With the recent news about litellm being compromised, I was wondering what techniques other people use (if any) to sandbox their applications to protect themselves. Up to this point, the only sandboxing I've done is with docker on my coding agents like pi. Not really so much for malware reasons, it'
