FBI warns ARTEX AI hacking tool is powering global cyberattacks
The FBI has issued a FLASH advisory warning that the AI-powered ARTEX penetration testing framework is being used by hackers to automate and accelerate cyberattacks globally, including a recent ConnectWise supply-chain breach.
Intelligence analysis by Gemini 2.5 Flash

ARTEX is an open-source, AI-driven framework that automates reconnaissance, vulnerability exploitation, privilege escalation, and data theft by integrating with external large language models. It significantly reduces attack times, enabling threat actors with minimal technical expertise to conduct complex, multi-target operations simultaneously across various sectors worldwide.
Imagine a super-smart robot that can learn how to break into computer systems really, really fast, even if the person controlling it isn't a computer genius. That's kind of what ARTEX is. Bad guys are using this robot to find weak spots and sneak into many different companies' computers all over the world, sometimes in just a few hours, making it hard for the good guys to stop them in time.
Analysis
The FBI's recent FLASH advisory underscores a significant shift in the landscape of cyber warfare, driven by the emergence and weaponization of the ARTEX AI-powered penetration testing framework. This tool, initially released as open-source, has been adopted by various threat actors since at least August 2026, enabling them to automate complex cyberattacks across critical sectors including financial, government, technology, defense, and healthcare. The core concern is ARTEX's ability to drastically reduce the time from initial targeting to privileged access, often achieving this in under three hours, a speed that severely challenges traditional defensive response times.
ARTEX Framework
ARTEX operates as an autonomous penetration testing framework, written in Go, that leverages artificial intelligence agents to execute a full spectrum of intrusion activities. Rather than developing its own AI model, ARTEX integrates with established large language models (LLMs) such as Anthropic Claude, OpenAI GPT, DeepSeek, GLM, Grok, and Kimi. This integration allows attackers to define their objectives, which ARTEX then translates into commands for its specialized AI agents, automating tasks like vulnerability discovery, system exploitation, and data exfiltration. This modular approach makes the framework highly adaptable and potent, capable of identifying weaknesses in often-overlooked systems like partner portals and legacy services.
Global Deployment
The framework's impact is already global, with researchers identifying over 1,600 publicly exposed ARTEX consoles by early November, suggesting a much higher actual deployment count. Despite its original developer, Autumn-27, removing the GitHub repository in October 2026 due to abuse reports, ARTEX remains widely accessible through community forks, archived copies, localized versions, and even a commercial edition known as ARTEX PRO. This widespread availability, combined with its ability to empower threat actors with minimal technical skills, significantly expands the operational scale of cyberattacks. The FBI's 13-page advisory provides crucial indicators of compromise (IOCs) to aid defenders, complemented by an open-source detection signature pack released by cybersecurity researcher MalwareMustDie.
ConnectWise Breach
One of the most prominent incidents linked to ARTEX is the October ConnectWise Cloud Backup supply-chain attack, which led to unauthorized access to Microsoft 365 customer environments. In this attack, ARTEX was utilized by a China-state-aligned threat actor, Storm-0940, to accelerate reconnaissance, password spraying, and the abuse of cloud application permissions. After compromising ConnectWise application identities, ARTEX was instrumental in attempting authentication against over 2,250 downstream Microsoft 365 tenants, with some intrusions progressing to credential modifications and persistent access attempts. The FBI also cited attacks against South Korean financial institutions, including Shinhan Bank, KB Kookmin Bank, and Hana Bank, where a Chinese-speaking threat actor combined ARTEX with Claude Code agents to expose personal information of approximately 67,000 individuals.
Key points
- The FBI warns that the AI-powered ARTEX penetration testing framework is being used by hackers for global cyberattacks.
- ARTEX automates reconnaissance, exploitation, privilege escalation, and data theft by integrating with external LLMs, significantly accelerating attack velocity.
- Attacks can progress from initial targeting to privileged access in under three hours, enabling threat actors with minimal technical expertise.
- The framework was linked to the October ConnectWise Cloud Backup supply-chain attack, compromising Microsoft 365 customer environments.
- ARTEX remains available through community forks and commercial versions despite its original GitHub repository being removed.
The FBI's public warning and the release of detection signatures by cybersecurity researchers like MalwareMustDie offer a crucial opportunity for organizations to bolster their defenses. Increased awareness and the implementation of recommended security measures, such as securing internet-facing applications and monitoring for automated requests, could help mitigate the threat posed by ARTEX.
The rapid automation capabilities of ARTEX, combined with its continued availability through various channels, mean that even less skilled attackers can launch sophisticated and widespread cyberattacks. The speed at which ARTEX can achieve privileged access leaves defenders with minimal time to react, potentially leading to more frequent and damaging breaches globally.


