I’m seeing the same thing. Of the four spam I’ve had to manually moderate in the last week, two have contained just “IW” and “MD” as the body text and the spammy URL in the URL field.
(The other two were “a string of Chinese followed by a non-linked URL” and “Five spammy links and no other text”.)
Is there an easy way to hook a custom “Is this spam?” function into Antispam Bee? The filter I use on my hand-written sites would have caught all of these.
(Among other things, it requires a minimum of three non-URL words in the message body, at least two non-URL words per URL, and at least 33% of non-URL text to consist of words containing only ASCII letters as a privacy-preserving way to detect whether the post either is in English or contains an English translation. The design is focused around refusing things I wouldn’t want from a real human either.)
…granted, your honeypot markup is more advanced. I haven’t needed to implement that yet because, for my contact forms, I just refuse all attempts to use link markup of any kind and use the URL field as the honeypot.
