Staff Reports
Microsoft just went public with a clear and alarming threat report. Its Threat Intelligence team says a China‑state‑aligned group it tracks as Storm‑0940 used an open‑source agentic AI framework called ARTEX to automate a large supply‑chain style campaign that abused ConnectWise application identities. This is not a techy parlor trick. Microsoft says ARTEX helped the hackers try thousands of sign‑ins per minute and in some cases gave them privileged access in under three hours.
What Microsoft found about Storm‑0940 and ARTEX
The new report says Storm‑0940 used ARTEX to run reconnaissance, password‑spray attacks, OAuth/application abuse, and follow‑on Microsoft Graph activity. At peak, those ARTEX‑driven attempts hit more than 2,250 Microsoft 365 tenants at roughly 2,900 sign‑ins per minute. Microsoft recorded a related campaign with more than 151,000 failed logon events. The company stresses most attempts failed, and that Microsoft corporate systems were not breached. ConnectWise confirmed unauthorized access to some application identities for its Cloud Backup product and began suspending connections and restoring services.
Why agentic AI in cyberattacks is a game changer
Agentic AI frameworks like ARTEX can plan, act, and run multiple tasks at once. In plain English: give the system a job and it will try all the ways to finish it, 24/7. That means attackers can scale far faster than human crews. For supply‑chain attacks, where a single vendor tool can touch hundreds or thousands of customers, the math is ugly. This is national security risk, not just an IT headache. When the targets include government, defense, energy, and finance, our exposure is strategic — and the bad actors are using automation to compress discovery and compromise into hours instead of weeks.
What companies, MSPs and agencies should do now
Microsoft published detections, hunting queries, and mitigation steps that every affected tenant should follow: keep impacted application connections disabled, inventory service principal sign‑ins, preserve audit logs, and rotate or revoke credentials as needed. MSPs and vendors must audit OAuth grants by application ID — not just display names — and force credential rotation for third‑party apps. In short: don’t wait for a media statement before fixing a gaping access hole. If you run an MSP and think “that won’t hit me,” it already did — the attackers were casting a very wide net.
This episode should teach two lessons. First, supply‑chain risk needs stronger rules and real penalties when vendors fail to protect customers. Second, we must treat state‑aligned cyber campaigns as the geopolitical threat they are and respond accordingly — tighter defenses, smarter sanctions, and a firewall between critical services and foreign influence. China’s embassy has denied the charge, which is predictable. Meanwhile, Americans and businesses should stop hoping tech platforms will handle everything for them and start demanding real accountability and muscle from the people who run our national cyber defenses.
