One practical key difference though is that the compute environment for the JavaScript is already built under an assumption that none of the software it's runni

One practical key difference though is that the compute environment for the JavaScript is already built under an assumption that none of the software it's running can be trusted. So practically speaking, malicious JavaScript could, perhaps, exploit weaknesses in your sandbox but it can't save files the browser doesn't allow, read state the browser doesn't allow, make network connections the browser or the hosting machine's network configuration don't allow, etc., etc., etc. It's... An odd hill to choose to die on.