Definitely Not Cerber

At the beginning of last week we noticed a spam campaign delivering a double zipped JScript file. The campaign started on September 8th. The email had the subject line of “RE: [name of recipient]” with an empty body, and an attached zip file named “[recipient name][a-z]{4}.zip”.

The characteristics of the mail, naming of the attached item, and obfuscation used in the sample were similar to what has been previously seen with the distribution of Cerber ransomware. Testing one of the samples lead to an unpleasant surprise looking nothing like Cerber.


Definitely not Cerber

The final payload of that particular sample was Locky ransomware. It was an odd discovery, especially as Locky is known to be distributed by the Necurs botnet in totally different campaigns with higher prevalence. This campaign spanned over a week, with no more than a few dozen samples per day. Further analysis of the campaign revealed minor tweaks and updates to the attached item during the week.


The first delivered attachment type on the evening of the 8th was an obfuscated JScript downloader. Distributing this type continued for few days. The next surge two days later delivered a similarly obfuscated JScript downloader in a JScript encoded script file (.jse). Later, the campaign continued by spamming encrypted JScript files, but changed the obfuscation to support custom XOR encryption on critical strings. In the last update the size of the downloader was doubled with comments, and the distribution spiked a little.

The contacted URLs were also following the format observed in previous Cerber campaigns. In total, the samples contacted 7 domains registered under the .top domain (TLD), resolving to two IP addresses, each with 7 different query parameters in format of ?f=[1-7]{1}.bin. The query was hard-coded on the distributed samples, and 25% of the samples were contacting the domains with query parameter 1. (By comparison, if the parameters were randomly generated the distribution share would be 14% instead of 25%.)

Further analysis on the URLs revealed that same sample of Locky was delivered on all domains with query parameters from 2 to 7. Query parameter 1 was allocated to serve Cerber ransomware.


Probably Cerber

This is not the first time Cerber has been distributed in the same campaigns with other nasty malware. Last May Cerber shared distribution framework with Dridex banking trojan. Though the campaign seems to be on a test phase based on the multiple minor updates on the dropper during the week, so far seeing two different ransomware on same campaign is unusual.


Seriously, Put Away The Foil

I was scanning the headlines this morning, as I do, and came across this article by YLE Uutiset (News). — “Finnish police: Keep your car keys in the fridge”

Finnish police: keep your car keys in the fridge

From YLE’s article:

“These so-called smart keys work by emitting a signal when the driver touches the door handle. The lock opens when it recognises the key’s signal. Criminals have technology that can strengthen that signal even from a hundred metres away—well inside the residential property where most owners keep their keys, according to Eero Heino of the If insurance company.”

So, should you keep your keys in a refrigerator?

Car key in a fridge

Don’t. (Cold can damage some batteries.)

Well, what about foil?

No. Put away the foil…

Look, if you have a car that’s actually valuable enough to be concerned about – get yourself a Faraday bag. Here’s one designed to fit a phone.

Car key in a Faraday bag

A very handy item to have when traveling abroad to “certain countries“.

Wickr branded Faraday bag

I got mine from the fine folks at Wickr. A quick search on Amazon yields results starting at about 10 bucks.

Or hey, here’s an idea, perhaps insurance companies could start giving customers Faraday bags when insuring an expensive car?

Just a thought.

0ld 5ch00l MBR Malware

I recently installed Audacity, an open source audio editor

Audacity UI

And while verifying the current version to download, I came across an interesting security notification. Before I read the details, I fully expected to discover yet another case of some crypto-ransomware group hijacking and trojanizing an application installer.

But not so!

Audacity’s download partner was infiltrated via compromised accounts and Audacity’s Windows installer was replaced by purely destructive malware, an MBR-overwriting trojan. That’s really something of a throwback in this age of malware-for-profit.

Those who installed the trojanized installers saw this message on reboot.

MBR message: It is a sad thing your adventures have ended here!

Classic Shell was also affected, here are file details from its forum.

And here’s a video by @danooct1 demoing the malware, and how to repair the overwritten MBR.

Infected Classic Shell/Audacity Trojan

Great stuff. And check out the view statistics… it seems there’s a decent audience for malware analysis video.

What’s The Deal With Machine Learning?

We’ve recently received quite a few questions regarding the use of machine learning techniques in cyber security. I figured it was time for a blog post. Interestingly, while I was writing this post, we got asked even more questions, so the timing couldn’t be better.

