On Fri, 2004-03-26 at 21:06, Hackworth, Keith A wrote: > I need some suggestions building a rule for this one... Only thing that > hit was bayes - 40-44% probability of spam. Of course, I trained bayes > already. > > I'm running Postfix/Amavisd (with Spamassassin module). Only cf file I > have is my local and bigevil. <snip>
A similar spam come through on one of the mailing lists I subscribe to -
it got tagged as spam:
X-Spam-Status: Yes, hits=8.6 tag1=-999.0 tag2=3.0 kill=5.0
tests=BAYES_50,
L_RCVD_IN_MANY, PYZOR_CHECK, RCVD_IN_BL_SPAMCOP_NET,
RCVD_IN_DSBL, RCVD_IN_NJABL, RCVD_IN_NJABL_PROXY, RCVD_IN_SORBS,
RCVD_IN_SORBS_SOCKS
X-Spam-Level: ********
All of them are network checks however, so I assume that you didn't have
them turned on. If you get SA to learn it as spam, bayes should kick in
the next time something like this appears.
You could also try Chris's new anti-porn-spam rules...
HTH
--
-jamie <[EMAIL PROTECTED]> | spamtrap: [EMAIL PROTECTED]
w: http://silverdream.org | p: [EMAIL PROTECTED]
pgp key @ http://silverdream.org/~jps/pub.key
20:30:02 up 6 days, 6:32, 2 users, load average: 0.42, 0.63, 0.91
signature.asc
Description: This is a digitally signed message part
