Technology
Chinese Hackers Use DeepSeek AI to Automate Cyberattacks
cyberpress.org1h
THE BRIEF
1
Palo Alto Networks Unit 42 identified a threat actor using DeepSeek AI via the Hermes Agent framework to autonomously identify vulnerabilities and launch cyberattacks with minimal human intervention.
2
The AI agent scanned over 647,000 internet-exposed instances of the n8n workflow platform and targeted CVE-2026-33017, though autonomous exploitation attempts failed due to security configurations.
3
The actor successfully exfiltrated data from three Citrix NetScaler targets and executed commands on 11 Marimo notebook instances using manual techniques, impacting over 460 systems in total.