It seems that there are quite a few companies out there making noise about using machine learning techniques in their security products like it’s a new thing. It’s not. We’ve been using machine learning techniques since 2005, and nowadays you’ll find machine learning being used almost everywhere.

Machine learning techniques were first used by the security industry to train anti-spam engines. That fact prompted us to experiment with machine learning in an attempt to identify malicious files. In late 2005, we developed an engine designed to rate the suspiciousness of files based on both structural and behavioral characteristics. This engine was originally designed to suppress false positives generated by our new behavioral blocking technology, but since then has cemented itself as a solid piece of detection technology. Both of these components were introduced into our product line in 2006.


I couldn’t resist. (Source:

As I mentioned, we’re using machine learning all over the place. Here are a few examples of what we’re doing with it.

Sample analysis and categorization – We’re using expert systems and machine learning to automatically categorize the 500,000 new samples we receive each day. These systems generate a lot of high-quality metadata that is transformed into actionable threat intelligence.

URL reputation and categorization – We feed content from URLs into a machine learning system in order to categorize sites both for maliciousness and type of content (such as adult content, shopping, bank, et cetera).

Client-side detection logic – We use machine learning to train client-side components to identify suspicious files based on file structure and behavioral characteristics. We refer to these components as heuristic engines.  On August 25th, Sven Krasser at CrowdStrike published an informative and detailed blog post on how these techniques work that I recommend reading if you’d like to know more.

Breach detection – This is something I haven’t covered much yet, but plan to in the future. We use machine learning techniques to identify suspicious behavior on networks. These signals are sent to security experts working in our Rapid Detection Center, who investigate the incident and alert the customer if the information is valid. Naturally, the same techniques that uncover signs of breaches can also alert us to malicious insider activity.

Machine learning can be quite false-positive prone. This is why we prefer to use a hybrid approach that utilizes both human and machine. Combining machine learning with expert-developed rules and extensive automation allows us to reduce false positives and make much more accurate determinations of threats and suspicious behavior. For instance, in our sample categorization systems, machine learning techniques do a good job clustering incoming samples. However, for new samples it’s never seen before, we still use real humans to identify, label and categorize those clusters.

We’ve found machine learning to be extremely useful. However, it’s not a substitute for real human expertise just yet. As one colleague of mine put it, if you treat machine learning as a silver bullet, you’ll very quickly find that bullet in your foot. And that’s our advice to everyone out there – it’s critical that you don’t rely solely on machine-learning to protect your systems, and especially not solutions that can only identify file-based threats.

And there’s a couple of reasons why you shouldn’t do that. Firstly, you’ll not be protected against scams, phishing, and social engineering. For that, you need a URL blocking component. If you don’t have one, you can still easily end up on a site designed to steal your credentials, identity, or banking information. A solution designed to identify malicious files won’t be enough to keep you properly protected on the Internet.

Secondly, you definitely want protection against exploits. Exploits are the choke-point in the kill chain. There are hundreds of thousands of compromised or malicious sites out there, and hundreds of thousands of unique malicious files. However, there aren’t all that many unique exploits. Blocking all known exploits is much easier than ensuring every bad site out there and every single payload is handled. Here at F-Secure, we frequently gather the threat intelligence needed to find these exploits from in-house automation that relies on machine learning. However, the rules are still hand-written by our experts. This is one example of a client-side protection technology that simply doesn’t lend itself all that well to machine learning.

Finally, here are some questions @kevtownsend asked us, and my answers.

Will machine learning make jobs in the cyber security industry obsolete?

Absolutely not! Attackers, be they malware writers or actors looking to breach corporate networks, are humans. They think creatively and design attacks that can easily bypass purely automated solutions. Because of this, defenders need to be able to think creatively, too. Until artificial intelligence is capable of human-level creativity, humans will continue to be crucial in the field.

If machine-learning engines can be integrated into Virus Total, why can’t behavioral analysis engines be integrated?

Behavioral engines are difficult to integrate into Virus Total’s system. Every sample run through their system would need to be executed in an environment containing each vendor’s protection solution. Practically speaking, this means bringing up a virtual machine, installing or updating a vendor’s product, injecting the sample into the VM, executing it, extracting the product’s verdict, and then destroying the VM. This all has to happen under special network conditions to ensure malware is not spread further.

This whole process is not only super-resource intensive, it’s hell to maintain, especially when you consider that VT’s systems already contain over 50 products. Even if VT had the infrastructure available to do this for 500,000 samples times 50 vendors per day, they’d still need to hire a fleet of people to maintain the environment and keep the products up to date.

