in reply to Re^4: Postfix: Piping an email into a PERL script
in thread Postfix: Piping an email into a PERL script

Since the transport is insecure, armor the message. See gpg (or the like) as one possible way to assist. It could be set up so that only a validly signed message from an allowed user can pass through.

I am not saying you have to go this route, just pointing out some weaknesses in the area of security, and this will probably be the last I say on this particular thread.

--MidLifeXis

  • Comment on Re^5: Postfix: Piping an email into a PERL script

Replies are listed 'Best First'.
Re^6: Postfix: Piping an email into a PERL script
by Zzenmonk (Sexton) on Apr 26, 2013 at 15:51 UTC

    Unless I am wrong, gpg protects the content of the mail, the so called body not the header sometimes called the envelope. Accordingly all header fields can be spoofed. Just imagine I spoof the Reply To: field of a mail. Depending on the case, you would respond to me and not to the sender of the email. Gpg will not protect you against this. It will just prevent me to read the message.

    My best knowledge of this matter is: IPV6 is designed to protect email traffic. It supports encryption and certificates at protocol layer level. Unfortunately the only implementation of IPV6 I have seen until now are at Internet-backbone level. Most of the mail server still use IPV4 and do not implement encryption and certificates for a simple reason: you partially close your inbox if you do and so you might miss some emails of potential clients.

    The best medicine against depression is a cold beer!
[reply]
      No, GPG does more than that.

      By using public/private keys, GPG is able to tell who the message is actually from (assuming private keys were not stolen). The incoming message would be signed (encryption of the message isn't necessary) by the user's private key, and you'd then use that person's public key to verify that it was actually sent by that person.

      Thus, by using GPG, you don't need to implicitly trust the headers, you can verify if the header has been spoofed or not.

      -Scott

[reply]

        Ok! I will make it short. My post is about piping an email into a Perl script without fooling around with the local mailer definition of postfix (see the other post here). Basically my solution simulates the smrsh feature of sendmail. I do not intend to discuss more than this here.

        The topic of email security is of an other order of magnitude. As stated you can only protect your emails by encrypting of the communication link like https. This is only available if you use IPV6.

        Facts are security is matter of time and resources. With an infinite amount of time or infinite resources you can breach any security feature, even PGP.

        Now to round up the case, to read the content of an encrypted email, you first capture it from the Net, for example with tcpdump. Next you write a program and you crack it.(dot) Depending on the encryption used, this may last several centuries. Of course you can make like EFF, that just created a machine dedicated to crack the DES encryption in 1998.(dot again)

        If the communication link is not encrypted it will be pretty easy to capture interesting emails and to proceed with the above classical procedure. If the link is encrypted, it will be very difficult to find the mails your are looking for, potentially it will be impossible. (last dot)

        K

        The best medicine against depression is a cold beer!
[reply]

In Section Cool Uses for Perl