Threat Lab

If you live in the US, you may have played sports, barbequed, or enjoyed the last long weekend of the summer outside doing something fun outdoors. Unfortunately, that wasn’t an option here in Boulder, where a large wildfire generated a thick plume of smoke and ash. So, what’s a malware analyst to do indoors on a beautiful day with toxic smoke outside? Why, spend some quality time with Koobface, of course.

I took a closer look at the worm’s behavior and also noted that, since the Migdal keylogger site went dark for the Koobface crew, they’ve switched to using a new domain as the dead drop for credentials stolen by the Koobface password stealer payload: m24.in, the Web site of some sort of media company based in India. The behavior I saw by the keylogger was virtually identical to that used by the Migdal variant, reported in a previous post. The payload is even named m24.in.exe, just like the Migdal payload was named after the domain where it posted stolen passwords.

It’s been a while since the worm changed its primary method of infection: For nearly its entire existence, Koobface has spread by manipulating the social network accounts of infected users so it appears the user posted a link to a video. Of course, the worm does the posting in the name of the user, and the link points to a page which purports to be some sort of streaming video, but actually pushes the malware on anyone who visits.

And, in order to take on the appearance of a real online video, it uses Flash.

read more…

Is there no honor among thieves anymore?

The other day I was looking at a remote access Trojan written in the PHP scripting language. The bot loads into memory on a victim’s computer when an unsuspecting user, for example, stumbles upon an iframe pointing to the PHP script embedded in a Web page. The code is  nicely appointed with such desirable features as the ability to execute shell commands on the host server, send a flood of data packets at another computer, and scan remote computers.

Once loaded into a victim’s browser, the bot connects to, and is capable of executing commands issued by, a botnet server–until the victim reboots their computer. But for most users, that’s probably long enough. If an attacker can execute commands on an infected user’s computer, installing more Trojans is just child’s play.

But someone appears to have embedded a surprise into this PHP backdoor: It’s another backdoor within the backdoor.

read more…

Is the team behind the Koobface worm taking a stance on the Israeli-Palestinian peace talks, or is this notorious worm’s most recent, bizarre twist just a coincidence?

We’ve seen Koobface hijack legitimate Web sites for more than a year, using them not only to host malicious payload files, but also to work as proxy command-and-control servers for the botnet. One such hijacked Web domain, migdal.org.il, popped up in a number of blog posts and on Web sites which list the domains used to host malware, as far back as this past May, when the Koobface crew began using a slew of new hijacked servers as distribution points for its malicious files.

And since the summer, Koobface has been delivering a password stealing Trojan among the several payloads it brings down to an infected computer. That Trojan’s name is migdal.org.il.exe, and the stolen passwords it scrapes from infected computers are sent right back to the migdal.org.il Web server, which is physically located at an ISP in the UK.

Migdal also seems to be (if you can believe the content posted to the Web site) a French jewish organization that provides aid and resources to Israeli children and border guards, and whose leadership opposes many of the Israeli concessions that Palestinian negotiators have requested during the long peace process. Have the Koobface gang gone political, or are they just capitalizing on a convenient situation with an abandoned Web site?

(Update: The site went down on September 3rd, the day after this post went live. Thanks, helpful ISP who shall remain nameless.)

read more…

A novel and pretty sneaky Trojan designed to steal financial data appeared on our radar screen last week. The Trojan, once installed on a victim’s computer, rootkits itself to prevent detection, then watches the victim’s browser for any attempt to connect to the secured, HTTPS login page of several online banks. When the victim visits the login page the Trojan has been waiting for, the Trojan generates a form that “hovers” over the login page asking for additional verification information.

“In order to provide you with extra security, we occasionally need to ask for additional information when you access your accounts online,” reads the popup window. Everybody needs extra security, right?

Of course, the additional information that the bank appears to be asking for is all information the bank already should have if you have an account there: The number on your credit and debit cards; a Social Security number; your date of birth and mother’s maiden name; The PIN code for your debit card and the security code printed on the front of any credit card issued by the bank.

