CrowdStrike: China-nexus group VAPOR PANDA used the ARTEX AI agent framework with DeepSeek against more than 200 organisations from 17 October, with post-exploitation confirmed at nine
CrowdStrike said in a blog post dated 29 October 2026 that its OverWatch team identified a VAPOR PANDA campaign using the ARTEX agentic penetration-testing framework alongside the Chinese-developed DeepSeek large language model and a second, subscription-based commercial LLM it could not confirm.
CrowdStrike says that beginning on 17 October 2026 the group used ARTEX against more than 200 distinct organisations worldwide, including foreign ministries in Latin America, Europe, Asia and the Middle East; semiconductor, defence and think tank organisations in the US; and telecommunications providers in multiple regions, with activity from a single ARTEX controller against entities in at least 14 countries over an 11-day window.
CrowdStrike says it confirmed post-exploitation activity at nine organisations, where the group deployed the BLOODALCHEMY backdoor and carried out credential spraying, network scanning and reconnaissance, and saw password-spraying and credential-scanning against at least 12 more.
According to CrowdStrike, a human operator issued instructions in Chinese and steered planner and worker agents, which wrote web shells, extracted domain administrator passwords, chained an unsigned SAML assertion with an XXE flaw to reach a cloud Instance Metadata Service, read LAPS passwords, forged a mail server administrator token with keys not rotated since a 2022 export, and looked for Windows machines without the Falcon sensor; in a single intrusion the planning agents consumed on the order of 1.4 billion input tokens.
CrowdStrike says the campaign is separate from the ARTEX activity reported under the Storm-0940 designation, that VAPOR PANDA has since 2021 primarily targeted government, think tank, energy and extractive organisations across Europe, North America, Oceania and the UK, and that this is its first observed use of AI-driven intrusion tooling. CrowdStrike publishes BLOODALCHEMY C2 domains, IP addresses and file hashes.
The Hacker News reported the research on 30 October. (sources: CrowdStrike; The Hacker News)
- Product
- ARTEX agentic AI intrusion framework (with DeepSeek and a second commercial LLM); BLOODALCHEMY backdoor
- Versions
- n/a — threat campaign
- Exploited in Australia?
- unknown
- Patch to
- Follow CrowdStrike's advice: rotate signing keys, preauthentication keys and service credentials after any suspected data exposure, including keys in old backups and directory exports; enforce SAML signature validation, require IMDSv2 and limit cloud roles' command-execution rights; hunt for ARTEX markers (command lines containing artx or artex, artx-prefixed cloud roles, art18- administration-command comments, the C:\ProgramData\artx5v\ directory); patch internet-facing appliances; run EDR on all hosts, including servers thought unreachable from the internet; and block the BLOODALCHEMY indicators in CrowdStrike's post (for example nginxmysql[.]com, sqlserver-sqlite[.]com and www[.]ddduuuppp[.]com).