Anthropic Warns Claude User of Infostealer Abuse, Security Panel Discusses

Summary

A security panel discusses a reported Anthropic response to infostealer-related Claude account abuse, identity-data sales, software vulnerabilities and cuts to CISA assessments.

Key points

A Reddit user reported that Anthropic warned them of attempted Claude account abuse, signed out sessions and removed the payment card on file after an infostealer infection.
The panel discussed an FBI investigation into a service reportedly selling driver's license records and the risks of collecting identity documents through third parties.
Shiny Hunters claimed a Florida DMV breach; the speakers said the claim was not confirmed.
The panel urged patching an authentication bypass in JROG Artifactory, an unauthenticated root-access issue in Proxmox and an unauthenticated remote-code-execution vulnerability in Langflow.
Speakers discussed the reported end of six free CISA security assessments and the potential impact on smaller critical infrastructure organizations.
The panel highlighted an Outflank NetNTLMv1 rainbow table of under four terabytes and an accompanying tool for using it.

Article Details

Topics: [08:12] Reported AI agent activity and incident response considerations
[18:35] Reported Anthropic notification and remediation of Claude account abuse
[30:30] Driver's license data sales and identity verification risks
[33:58] Unconfirmed Shiny Hunters claim involving the Florida DMV
[39:20] JROG Artifactory, Proxmox and Langflow vulnerabilities
[48:34] Cuts to CISA security assessment services
[53:43] Passkey research and NetNTLMv1 rainbow tables