Method shelters are an alternative concept to limit the consequences of monkey patching in ruby to a specific context. Method shelters are an alternative approach to refinements, which was suggested at the end of 2010.

Method shelters were proposed by @flexfrank and you can check out a presentation from the end of 2011, that explains the concept of method shelters and different concepts to solve the same problem. You may also have a look at a prototype implementation at github, based on ruby 1.9.2, and a little readme explaining the concept.

The author gave a presentation at the AOSD conference. Personally I like method shelters better than the proposed refinements, what do you think?