I'm not a fan of the reason weights being in Smokey's messages.
It's misleading if you don't know what it's telling you. "Potentially bad keyword in answer (1)" reads like there is one bad keyword. "Bad keyword in link text in answer (78)" reads like there are 78 bad keywords. Not that it's a reason weight.
I don't think adding the weight adds much; it takes up space in the message; the people that care have scripts to add it; and it's confusing to new users.
hm, not sure, Smokey uses a bayes too, and VLQs are at least somewhat likely to have contained offensive content in the very near past measured in seconds, so I guess it should be trained for that too (Smokey doesn't fetch previous revisions)
@SmokeDetector That's actually spam; it looks like it unsolicited-ly advertises some "capital market" service (user profile has website "stephaniecurrycapitalmarket.com").
@SmokeDetector While not spam per se, I'd prefer to see a system level block on the content. Given the content, I'd also like the user on the blacklist, until they contribute some appropriate content. Thus, tpu.