Agent based policy scanning is not something that can be done ad-hoc. When you enable a policy for agent based policy scanning it goes into a queue that revolves every 24 hours performing the assessment on those assets.
If you’re looking for ad-hoc policy scanning you would need to use the scan engine and a policy template. This of course requires authentication so either username:pass or the scan assistant.
And of course, if your target is a remote desktop not connected to the network then you can’t scan with the engine so the only option is to wait for the policy to re-assess the asset on it’s 24 hour interval.