The auto-discovery for search boxes in unknown websites is very interesting. You don't have to register special rules for websites anymore. Only a concern, this seems to support character-encodings for Japanese language (Shift_JIS, EUC-JP, etc.). I hope this will be updated for other languages, e.g. Chinese.
I would love if searches performed from webpages would be added to the searchbox history, like the searches made from it directly, so I can repeat them at will (I have it set to show the searches chronologically and often do different searches simultaneously).
I tried Conform Searchbox (CS) as a possible replacement for the search box sync (SBS) addon. However, I like the latter better because SBS supports (unlimited) undo for the search box contents. Both SBS and CS are broken when Search WB is installed though. Bummer.
Useful, but has some bugs. Eg: can't have a space in the regexp, otherwise entries will be messed up (can be corrected by editing directly in about:config).
NB: using %20 might be a good workaround if you need to use the space character, eg: [&?]q=is\%3Astarred\%20([^&]+) will work to recognize keywords from a google bookmarks search url, eg: https://www.google.com/bookmarks/l#!view=threadsmgmt&q=is%3Astarred%20firstword%20secondword
Very useful extension and much easier to work with than Searchbox Sync.
But it has some quirks. Some websites have default text in their search fields. Something like "Search..." or "Search our site". By default Conform Searchbox copies this garbage to Firefox searchbox each time website is opened. This problem has long solution, but I am not sure technically inexperience will be able to modify website's regular expressions that easily.
Options also may use some work. I would love to be able to sort websites.