Maybe it is, but in the current age of AI-Vulnerability research I am not going to trust that there are no obscure exploits that breaks that sandbox.