Artex is Criminals’ Favorite New Point-and-Shoot AI Hacking Tool

During the past Last year, security researchers warned that AI agents could significantly accelerate cybercriminals’ ability to launch cyberattacks, hack businesses, and ultimately steal money, data, or sensitive information from these compromised systems. For at least a subset of hackers, it seems that era has now arrived.

Since around late August, many businesses, government agencies and banks around the world have been targeted by cybercriminals who have adopted a new wave of AI-based penetration testing tools. These setups, many of which mimic legitimate defensive penetration testing techniques, use LLMs to autonomously scan potential hacking targets and attempt to breach poorly secured networks, then execute attack plans written by human operators.

Dozens of these AI penetration frameworks or “harnesses,” which detail how an AI model should perform cybersecurity testing or attacks, are increasingly linked to real-world attacks. The tools do not introduce new ways to attack websites or servers and are not sophisticated, according to alerts from a half-dozen security firms. Instead, they allow cybercriminals to launch attacks at speeds that would be impossible for humans.

Hackers also don’t need to be skilled to use these tools, researchers say, and attacks that would have taken days are carried out in just hours. “Agentic frameworks have evolved from experimental use to sustained, multi-objective offensive operations that execute with minimal human intervention,” AWS Chief Information Security Officer CJ Moses recently wrote in a company blog post.

The most important of these tools is Artex. Written in the Go programming language, the open source framework was first released on GitHub in July by a pseudonymous developer using the handle Autumn-27, and has since spread across some hacker communities. Artex was used in all aspects of the attacks, from conducting initial analysis of networks to gaining access to moving through systems and preparing data for export.

“It’s really an all-in-one product,” says Jacob Larsen, offensive security researcher and founder of security company Cypen. “Previous iterations of AI agent penetration testing tools only did reconnaissance, like scanning and scanning (of targets), and some only did testing. It’s really a point-and-shoot weapon.”

As with other AI hacking frameworks, Artex is not itself an AI product. Instead, it is essentially a guide that directs external, open source frontier AI models (which can be connected to it via APIs) towards a series of existing penetration testing tools and processes. People using Artex or other similar setups can send simple natural language prompts to begin their operations. “Launch an intrusion attack on (victim redacted) to gain access to the site’s admin panel, database, or server,” prompted a group of hackers at Artex, according to researchers at security firm Check Point.

Gn bussni

Logo