Researchers have found hundreds of copies of an AI-driven hacking tool sitting open on the internet. Artex, a Chinese-made autonomous penetration testing tool, has been linked to several cyberattacks since its release in July, and security firm Censys says its deployments are reachable from the web without a password, according to TechCrunch and Censys’s own blog post.
What Artex is. Microsoft has described it as a system of AI agents that can carry out attacks from reconnaissance through data exfiltration with little human interaction, according to TechCrunch. Penetration testing tools are normally used by authorized testers to find weaknesses, which is why the same software can be used for defense or for attacks. TechCrunch calls Artex open source and Chinese-made but does not identify its developers.
What Censys found. TechCrunch says Censys found more than 800 distinct Artex deployments reachable without a password. Censys’s blog puts the total at 1,465 hosts running an exposed console as of November 4, and about 815 distinct-looking deployments after duplicates are removed, according to other coverage. The figures differ because of how duplicates are counted. Censys also saw 32 open directories on Artex hosts that hold attack kits, according to those reports.
A caution from the researcher. Silas Cutler, a principal security researcher at Censys, said not every instance is malicious and many are used by authorized testers. Cutler said it is unclear why the systems are exposed, since the software does not require internet access. Exposed systems can also expose implants, the hidden malicious software left on victims’ systems, according to TechCrunch.
The attacks linked to it. In September, attacks on South Korean financial firms were attributed by CrowdStrike to a Chinese-speaking hacker who used an Artex instance and left an open directory hosting the system, according to TechCrunch. Censys says the instance had been exposed for months, and its data showed attack data reaching well beyond Korean banks, including payment and gaming companies. Earlier in October, Chinese-linked hackers used Artex to spy on thousands of organizations whose data is hosted in Microsoft’s cloud, after a compromise of ConnectWise, according to the same article.
The common weakness. Many of the attacks exploited weak or known passwords and missing multi-factor authentication, according to TechCrunch. Microsoft said the attackers took data so quickly that defenders had little time to detect and respond.
What changes next. TechCrunch says Artex announced it will go closed-source, limiting future releases to paying customers, while dozens of copies of its code remain on code-sharing sites and can still be used. One outlet reports this only in secondary coverage, so it should be treated as unconfirmed.
Why it matters. AI agents can now run much of an attack on their own, and the tools are spreading faster than defenses. For organizations in the US and UK, the basic steps still help most: use strong passwords, switch on multi-factor authentication, and apply updates quickly. Security teams can also check that their own testing tools are not reachable from the internet.
Sources: TechCrunch and Censys, with CrowdStrike and Microsoft findings as reported by TechCrunch. Figures are as reported and counts differ between outlets.
Cybersecurity illustration (representative image), Wikimedia Commons, CC0
