Spam filtering

On this page I give some hints on how to filter spam. I am quite conservative in this respect as I don't want to to loose any mail which is not spam. That's why there are currently only a few hints.

Filtering Spam from Asia

As I get a lot of chinese and korean spam, I started to filter the following header entries (procmail syntax):

* ^Content-Type: .*ks_c_5601-1987
* ^Content-Type: .*[gG][bB]2312
* ^Subject: \=\?GB2312\?.*
* ^Subject: \=\?big5\?.*
* ^Content-Type: text/html; charset=euc-kr
* ^Content-Type: text/html; charset="euc-kr"
* ^Subject: \[[gG][bB]2312\].*
* ^Subject: \[[Bb]ig5\].*

Filtering Spam Mailers

I do also filter for Spam-mailers that identify themselves with an "X-Mailer" header entry. I have no guarantee that these tools are only used for spams, by feel free to take a look at my list of tools from which I already got some spam: spam_mailers.asc.

You can use the file for example with procmail to filter out some of the spam, e.g. with the following rule:

BIGMAILER=`formail -zXX-Mailer:`
FILTER_FOLDER=/home/<HOME>/mail/FILTER

# Check for X-Mailer headers hinting for spamming tools.
:0
* ? echo "$BIGMAILER" | fgrep -f $HOME/spam_mailers.txt
{
  EXITCODE=77
  LOGABSTRACT=all
  LOG="Filter: Sender using a spam-tool. You are probably a spammer. "
  :0:
  $FILTER_FOLDER
}

This filters quite some chinese spam and I never had a false positive.

Last change: 22-May-2002

felix@nice.ch