I think part of the confusion is that the program 'usearch61' does chimera filtering, as does chimeraslayer. If you're opting for chimeraslayer, the qiime devs typically recommend running
identify_chimeric_seqs.py on the rep_set_aligned.fasta file.
I'm not sure about your bigger question of recommended chimeraslaying approaches. I will note that pick_open_reference_otus.py mitigates some of the chimera problems by comparing against a known database of sequences, such as greengenes.