YAML is fine if you use a subset (don't use the advanced features - not like you know those anyway) and use explicit strings (always add " to strings), otherwise things may be cast when you did not intend values to be cast.
Example:
country: NO (Norway) will be cast to country: False, because it'll cast no (regardless from casing) to false, and yes to true.
That's a thorny question: Do comments need to be in the base standard, or can that be offloaded to those building on it? It doesn't look like it would be hard to have (comment "foo bar baz") in an expression and have a re-parser throw that out.
Is the complaint that no two groups of people will use the same comment standard if left to their own devices? It's not like the other data from different sources will always match up. What's one extra, and fairly easy to handle SNAFU?
That said, yes, I think I'd be more comfortable knowing there was an accepted comment format. The aesthetic seems to be Lisp-like, and I notice that the Lisp comment marker, the semicolon, is currently a reserved character, that is, it's illegal to use it unquoted. Maybe they're thinking of adding that at some point.
It doesn’t look like it would be hard to have (comment "foo bar baz") in an expression
That is pretty much what the official "solution" is for comments in package.json - add "//": "foo bar baz" keys to dictionaries and NPM will ignore them.
In practice it's terrible. You need real comments.