It all depends on if said generality is even wanted. There is a gigantic threshold between being able to introspect the inputs or parent transactions in a meaningful way or not. It’s like two different universes.
The point of this exercise if you like, is to show what you can do by staying within the exact set of fields CTV commits to, just with more flexibility or programmability than you get with something like a singular use of TXHASH.
And you get a lot of utility out of it without really violating said threshold boundaries.
edit: See Optimistic ZK verification using MATT!