auto learn spam for particular address?

[Steve Thomas]

On Thu, Apr 08, 2004 at 06:57:24AM -0700, Chris Yuzik is rumored to have said:
>
> I have a few email accounts on the system that have been seeded to a
> number of "get free stuff" websites and also used on the newsgroups.
> These addresses receive a lot of spam, and as one would guess, 100% of
> it is spam.

When signing up to "get free stuff", did the agreement that you accepted say that they would be sending you advertisements and/or providing your address to 3rd parties? If so, the messages you're getting from them or their partners isn't technically spam. You may not want it, but you agreed to receive it and they could be legitimate bulk mailers with proper opt-out procedures whose messages you're improperly (IMHO) training your bayes db with.

Seeding newsgroups is another story...


> Some of the spam received by this account gets pretty low SA
> scores,

Could be because it's not really spam...


> Any thoughts?

I don't know of a way to do this within MailScanner (could be one; I'm not a MS expert), but if you're using procmail, you could just create a recipe for that user which feeds the message to sa-learn.


--
"Being on the tightrope is living; everything else is waiting."
- Karl Wallenda