Zero Day Room
Live
Threats

Storm-0940 Uses AI Agents to Breach ConnectWise Supply Chain

China-aligned threat actor Storm-0940 compromised the ConnectWise Cloud Backup supply chain, using the ARTEX AI framework to launch a massive automated attack against over 2,250 Microsoft 365 tenants.

China-aligned threat actor Storm-0940 compromised the ConnectWise Cloud Backup supply chain, using the ARTEX AI framework...

A China-state-aligned threat actor compromised the ConnectWise Cloud Backup supply chain to access Microsoft 365 customer environments. Microsoft identified the group as Storm-0940, which used AI agents to automate a credential attack against more than 2,250 tenants.

Storm-0940 leveraged stolen ConnectWise application credentials to authenticate to customer environments without interactive sign-in. The attack exploited trusted third-party software, providing access beyond the application's intended purpose. Sectors targeted included aerospace, defense, automotive, energy, financial services, government, legal, manufacturing, media, military, semiconductors, software, technology, think tanks, and non-profit organizations.

AI agents automate credential abuse at scale

Automated AI agents performed rapid password spraying to maintain persistent access across hundreds of organizations. The campaign, active between October 10 and 14, 2026, targeted 327 accounts across eight organizations.

Microsoft telemetry captured over 151,000 failed logon events, with a peak of 2,900 failures per minute on October 11. The AI tool fired 2,900 sign-in attempts per minute. Forty-eight accounts received more than 1,000 failed login attempts. Lockout responses represented nearly 98% of all activity. No successful authentication, token issuance, Graph call, or directory change was observed during this password-spraying phase.

On October 15, AI agents gained access to ConnectWise cloud application identities and began authenticating to customer environments using the same IP range. Microsoft observed that some attacks resulted in privileged cloud access within less than three hours of initial targeting.

ARTEX framework enables autonomous attacks

Storm-0940 used ARTEX, a Go-based agentic AI penetration testing framework, to automate scanning, reconnaissance, password spraying, and OAuth abuse with minimal human interaction. ARTEX is an open-source, multi-agent LLM system written in Go with a Next.js frontend.

The threat actor has been active since at least 2021 and first detailed by Microsoft in November 2024. Its use of ARTEX in AI-driven attacks increased from early October 2026. The developer made ARTEX closed source on October 8 after public reporting linked it to attacks on South Korean financial institutions.

Security researchers found the attack footprint extended globally using a custom ARTEX deployment. Prompts directed AI models to divide investigation tasks and review findings. Analysis of prompt histories revealed requests for bulk personal data, selling stolen data, extracting payment gateway API keys, exploiting NFT minting workflows, and gaining unauthorized administrative privileges.

AI agents progressed to Microsoft Graph and directory operations, including reconnaissance and credential modifications. These modifications enabled persistence through other identities and cloud services, even after the original compromised application was mitigated.

Microsoft response and mitigation guidance

Microsoft issued specific recommendations for customers to investigate compromised credentials and monitor Microsoft Graph activity. The company recommends reviewing service principal sign-ins and Microsoft Graph activity from October 1, 2026.

Customers should review Microsoft Entra audit records for application credential additions and privilege changes. They are also advised to review Microsoft 365 audit records for unusual mail or file access. Microsoft recommends enabling threat protection for AI services in Microsoft Defender for Cloud and applying least privilege to multi-tenant application permissions.

Monitoring credential additions and anomalous app-only activity is critical. Microsoft advises preserving evidence before removing unauthorized credentials. Microsoft Defender XDR provides detections for suspicious Microsoft Graph activity.

ConnectWise issued an advisory on October 16 detailing unauthorized access attempts related to Cloud Backup. The company temporarily shut down Cloud Backup, SaaS Security, and Cloud Management services. It began restoring them in phases on October 17, a process expected to take several days. ConnectWise completed a full reset of accounts and credentials on October 19. Huntress reported 173 unauthorized logins across 173 tenants as of October 17, with the actual number likely higher.

Customers should review Microsoft 365 audit records for unusual mail or file access as recommended by Microsoft.

Related coverage

More from Threats