Ok, new issue I have come across - no biggie but for that annoying Goodreads thoughput threshold rearing it's head again...
The API to add to a shelf requires a Goodreads bookid for the book. So for every book you first have to do an API call giving the ISBN to retrieve their bookid (if it is known). Then a second call to actually add the book to the shelf.
If we were running the calls through a centralised server, you could store a database cache of ISBN to Goodreads id mappings, and over time reduce the throughput. However for now at least we aren't...
Given the similar issues with LibraryThing currently being discussed in the main forum this plugin could fairly quickly share a similar demise...