I found the following sentence online; it's a comment about three players A, B, and C that regularly compete against each other.
Aに勝てるのがBだけでBに勝てるのはCだけなの良いな
To me it seems like the structure of both subclauses (Aに勝てるのがBだけで and Bに勝てるのはCだけなの) is really similar, so what warrants the change from が to は in this scenario? Is it the proximity of the latter subclause to the main predicate?
This is a case where multiple means of emphasis, namely the cleft sentence structure and だけ, and multiple possibly contrasting statements are making the choice of the right particles hard, or in a way easy as all combinations seem fine.
Aに勝てるのがBだけで、Bに勝てるのはCだけだ。
Aに勝てるのがBだけで、Bに勝てるのがCだけだ。
Aに勝てるのはBだけで、Bに勝てるのはCだけだ。
Aに勝てるのはBだけで、Bに勝てるのがCだけだ。
The first pattern, with が and は in that order as in the original sentence, sounds to me like it is explaining the rules centering around B. It first states a characteristic only B possesses, and then goes on to explain another characteristic of B in terms of its power balance with other things.
The second and third sentences seem to give equal weight to B and C (which come after が in the second), and A and B (which come before は in the third), respectively.
Only the last one sounds slightly off to me. The first part seems to focus on A, but it is not even mentioned in the second part. If anything, the sentence sounds like an explanation about what C is like, with additional information about what it is not: It is not the one that beats A because that’s B.
Without だけ, the first combination sounds a bit weird to me, at least unless someone has asked what B is like to solicit this statement. Otherwise, the exclusive focus on B in the first part of the sentence seems unwarranted.
Aに勝てるのがBで、Bに勝てるのはCだ。
I think all four possible combinations are fine in this instance. When you use は, you create a slot that needs to be filled, and then you fill it with the answer. When you use が, you describe a characteristics of something and then reveal its identity with your answer (I hope this makes some sense).
For example, if you say
ご飯にかけられるのは、ふりかけなど
You describe something that needs to be filled (what can be put on rice) and then you fill it (ふりかけなど).
If you say
ご飯にかけられるのが、ふりかけ
You describe a characteristic (ご飯にかけられる) and then reveal its identify (ふりかけ).
This is why this next sentence is slightly odd
ご飯にかけられるのが、ふりかけなど
Because you are revealing an identity but it is ambiguous (ふりかけなど).
Now coming back to your original sentence, both work because each thing (A, B, C) is unique.
If we compare
(1) 空を飛べるのが鳥類などで、海を泳げるのが魚類などで、陸を歩くのが哺乳類など
(2) 空を飛べるのは鳥類などで、海を泳げるのは魚類などで、陸を歩くのは哺乳類など
IMO (1) is slightly unnatural because you expect something unique (and not have "etc"), though I wouldn't say it's downright wrong. In contrast (2) has no issues IMO.
If you make the groups unique, it again reads naturally:
(3) 空を飛べるのが鳥類、海を泳げるのが魚類、陸を歩くのが哺乳類
