Built for authorized penetration testing, ARTEX allows operators to direct AI agents in ordinary language. The agents can carry out sequences of tasks that would otherwise require an attacker to perform steps manually.
Microsoft attributed a campaign against hundreds of organizations to Storm-0940, a China-state-aligned group. After a ConnectWise compromise, attackers used stolen application credentials to attempt sign-ins to more than 2,250 Microsoft 365 tenants. That figure represents attempted access, not confirmed breaches. In a separate password-spraying operation, more than 151,000 login attempts against 327 accounts across eight organizations failed, Microsoft said.
Google researchers reported ARTEX activity affecting organizations in the United States and Europe, including semiconductor, telecommunications and enterprise software companies. They observed financially motivated attackers and suspected espionage groups. In one intrusion, agents generated 240 log events; researchers also noted repeated errors and signs that agents interfered with each other.
Barracuda traced an intrusion into its cloud environment to a cloud access key that had not been revoked after an AI vendor finished its work. The attacker used the key on Nov. 1, then spent about 44 hours reaching around 10 cloud accounts and searching compromised servers for Ethereum keystore files, extended private keys and seed phrases. Barracuda said no wallet material was present and acknowledged its failure to revoke access: “That is on us.”
Amazon described attempts to bypass identity checks at two crypto exchanges using deepfake videos and altered documents. Its evidence, collected from Oct. 3 to Nov. 5, included recorded failures: one verification vendor rejected every deepfake attempt, including those that completed the required face-movement sequence. Amazon assessed ARTEX involvement with moderate confidence and said it could not confirm activity against every organization mentioned in the records.
Other incidents included an ARTEX-related attempt to collect AI-service credentials and use stolen keys to fund further processing. GitBook also told customers to replace integration secrets after an AI-agent attack exposed infrastructure credentials. In South Korea, Everspin said it detected ARTEX-related attempts in financial service environments.

Comments (0)
No comments yet. Be the first!