Is it only me that just hates the idea of "marking" someone up-front, rather than deriving what you think of them from their actions? My point being that the way we interpret things should be "this thing is bad" -> "the person is bad", not the inverse "this person is marked as bad" -> "they said/did something" -> "that thing must be bad".

I feel like an extension like this is built around facilitating the latter: Judging the content or actions based on whether the sender is tagged as "good" or "bad", rather than looking at the words and actions themselves.