Claude Ai

Malicious npm package mouse5212-super-formatter stole files from Claude AI's working directory. The attacker's own GitHub token was exposed in the code, allowing researchers to trace exfiltration.

Varonis joins 27 other security vendors integrating Anthropic's Claude Compliance API, enabling enterprises to monitor AI conversations, detect data leaks, and enforce governance policies in real time.

Attackers are distributing PlugX malware through phishing campaigns impersonating Anthropic's Claude AI. The fake installer abuses a legitimate G DATA binary for DLL sideloading.