
The FBI has issued a warning about an AI-driven hacking tool called Artex that can breach networks in under three hours. This open-source software automates attacks, a task that previously required skilled human operators.
Artex connects to large language models and handles scanning, exploitation, and lateral movement independently. It has been linked to breaches at ConnectWise, Barracuda, and several South Korean financial institutions.
How Artex Works
Artex belongs to a growing category of open-source AI penetration testing platforms. It can connect to hosted models from Anthropic and OpenAI, as well as open-source or non-US-built models.
The tool typically begins by scanning a target network for vulnerabilities. It then exploits simple flaws, such as weak protections or exposed passwords, to gain an initial foothold. From there, it moves toward databases, mailboxes, and other sensitive information stores.
The Speed and Noise of Artex Intrusions
The FBI describes Artex intrusions as “noisy”, generating large volumes of traffic and data that can trigger defensive alerts. However, its speed often allows it to spread through a network before defenders can respond.
In South Korea, attackers used Artex to target financial institutions, exploiting weak security measures such as missing multi-factor authentication. This resulted in the theft of personal data belonging to at least 68,000 people.
Notable Breaches and Responses
The largest known operation involved ConnectWise, where hackers used Artex to breach the company and access the Microsoft 365 cloud environments of at least 2,250 customers.
In the case of Barracuda, attackers found an exposed credential from a third-party service that had not been revoked. This allowed them to access stores of credentials used to collect data from customers’ systems.
Arve Kjoelen, Barracuda’s chief information security officer, noted that traditional defenders often arrive after the attacker has finished. The company rotated all customer credentials as a precaution.
Artex’s Impact and Ongoing Risks
After the attacks, Artex’s developer made the software closed source to prevent further misuse. However, the FBI notes that the open-source version remains widely accessible, posing an ongoing threat.
Researchers also discovered that many Artex systems were exposed on the public internet without passwords, likely due to misconfiguration. One such system was linked to the South Korean attacks, highlighting the risk of outsiders taking over ready-made hacking setups.
Artex’s Exploitation of Exposed Credentials
A spokesperson confirmed that a minority of XDR customers were directly notified to recreate their integration credentials. The exact number of affected customers remains undisclosed.
