RC RANDOM CHAOS

Chinese Threat Actors Scale Attack Using AI-Powered ARTEX

· via The Hacker News

Original source

Chinese Hackers Use AI-Powered ARTEX to Target Over 200 Organizations Worldwide

The Hacker News →

Chinese state-sponsored attackers designated Vapor Panda launched a campaign beginning mid-October targeting over 200 organizations worldwide using ARTEX, a Go-based AI framework for penetration testing. Targets included foreign government ministries across multiple regions, U.S. defense and semiconductor firms, pharmaceutical companies, and international telecom providers. Nine organizations suffered complete compromise with attackers deploying BLOODALCHEMY, a ShadowPad derivative, following initial access.

ARTEX operates through coordinated AI agents overseen by human operators, integrating multiple language models including DeepSeek to automate reconnaissance, vulnerability identification, and exploitation. A single intrusion consumed 1.4 billion input tokens, indicating the scale of automated activity across the broader campaign. This approach enables minimal human involvement while executing parallel attack paths across web applications, cloud services, containers, and directory services—automatically pivoting when exploitation vectors fail.

Attackers combined multiple techniques including SAML assertion forgery, XML external entity injection, web shell deployment, container escape, and credential reuse to penetrate infrastructure. The rapid adoption of ARTEX across the threat landscape prompted its developer to restrict access. Researchers concurrently documented four additional Chinese state-sponsored groups, including Overcast Panda, which deploys the FlowCloud backdoor through direct physical access in China, and Titan Panda, which exploits hard-coded keys in legal billing software targeting law firms.

Read the full article

Continue reading at The Hacker News →

This is an AI-generated summary. Read the original for the full story.