We put LaTeX (and Asciimath and PlantUML) parsing into Asciidoc kinds and it works well.

Running the LaTeX parser over Markdown is a total shitshow, of course, but we all knew that. Use at your own risk. No, the other people don't do it any better. They just don't test as much and don't even notice that sections disappear or are garbled, or that prose text gets mathified.