Rated 4 out of 5 stars
Variable syntax is weak and a little haphazard (couldn't stop on $ or %), on the bright side regexp in target folder name was a good idea (e.g. can save wikis into /%asf_rf%(\w+\:)%asf_rf%/).
Its great potential is almost unused, though.
After those saving file name updates of FF, ability to explicitly set name would be a good start. Just add one line for target file name and make at least title from current page's content accessible as %asd_ct% or something, so between it and file names it can be set case-by-case. Of course, since this means title gets grep'd now, it would go beyond workaround and be a great option in itself.
From bigger things, needs at least "if empty" fallback syntax or better true if-then condition. Add this and ability to get arbitrary META, and it could do just about anything short of julienne fries. :]
Most of all, needs updates - right now, last fixed beta from its site would be okay, especially since after first run got broken.
Rated 5 out of 5 stars
Pro: it can auto-create structure and filters are easier to access via toolbar button. (unlike Download Sort)
Con: ASF doesn't have per-filter memory, thus "it's root or nowhere" problem with most sites with unstructured URLs - like on xyzTubes and "thread only" subset of phpBB forums.
Proposed solution: to extract data from current page's elements the way SiteDelta does. This would easily catch Tubes' categories and forum structures even on most unknown phpBB.
Bug: currently help claims (\w) to be a word, but in fact $1d from it gives only 1 word character; (\w+) works, though.