JFrog says six malicious npm packages used hidden install-time execution, JSONKeeper fetches, and sandbox checks to enable remote access.
Anthropic's most advanced AI is back online, but it's a shadow of the model that launched in June.