Industry Intel

You’re probably familiar with some of the most common requirements for creating passwords. A mix of upper and lowercase letters is a simple example. These are known as password constraints. They’re rules for how you must construct a password. If your password must be at least eight characters long, contain lower case, uppercase, numbers and symbol characters, then you have one length, and four character set constraints.

Password constraints eliminate a number of both good and bad passwords. I had never heard anyone ask “how many potential passwords, good and bad, are eliminated?” And so I began searching for the answer. The results were surprising. If you want to know the precise number of possible 8-character passwords there are if all of the character sets must be used, then the equation looks something like this.

A serious limitation of this approach is that it tells you nothing about the effects of each constraint alone or relative to other constraints. (I’m also not sure if there were supposed to be four consecutive ∑s or if the mathematician was stuttering.)

We choose to use a Monte Carlo simulation to analyze the mathematical impact of the various combinations of constraints. A Monte Carlo simulation uses a statistical analysis approach that provides a close approximation of the answer, while also providing the flexibility to quickly and easily measure the impact of each constraint and combination of constraints.

A look at minimum length limits

To start, let’s look at the impact of an eight-character length constraint alone. There are 95^8 possible combinations of 8 characters. 26 uppercase letters + 26 lowercase letters + 10 numerals + 33 symbols = 95 characters. For a length of 8 characters, we have 95˄8 possible passwords.

If a password must be at least 8 characters long, then there are also about 70.6 trillion otherwise viable passwords you are not allowed to use (95+(95^2 ) +(95^3 ) +(95^4 ) +(95^5)+(95^6 )+(95^7)). That’s a good thing. It means you can’t use 95 one character passwords, 9,025 two character passwords, and so on. Almost 70 trillion of those passwords you cannot use are seven characters long. This is a great and wholly intended effect of a password length constraint.

The problem with a lack of constraints is that people will use a very small set of all possible passwords, which invariably includes passwords that are incredibly easy to guess. In the analysis of over one million leaked passwords, it was found that 30.8 percent passwords eight to 11 characters long contained only lowercase letters, and 43.9 percent contained only lowercase letters and numbers.  In fact, to perform a primitive brute force attack against an eight-character password containing only lower case letters, it’s only necessary to try about 209 billion character combinations. That does not take a computer very long to crack. And, as we know from analyzing large numbers of passwords, it’s likely to contain one of the most popular ten thousand passwords.

To beef up security, we begin to add character constraints. But, in doing so, we decrease the number of possible passwords; both good and bad.

Just by requiring both uppercase and lowercase letters, more than 15 percent of all possible 8-character combinations have been eliminated as possible passwords. This means that 1QV5#T&|cannot be a password because there are no lowercase letters. Compared to Darnrats,which meets the constraint requirements, 1QV5#T&|is a fantastic password. But you cannot use it. Superior passwords that cannot be used are acceptable collateral damage in the battle for better security. “Corndogs” is acceptable, but “fruit&veggies” is not. This clearly is not a battle for lower cholesterol.

As constraints pile up, possibilities shrink

If a password must be exactly eight characters long and contain at least one lower case letter, at least one uppercase letter and at least one symbol, we are getting close to one-in-five combinations of 8 characters that are not allowable as passwords. Still, the effect of constraints on 12 and 16 character passwords is negligible. But that is all about to change… you can count on it.

Are you required to use a password that is at least eight characters long, has lower and uppercase letters, number and symbols? Just requiring a number to be part of a password removes over 40 percent of 8-character combinations from the pool of possible passwords. Even though you can use lowercase and uppercase letters, and you can use symbols, if one of the characters in your password must be a number then there are far fewer great passwords that you can use. If a 16 character long password must have a number, then 13 times more potential passwords have become illegal as a result of that one constraint than the combined constraints of lower and uppercase letters and symbols caused. More than one-in-four combinations of 12 characters can no longer become a passwords either.

You might have noticed that there is little effect on the longer passwords. Frequently there is also very little value in imposing constraints on long passwords. This is because each additional character in a password grows the pool of passwords exponentially. There are 6.5 million times as many combinations of 16 character pass words using only lowercase letters than there are of eight character passwords using all four character sets. That means that “toodlesmypoodles” is going to be a whole lot harder to crack than “I81B@gle”

Long and simple is better than short and hard

People tend to be very predictable. There are more symbols (than there are in any other characters set. Theoretically that means that symbols are going to do the most to make a password strong, but 80 percent of the time it is going to be one of the top five most frequently used symbols, and 95 percent of the time is will be one of the top 10 most frequently used symbols.

Analysis of two million compromised passwords showed that about one in 14 passwords start with the number one, however for those that started with the number one, 75 percent of them ended with a number as well.

The use of birthdays and names, for example, make it much easier to quickly crack many passwords.

