I have created a fix for this so that discobot checks for the correct flag type. If the incorrect flag is used, it will reply letting the user know they didn’t quite do it right and it will remove the bad flag so that the user can try again.
Here is a demo of how it behaves now:
https://github.com/discourse/discourse/commit/18dad4cfbd8b26182561516c52f8b3bf233b3dcf