One AI swarm targeted 200-plus at once — and got into nine

CrowdStrike reported on October 29 that the China-based adversary it tracks as VAPOR PANDA used ARTEX, an AI-powered penetration-testing framework, against more than 200 distinct organizations in at least 14 countries over 11 days starting October 17 — including foreign ministries in Latin America, Europe, Asia and the Middle East, 19 U.S. organizations in pharmaceutical, semiconductor, defense, nuclear and financial sectors, and telecom providers in several regions.

The operation ran as a swarm, not a single break-in: a main agent coordinated planner and worker agents sharing a graph of findings and goals, steered by a human operator who directed them to write web shells, extract domain administrator passwords and map endpoint defenses. It ran mainly on DeepSeek plus a second subscription-based commercial model CrowdStrike could not identify, and the planning agents in a single intrusion consumed on the order of 1.4 billion input tokens.

The count that matters is narrower than the headline reach: CrowdStrike confirmed post-exploitation activity at nine organizations, where the attackers deployed the BLOODALCHEMY backdoor and moved to credential spraying, network scanning and reconnaissance, with password-spraying or credential-scanning against at least 12 more. CrowdStrike says China-linked adversaries are highly likely to keep adopting agentic frameworks like ARTEX over the next 12 months — the same framework, it notes, is already being picked up by multiple China-linked actors.