The process looks fine, if a bit overspecified. I have some concern
about the choice of ombudsman*, given the extremely poor experience I
had with ICANN's ombudsman a few years back, although the lack of pay
will certainly discourage careerism.
My larger concern is that you need at least some general descriptions
of what would constitute harassment to keep it from being too
arbitrary. It seems to me that the essence of harassment is unwanted
attention directed to an individual or small group of individuals. If
the target told the harasser to stop and he continued, that makes the
claim more credible, although there are surely situations where the
situation is threatening enough that it's reasonable not to respond at
the time.
A few general guidelines like that would make the policy a lot more
concrete, and avoid "how could I know X would take that as
harassment?" issues.
R's,
John
* - ombudsperson is unacceptably species-ist, while ombudsman is a
Swedish legal term from the 1950s