infrastructure, the same functionality can be misused.
Closing the source does not remove copies that already exist, but it signals that developers are beginning to reconsider how offensive-capability agents should be distributed.
The bigger picture
Cybersecurity is becoming one of the earliest places where agent governance is being tested in practice.
The market will likely move toward stronger access controls, audit trails and usage monitoring for powerful offensive tools.
ARTEX is a clear example of why agent capability and agent distribution cannot be treated as separate questions.