I kept at it today, but running out of my mini vacation so unsure how much time I'll have to finish the last couple pieces. The only things really remaining is to add all CEL tokens to the Lexer so it can skip caveat function bodies, and some handling for arrow expressions and subject relations. I've tried to explain how to attack those problems as best I could if anyone is interested in taking a stab at it https://github.com/swantzter/authzed-ts/issues (assuming I've been able to wrap my head around how those things actually work, it's getting late) Admittedly I am considering pushing it out as-is as an alpha as it covers all my own use-cases. But hey, it shows the direction I was going for the simpler abstractions I have out in my projects, so goal achieved I guess