I think just checking that the contents of some of the docs are contained in the README would be better than checking the contents are strictly matching some assembly of multiple file + hard-coded strings. It will help make contribution more streamlined. At the moment it is causing too much pain.