Some spammers found a sneaky way to bypass Discourse security. But not for long. Here’s what community moderators need to watch for:
New users with proper profile, three legitimate but low quality posts (or AI-generated content) with autobiographer, first like, first emoji, first reply badges
Next post is longer, seemingly more elaborate, and contains at least two links: one of them is legitimate and one of them is SEO spam.
The SEO spam link has a short anchor that is hidden next to the legitimate link.
In other words, be watchful about seemingly enthusiastic new users who post short first contributions, and watch out for obfuscated links.
One of the trigger warnings is the discrepancy between email username (at usually some large provider) and the actual username; profile picture looks legit (and was probably stolen from an actual account.)
No, I didn’t. Discourse trust levels have been working very well so far, and I think it’s important not to rely on more code for most of the use-cases. Minimalism is an important feature for me, and for future life on our planet.
I would suggest harder regulation against spammers (and the advertising industry in general) but this is another topic.
We have Akismet activated. Akismet does not flag it. See background on Are you experiencing AI based spam? - #13 by guidoleenders. In our case, using hotmail etc. is not possible, but then they revert to using temporary domains. Messages are quite well-crafted and seem realistic and relevant probably for most readers not looking for spam.
In the cases we have seen there are indeed sometimes replies without links and sometimes a double link is contained on the first reply.