
Microsoft has identified a financially motivated cybercriminal group using an agentic artificial intelligence framework called ARTEX to attack financial institutions, primarily in Southeast Asia.
The activity involves automated reconnaissance, vulnerability exploitation, credential attacks and the deployment of remote access malware.
In a November 6 report, Microsoft Threat Intelligence attributed the campaign to Storm-3358, which it describes as a Cambodia-based intrusion group. The company observed ARTEX being used in attacks against financial-sector organizations beginning in October 2026.
Microsoft said the group primarily targets financial institutions in Southeast Asia, although it has also pursued a smaller number of large financial organizations elsewhere.
The report does not identify the affected institutions or provide figures for financial losses or stolen records.
ARTEX is an open-source penetration-testing framework that uses multiple AI agents to coordinate offensive security operations. Rather than requiring an operator to manually direct every step, the framework can assign tasks to specialized agents, execute security tools, assess their results and determine subsequent actions.
Storm-3358 has used the framework to scan internet-facing systems and identify weaknesses that could provide access to target organizations.
Microsoft identified anonymous file uploads and server-side request forgery (SSRF) among the vulnerabilities exploited during the observed activity. An improperly secured file-upload function can allow unauthorized users to introduce files into a server, while SSRF can be abused to make vulnerable applications send requests to internal systems.
After gaining access, the attackers deployed three remote access and command-and-control malware families: WISP, vshell and AdaptixC2. Microsoft also identified the use of conventional command-line utilities, including netcat and curl, to download malicious tools onto compromised systems.
The group’s activities extended to credential attacks, API enumeration and database dumping. Microsoft observed brute-force attempts, password spraying and the use of compromised credentials to access additional systems. These operations can help attackers move laterally through an organization’s infrastructure after obtaining an initial foothold.
Microsoft also documented the use of web shells on compromised servers. These malicious scripts allow attackers to execute commands remotely through web applications and can provide continued access after the initial vulnerability has been exploited.
The reported activity highlights how ARTEX can combine AI-driven decision-making with established penetration-testing tools and malware. Its agents can coordinate several stages of an intrusion, reducing the amount of manual work required from an operator.
However, the framework does not eliminate human involvement, particularly when attackers must select targets or provide credentials.
Microsoft published eight IP addresses associated with the Storm-3358 investigation, classifying them as ARTEX servers, command-and-control infrastructure or WISP nodes. The company reported that the addresses remained active as of November 6 and cautioned that attackers can use proxies and VPNs to change their apparent locations.
Discover more from Aree Blog
Subscribe now to keep reading and get access to the full archive.

