VAPOR PANDA AI-Driven ARTEX Global Espionage Campaign
VAPOR PANDA is using the ARTEX agentic AI framework and DeepSeek LLM to conduct simultaneous global espionage against foreign ministries and critical infrastructure sectors.

Browse public community intelligence reports, source analysis, and threat research.
1,319 intel reports
VAPOR PANDA is using the ARTEX agentic AI framework and DeepSeek LLM to conduct simultaneous global espionage against foreign ministries and critical infrastructure sectors.
Microsoft Outlook vulnerability CVE-2026-124194 allows unauthenticated attackers to perform network spoofing and steal NTLMv2 hashes by bypassing external-content warnings.
China-nexus group CopperDragon operates the Sparrow malware ecosystem while Blind Eagle targets Latin America with RATs, alongside significant AI infrastructure breaches and high-volume Microsoft patching.
SignDrop is a local privilege escalation exploit targeting the Windows Common Log FileSystem (CLFS) authentication mechanism to gain NT AUTHORITY\SYSTEM privileges.
Adversaries are exploiting CVE-2026-84869 and hijacked ScreenConnect sessions to deploy a multi-stage VBScript chain that enables worm-like propagation, credential theft, and cryptomining.
China-nexus actors and other threat groups are increasingly utilizing autonomous agentic AI frameworks like ARTEX and CyberStrikeAI to automate large-scale reconnaissance and rapid exploitation of cloud and on-premise environments.
An obfuscated JavaScript loader uses environment fingerprinting and multi-stage decryption to deploy ScreenConnect RMM software via spoofed Adobe branding infrastructure.
A Russian-speaking threat actor uses XaocMail phishing and ClickFix lures to deploy the Phantom53 DoH framework and Havoc post-exploitation agents.
The ARTEX LLM-driven offensive framework was observed targeting Oracle WebLogic, Langflow, and Flowise using adaptive exploitation techniques and distinct framework markers.
NETRYX RAT is an emerging Rust-based backdoor featuring experimental Large Language Model (LLM) integration for operator-assisted tasks and standard persistence mechanisms.