Security researchers have identified a new wave of cyberattacks driven by an open-source framework called Artex. Since late August, multiple companies, government agencies, and banks around the world have been targeted by cybercriminals using this AI-powered penetration testing tool. The framework allows attackers to autonomously scan networks and attempt to force their way into poorly secured systems, significantly speeding up the attack process.
Artex AI hacking lowers the barrier for cybercriminals
Artex is written in the Go programming language and was first published on GitHub in July by a pseudonymous developer known as Autumn-27. The tool does not introduce new methods for attacking websites or servers. Instead, it acts as a guide that connects external artificial intelligence models to existing penetration testing tools. Users can issue simple natural-language prompts to initiate attacks, and multiple AI agents then break down the plans into tasks and execute them.
“It’s really an all-in-one product,” said Jacob Larsen, an offensive security researcher and founder of Cypen. “Previous iterations of AI agent pen-testing tools would either only do reconnaissance… and some would only do testing. This is really a point-and-shoot weapon.”
The simplicity of the tool has lowered the technical barrier for entry. Researchers note that attacks that previously required days of skilled labor can now be completed in hours by individuals with little to no cybersecurity experience. Guillaume Valadon, a researcher at GitGuardian, stated that the tool is so easy to use that even unskilled individuals can attempt to breach systems.
Global targets include finance and defense sectors
The FBI issued an alert last week warning that a variety of malicious actors have used Artex to conduct campaigns targeting organizations globally. The affected sectors include financial services, government, technology, defense, semiconductor, energy, health care, manufacturing, telecommunications, and legal industries. The attacks are often linked to Chinese-speaking cybercriminal groups seeking financial gain through the theft of personal information, cryptocurrency, or exposed login details.
In October, China-linked hackers known as Vapor Panda used Artex to target more than 200 organizations. These targets included think tanks, semiconductor and defense firms in the United States, and government foreign ministries in Europe. Microsoft also identified a group it tracks as Storm-0940 using the framework to attempt unauthorized access to Microsoft 365 setups following a supply chain attack. In one instance, hackers gained privileged access to systems in less than three hours.
On October 8, the developer Autumn-27 removed Artex from GitHub, stating they were not associated with the malicious attacks and that the tool was developed for educational purposes. However, multiple versions of the framework have been copied and continue to be used alongside other similar tools such as CyberstrikeAI, Hexstrike, CAIRN, and PentAGI.
Defending against AI-driven cyberattacks
Despite the speed and efficiency of these attacks, security experts note that the tools are currently not stealthy. The automated nature of the AI agents often creates significant noise in network systems, making thousands of probes in quick succession. For example, the agentic coding platform Dosu reported that Artex was used to send 80,000 requests to its systems in just 90 minutes.
Silas Cutler, a principal security researcher at Censys, emphasized that basic security fundamentals remain the best defense. Patching software, avoiding default credentials, and monitoring for unusual network activity can help organizations detect and block these automated attempts. Cutler also noted that thousands of instances of the systems being set up have been exposed to the internet, allowing researchers to inspect how operators are using the tool.
Recent analysis by Censys and the Shadowserver Foundation revealed that some exposed Artex servers appeared to show Chinese native speakers targeting other Chinese entities. In one case, a university was targeted with a password spraying attack, suggesting that some of the activity may be driven by individuals rather than organized criminal groups. However, Cutler warned that a significant portion of the activity is likely organized, sophisticated, and criminal in nature.
Source: Wired

