On 7/20/2013 1:35 AM, Andrea wrote:
Hi all.
Since a few days ago I'm being buried under spam messages that slip
through my amavis/SA setup.
The messages all look alike: plaintext with random junk + URL in the body.
Pastebin with a few examples here: http://g2z.me/ed64d
I've tried running a sa-update but I don't have enough samples
(yet). The thing that bothers me is that all the messages have been
classified as HAM by the auto learn (which I have now disabled).
What could be an effective rule/ruleset to block emails like this?
I assume you meant to say "sa-learn" rather than "sa-update"?
The main problem that I see with the scoring is that it is hitting on
BAYES_00. This may have been caused by auto-learn. You need to
manually learn these as spam using sa-learn to couteract that.
--
Bowie
