I don't really see the point of taking a sandbox and then explicitly granting an app inside the privilege to step out of it.
Without the source code and large amounts of resources it is not possible to determine whether an app is malicious or not. The only way would be for Apple to trust that developers actually do what they say they do. The unfortunate reality is that they cannot be trusted. I don't blame Apple for not taking that responsibility. Unfortunately that means the responsibility ends up with the user who cannot make the determination either. Such is life.