.. is it worth the effort to have all messages in this command-line suite be separated out into a message catalog?

My experience in dealing with developers in India (i.e., with a very biased sample with respect to the entire population) is that English is a non-issue. I have no experience with working with developers in China (from, but not in).

Extracting messages into a catalog isn't too big a deal. It's one of those problems you just throw time at. Getting those messages translated (and the translations checked), can be a large, long, frustrating undertaking, even for large enterprises that do it on an ongoing basis. You either have to impose early "language change" freezes so that translations can all be done at once, or you stagger releases as translations become available. Neither is fun.

I'd run the question by your product manager. Not phrased as "should we do this?", which just invites a "Yes!", but phrased as "If we do this, it'll cost us X$ to do the message catalog, and Y$ (where Y might be "unknown, and we'll need outside help") to get the translations done. Where should that fit on the list of priorities?"