ConnectWise and its customer organizations agentic AI attack

Microsoft has identified the China-aligned threat actor Storm-0940 as the force behind a supply chain compromise targeting the ConnectWise Cloud Backup application. The attackers utilized ARTEX, a Go-based agentic AI penetration testing framework, to automate reconnaissance and credential abuse at machine speed, significantly reducing the time required to move from initial targeting to privileged access.

What happened

Beginning in September 2026, Storm-0940 deployed the ARTEX framework to conduct large-scale, automated attacks across various sectors, including aerospace, defense, and financial services. The AI agents were configured to utilize multiple models, including DeepSeek, GLM, and Grok, through an OpenAI-compatible gateway. These agents were instructed to divide investigation tasks and review each other's findings to optimize offensive workflows.

On October 15, 2026, the agents successfully acquired the identities of multiple ConnectWise cloud applications. By leveraging these stolen credentials, the agents authenticated to downstream customer environments, bypassing the need for interactive user sign-ins. Once inside, the agents performed follow-on activities, including reconnaissance, modification of application credentials, and changes to privileged roles to establish persistent backdoors. In some instances, the time from initial targeting to achieving privileged access was less than three hours.

Prior to the successful compromise, the threat actor engaged in a massive password-spraying campaign. Between October 10 and 14, 2026, infrastructure associated with Storm-0940 attempted to sign in to over 2,250 tenants, peaking at 2,900 failed logon attempts per minute. Microsoft noted that the use of AI agents allowed the adversary to operate continuously, attempting multiple breach paths simultaneously without the constraints faced by human operators. ConnectWise temporarily suspended services to address the breach and initiated a phased restoration process. Microsoft emphasized that the incident highlights the growing risk of nation-state actors using agentic AI to exploit trusted software supply chain connections at scale.