ARTEX AI agents automate enterprise attacks with Nmap, PsExec and WinRM
Original title: ARTEX AI Agents Use Nmap, PsExec, and WinRM to Automate Enterprise Network Attacks
Tests found agents retry after failures, while Microsoft saw attempts against more than 2,250 Microsoft 365 tenants.
Reveal Labs tested ScopeWeaver, a localized fork of the ARTEX penetration-testing framework, and found agents switching among Nmap, PsExec, WinRM, RDP, and related tools after failed access attempts. Researchers supplied credentials, so the lab work evaluated execution rather than an unaided compromise. ExtraHop said the findings follow a November 13 FBI FLASH warning. Microsoft separately documented ARTEX-assisted authentication attempts against more than 2,250 Microsoft 365 tenants using compromised ConnectWise application credentials.
Unconfirmed: Exposed consoles and user-agent strings do not by themselves prove malicious ARTEX activity, Censys and ExtraHop caution.
- The original repository was removed on October 8, but forks remain available.
- ScopeWeaver once declared success after only confirming WinRM reachability.
- In another simulation, Impacket PsExec opened a shell and set NoLockScreen to 1.
- Censys reported 1,465 exposed consoles and warned templates inflate apparent diversity.
- Google observed execution bursts lasting roughly two to 25 seconds.
Impact: Automated intrusion attempts using supplied or compromised credentials
- 1.Correlate scanning, exploitation attempts, and technique changes rather than isolated alerts.
- 2.Test candidate user-agent and JA4 signatures against historical traffic for false positives.
- 3.Validate agent-reported success against supporting evidence before treating an objective as achieved.
- ATT&CK
- T1046 Network Service DiscoveryT1021.002 SMB/Windows Admin SharesT1021.006 Windows Remote ManagementT1021.001 Remote Desktop ProtocolT1078 Valid Accounts
- Vendors
- MicrosoftConnectWise
- Products
- ARTEXScopeWeaverNmapRevealX
- Organizations
- Reveal LabsExtraHopFBICensysGoogle
From the original article
3 min read at gbhackers.com
ARTEX-derived AI agents orchestrate enterprise intrusion attempts using Nmap, PsExec, and Windows Remote Management. Controlled testing showed agents switching techniques after failures, generating detectable network activity even when they failed to achieve their objectives.
By Mayura Kathir · gbhackers.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.