I agree with you concerning the evolution of agile methods with the rapid pace of scientific advances, and I think, although the popularity of some agile methods such as Scrum, XP, ASD, Kanban, Crystal, etc. they cannot support all project aspects (modelling, organizational, technical and quality assurance) and it is difficult that an agile process can be used for any kind of projects. I think that is time for a coalescence of those lightweight methodologies towards a unified agile process as it was made previously with UP in classic and heavyweight family of methodologies. A deeper evaluation of the practices supported by agile methods could be useful before evolution or coalescence (union). Within this context, you can take a look in this paper :
Multicriteria decision aid for agile methods evaluation using fuzzy PROMETHEE, Journal of Software: Evolution and Process, Wiley, 30(12), 1-21.