What's the reason for using different approach for Codeforces and Codechef? Codechef is spawning a page that periodically uses XHR to get the status, whereas Codeforces parses the my submissions page. That means that if I close the status page, I will not get a notification (which is arguably a minor issue, because it's not worse than without the extension).

Codechef uses a simple polling mechanism to poll for the result of submission, so I can easily poll the endpoint from the extension whereas Codeforces uses websocket connection and I couldn't figure out what the packets were doing exactly. So, I lazily decided to go with simple parsing. If someone can figure out an endpoint (they must have it for backward compatibility), and tell me about that, I would surely make an update.

Another small issue is that due to the server load, it sometimes takes a while to display the status page, and the verdict is present there when this happens. As such, the submission is never in the "waiting" state, and the script ignores it.

I didn't consider that. I can just make it say the topmost result, but that would be irritating if you want to go through your old submissions for a contest. Moreover, I haven't experienced that with the few contests I have participated in.

Will this also work for the system test?

I will have to wait for a contest to try that out.

Any plans to support more judges?

Not currently. But, I would love to get some pull requests which add support for other judges :)