Yeah, that had me confused for a bit. And you would never use "charset=wtf-8" (as in the title for this page) because the spec says:

"Any WTF-8 data must be converted to a Unicode encoding at the system’s boundary before being emitted. UTF-8 is recommended. WTF-8 must not be used to represent text in a file format or for transmission over the Internet."

Some specs like to claim things that are out of their jurisdiction. A format spec has no say on where that format is used. It's best to ignore such hubris.

And in this particular case it doesn't even make sense because the entire point is to round trip WTF-16. If that requires one "system" to communicate WTF-8 with another "system" (which is really an arbitrary boundary) then so be it. And anything that expects UTF-8 will need to deal with "invalid" data in a usecase-dependent way anyway.