If I read the GitHub issue correctly, Firefox 105 returns "text/plain; charset=utf-8", Chromium 108 Dev Channel returns "text/html; charset=utf-8" when running this at console
var blob = new Blob([]);
var url = URL.createObjectURL(blob);
fetch(blob).then((r) => r.headers.get('content-type'))
.then(console.log, console.error);
The "text/plain" I can understand, an empty string. I don't understand "text/html".
I don't know how"application/x-unknown-content-type" got returned anywhere in the test in the GitHub issue. Are you sure that is what is returned by Firefox; it looks like that was just a parameter passed to the function?
I don't think the browser should return an empty string when it doesn't recognize the MIME type. I haven't seen that brought up; mentioning here so whatever decision is made re no MIME type overflows to unrecognized MIME type get converted to empty string.