AI Agents Speed Up China Hack

Microsoft revealed on its security blog that a Chinese state aligned group known as Storm-0940 leveraged an open source AI framework to execute a large scale attack on the ConnectWise platform. The operation involved unauthorized access to a multi tenant application, allowing the attackers to attempt sign-ins at more than 2,250 Microsoft 365 tenants. In some cases, the group achieved privileged access within less than three hours of initially targeting a victim. The attackers used stolen credentials from the ConnectWise app to authenticate without interactive user sign-ins, bypassing typical login screens.

The group utilized ARTEX, a system of large language model driven agents that perform autonomous penetration tests. This technology accelerated scanning, reconnaissance, and the abuse of OAuth applications. The agents operated around the clock with minimal human interaction, creating a pace that left defenders with little time to detect or respond. While the exact method of credential theft remains unclear, the efficiency of the AI tools significantly compressed the timeline of the intrusion.

Storm-0940 has been active since 2021 and is described as a China state aligned actor with no known industry aliases. The name follows a convention for developing actors. This specific campaign targeted industries including aerospace, defense, semiconductors, government, and think tanks, with fourteen sectors listed in total. Microsoft confirmed that the activity did not stem from a compromise or vulnerability in its own systems, products, or services.

The attack also included automated password guessing, targeting 327 accounts across eight organizations. This resulted in more than 151,000 failed logons with no successes. At its peak on October 11, the password spraying campaign saw 2,900 attempts per minute, totaling over 111,000 for the day. Nearly 98 percent of this activity led to lockouts. However, the attackers also used application credentials to sign in from the same IP range, with more than 200 sign-ins occurring in under four minutes using just four application credentials.

Within affected tenants, the AI agents performed directory operations, reconnaissance, and changes to app credentials, identities, and groups. On October 15, Huntress, a security firm, observed an app adding itself to the top administrator role in one tenant. This triggered a Microsoft Defender alert for an application self assigned Global Administrator role. The attackers also planted new secrets set to expire in 2028. By October 16, Huntress confirmed unauthorized sign-ins for 162 tenants. A day later, nearly 1,000 Entra tenants showed sign-in activity from the attacking IP, with over 300 succeeding.

The pattern of the sweep, moving through tenants in sorted order at intervals of a few seconds, indicated a script working from a target list. This method contrasts with many traditional attacks due to the use of the open source pentest framework. Microsoft notes that ARTEX can use any AI model. Threat actors in multiple regions are now using this technology. Public reporting this month linked similar breaches at South Korean financial institutions to the same tools. CrowdStrike stated the suspect in those cases was likely based in China and allegedly used ARTEX with Anthropic's Claude Code.

The developer of ARTEX, identified by the handle Autumn-27, announced on GitHub that the project will no longer be updated and will become closed source, according to Reuters. This development highlights the rapid evolution of offensive AI tools. The shift to closed source may limit the availability of the framework for other threat actors, but the impact of the current campaign remains significant for the targeted industries.

ConnectWise suspended Cloud Backup, SaaS Security, and Cloud Management as a precaution following the unauthorized access. The company has begun a phased restoration, with recovery expected to take several days. There are currently no indicators that partner backup data was compromised. Microsoft advises customers to keep affected app connections disabled until formal restoration instructions are issued. Users should also review activity from this month where logs allow to identify any potential anomalies.

The situation underscores the growing threat posed by AI driven cyber operations. The speed at which the attackers moved from initial access to privileged roles demonstrates the need for faster detection capabilities. Organizations are advised to monitor their Entra tenant logs closely for signs of unusual sign-in patterns or role changes. The incident serves as a reminder that supply chain attacks can leverage new technologies to bypass traditional security controls.