Since the SDK depends on crawlee, users might configure their dependencies so those two don't fit each other, which ends up with multiple @crawlee/core installs, which in turn might cause weird issues, since there will be two global configs. This could potentially result in the crawlee storage methods (like Dataset.pushData) working with the memory storage even when on platform.

We should validate this in the Actor.init call, there needs to be a single crawlee installation.