I would try the following:

in "Scan Rules" I would clear the text box;
I would enter -*;
then +*www.mysite.com/test*

You would probably also get anything linked from the
root/test* files.  I think that you'll have to delete any
files that you don't want after the mirror completes.

I don't think that httracker downloads specific directories,
although this can be somewhat forced using the scan rules. 
Rather, httracker seems to follow the links in the anchor
tags (sometimes in javascript code) and scans/downloads
those specific files.