Password strength: It’s length, not complexity that matters

As covered above, all four character sets (95 characters) in an eight character password allow for about 6.634 quadrillion different password possibilities. But a 16 character password with only lowercase letters has about 43.8 sextillion possible passwords. That means that there are well over 6.5 million times more possible passwords for 16 consecutive lowercase letters than for any combination of eight characters regardless of how complex the password is.

My great password is “cats and hippos are friends!”, but I can’t use it because of constraints – and because I just told you what it is.

For years password experts have been advocating for the use of simple passphrases over complex passwords because they are stronger and simpler to remember. I’d like to throw a bit of gasoline on to the fire and tell you, those 95^8 combinations of characters are only  half that many when you tell me I have to use uppercase, lowercase, numbers, and symbols.

Latest Windows 10 Update Removes User Files

Microsoft recently pulled its latest update, version 1809, after several users complained about personal files being deleted. While some users were able to use third-party software to retrieve deleted files, users whose files wnet missing from the Documents folder are having a much trickier time without restoring from backups. Since hearing of the issue, Microsoft has paused the automatic update until they can find a resolution.

Magecart Campaign Continues Its Spread

Following the attacks on British Airways and Ticketmaster, Magecart skimmer techniques have been discovered on Shopper Approved, a collective of several online stores. Fortunately, the company was able to identify the altered JavaScript code and contact affected vendors. The malicious code itself was targeted at the checkout pages for the affected stores with specific URL keywords, leaving the remainder of the thousands of online retailers unaware anything had occurred.

Vulnerabilities Found in Millions of Chinese Electronics

A new wave of vulnerabilities has been spotted in nearly 9 million devices made by Chinese-based Xiongmai, leaving them susceptible to attack. Serious issues include default admin passwords without a prompt to immediately change it, no encryption when connecting to their cloud servers, and a lack of authorization checks when searching for updates. Many of these devices were known to be compromised during the Mirai botnet attacks, though the access points used for that have since been patched.

FCC To Block Illegal Spam Calls

Most people have received at least one unwelcome call on their mobile phone from a robotic auto-dialer. Now the attorneys general from 35 states are coming together in hopes the FCC can do something about those annoying calls. These types of spam calls seem to have increased in volume in recent years, even after the 2017 Call Blocking Order aimed at stopping them,  forcing customers to block calls themselves. With an estimated 40 billion robocalls this year alone, it’s no surprise so many states are interested in putting a stop to this nuisance.

Google+ Goes Out on Low Note

After constantly struggling with low adoption, Google’s response to more popular social media platforms like Facebook has officially reached its end of its life. Several months ago an API bug was spotted that allowed unauthorized access to thousands of Google+ user accounts. The bug was patched but remained undisclosed until recently. With new GDPR regulations on breach disclosure, even the possibility of low volumes of affected clients could still be trouble for Google.

Brazilian Bank Traffic Rerouted by Massive Botnet

A botnet containing more than 100,000 routers and other devices was recently spotted hijacking traffic destined for several Brazilian banks. The hijacking victims are then sent to one of at least 50 confirmed phishing sites that will attempt to steal any information the user will provide. Backing this ever-growing botnet are a small collection of tools used to brute-force weak passwords and continue to search for other devices with poor security.

Cyber Attack Shuts Down Canadian Restaurants

A major Canadian restaurant chain announced several of their restaurant brands had suffered a ransomware attack that affected nearly 1,400 stores in recent days. While many of the IT systems were quickly taken offline to prevent further spread of the infection, customers were met with non-functioning payment systems or just closed doors. Fortunately, the company keeps regular backups and was able to restore their systems without paying a ransom.

High-Profile Instagram Accounts Being Hacked

Several high-profile Instagram accounts were hacked and held hostage recently, with some accounts being deleted even after a payment was sent. Though many victims have contacted Instagram multiple times regarding access to their accounts, some were sent automated responses while others regained control of their accounts without hearing from the company.

Google Chrome Cracks Down on Extensions

With dozens of new extensions being added to Google’s Chrome Web Store every day, it has become increasingly difficult for Google to police for malicious apps. That’s why, accompanying the release of Chrome 70, will be the ability for users to restrict browser extensions to a single site and limit the amount of permissions the extension has over the pages viewed. Additionally, Chrome has implemented 2-step verification for all developer accounts to curb the volume of hacked apps made available.

Port of San Diego Hit by Ransomware

It was revealed last week that the Port of San Diego, which controls over 34 miles of coastline, suffered a ransomware attack that temporarily knocked out their computer systems. Fortunately, most routine port operations remained able to function normally while systems were offline. There is still no information on whether the ransom has been paid or how the infection occurred.

Firefox Vulnerability Leads to Crash

A new denial-of-service (DoS) attack has been created with the ability to cause desktop versions of the browser Firefox to freeze or crash. Upon visiting sites where the malicious script is present, the user’s browser forces download requests for a massive junk file that can cause the IPC channel for the browser to crash. Luckily, the researcher who created the attack method has contacted Mozilla about the issue, and there’s hope for a swift resolution.

