Document sorter, in perl, written with fuzzy searches, data extraction, and Baysean filtering, along with various other tricks that I haven't seen used in other document categorization systems. Should help with my paper load.

Also, and XML based (actively using namespaces/xpath/xslt, so json wouldn't cut it) network documentation and configfile creation system, as everything else I've seen out there sucks or is limited/nonextensible.