> This is not possible for the moment, and it would require 
> some coding (collecting all caches and links), as httrack 
> is not well dimensionned for that. Cross-linking is handled 
> per projects, but not yet globally .. 
> 
   That's where you misunderstood.  The object is to create
a simple master list text file.  And load, check and add to
that file each time something from that category is downloaded.

   Say creating a "pokemon.txt" file for all sites that you
would download with pokemon stuff on them.  And once a site
was gotten the url would be added to the pokemon.txt file,
so that httrack would check this list before downloading any
site that would call to check the pokenmon.txt file.
 
   If this was stored in the main directory of httrack
rather than in the projects directory updates shouldn't be a
problem.  Loading and checking a simple textfile shouldn't
be a problem either.

   I am not sure how much work it would be to add a text
file url list option or to update it after every use of it
during a download, but it seems that all the pieces are
there for this, so as long as it wasn't collecting the data
on each run from seperate project directories it should be
able to be implemented at some point.