Code I write, be it python, php, HTML or JS is still plain text as a format. I can open any of these code documents in any dumb text editor.

The fact that some other tool can do more with these files, parse them, execute them, "extract" some logic from them isn't relevant to the fact that I can open, read and edit these files as a pure and plain text file.

Therefore adding parsable parts/logic to files like YAML headers (or any self created style) doesn't make these files 'non plaintext'.

Following your logic would mean that doing any markup in a text file would make it somehow 'non plaintext'. I remember old Readme files with headlines being underlined by adding a new line full of - - - - (dashes). And stuff like this. Because these could theoretically be used to parse the files and extract headlines from the text.