Huxley now supports a subset of the staff version of the Darwin API, which provides significantly more detail over the basic one.
The client side access token is the same for both APIs and Huxley will choose the correct Darwin access token to use. If you pass through your token then you will need to use the correct one for the API being called.
There are a bunch of extra fields on the staff API, some have been renamed and some have a different type. For example, the times are all now real DateTime objects, which makes them easier to use, rather than simply display (or do complex parsing). Something to keep in mind with the staff service details is that you don't just get calling points. You get all locations, including passed stations and junctions (for example, Borough Market Jn. and North Kent East Jn. - but you won't get a proper location name for these, just a TIPLOC with whitespace trimmed).
There are also some improvements to the service API. Huxley now returns the service ID in multiple forms (original Base64, percent encoded, GUID and URL safe Base64). The service endpoint will accept any of these and it will also attempt to patch up any malformed IDs. Plain Base64 strings don't always play nicely in URLs (particularly if used in the path, rather than as a query string parameter).