Wikitext is not parseable. I've posted a bunch of notes here maybe 2 months ago highlighting how absurdly horrible it is. There is no spec but even if there was one it would be unimplementable. There is only one parser: the one that powers Wikipedia, it's a mixmash of cursed PHP that is not portable. I don't know how you can think that is compatible with anything.
It's compatible with Wikipedia and most other wikis on the web.
Markdown is cool, but a NIP that just uses nostr to back up and synchronize articles across MediaWiki instances seems like it would be better than an asciidoc wiki platform that isn't cross-compatible
I wouldn't be against that if you want to do it, but it would be a completely different thing than NIP-54.