The second part of 6.3. says:

To convert from WTF-8 to potentially ill-formed UTF-16, run these steps:

Decode from well-formed WTF-8 to code points [link to 6.2.]
Encode from code points to potentially ill-formed UTF-16 [link to 4.1.]

Note: This conversion never fails and, if the input is well-formed in WTF-8, is lossless.

If "WTF-8" in the first quoted line means "potentially ill-formed WTF-8", then the provided steps cannot be used since the algorithm referred to by the first step is defined only for well-formed WTF-8.

If, on the other hand, it means "well-formed WTF-8", then the bolded part of the note in the last quoted line is extraneous and misleading.