That's actually kind of an interesting idea. I don't know of anything offhand but you probably could repurpose and/or reuse one of the Firefox web dev packages like Firebug or XPather. You'd use those to find the xpath that describes the parts you're interested in and then plug that into something that pulls the relevant chunks out of the tree.

At the least something like that might let you pawn off the tedious "dig through the page's source and figure out what you want" stage off on a non-technical person and you just tweak a standard skeleton as necessary.