> From: "Apolinaras 'Apollo' Sinkevicius (Carmel Music & Entertainment, LLC)"
><[EMAIL PROTECTED]>
>
> How do I make SA to send messages with certain score or origin to
> [EMAIL PROTECTED]
> I am done blocking those a**holes, so it is time to get government involved.
>
> Also, what do I have to enter in conf file (for whole system) to have
> all e-mail originating from any host in RBL list to be sent to [EMAIL PROTECTED]?
You can't make SpamAssassin do it, but you could make procmail take a message
which exceeds a certain threshold (via the number of *'s, not via numeric
comparisons) be sent to a program which would then send it to any address
you want.
However, you should consider how ethical (or, rather, unethical) it might be
to automatically submit someone's messages to an external agent without
verifying for yourself that it was indeed spam, both in terms of how it
affects possible innocent senders whose email is a false positive (raising
the threshold merely makes this less likely, not impossible), and in terms
of how it affects suddenly flooding that external agent with messages you
haven't inspected yourself.
Instead, I would recommend that you have tagged messages all go to a
particular mail folder, which you can inspect directly. If a message turns
out to actually be spam, deposit it into a second folder. Then, have a cron
job run which takes messages in the second folder and does things like: adds
the senders to your personal blacklist, reports the message to external agents
(RBLs, SpamCop, FTC.GOV, the Razor, etc), and any other machinations you want.
Then, last, have all of your mail which trips the "BLACKLIST" rules be
automatically deleted.
That way you're not seeing repeated spam messages, but you're not doing
something unethical, either. (nor do you have to craft report messages over
and over, since your reports to the external agents are being done
programatically)
My own strategy is a little different:
1) if a message trips a SA blacklist rule, and not a SA whitelist rule, then
procmail automatically deletes the message.
2) If SA says its spam AND it didnt' trip a whitelist rule, procmail sends
the message to a program which:
a) sends back to the sender saying they're being added to my personal
blacklist and detailing how they can appeal the process,
b) appends them to my personal blacklist (not an external agent)
c) deletes the message
3) If it's not marked as spam by SA, but still meets other spam definitions
(such as, it wasn't to/from me, wasn't to/from any mailing lists I have
acknowledged in my procmail rules, nor to/from any individual senders I
have set up in my procmail rules), then it gets deposited into an "other"
folder.
4) If mail in ANY of my folders is spam (stuff SA didn't catch), then I put
it in my "blacklist" folder. I have a cron job that comes through every
night and automatically adds those people to my blacklist (without reporting
it to anyone, nor sending the sender a warning: I know they weren't a
false positive because I inspected it myself, so I don't care to give them
the opportunity to appeal). I've thought about having this step ALSO
report them to an external agent, but I haven't implimented that yet.
But this is the only place I would do so, because it would be personally
inspected messages, not automatically processed messages.
(and, the programs in step 2 and 4 are the same program, invoked with
different CLI args)
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