Is there an intrinsic difference between machine learning detection engines and behavioral detection engines?

This is an apples and oranges comparison. Machine learning techniques are used to “train” client-side detection logic. The actual machine learning process is run on heavy back end infrastructure, since it requires large volumes of samples and a significant amount of processing power. The logic bundle, once generated, is delivered to the client via product updates. Although some vendors don’t specifically talk about rules, signatures, or databases, you can be sure their products do contain them, one way or another. If a database is bundled into the binary itself, it’s still a database. Machine learning can be used to train logic designed to detect suspiciousness based on the structure of a file or its behavior, or both.

We strongly warn people against reading into the marketing hype out there. Most “AV” vendors have been using machine learning techniques to create rules and logic for years already.

Coming Soon: iOS 10

I’ve been testing iOS 10 Beta for several weeks (on a secondary iPad mini 2 of mine) and so far, so good. I’m enjoying Swift Playgrounds and looking forward to the final release.

Most of the changes I’ve noticed have been surface (i.e., UI) changes. But today I read an interesting blog post by @nabla_c0d3, regarding iOS 10 security and privacy. Under the hood stuff that sounds very promising.

Full post here: Security and Privacy Changes in iOS 10

If you don’t already use “Limit Ad Tracking”, you’ll find the option from: Settings > Privacy > Advertising > Limit Ad Tracking.

Enabling the option in iOS 10 will cause apps to see your Advertiser ID as all 0s, putting a limit on third-party tracking.

Apps on iOS have long been designed to ask for various permissions as needed, rather than all up front (à la Android), but with iOS 10, Apple will enforce the use of “purpose strings” which should be used to communicate a reason for why the permissions is needed.

Got Ransomware? Negotiate

ICYMI: we recently published a customer service study of various crypto-ransomware families. Communication being a crucial element of ransomware schemes, we decided to put it to a comparative test.

The biggest takeaway? If you find yourself compromised – negotiate.

Our Findings – In A Nutshell

You have little to lose, the majority of extortionists appear to be willing work with their “customers”.

Our report (download) also contains a fascinating email conversation as an appendix…

NanHaiShu: RATing the South China Sea

Since last year, we have been following a threat that we refer to as NanHaiShu, which is a Remote Access Trojan. The threat actors behind this malware target government and private-sector organizations that were directly or indirectly involved in the international territorial dispute centering on the South China Sea. Hence, the name nán hǎi shǔ (南海鼠) which means South China Sea rat.

Based on our observations, the timings of the attacks indicated political motivation, as they occurred either within a month following notable news reports related to the dispute, or within a month leading up to publicly-known political events featuring the said issue.

Timeline of events

Timeline of events

The white paper is a culmination of our research to understand the motivation behind NanHaiShu. To know more about our analysis and other interesting details, please read our white paper from here.

nanhaishu whitepaper cover

Bye Bye Flash! Part 2.5. Microsoft Edge Is Going “Click To Flash”

After last Thursday’s article on how Firefox will start reducing support for Flash, I received some comments pointing me to an announcement from Microsoft, back in April, where they stated that their Edge browser would also move towards a “Click to Flash” approach. The announcement notes that Flash plugins not central to the web page will be intelligently paused, and that content such as games and video will continue to run normally. This change to Edge will be delivered in the anniversary update of Windows 10.

I’d like to point out that we did notice this news back in April, and kudos to Microsoft, and the Edge team, for making this happen.

Microsoft Edge Logo

Microsoft Edge Logo (source:

Why didn’t we talk about this at the time? Well, Edge only works on newer Windows versions. It seems that Microsoft won’t make their 1 billion target for Windows 10 installs, and at current count, Windows 7 still has about 50% market share. So, we’re still waiting for that all-important announcement about Flash and Microsoft Internet Explorer.

Bye Bye Flash! Part 2 – Firefox Plans To “Reduce” Support For Flash

Earlier this year, in our 2015 Threat Report, our own Sean Sullivan predicted that Chrome, Firefox, and Microsoft would announce an iterative shift away from supporting Flash in the browser by 2017. Last month, we covered the announcement made by Google.

As predicted, just yesterday, the Firefox developers made a similar announcement on their blog.

Mozilla Firefox logo.

