Honestly was stumped that there was no more explicit mention of this in the Anthropoc docs after reading this post couple days back.
Sandbox mode seems like a fake sense of security.
Short of containerizing Claude, there seems to be no other truly safe option.
Honestly was stumped that there was no more explicit mention of this in the Anthropoc docs after reading this post couple days back.
Sandbox mode seems like a fake sense of security.
Short of containerizing Claude, there seems to be no other truly safe option.