> 1) Don't do extended parsing -%P

No that didn't work. For those using WebHTTrack that is removing the tick from
the checkbox for "Try to catch all URLs (even in unknown tags/code)" under the
Links options.

> or perhaps
> 2) -*greybox/*[]

That didn't work either. I already had -* at the beginning so another -*...
isn't going to make a difference.

Since it only affected about 30 files I simply used Unix text processing tools
to correct the problem after the capture.

I think this is a bug with httrack. It should have some means of identifying a
path and not modify them. If it's referring to a file then yes you need to
alter its path and filename to match the local copy but for a path only it
should be left alone.