Let skill determine between an invocation on the app, and physical device
Currently, the Alexa app will resolve a listen() call within 1-2 seconds, requiring the user to make an immediate reply.
A physical device will wait 8 seconds.
By knowing on which type of device I'm running I can add a few seconds of silence to the playback, to give a good user experience, without them having to upgrade the app. (Assuming a fixed version of the app exists.)