Mozilla Firefox logo. Source:

Firefox will begin dropping Flash support by blocking specific SWF files via a blocklist. The list will initially contain just plugins designed for “fingerprinting”. As stated by the Firefox developers, the criteria for adding content to the blocklist are:

  • Blocking the content will not be noticeable to the Firefox user.
  • It is possible to reimplement the basic functionality of the content in HTML without Flash.

The blocklist will be expanded to cover more types of content throughout this year, and by the beginning of next year, Firefox will require click-to-activate approval from users before a website activates the Flash plugin for any content. The next major Firefox ESR (Extended Support Release) release, scheduled for March 2017, will, unfortunately still continue to support plugins such as Silverlight and Java until early 2018.

The guys at Mozilla state that these changes will improve browsing stability, battery life, and performance. For us, the great news is that these changes will improve browsing safety, by greatly reducing the attack surface exploit kits have to work with.

And with that announcement, it’s two down, one to go.

Malware History: Code Red

Fifteen years (5479 days) ago… Code Red hit its peak. An infamous computer worm, Code Red exploited a vulnerability in Microsoft Internet Information Server (IIS) to propagate.

Infected servers displayed the following message.

Welcome to !

Description: Worm:W32/CodeRed

See @mikko‘s Tweet below for a visualization.

A New High For Locky

After seeing a drop during first weeks of June, the spam campaigns distributing Locky crypto-ransomware has returned as aggressive as ever. Normally we have seen around 4000-10,000 spam hits a day during spam campaigns. Last week from Wednesday to Friday we observed a notable increase in amount of spam distributing Locky. At most we saw […]


Black Hat USA 2016 Briefings

We get a fair amount of requests from journalists and media organizations asking our opinion on a whole range of tech topics. And when Black Hat rolls around, the pace of those requests often picks up considerably. So, I spent some time last week reading through the Black Hat USA 2016 briefings. That was a […]


What’s The Deal With Detection Logic?

Detection logic is used by a variety of different mechanisms in modern endpoint protection software. It is also known by many different names in the cyber security industry. Similar to how the term “virus” is used by laypeople to describe what security people call “malware” (technically, “virus” is the term used to describe a program […]


What’s The Deal With Network Reputation?

Drive-by downloads or, more accurately, drive-by installations are some of the scariest threats on the Internet. Exploit kits provide the underlying mechanisms for this behavior. They work by examining your browser’s environment – browser type, browser version, installed plugins, and plugin versions, looking for a vulnerable piece of software. If the exploit kit finds any […]


Out of Office OPSEC

A “found object” from my Inbox (with sundry modifications). A vacation greeting from our CSS OPSEC experts! It’s absolutely fantastic that you’re soon going on holiday and are not at the office. And we’re sure it’s very well deserved! But before you go, consider this – you don’t have to tell the world where you […]


What’s The Deal With Threat Intelligence

The term “threat intelligence” is quite trendy right now. For many, threat intelligence is a term used to describe IOC feeds that are plugged into security infrastructure to identify suspicious or malicious activity. For us, it describes a whole lot more. As a company, we’ve been actively gathering and assimilating threat intelligence for over 25 […]


What’s The Deal With Prevalence

We use the word “prevalence” a lot at F-Secure Labs. And what’s prevalence? The prevalence of an executable file is defined as the number of times it’s been seen across our entire customer base. Malicious executables tend to be rare over time, most live and die quickly, and thus the number of times we’ve seen […]


Qarallax RAT: Spying On US Visa Applicants

Travelers applying for a US Visa in Switzerland were recently targeted by cyber-criminals linked to a malware called QRAT. Twitter user @hkashfi posted a Tweet saying that one of his friends received a file (US Travel Docs Information.jar) from someone posing as USTRAVELDOCS.COM support personnel using the Skype account ustravelidocs-switzerland (notice the “i” between “travel” […]


“UltraDeCrypter” Wants To Speak Your Language

There’s a new crypto-ransomware brand in-the-wild called “UltraDeCrypter”. It’s an evolution of CryptXXX that is being dropped by the Angler exploit kit. In our tests, using an older CryptXXX “identification code” with UltraDeCrypter’s decryption service portal redirected to an older CryptXXX portal. So there’s evidence the back ends are interlinked. Regarding the payment support pages… […]


IC3’s Internet Crime Report

I’ve spent part of my day reading through the Internet Crime Complaint Center’s 2015 Internet Crime Report, and the numbers… are impressive. There were 288,012 complaints received by IC3 in 2015 and more than one billion dollars in losses reported. Hot topics? Business Email Compromise (BEC), Email Account Compromise (EAC), and ransomware. On a positive […]