After studying the new.log, I believe I've found the source of the problem. 
I'm passing encoded HTML into a javascript function for display in a
"lightbox" window with an image.  I believe httrack is seeing:

<a href="/foo.html">

and is dutifully trying to parse the url.

Short term solution (for my httrack script) is it use the following filter to
exclude any URL that contains single or double quotes.

-*\'* -*\"*

But I will have to rethink how to encode and display these URLs passed to the
javascript to solve the problem for external httrack requests.