The problem is, the form completely blocks the full page, preventing you from logging in — until you fill in all the fields in the form it displays. Then it sends that information (encrypted with SSL, mind you) to a server at the IP address 121.101.216.234, part of the address space allocated to Beijing Telecom.

Your bank may outsource some of its customer service tasks, but stealing your financial identity isn’t part of the normal services your bank provides.

read more…

Dear Customers: Please be aware that a crew of Russian malware distributors are circulating a spam message which looks like a subscription renewal confirmation from Best Buy, allegedly for one of our products.

The linked text in the message, however, leads to a Web site which performs a drive-by download. Please don’t click the links in the message; If you have any questions about your subscription, please contact support.

The spammers appear to have done some homework. Some, but not enough. Best Buy currently sells our products through their online software subscription service. Note to spammers: If you’re going to try to hijack our trademark, the least you could do is get the name right. Best Buy doesn’t sell anything called Webroot Spysweeper with Antivirus Product. Nor do we.

The email message claims it is a notice that your subscription has been renewed, and includes a serial number (which doesn’t work) and a transaction date of July 17.

The link in the message leads to the Web site of a small bed and breakfast in New Zealand, which has been compromised. We’ve informed the owners of that Web site of the spam campaign and asked them to take down the page referenced in the spam message.

I guess we struck a nerve, hurt some sensitive malware author’s pwetty widdle feewings, and ended up a target for attack, one that falls down. Too bad, so sad.

read more…

Yesterday, a few of the Threat Research folks and I had a little fun playing with a hack that had, for one day at least, pretty much decimated Google’s Image Search feature. One researcher, who stumbled into the attack purely by chance, found that a Google Images link to a map of the United States was, instead, redirecting hapless Web surfers to pages that deliver an installer of a rogue antivirus in the Security Tool family of fine, fraudulent products.

What really caught our interest was how the hack behaved, depending on the operating system and browser you used. With each different browser configuration, we were treated to one of several different, specially crafted malware delivery Web pages.

I’m not sure when the attack started, but we started analyzing it at around 10am, Mountain time. By late afternoon, the sites were offline and the attack no longer worked.

To test the extent of the hack, we played around with the manipulated search results using five different browsers: Internet Explorer 6 and 8, Safari 5, Google Chrome, and Firefox. All the browsers were set up with default settings in an otherwise identical installation of Windows XP SP3. We then searched for USA Map and clicked the second result that appeared under the header “Images for usa map.” (All but the first image result that appeared on that first page of results linked to the malicious Web site.)
read more…

As recently as a few months ago, malware distributors went to what looked like great lengths to craft complex, sophisticated Web pages designed to trick visitors into believing they were visiting a page with an embedded video and — oops! — you need to update your copy of Adobe Flash in order to view it.

Well, those days of hard work seem to have faded into memory. All we’re left now is this.

In a recent attack that came to my attention, the guys behind the attack didn’t bother to build a sophisticated Web page. Well, nothing along the lines of pages we’ve seen before, with cool graphics, slick design, or interesting programming. In fact, they hardly built a Web page at all.

In this case, the unknown person or people created an HTML file that loads someone else’s graphic, which happens to be a warning about an outdated version of Flash, that is located elsewhere. Specifically, they load a graphic that just happens to be hosted on the Coca-Cola company‘s Web server. This isn’t a site hack against the Coke people — the graphic is probably legitimate, considering how Flash-heavy the Website is — just an example of how pathologically lazy or incompetent some malware distributors can be.

read more…

While some members of our Threat Research group are attending talks at the Black Hat Briefings, the rest of the team is back at our offices, hard at work watching for novel threats.  That’s good news for gamers, and bad news for malware distributors who might try to take advantage of a confluence of events where many elite members of the security community are temporarily turned away from monitors while they attend the conference. I received a warning about one potential threat facing gamers who might turn to piracy to get a copy of Blizzard’s new real-time-strategy game, Starcraft II.

Apparently, there are a flood of torrents where gamers can download purportedly pirated versions of SC2. While your less ethical gamer might cheer this news, you might be less pleased to find out that some of the SC2 torrents appear to bring along a side order of malware. One of the torrents, for example, touted as a custom game launcher, drops the Zbot keylogger Trojan—albeit a variant we can easily detect and remove.

