CrowdStrike: Vapor Panda Used AI Framework ARTEX Against 200+ Organizations
The Hackers NewsAI & LLMs

CrowdStrike: Vapor Panda Used AI Framework ARTEX Against 200+ Organizations

Chinese state-sponsored hackers have been observed using an AI-powered penetration testing framework called ARTEX to hit high-value targets at scale, according to a CrowdStrike analysis. The activity is attributed to a China-based targeted intrusion adversary the firm tracks as Vapor Panda, also known as VexatiousTemper, first identified in February 2021 and overlapping with a cluster Trend Micro calls Earth Entries.

CrowdStrike says Vapor Panda used ARTEX against more than 200 distinct organizations worldwide, including foreign ministries in Latin America, Europe, Asia and the Middle East, U.S. semiconductor, defense and think tank organizations, and telecommunications providers in multiple regions. Post-exploitation activity was confirmed at nine organizations, where the attacker deployed BLOODALCHEMY, a ShadowPad derivative, followed by credential spraying, network scanning and reconnaissance. Password-spraying and credential-scanning activity was also seen against at least 12 additional organizations.

ARTEX was paired with DeepSeek and an as-yet-unidentified subscription-based commercial large language model, using planner and worker agents that coordinate to move from initial access to host reconnaissance and lateral movement. A human operator oversees the operation, directing agents to write web shells, extract domain administrator passwords and enumerate endpoint sensor coverage. CrowdStrike said planning agents consumed on the order of 1.4 billion input tokens in a single intrusion, indicating the scale of automated activity.

Recovered transcripts show ARTEX combined with established techniques: chaining an unsigned SAML assertion and an XXE flaw to query the cloud Instance Metadata Service, deploying web shells and abusing an unauthenticated orchestration service for root access in a production container environment, and using authentication keys unrotated since 2022 to forge an administrator token and read mailboxes. CrowdStrike said the agents pursued multiple attack paths in parallel across web applications, cloud services, containers, Active Directory, and mail and VPN infrastructure, shifting approaches when a path failed.

The disclosure coincides with CrowdStrike detailing four previously undocumented China-linked groups: Hollow Panda, Overcast Panda, Titan Panda and Veiled Panda. Overcast Panda has been linked to close access intrusions using its FlowCloud backdoor, deployed via DLL sideloading when targets travel to China. Titan Panda has targeted legal billing software, exploiting a hard-coded IIS machine key vulnerability in Intapp Time Server; attacks first detected in April 2026 led to deployment of an in-memory framework called IceGiant.

Source: The Hackers News

← Back to News