Threat Intelligence
Adversary methods, active campaigns and evidence-based threat research. No fear. Just analysis.

EU Commission Confirms AI-Assisted October Cyberattack; Data Impact Under Investigation
EU officials confirm an AI-assisted cyberattack detected October 22, with several applications taken offline. No staff-wide breach has been established.

Coinbase Says It Blocked $4.7 Million in Transfers to an Elon Musk Deepfake Giveaway
Coinbase describes stopping transfers to a fake Tesla livestream giveaway. The case shows why an authorized payment can still be driven by deception.

Exposed Ollama Inventories Reveal Signs of Outside Activity, Rapid7 Finds
Rapid7 finds suspicious model-name artifacts on exposed Ollama services. Its read-only study distinguishes possible outside activity from proven compromise.

Talos Tracks UAT-12643 Using Solana Memos to Keep Malware Connected
Cisco Talos describes UAT-12643 malware using Solana transaction memos, credential theft and clipboard hijacking, with North Korea attribution uncertain.

EggStreme Passive Backdoors Shift Attention to Inbound Connections
Bitdefender details rebuilt EggStreme malware, including passive backdoors, and explains why missing outbound beacons cannot establish that a host is clean.

Bulgaria Says It Ended Known Access in Five-Year Government Cyberespionage Case
Bulgaria says attackers accessed government networks from 2021; its November 18 update reports known access cut off and 90 days of heightened monitoring.

Entrust Reports 165% Rise in Deepfake Detections as Identity Fraud Changes
Entrust’s new identity-fraud report tracks rising deepfake and injection detections. We explain the dataset, its limits and the lessons for verification.

Malicious Repository Instructions Turn AI Coding Assistants into a Credential-Theft Route
Bitdefender describes malicious repository instructions and startup hooks targeting AI coding environments. Review project trust and reachable credentials.

Container Honeypots Show Exposed Cloud Keys Being Used Within Eight Seconds
An ACM CCS 2026 study tracks deliberate credential exposure in container images, showing why release checks and rapid revocation matter for cloud secrets.

Cyderes Links ‘tcherber’ to Stolen VPN Access and a Malware Service Operation
Cyderes connects the tcherber handle to VPN credential records and malware infrastructure. Its findings warrant access reviews, not assumed victim counts.