Cybersecurity: AI agents accelerate supply chain breach of 2,250 organizations

Microsoft has identified a China-aligned threat group called Storm-0940 as the actor behind a massive supply chain attack that compromised more than 2,250 Microsoft 365 tenants. The intrusion leveraged an open-source, AI-driven penetration testing framework named ARTEX, which allowed the attackers to move through systems with unprecedented speed. In some cases, the group achieved privileged access to corporate networks within three hours of their initial targeting.

The operation relied on stolen credentials belonging to ConnectWise applications. By authenticating as these apps rather than individual users, the attackers bypassed standard interactive sign-in procedures. Microsoft described ARTEX as a system of large language model driven agents that conduct autonomous penetration tests. This technology accelerated scanning, reconnaissance, and password spraying, enabling the agents to operate continuously with minimal human oversight.

Storm-0940 has been active since 2021 and currently lacks industry aliases, fitting the profile of a developing threat actor. The campaign targeted 14 industries, including aerospace, defense, semiconductors, and government entities. Microsoft clarified that the attack did not result from a vulnerability in its own products but rather from the compromise of a third-party multi-tenant application.

A separate password spraying effort targeted 327 accounts across eight organizations, resulting in over 151,000 failed logon attempts. At its peak on October 11, the attackers made 2,900 attempts per minute. Nearly 98 percent of this activity triggered lockouts, and no successful logins were recorded in this specific phase of the attack. The high volume of failures indicates a brute force approach that was largely contained by standard security controls.

The more effective phase involved the takeover of ConnectWise cloud applications. Using four distinct application credentials, the AI agents signed into customer tenants from the same IP range, logging over 200 sign-ins in under four minutes. Once inside, they performed directory operations, changed application credentials, and modified privileged roles. Security firm Huntress observed one tenant adding itself to the Global Administrator role on October 15, a behavior that triggered specific Defender alerts.

By October 16, Huntress confirmed unauthorized sign-ins at 162 tenants. The following day, the count rose to nearly 1,000 Entra tenants with activity from the attacking IP, with more than 300 successful sign-ins. The attackers moved through the tenant list in sorted order, with actions occurring seconds apart. This pattern suggested a script working from a predefined target list rather than manual exploration.

The use of ARTEX distinguishes this campaign from traditional attacks because the framework can operate with various AI models. Threat actors in multiple regions are using the tool, and recent reports linked it to breaches at South Korean financial institutions. Reuters reported that ARTEX developers have stated the project will no longer be updated and will become closed source, potentially reducing the tool's availability to other attackers.

ConnectWise detected the unauthorized access and suspended its Cloud Backup, SaaS Security, and Cloud Management services as a precaution. The company has begun a phased restoration process expected to take several days. It reported no indicators that partner backup data was compromised, though the situation remains under active investigation.

Microsoft advised customers to keep affected app connections disabled until formal restoration instructions are issued. The company noted that a successful sign-in does not necessarily mean data was accessed or exfiltrated. Customers were urged to review activity from the current month where logs allow to identify any anomalous behavior related to the incident.