Including the contribution from both transformers and motors what would be the current at fault levels at the nominal circuit voltage and the current that is available at the line terminals of either main or tie breaker. Even if it isn't what is interrupted, it's what is available at the line terminals.
Well, for the mains, there is only 10KA available at the line terminals. For the tie there is 10KA on each side, but they two are not additive. But that is okay, I'll not pick at those.
As you may note from the "history", I'm very much in favor of permitting Secondary Selective systems to have an "underrated" interrupting duty for the duration of an automatic transfer. The transfer would not precipitate a downstream fault and the likelihood of a simultaneous downstream fault occuring durring a closed transition is negligible. Secondary Selective systems are commonly applied in many continuous process facilities, like refineries; which is why API supported it. I would have preferred wording similar to the 110.9, second paragraph to be applied. Nevertheless, the first paragraph “says what it says” and the “line side” of all breakers will see the current contribution of all sources – even if they don’t actually have to interrupt them.
While bracing is relevant, Code Making Panel 1 (CMP1) never spoke to the issue, nor does Section 110.9; it only speaks to determining the interrupting rating of the OCP involved. Frankly however, unless a downstream fault occurs simultaneously with the closed transition, bracing should be no more a concern than the interrupting capability.
What's really scary is when they use a "key-interlock" system and all three devices have a key. My initial Proposal advocated automatic only and the second Proposal's Substantiation specifically rejected manual.
