Hi Robert,
The code is there and should work but has not been thoroughly tested and is indeed not documented.
Specifying a sampling distribution such that generalised stepping stone works well is the part that is especially badly tested/documented.
Are you interested in developing the method further, or just apply the method? If the former, it may make a good starting point, but if the latter it is not recommended at this stage.
Cheers,
Remco