The example in LDML Version 35 Part 7 Section 5.19 is riddled with
1) U+1A21 is called TAI THAM LETTER HIGH KHA; calling it 'ka' is wong.
I suggest 'kha' or, better still, 'high kha'.
2) (a) 'Asat' is the name of a printing Burmese character,
U+103A MYANMAR SIGN ASAT; the corresponding characters in the Thai Tham
script are U+1A7A TAI THAM SIGN RA HAAM and U+1A7C TAI THAM SIGN
(b) The character used in the example is the invisible stacker
U+1A60 TAI THAM SIGN SAKOT - I suggest you label it 'sakot'.
3. (a) Only U+1A75 to U+1A79 are tone marks. As noted in 2(a), U+1A7A and U+1A7C are something else. Typing them in in the sample sequences would lead to garbage in, garbage out.
(b) Applying those rules to U+1A7B TAI THAM SIGN MAI SAM, which is a repetition mark. Both <1A7B, 1A60, 1A45, 1A76> and <1A60, 1A45, 1A76, 1A7B> are legitimate sequences, with rather different phonetic
effects. But typing the former leads to the latter by the rules given.
(c) In short, reduce [\u1A75-\u1A7C] to [\u1A75-\u1A79]in both places where it occurs.