I've just sent in a PR for a new draft that fixes this.

On Fri, Jun 5, 2020 at 12:04 AM Shiro Kawai <xxxxxx@gmail.com> wrote:
The spec says it is the inverse of maybe->lisp-values, but what follows is the same as values->maybe and doesn't seem like the inverse of maybe->lisp-values.
If it is the inverse, producer must return two values, the second one is a boolean to indicate nothing/just.  Don't know if it's useful, though.