> Yes, I did want to allow downloading of linked
> images/JS/CSS - but it is also pulling html files,
> really indexing entire sites linked to the one I
> want. It is like there is no boundary, and I could
> end up with google on my desktop.

Try setting your filters (scan rules) to the following, correcting "myfriend"
as required:

-*  +www.myfriend.com/*
+*.png +*.gif +*.jpg +*.css +*.js