I have had some success running spatially explicit CMR analyses in a Bayesian framework, using BUGS or JAGS. These approaches can be useful when random effects or hierarchical model structure is involved. I have also found them helpful in cases where ML approaches like secr crash (e.g., due to few or no recaptures on multiple detectors), in which case putting a reasonable prior on sigma allows reasonable looking results to be produced. However, I emphasize that fancy analytical approaches are not a panacea for decent study designs (e.g., proper detector spacing in relation to the animal's scale of movement).
SCR using BUGS/ JAGS in R (Additional material from a recent workshop)