Microsoft details how AI agents accelerated a complex supply chain breach
A China-linked threat group leveraged an automated testing framework to probe thousands of Microsoft tenants, with some breaches occurring within hours of initial access.
By Olivia Miller ·
For many businesses relying on cloud-based management tools, the recent security incident highlights a new level of speed and automation in cyberattacks. Ordinary users and IT administrators in Texas and the surrounding region should be aware that threats are no longer solely manual. Attackers are now deploying artificial intelligence agents to navigate systems and test defenses at a pace that leaves little room for human reaction. This shift changes how organizations must prepare for potential breaches in their digital infrastructure.
Microsoft disclosed these details in a recent security blog post. The incident involved a threat group known as Storm-0940. This group utilized ARTEX, an open-source framework designed for autonomous penetration testing. The tool uses large language model driven agents to perform security scans and attacks. By automating the process, the group significantly reduced the time required to identify vulnerabilities and move through a network.
The numbers in this attack were substantial. The threat actors attempted sign-ins at more than 2,250 Microsoft 365 tenants. They used stolen credentials from ConnectWise applications to authenticate without an interactive user login. In some cases, the attackers reached privileged access levels in under three hours. Additionally, automated password guessing targeted 327 accounts across eight organizations, resulting in over 151,000 failed logon attempts. No successes were recorded in those specific password spraying efforts.
Storm-0940 has been active since 2021 and is described as a China state aligned actor. The group does not have a known industry alias. Their targets included the aerospace and defense, semiconductor, government, and think tank industries. Fourteen specific industries were listed in the report. Microsoft confirmed that the attack did not stem from a vulnerability in its own systems or products. Instead, it was a supply chain issue originating from the ConnectWise application.
The peak of the password spraying activity occurred on October 11. During that day, the system made 2,900 attempts per minute, with over 111,000 failures total. Forty-eight accounts each logged more than 1,000 failures. Nearly 98% of this activity resulted in lockouts. However, the AI agents managed to sign in to customer tenants using the compromised application credentials. More than 200 sign-ins occurred in under four minutes using just four sets of credentials. The agents performed directory operations, reconnaissance, and changes to user roles and groups within these tenants.
Security firm Huntress observed specific behaviors that raised alarms. On October 15, an application added itself to the top administrator role in one tenant. This action triggered a specific alert in Microsoft Defender. The attackers also planted new secrets set to expire in 2028. By October 16, Huntress had confirmed unauthorized sign-ins for 162 tenants. The following day, almost 1,000 tenants showed sign-in activity from the attacking IP address, with more than 300 of those attempts succeeding. The pattern of the attacks, moving through tenants in sorted order seconds apart, suggested a script working from a pre-made target list.
The use of ARTEX distinguishes this event from many traditional attacks. Microsoft noted that the framework can utilize any AI model. Threat actors in various regions are adopting this technology. Public reports this month linked similar breaches at South Korean financial institutions to this tool. In those cases, the suspect was likely based in China and used the framework with Anthropic’s Claude Code. The developer of ARTEX, known as Autumn-27, announced that the project will stop receiving updates and will become closed source.
ConnectWise has taken steps to address the unauthorized access. The company suspended its Cloud Backup, SaaS Security, and Cloud Management services as a precaution. A phased restoration process has begun, with recovery expected to take several days. So far, there are no indicators that partner backup data was compromised. Microsoft advises customers to keep affected app connections disabled until formal restoration instructions are issued. It also recommends reviewing activity logs from this month where possible to check for any anomalies.
This incident underscores the changing landscape of cybersecurity for non-experts. The integration of AI into offensive operations means that defenders have less time to detect and respond. The speed at which these agents operate can bypass traditional manual checks. For organizations in the South Central states, reviewing their supply chain security and monitoring for rapid, automated access patterns is now more critical than ever.
