After being implicated in an AI cyberattack on South Korean banks, the ARTEX author announced a move to closed source, and the GitHub repository has been taken down.

Beating AI News Flash: After the open-source AI penetration testing tool ARTEX was found to be used in attacks on multiple South Korean financial institutions, developer Autumn-27 announced on October 8 that the project would be closed-source, updates would stop, and no new versions would be publicly released or maintenance support provided. Subsequently, ARTEX's original GitHub repository was also taken down.

ARTEX is a multi-agent system that can connect to different large models, automatically find vulnerabilities, and execute penetration tests. Previously, cybersecurity company CrowdStrike found that attackers associated with the South Korean bank attacks used ARTEX and also left tool configurations and Claude Code chat logs on the servers.

In a statement, Autumn-27 said ARTEX was originally developed for learning, research, and authorized enterprise security testing. The tool being used for illegal attacks completely goes against the original intent of development. The author condemned unauthorized use and said they do not bear related responsibility.

However, ARTEX's source code has already been backed up by other developers. Even if the author deletes the original repository and stops maintenance, previously public versions can still be used and modified by others.