Today, I can proudly announce the initial release of the myND planner for fully and partially observable nondeterministic planning tasks.
The code as well as some rudimentary documentation in the form of a wiki can be found at Bitbucket.
More technical and detailed descriptions of how the planner works algorithmically can be found in a number of conference publications. They mostly focus on the heuristics used to guide the search, but a brief description of the search algorithm is also given [MOHB10]. The earlier publications [BM08, BM09, MOHB10] deal with fully observable problems, whereas the most recent one [OM13] deals with partial observability. Regarding the heuristics used to guide the search, one publication [BM08] is about a delete-relaxation approach, whereas the others [BM09, MOHB10, OM13] use abstraction heuristics.