Info:  Note: due to www.thecoverproject.net remote robots.txt rules, links
begining with these path will be forbidden: /includes, /images (see in the
options to disable this)

Could this be the reason? It's the only error I got after the dump.

Mimetypes are as default (I leave untouched the first filtering line).

This is the aspect of the folder I got loading first 10 cover id sites (from 1
to 10):

<http://img187.imageshack.us/img187/8913/66095658.png>

Those php files are the jpgs :S