# -*- mode: org -*-

Something like spamgourmet.

* WHY THIS EXISTS
Having a public email address is an invitation to spam. Private email
addresses allow you to filter your spam based on how much you receive
and from whom it came. Instead of giving your main email address to
every web site you run across, you can give a private email address to
each web site (e.g., "facebook@myspam.example.com" for Facebook). That
way if you no longer want to receive email from a given sender, you
can easily filter it out.

Additionally, by having site-specific email addresses you can also
tell when a breach of security happened. If I tell Home Depot my email
address is home-depot@myspam.example.com and I start receiving spam
from Nigeria on that address I know that Home Depot has had a security
breach. I can use that to inform Home Depot, Twitter, my friends that
there's been a breach, in addition to checking credit cards for
possible fraudulent charges.

In short, having private email addresses per organization allows you
to keep tabs on your private life and help keep you more secure.

It would be remiss to not mention spamgourmet.com, a wonderful service
that was the inspiration for this code. I've been a happy spamgourmet
user for many years, but unfortunately it has some problems this
software attempts to address:

1) spamgourmet.com and its various addresses are too well known. By
   hosting this on your own domain you can prevent, for instance,
   Facebook from filtering your email address.
2) spamgourmet addresses are very long and hard to communicate over
   the phone or when filling out forms at, say, a doctor's
   office. This is attempting to provide much shorter addresses that
   will raise fewer eyebrows and can be communicated more easily.

* HOW IT WORKS
Once the software is installed (see INSTALLATION, below), it will
start receiving all email directed at it (such as by procmail or
sieve), and it will examine the destination address (i.e., "To"
header) to determine whether it should deliver the email to the final
recipient or not.

SpamCat relies on your MTA delivering mail from your spam domain to a
real email address that can be used for IMAP/POP/etc. This can be
accomplished in any number of ways based on your MTA/LMTP/LDA setup;
an example for Postfix/Dovecot is below.

When SpamCat receives a message it looks at the username (the
left-hand of the "@" in an email address) to determine the
sender. Thus, "foo@bar.com"'s sender is "foo".

Once the sender is determined SpamCat will see how many messages are
remaining for that sender, and if the sender has no remaining messages
then the message is thrown away.

If the sender has messages remaining then the message is delivered
as-is with the exception that the Subject header shows how many
messages are remaining and the total number of messages available
(e.g., "Subject: XANAX FOR NIGERIA - [19/20]" shows that there are 20
total messages from this sender, and there are 19 remaining).

* RECOMMENDED USAGE
1) Create your own spam-handling domain (e.g., "spam.example.com")
2) Use private email addresses in that domain anywhere you're asked
   for an email address (e.g., giving ShopRite an address of
   "shoprite@spam.example.com")

* PREREQUISITES
1) Perl 5
2) Email::Simple
3) SQLite 3
4) DBD::SQLite

* INSTALLATION
1) See INSTALL to install the library and spamcat executable.
2) Once the spamcat executable is installed you'll need to add it to
   your procmail, sieve, or some other similar device that writes
   email to standard input and receives the transformed email on
   standard output.
3) You will also need to create a config file for spamcat (see
   config/spamcat.conf for an example) somewhere that can be read by
   spamcat.
4) Make sure you pass in the spamcat.conf file to spamcat via the "-c"
   argument (e.g., "spamcat -c /home/me/etc/spamcat.conf").

* EXAMPLES
** Postfix
*** main.cf
Postfix needs to know that the spam domain should be handled locally.

#+begin_example
  Virtual_mailbox_domains = spamcat.domain somewhere.else
#+end_example

*** virtual aliases table
Make sure postfix can route the spam domain to the user who should
receive it.

#+begin_example
  @spamcat.domain  hidden-address@somewhere.else
#+end_example

** Dovecot/Sieve
The sieve file for the user which receives mail for the spam domain
should pipe the email into spamcat, which will handle final delivery
for the spam domain.

#+begin_example
  require ["fileinto", "envelope", "vnd.dovecot.filter"];

  if address :contains "to" "@spamcat.domain" {
    filter "spamcat" ["-c", "/path/to/spamcat.conf"];
  }
#+end_example

** DNS
The MX record for your spam domain should point to the MTA that hosts
spamcat.

#+begin_example
  spamcat.domain MX 10 yourmta.domain.
#+end_example