While this isn’t exactly new, we’re finding that the incredible demand for this game is driving malware distributors to supply something that looks like what the gamers want. We’ll keep an eye on this trend, and update the post if necessary with more details as they become available.

And if you want a copy of the game, just go out and buy it. It may not be the most thrifty use of your money, but it’s the ethical thing to do, and the safest way to get a copy of the game.

(Starcraft 2 logo courtesy of Blizzard Entertainment)

The Threat Research group sat in on a talk by HBGary CEO Greg Hoglund yesterday where the regular speaker discussed some research he’s been doing over the past year that he hopes will help connect malware samples to known groups of malware creators. While that sounds promising for law enforcement, it’s actually not as helpful for tracking down originators of malware for prosecution as it is for security researchers to preliminarily group and classify the masses of outwardly-dissimilar Trojans we see every day.

In most conventional methods of classification, researchers look for programmatic similarities or behavioral characteristics as a way to group similar pieces of malware into definitions, which then simplify the task of an antivirus tool to clean up an infection. In Hoglund’s talk, he proposed another set of criteria antimalware researchers can use to make these kinds of classifications: the “tool marks” left behind inside of malware samples as a result of compiling tools, languages, and even sloppy coding habits employed by malware creators.

On a technical level, Webroot’s Threat Research team has been using these “tool marks” as guides for some time when they perform manual analysis of malicious files. Hoglund’s talk introduced a tool he created, called Fingerprint, which can process a malware file and, in an automated fashion, provide malware researchers with simplified output they can then add to a database. With a sufficiently large sample set, surprisingly good clustering seems to appear, as shown in the photograph above, which is a snapshot of one of Hoglund’s slides.

While the characteristic “tool marks” alone are probably not sufficient to establish that an arbitrary, unknown file is malicious, it can be a good indicator that the unknown file is related — possibly in several significant ways — to files that have been established to be malicious. It is this predictive ability of the fingerprint that may be its greatest strengths…at least, until the malware authors catch on, and strip this identifiable information out of their files. For the meantime, however, laziness on the part of malware creators, and the difficulty of completely re-coding new malware, means identifiable tool marks should persist for a while, which means this fingerprinting method may remain effective for some time.

I’m at the Black Hat Briefings this week, the annual confab of the best and brightest in computer security, catching up on the trends and tricks malware authors and data thieves employ. I just saw an impressive demo by a pair of security researchers who took a deep dive into the behaviors of four pieces of highly targeted malware.

The researchers, Nicholas Percoco and Jibran Ilyas of Trustwave, ran a live demonstration of four Trojans designed to steal sensitive information and surreptitiously exfiltrate that data to the criminals. Three of the Trojans had been found installed on the servers of retail businesses, and capture credit card information — including the magnetic stripe data recorded by point-of-sale devices (ie., cash registers). The fourth Trojan, found on the computers of a large military contractor, was designed to steal any files in the My Documents folder, as well as any saved passwords on the system.

Of note was the highly targeted nature of the Trojans. In the case of the military contractor, for example, the criminals had obviously done their research, because the attack had targeted several high-level executives within the firm. According to the researchers, the attack started when a maliciously crafted Adobe PDF file was emailed only to the executives in a forged message that appeared to come from the CEO of the company. The forged message even included the CEO’s customized mail “signature” and the message text sounded convincingly similar to the language the CEO might have used.

Most importantly, all four Trojans did an outstanding job of remaining undetected for a significant period of time, which gave them more time to get the job done. Although one Trojan, which used a rootkit driver, had a tendency to “blue screen” their test machines, even a crash might not alert a victim that their computer hosts an infection. After all, Windows can crash for all kinds of reasons, and a crash isn’t necessarily an indication of a malware infection.

I’m looking forward to seeing more talks from other researchers over the course of the coming week. Of particular interest is a talk being given by Greg Hoglund about identifying the perpetrators of malware infections and even the creator(s) of widely distributed types of malware.

