Posted by Virus Bulletin on May 7, 2013
Flaws in IBM Notes and Exim/Dovecot easy to mitigate.
Two recently discovered vulnerabilities in mail processing software could give an attacker access to a targeted system without the need for any links to be clicked or attachments to be opened.
When email security experts talk about "malicious emails", they usually mean emails with malware attached, or emails that link to websites hosting malware. Both are commonly seen in today's spam, and are also regularly used in targeted attacks.
In fact, the term "malicious email" is something of a misnomer here: there is nothing malicious about the emails themselves, and the simple action of a mail server accepting these emails, or a user viewing them in an email client, won't cause any harm. As a consequence, only static analysis of the attachment and URL tend to be performed by security software, thus making it important for the files and URLs to be scanned when they are accessed.
However, two recently discovered vulnerabilities demonstrate that in some cases the emails themselves can actually be malicious.
As most such attacks ultimately exploit vulnerabilities in Java, allowing Java applets to run inside an email client is inadvisable - especially for a client, like Notes, that usually has access to various internal databases. IBM has promised to come up with a fix, and in the meantime, an advisory already offers two ways in which these vulnerabilities can be mitigated.
Another vulnerability, in a default configuration for Dovecot and Exim, could allow for remote code execution using specially crafted emails and wouldn't even need the emails to be opened by the recipient.
Exim is a popular mail-transfer agent and Dovecot is an IMAP/POP3 server, with the two commonly used together on Unix-based systems. So commonly, in fact, that Dovecot provides a default configuration file for Exim to make the two packages work together. Researchers at RedTeam Pentesting have discovered a vulnerability in this configuration that could allow for remote code execution.
To exploit this vulnerability, an attacker would merely have to embed the commands in the SMTP envelope's From address, which would then be executed on the server running Exim. The vulnerability can easily be fixed by making a small configuration change in Exim. It is unclear how many mail servers use this particular setting, but the Immunity blog does some back-of-the-envelope calculations to suggest there might be quite a few of them.
It is good to keep in mind that both cases are exceptional and easy to mitigate: there is currently little reason for end-users to fear that merely viewing or receiving an email can be dangerous. There is no evidence in either case that the vulnerabilities have been exploited in-the-wild. Still, they are a reminder of the fact that in email the sender has some control over what happens - and that in some cases this could have unwanted consequences.
Posted on 07 May 2013 by Martijn Grooten