Intrepidus Group

Identifying Malware via User-Agent Headers

Posted: June 16, 2010 – 10:17 am | Author: | Filed under: foo, Phishing, Security Management, Techno, Web Apps

Can you tell if a host is remotely infected just by a single HTTP request? For some malware the answer is yes.

By now, I think our readers are pretty familiar with PhishMe. As you can imagine, we see a lot of hits to PhishMe from a variety of browsers. And even better, we see a lot of hits to PhishMe from a variety of browsers where the user is likely to click on things. Each time a user makes a requests a website, the user’s browser sends a “user-agent” string to the web server as part of the request. A simple user-agent string looks like:

Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1)

Here’s a quick break down of what this string tells us. The Mozilla/4.0 portion indicates a Mozilla-based browser. This user is running Internet Explorer 7.0 and Windows NT 5.1 (Vista). You can check your user-agent here.

Now for Internet Explorer, it’s pretty easy to append information to this user-agent string by editing the registry. You will typically see a number of .NET related items coming from a normal user-agent header on a Windows system.

Windows UserAgent Registry

Where it gets interesting is when we see user-agents like these next ones. It seems that some viruses and malware (or “potentially unwanted software”) insert their name or a token into the user-agent string. Here’s some examples we found:

Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; AntivirXP08; .NET CLR 1.1.4322) 
Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1; Trident/4.0; .NET CLR 1.1.4322; PeoplePal 7.0; .NET CLR 2.0.50727) 
Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; FunWebProducts; .NET CLR 1.1.4322; .NET CLR 2.0.50727)

This last user-agent string shows the user’s browser is infected with FunWebProducts which are not nearly as much fun as they sound.

If the malware instead appends a token to the user-agent string, this token could be used to track the user from site to site or to trigger certain behavior on malicious websites. We identified several pieces of potentially unwanted software and tallied the number of infected users using PhishMe. The graph below shows the most common pieces of malware found in user-agent strings:

Infected User-Agents Pie Chart

We looked at IE 6, 7, and 8. Using this total number of “infected” users, we broke down the infections into browser version and divided by the total number of users running each browser version to get the percentage of each version’s population which is infected. As it turns out, the portion of infections is pretty similar across all IE versions. Isn’t IE 8 suppose to protect users much more than IE 6? This is a bit of a surprise, but suggests something we’ve known about the current state of attacks. You can have strong software controls, but security still depends as much on the user operating the software safely. Even given a browser that is relatively hardened against threats, users must know how to identify sites with malware and phishing schemes in order to stay safe. Patching and updates are important, but so is user awareness.

PhishMe clients can contact our support team for an analysis of your user base.

Both comments and trackbacks are currently closed.

One Trackback

  1. [...] This post was mentioned on Twitter by cubitouch, xanda. xanda said: Identifying Malware via User-Agent Headers [...]


This site is protected with Urban Giraffe's plugin 'HTML Purified' and Edward Z. Yang's Powered by HTML Purifier. 24799 items have been purified.