When it comes to spam messages, conventional wisdom dictates that you shouldn’t follow links or call phone numbers in the message, order products from the spammer, or open files attached to the email. We all should know by now that you should never open attached executable files, and spam filters now treat all .exe files as suspicious. When spammers began flooding inboxes with .zip files containing executables, we caught on pretty quickly as well.

Phishing has evolved. Learn all the ways hackers are angling for your data with our 11 Types of Phishing eBook.

But HTML isn’t executable — it’s just plain text — so does that mean it’s safe to open attachments when they’re just HTML files? Hell no! Case in point: this doozy that came through our spam bucket last week.

The message subject reads Your Funds Will Be Transfered and the body helpfully informs the recipient that I am able to complete the funds transfer late night — I hope that doesn’t mean someone sent Jimmy Fallon $28,126 from my bank account. It continues, Copies of the payment is being attached, and the message indeed has an attachment named Copies of the payment.htm which I can open and…

…uh oh. That’s where the trouble begins.

The end result: Three pieces of malware installed; Two password-stealing copies of the Zbot phishing trojan, and a remote-access backdoor to boot. Considering Zbot’s propensity for stealing bank account logins and other sensitive credentials, I suppose the subject line was correct after all. Your funds will be transferred. Just not where you thought.

A more careful look at the attachment reveals that it contains some lightly obfuscated Javascript code. When rendered in a browser, the code instructs the browser to load a page on a Web site called DreamLifeAsia, which is (suprisingly, considering the name) not a pornography site. That page contains instructions for the browser to load a one-pixel-square iframe from a page on the raceobject.ru Web site.

The raceobject.ru page contains scripting that instructs the browser to download and execute a payload from yet a third Web site, problemdollars.ru. It just so happens that these fine, upstanding Russian-registered Web sites are hosted on the same server, located at the same IP address.

Whomever scrabbled together this row of dominoes must have thought they were being really clever when they made their code hard to read. Take a look at the line below, for example. At first glance, you might not be able to determine what it’s supposed to do. It looks like gobbledygook, right?

Well, take out every other character in the line encircled in red, and the cloying concatenation code, and this is what you get:

Yeah, genius, we figured out your awesome trick. It’s the “insert a random letter between every character” encryption algorithm, first invented in the year one by a five year old girl who wanted to pass a clay tablet to her friend in class without the teacher being able to read her Aramaic. The only significant difference is that there are no little hearts drawn over the “i”s. RSA and Bruce Schneier, eat your hearts out. These guys are obviously l33t.

In the end, the page directs a vulnerable browser to download the backdoor to the victim’s computer. The backdoor pulls down Zbot. Zbot contacts its command and control server, pulls down a data file with instructions, and makes another copy of itself on the infected computer’s hard drive.

In addition, the backdoor changes all the Security Zone settings in IE, clears your cookies, changes the settings for the Windows firewall to permit Windows Explorer (not IE) to receive data from anyone on the Internet, and makes other modifications to security settings that make it far more likely a victim will further infect him- or herself.

Bottom line, it doesn’t matter what kind of attachment you get. Don’t open anything attached to a spam email in a browser. After all, it’s pretty obvious that this is just a ruse. Jimmy Fallon doesn’t need your money, just some decent Nielsen numbers in his time slot, and that’s one thing Zbot can’t steal.

I just want to take a moment to thank the malware author who posted a spam comment to the Webroot Threat Blog blog the other day. You guys make my job so easy.

The spam comment, which reads Hello. I the beginner. I wish to show to you,scandal story and links to a drive-by download site, is a tremendous help to our researchers, who are always on the lookout for new threats.

Of course, the malware distributor could have employed a more effective hook to convince someone to click a link than the one he used.

The link claims to point to a page hosted on the free Blogspot blog site to a nude video — not of Paris Hilton, Venus Williams, or Erin Andrews — but of…Diane Sawyer, the respected, award-winning anchor of ABC’s World News Tonight.

“Diane Sawyer Nude” — seriously? News anchor porn? Whatever happened to malware authors touting nude photos of starlets as an enticement?
read more…