Yaml

opass-yaml already exists, and this PR just makes opass a frontend to opass-yaml.

The main reason to do this is just to have one front-end for opass things. Also, there might be some benefit in that once a user sets up opass and twiddles envvars for it, it would be nice if that flowed through to opass.

I'm asking for review here because I don't use opass, so an actual user of it might see a mistake I missed.