A sample record set can be had here:
http://about.scriblio.net/downloads/ScriblioSampleMarcRecords.mrc.zip

The most difficult challenge in importing MARC records is in identifying a field that can represent the source id the record. That is, a field that uniquely identifies that record in a way that can prevent duplicates. The field generally varies based on the system the records are coming from, but you can get a view of the data in the record by ticking the “debug mode” checkbox. The example’s offered on screen there may help:

to use barcodes from Infocenter: [852][0]->subfields['p']
or to use the the Project Gutenburg ID: ['000'][0]->data
or to use the LCCN: ['001'][0]->data

The sample records are from Infocenter, so use the first example.

This recent message to the Scriblio mail list have a few more tips:
http://groups.google.com/group/scriblio/msg/9956c337d5c8802c

I just need to either be able to download a sample catalog to figure out what format to write my MARC records into (xml?) – or just to be told what file form my MARC records should be. (Sample would be best) However the links to the sample catalog are all broken or return “file not found.

http://groups.google.com/group/scriblio

Deletes are easy, as that can be checked when the location and availability information is checked (see http://about.scriblio.net/wiki/scrib_availability ). It’s not a serious factor now (and I haven’t coded anything for it), but the solution is at least pretty easy.

As for updating edited records, that hasn’t been much of a problem yet — I’ve had reasons to reindex the collection a few times already. Longer term, however, the question of even getting a list of recently edited records out of the ILS has seen considerable discussion on my vendor’s mail list.

Scriblio has all the same user management roles as WordPress. The dashboard image was taken from a user with “author” or “editor” role, that’s why you don’t see the Users tab.