Oh man. This is a super interesting question and one I've thought about a ton. I even ran a contest at one point the sole purpose of which was to try to get an idea of how to rate difficulty. I think the main conclusion I've drawn is that using a numeric ranking scale to rate difficulty is a pretty flawed method, and you're never going to get a satisfying result from it.
There are a couple of reasons for this. The main one is that difficulty isn't a one dimensional thing. There are different kinds of difficulty in DROD holds. Back in the AE days when the first user holds were coming out, a lot of the difficulty was execution based. Difficulty was more a matter of endurance. You had super long horde rooms and the difficulty came more from making small decisions about specific moves over the course of a 1000 move room. These days, a lot more of the difficulty comes from knowledge of the interactions and being able to think through their implications. The difficulty is more in reasoning out the correct overall approach to the room.
I think the community on the whole actually
doesn't use KDD as a benchmark for difficulty these days for this very reason. KDD has a kind of difficulty that just ins' very comparable to the kind of difficulty in TSS. There's some overlap sure, but for the most part they don't fit on the same scale.
Te other big problem with trying to put difficulty on a one dimensional numerical scale is that it's very subjective. Because there are different kinds of difficulty and different kinds of players, not everyone is going to rank holds in the same order in terms of difficulty. I may find One hold to be dramatically harder than another where you find it to be much easier.
I think the best solution is to try to use qualitative descriptions for describing difficulty whenever possible. "
Yeah, of the main holds, I think the first two (KDD and JtRH) are about equivalent in difficulty. It took me X amount of time to work through them. They start off fairly basic, but there are some very exacting horde rooms later on in both of them that required a lot of time strategically working out the best approach to dispatching all the monsters. The fifth game focuses a lot more on big picture puzzle solving and has some really devious puzzles that ask you to logically apply the tools you have to de-tangle an intricate knot of interactions."
I dont't know if that's what the person asking was looking for, but I feel like that's going to be more informative than any numeric answer ever could.
____________________________
Links to neat forum tools that I always have trouble finding:
Click here to view the secret text
I mostly agree with your ratings, except that for the official holds in particular there's an additional consideration: they might have different ratings depending on whether you play them in order, and therefore each hold is your first introduction to the elements it contains. For instance, TCB introducing pressure plates, lighting and arrow rotators for the first time would give it a rather higher "parsing" rating 
and KDD would have a higher "manipulation" rating than JtRH, for being the first introduction to goblins.
Similarly (going back to the original post), if you're going to rank the peak of TSS's difficulty as 21 out of 25, that doesn't leave much room for rating The Room of Humility, Gigantic Jewel Lost, Tarstuff Horticulture Facility....
