ARTEX AI credential-theft campaigns
Also tracked as Amazon; unnamed Chinese-nexus groups
0days in Daily CyberNews
2linked CVEs
3ATT&CK techniques
Relationships
Show 2 hops
exploits suspected named in the same story vendor · Select a box to open its page · All relationships
Context
Amazon Threat Intelligence (13/11/2026): MadPot sensors captured, from 03/10 to 05/11/2026, ARTEX AI-agent requests carrying the agents' own campaign records: mass theft of AI-model credentials from gateways and app builders (One API, LiteLLM, Dify, Langflow, RAGFlow) and deepfake bypass of identity checks at two crypto exchanges, with at least four agent sessions running at once. Amazon links the activity to disparate Chinese-nexus groups through simplified-Chinese prompts, China time-zone settings and tooling; it names no group and does not confirm a successful hit for each CVE, so both links are hedged Google GTIG (16/11/2026) lists ARTEX among autonomous multi-agent frameworks (with CAIRN, CyberStrikeAI, PentAGI) used by financially motivated and suspected China-aligned espionage actors, alongside LLM patch-diffing and resale of stolen LLM access through "LLM mixers" (APIMart, JetNight); no group or CVE named
Chinese-nexus (Amazon, moderate confidence)