Cybersecurity firm Censys reported more than 800 distinct deployments of Artex accessible from the web without needing a password to log in, allowing anyone to access systems designed to automatically probe for security weaknesses.[1] Artex, described by Microsoft as a system of AI agents that can conduct cyberattacks from reconnaissance through data exfiltration with little needed human interaction, has been linked to attacks on South Korean financial institutions and the IT support firm ConnectWise, where it enabled Chinese-linked hackers to spy on thousands of organizations.[1]

TechCrunch reported that the publicly accessible Artex systems expose attack tools, such as implants, and CrowdStrike attributed attacks to a Chinese-speaking hacker who left an open directory on his Artex instance.[1]

Censys identified the exposures through internet scans, with the United States hosting the largest share of the 815 distinct deployments, followed by Hong Kong and China, and many of the exposed systems running on default ports with Chinese-language interfaces.[1][2] The tool’s architecture involves a planner agent breaking down goals like gaining server privileges and worker agents executing steps, repeating the loop until the objective is met, according to analysis by GEroundtable.[2]

Artex has since announced it would go closed-source, effectively limiting further releases for paid customers, but there are still dozens of copies of Artex’s code available on code-sharing sites that can still be used in active attacks.[1] Silas Cutler, principal security researcher at Censys who wrote the blog post, said that not every instance of Artex is operated maliciously and many are used for authorized testers.[1]