Artex: Credentials Exposed to AI Intrusions

At Barracuda, an identifier that remained valid after a vendor's departure was exploited during an attack linked to Artex, according to the company. Installations of this AI-assisted hacking tool were also discovered unprotected by a password.
Barracuda Replaces Its Customer Identifiers
Barracuda attributes its compromise to criminals motivated by financial gain. The company explains that Artex detected an exposed identifier from a third-party service that remained valid after the collaboration with the vendor ended. This access allowed attackers to obtain sets of identifiers used to collect data from its clients' systems.
Due to insufficiently detailed logs to rule out access, Barracuda states it has replaced all its customer identifiers as a precaution. The company did not respond to a request regarding the number of affected clients.
Artex installations also present security flaws: last week, researchers discovered hundreds of Artex systems publicly accessible online without a password. This exposure is likely due to an error or misconfiguration, which could allow third parties to exploit these hacking capabilities. Among these installations was one used to target South Korea.
FBI Confirms Intrusions Targeting ConnectWise
The FBI confirms earlier findings from Microsoft: Artex was used to breach the American IT management company ConnectWise. Attackers accessed the Microsoft 365 cloud environments of at least 2,250 of its clients and stole data. Microsoft estimates that these hackers are likely Chinese. A spokesperson for ConnectWise did not respond to a request regarding the FBI's claims.
In the South Korean financial sector, the FBI reports the hacking of several institutions and the theft of personal data concerning at least 68,000 individuals. The agency primarily attributes these compromises to weak cybersecurity defenses and the absence of protections such as multi-factor authentication. More broadly, the sectors targeted in recent months also include government, technology, defense, telecommunications, internet service providers, and law firms.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Artex Can Compromise a Target in Less Than Three Hours
Artex corresponds to a solution dedicated to penetration testing, aimed at measuring a network's resilience against intrusions. It offers the possibility of being connected to large language models available online, such as those provided by Anthropic or OpenAI, as well as to open-source models or those from countries other than the United States.
The use of Artex often begins with a rapid network analysis phase to identify potential vulnerabilities. Insufficient protections or exposed or stolen passwords can then serve as an entry point. The tool allows movement towards databases, messaging systems, or other sensitive resources.
According to the FBI, some attacks using Artex transition from targeting to active compromise in less than three hours. The volume of traffic and data generated can trigger alerts, but the speed of execution makes human defense more challenging. Arve Kjoelen, Barracuda's Chief Information Security Officer, believes that teams waiting to detect malicious activity before reacting often intervene after the attack has concluded.
Closed Source Does Not Prevent Old Version from Circulating
The developer of Artex has closed the software's code after its use in cyberattacks to prevent further abuse. However, the FBI indicates that the previous open-source version is still circulating and remains widely accessible for conducting attacks. The agency reports several cyberattacks since the tool's first public release in July.
Its warning focuses on the use of autonomous AI tools, including Artex, to conduct cyberattacks in the United States and other countries. These attacks are increasing, particularly by exploiting easily identifiable vulnerabilities to breach companies.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.