About > Stigmergic antispam

Stigmergic antispam

I am developing a combined blacklisting and spamtrap system to spot spammers early before actual mail delivery occurs. This site is a testing facilty for this system.

In one line

Let's feed spammers with fake addresses (spamtrap), and block (blacklist) *any* mail from IPs who is trying to send mail to these addresses, or even block any IP traffic from them.

The basic idea

  • Spammers collect their email addresses mainly from web pages, where they are exposed.
  • SPAM comes from an information asymmetry: an MTA has to presume any incoming mail is not spam, so it has to check for content after having accepted the mail, or rely on errors (the typical spammish behaviour) from the sender. Current antispam techniques are have uncertain oucome and are time and CPU consuming. RBLs or other centralized techniques often arbitrarily deny access to sending email to legitimate hosts.
  • There is another information asymmetry which has never (AFAIK) been exploited: the spammer collects emails from web pages presuming they are good ones. If a potential recipient populates his pages with non existent addresses in its MX domain, only spammers or UCE senders will use them.
  • Exploiting this second type of information asymmetry allows a receiving MTA to discriminate good mail from spam.