Word list protection should report triggering message

When word list protection gets triggered, it also immediately redacts the offending message. This makes it difficult to detect false positives.

It would be nice if Mjolnir reproduced the entire offending message in its admin room whenever word list protection is triggered.

Start by locating the word list protection trigger and the code that reports moderation events to the admin room. Verify the behavior by checking that the entire offending message is reproduced there when protection triggers, while the original message remains redacted.

Written by the indexing model from the issue text.