Attackers automate intrusions with AI framework ARTEX
Original title: Hackers Use AI-Powered ARTEX Framework to Automate Cyberattacks Without Relying on Zero-Day Exploits
The FBI says misuse dates to at least August 2026, and Storm-3358 reached privileged access in under three hours.
Attackers are using ARTEX, an open-source framework that connects a large language model to scanning and exploitation tools, to press ordinary exposures such as weak passwords and loose permissions. The FBI said use began by August 2026. Microsoft tracked Storm-3358 scanning internet-facing systems and abusing issues such as anonymous file upload and SSRF, while other reporting describes AWS permission probing and Active Directory attacks. The original author removed the public code in October, but forks remain, including backdoored third-party Docker images.
Unconfirmed: TrustedSec’s backdoor finding covers third-party copies only, and WorkOS has not shown a container escape.
- TrustedSec says the flow still uses subfinder, httpx, Nmap, nuclei, and ffuf.
- Backdoored images named are tomicdd/artex and congdonglinux/artex.
- A changed login handler in those copies allowed unauthenticated command execution.
- A WorkOS benchmark agent reportedly reached an internal tool with a predictable dev credential.
- User-agent ArtEx/1.0 was listed among WorkOS indicators; file hashes were not published.
Impact: Privileged access via misconfigurations, not a named zero-day
- 1.Put exposed management APIs behind an identity-aware proxy, private access, or VPN.
- 2.Limit path enumeration, repeated probes, and bulk retrieval in the WAF.
- 3.Test ARTEX only on a disposable VM with temporary credentials and restricted egress.
- Threat actors
- Storm-3358
- Victims
- WorkOS
- Organizations
- FBITrustedSecMicrosoftMandiant
From the original article
4 min read at cybersecuritynews.com
Attackers are using an AI-driven testing tool called ARTEX to break into company networks. It needs no secret software flaws to do this. Instead, it hunts for ordinary mistakes such as exposed apps, weak passwords and loose permissions, and it keeps working at them far longer than a human would.
By Tushar Subhra Dutta · cybersecuritynews.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.