I have dropped this review as it has been overdue for quite some time. But since it still appears on my review-list, I took a look at it now in case this is of any interest.

I have little knowledge about TEEP and the rationale behind its design decisions. I trust that the author has that part figured out.
My interest was primarily in the requirements for HTTPS versus HTTP and how that was motivated.

A rather interesting observation in this regard was the attempt to "spice" the requirement language of the specification. See section 4:

"It is strongly RECOMMENDED that implementations use HTTPS."

This brings my thought to other interesting alternatives to spice requirements as defined in RFC 6919 like "OUGHT TO" ? ;)

But jokes aside, I'm not sure "strongly" is appropriate next to "RECOMMENDED".

But other than that I find no issues with the document.