Should we consider maintaining an official `SKILL.md` file(s) for this library?

Opening the discussion here. I know of other projects who maintain official `SKILL.md` files that devs can add to their repositories (although they're much bigger than this library, but still).

Looking into it a bit it appears that it could be useful, particularly because of the API churn we've had which can confuse models, or sometimes they just plain don't know that their information is stale, which is what the file help shortcut.

Please pitch in if you work with skills a lot and have an opinion either way!

Start by reviewing the official SKILL.md examples referenced in the issue and the library's recent API churn. Determine whether maintaining one or more files would address stale model information, then define their scope and ownership. Done means the project has a decided approach and, if accepted, clearly specified files and maintenance expectations.

Written by the indexing model from the issue text.