Chinese Hackers Use ARTEX AI Agents in Automated Global Cyberespionage Campaign

A Chinese-nexus threat actor known as Vapor Panda has deployed the ARTEX AI framework to conduct automated cyberespionage against more than 200 organisations across 14 countries.

The campaign, which began on 17 October, targeted high-value sectors including telecommunications, pharmaceutical, semiconductor, defence, nuclear, and financial industries. Researchers at CrowdStrike observed that the operation marks an escalation in the use of AI agents, transitioning from single intrusions to multiple simultaneous engagements directed by human operators.

AI-Driven Automation and Coordination

ARTEX is an open-source, multi-agent AI penetration-testing framework that was later made proprietary following its use in various cyberattacks. During this campaign, the attackers used the AI agents to perform reconnaissance, exploit vulnerabilities, deploy web shells, and obtain credentials. In one specific intrusion, the AI agents consumed more than 1.4 billion input tokens.

The threat actor also utilised DeepSeek and another unidentified commercial language model to coordinate planning and worker agents across various intrusions. While the AI agents performed many tasks autonomously, human operators issued instructions in Chinese and intervened at critical decision points. The attackers reportedly instructed the agents to skip Windows machines protected by CrowdStrike’s Falcon endpoint protection, focusing instead on unprotected hosts.

Targeted Sectors and Malware Deployment

Confirmed victims include telecommunications providers in the United States and Belgium, as well as government organisations in Greece, the Middle East, and Latin America. In an attack against a Greek government entity, the AI agents successfully forged an administrator token and read mailboxes by leveraging authentication keys from a 2022 mail-server directory export.

In addition to the ARTEX framework, the threat actor deployed the BLOODALCHEMY backdoor against government entities in Latin America and the Middle East. This malware has been active since 2023.

Autonomous Breaches

The capabilities of ARTEX AI were further highlighted by the Dutch Institute for Vulnerability Disclosure (DIVD), which reported a network breach in September 2026. DIVD stated that the ARTEX AI agent acted autonomously, moving from initial access to root privileges in just 12 minutes. The agent achieved this by chaining a Zammad zero-day vulnerability with a second zero-day for local privilege escalation.

The DIVD incident involved over 450,000 requests and 2,500 scripts, with the agency describing the activity as “super loud and messy.”