Microsoft ties ConnectWise breach to AI-driven campaign

Businesses that rely on ConnectWise tools faced a wave of account break-in attempts this month after a supply chain attack gave intruders a way to sign in as trusted applications rather than as ordinary users. Microsoft said the activity traced to Storm-0940, a group it describes as aligned with the Chinese state, and that the campaign leaned on an open-source tool called ARTEX, which runs automated penetration tests using large language model agents.

Microsoft's security blog said the intruders reached a multi-tenant ConnectWise application without authorization, then attempted sign-ins at more than 2,250 Microsoft 365 tenants. In some cases, the path from first targeting to privileged access took under three hours. Microsoft described ARTEX as "a system of LLM-driven agents that run an autonomous penetration test," and said it accelerates scanning and reconnaissance, password spraying and abuse of OAuth applications. The company also noted that its own systems, products and services were not compromised or vulnerable in the campaign.

A separate password-spraying push targeted 327 accounts at eight organizations and produced more than 151,000 failed logons with no successful entries. On Oct. 11 that effort hit a rate of 2,900 attempts per minute, all rejected, and logged more than 111,000 failures over the day; 48 accounts each saw over 1,000 failures, and close to 98 percent of the activity was locked out.

The application credential reuse proved more productive. Using four application credentials, the AI agents signed into customer tenants from a shared IP range and recorded more than 200 sign-ins in under four minutes. Inside some tenants they carried out Graph and directory operations, performed reconnaissance, and changed app credentials, identities, groups and privileged roles.

Security firm Huntress reported that on Oct. 15 it saw an app add itself to the Global Administrator role in one tenant, the behavior behind a Microsoft Defender alert labeled "Application self-assigned Global Administrator role." Those actions included planting new secrets set to expire in 2028. Huntress confirmed unauthorized sign-ins at 162 tenants by Oct. 16, then counted almost 1,000 Entra tenants with sign-in activity from the attacking IP the next day, more than 300 of them successful. The attackers moved through tenants in sorted order, seconds apart, which Huntress said suggested "a script working from a target list."

Microsoft said ARTEX can run with any AI model and that threat actors in several regions use the tool. Public reporting this month linked it to breaches at South Korean financial institutions, and CrowdStrike said the suspect in those separate attacks was likely based in China and allegedly used ARTEX with Anthropic's Claude Code. ARTEX's developer, who goes by "Autumn-27," posted on GitHub that the project "will no longer be updated" and would become closed source.

ConnectWise detected the unauthorized access and, as a precaution, suspended Cloud Backup (formerly SkyKick), SaaS Security and Cloud Management. It has begun a phased restoration expected to take several days and says it has no indicators so far that partner backup data was compromised. Microsoft said an affected app or even a successful sign-in does not by itself mean data was accessed or taken, and it advised customers to keep affected app connections disabled until formal restoration instructions are issued and to review activity from this month where logs permit.