Score:0

Multi-message anti-SPAM filter in SpamAssassin

tn flag

My vendor's database has been stolen and I started to receive spam messages in a huge amount. It is being sent by spambots / spam scripts, so sender and recipient email addresses are always fake. The only common thing is that these messages are:

  • Always delivered in three copies (to all three addresses of mine)
  • Always with exactly the same title and content

Since there is a very little chance that any legitimate sender will send three exactly the same copies of the same message, I was wondering, if there is any way filter out all messages that appears in Inbox that exists there in exactly three copies?

anx avatar
fr flag
anx
Wouldnt bayesian matching pick up the rest if you just pass a few identified ones to sa-learn? E.g. piping `doveadm fetch ... | sort | uniq -c | grep "^3 " | cut -f2 | xargs -r .. doveadm copy tempfolder ... && sa-learn --spam temfolder/`
trejder avatar
tn flag
@anx Thanks for your suggestion. I'll try to implement it. If there will be any valuable results, I'll surely post about them here, in an updated answer. But given the fact that each "three messages pack" is always different in any possible aspect (auto-generated sender, different content, auto-generated recipient) then I assume that no bayesian matching is able to cherry-pick each such "three messages pack" next time it comes to my Inbox.
I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.