The GNU coding standards specify a procedure for generating documentation (fetch gendocs; then run it) which means that the GNU Coding Standards themselves do not need a copy of the gendocs script. So a simple fix for this issue would be to remove gendocs.sh and gendocs_template from the GNU Coding Standards. (At least that way we'd have one less copy of these files floating around. :-)
No thanks, gnulib is 300MiB, asking people who maintain standards/maintain to fetch that blob is unreasonable and which is why there is now a copy of gendocs.sh in the gnustandards tree. gnustandards also does not use the template as is, and has it modified. This stuff should be simplified somehow, for someone that has the time to look into that. Indeed. As far as I'm concerned, this is not an issue with gnustandards.
