māchine · cybersecurity

FBI warns open-source AI hacking tool Artex can breach networks in under three hours

An FBI advisory links Artex to attacks on ConnectWise customers, South Korean banks and Barracuda. It says the AI tool can go from first targeting to active compromise in under three hours.

The FBI has warned that hackers are using autonomous AI hacking tools to run cyberattacks in the United States and abroad. TechCrunch reported the bureau's flash cybersecurity advisory on 16 November 2026. The advisory names Artex, an open-source AI penetration testing platform, and says intrusions using it can move from targeting a system to active compromise "in under three hours."

The advisory links Artex to a breach of IT management company ConnectWise, which reached the Microsoft 365 environments of at least 2,250 of its customers. It also links the tool to attacks on South Korean financial institutions. What these cases share is speed. The tool automates the steps of an intrusion and runs them faster than many defenders can react.

What Artex is and how it is used

Artex is an AI-powered penetration testing platform, a type of software built to probe a network's defenses. According to TechCrunch, hackers can connect it to hosted frontier large language models, such as those from Anthropic and OpenAI. It can also run on open-source and non-U.S. models.

The method the FBI describes is simple. Artex quickly scans a victim's network for signs of weakness. It then exploits basic flaws, such as weak protections and exposed or stolen passwords. Once it has a foothold, the attacker can use it to spread across the network to databases, mailboxes and other sensitive information.

The FBI says these intrusions can be "noisy." They generate large volumes of traffic and data that may set off a company's defensive alerts. But the tool moves at a speed that makes it much harder for human defenders to keep up. The FBI lists financial services, government, technology, defense, phone and internet companies, and law firms among the sectors targeted in recent months.

Artex first appeared publicly in July. After it was used in attacks, its developer made the software closed-source. The FBI says the earlier open-source version is still circulating and remains "widely accessible."

The known victims

The FBI confirmed earlier findings from Microsoft that attackers used Artex to break into ConnectWise. Through that breach they reached the Microsoft 365 cloud environments of at least 2,250 ConnectWise customers. Microsoft says the attackers are likely Chinese. They used the tool to steal data from the affected customers. ConnectWise did not respond to TechCrunch's request for comment.

In South Korea, attackers used Artex against several financial institutions. The breaches are blamed in large part on weak defenses, including missing multi-factor authentication. The attackers stole the personal data of at least 68,000 people.

Security company Barracuda has disclosed that it was breached in an Artex-linked attack by financially motivated criminals. The tool found an exposed credential that belonged to a third-party service. Barracuda had not revoked it when it stopped working with that vendor. With that access, the attackers reached stores of credentials that Barracuda uses to collect data from its customers' systems. Barracuda said its logs were not detailed enough to rule out access. It chose to "assume the worst" and changed all of its customer credentials. It did not tell TechCrunch how many customers were affected.

"Defenders who rely on spotting malicious activity and then responding to it will often arrive after the attacker has finished," wrote Arve Kjoelen, Barracuda's chief information security officer.

The tools are exposed too

The attackers' own setup has weak points. According to TechCrunch, researchers last week found hundreds of Artex systems that anyone could reach on the internet without a password. They were likely exposed by mistake or misconfiguration. One of them was the system used to target South Korea. An exposed console gives its capabilities to anyone who finds it, not only to the person who set it up.

What it means for work

Artex automates much of an intrusion: scanning a network, finding a weak or forgotten credential, and moving toward valuable data. These steps used to depend on an attacker's own skill and time. Now the work is done mostly by a tool and a language model. For security teams the main change is time. A team that waits for an alert and then responds may, as Kjoelen says, arrive after the attack is over. That gives more weight to routine work done in advance: enforcing multi-factor authentication, revoking credentials when a vendor relationship ends, and keeping logs detailed enough to show what happened. Two questions remain open. One is whether defenders can automate their response as fast as attackers have automated their attacks. The other is how testing tools built for legitimate work can be kept from people who would misuse them.

Sources

This article was researched, written and illustrated by AI, and published without human editing. Spotted an error? Let us know.