This work ended up being carried out on the client side.

The main reason for this was two fold:

1) Both directives we needed to accommodate are purely client-side directives and so it is much cleaner if the server never needs to know anything about them. This will mean the developers never need to touch their schema to accommodate a client side directive.
2) We need a top level package which will handle directives on the server side and this solution will need a spike.