Kodi Media Player Used to Spread Malware

Nearly 5,000 computers were recently compromised with cryptomining malware that was silently distributed either through malicious builds of the Kodi media player or from third-party add-ons used to enhance the player. Most of the infected computers were found to be mining for Monero and have already mined around $6,700 since the beginning of the campaign. When obtaining these types of add-ons, its best to visit official repositories rather than third-parties, as they tend to be more discerning of content they are hosting.

Online Fashion Retailer Breached

SHEIN has revealed a data breach from June that they themselves only discovered within the last month. Nearly 6.5 million customers could be affected, as the systems storing login credentials were compromised in the attack, the company stated in a recent press release. Fortunately for those customers, the company says they do not store payment data so a simple password change should be sufficient to protect their clients.

Scottish Brewery Hit by Ransomware

After publishing a job opening to their own site, Arran Brewery was able to successfully fill the needed position. Unfortunately for the Scottish brewery, attackers posted that listing on several international recruiting sites and received dozens of applications including documents embedded with ransomware, resulting in the company being locked out of crucial systems and a ransom demand of two Bitcoins. Arran Brewery opted to restore their systems from offsite backups rather than pay the ransom, but lost up to three months of data due to outdated backups.

DoorDash Customers Complain About Hacked Accounts

Several dozen people have contacted DoorDash regarding fraudulent orders placed on their accounts. DoorDash’s was confident they were not to blame for the breach, instead blaming “credential stuffing,” a tactic where attackers try using previous breach data from other sites hoping the same password was used multiple times. The company says it has no plans to implement further security measures such as two-factor authentication.

Massive Customer Database Left Exposed by Data Management Firm

A security researcher recently found a database containing customer information for nearly half a billion users of Veeam software on an unsecured AWS server. Most of the data was contact information spanning from 2013 to 2017 and was likely used by the Veeam marketing team’s automated customer contact functions. Fortunately, the database was taken offline within a week of the researcher contacting Veeam about the server.

Hacker Group Breaches British Airways

After last week’s reveal of the data breach affecting nearly 380,000 of the airline’s customers, it was discovered that the injection methods used were the work of known hacker group MageCart. By compromising third-party actors, the group can access hundreds of sites and begin passing any customer payment information back to their own systems. Even more toublesome, this particular attack appeared to be tailored for the British Airways systems specifically, but could very likely be readjusted for other applications.

Chinese Hackers Using Digitally Signed Drivers for Attacks

A long-active hacker group likely based in China has expanded their tactics to include a seemingly innocent network filtering driver (NDISProxy) to start their latest malware campaign. The driver itself has a signed digital certificate from a Chinese-based security software company, which was likely unaware their certificate was being misused. By injecting itself silently across the infected network, the fully functioning remote access Trojan can be used to execute malicious tasks with ease.

Scam Calls Causing Mobile Traffic Jam

The number of scam calls recoreded by the call management firm First Orion rose nearly 1000% over the past year, from 3.7% of total calls last year to 29% so far in 2018. The projections for the coming year project that number to rise to half of all mobile calls received in the U.S. Unfortunately, service providers have few options for slowing down the bombardment of phony calls facing their customers.

Latest MongoDB Attacks are Ransoming Empty Databases

While MongoDB attacks are nothing new, Mongo Lock has stepped up the game by identifying unprotected databases, exporting the data to their servers, wiping them clean, and leaving behind a ransom note instructing the victim to reach out via email rather than sending a Bitcoin payment directly to a crypto-wallet. Mongo Lock appears to operate via an automation script, though it has been known to fail, leaving the victim with both the ransom note and their original data.

Banking Trojans Still Appearing in Google Play Store

Multiple security researchers recently discovered a handful of banking trojans that have still managed to make their way into the Google Play app store, despite Google having increased its security to detect such apps. Many of the apps are disguised as astrology/horoscope software, but instead of reading the future, they steal SMS and call logs from the device, install unauthorized apps, and even seek out banking credentials based on other installed applications. Some of these apps had been installed by up to 1,000 individuals, many of whom are likely under the assumption that the app removed itself, after showing a fake error message claiming incompatibility with the device.

Obama-themed Ransomware Forges Dangerous Path

A new ransomware variant bearing the face of the former US president, Barack Obama, has been spotted in the wild performing some unusual encryption tactics. Rather than encrypting personal word documents and pictures, this variant focuses on encrypting executable files across the system, which could lead to the system crashing and other devastating results. It is still unclear if this methodology is the intent, or just an oversight by the ransomware’s authors, but this type of damage is unlikely to pay off if it renders the system nonfunctional.

Thousands of Online Stores Compromised

Due to security loopholes in eCommerce sites that use Magento as a host, nearly 8,000 sites have been confirmed to be hosting card-skimming malware, with up to 60 more being compromised every day. The breaches led to malicious scripts being added to the pages to record and upload any customer inputs in real time, rather than following a more complicated path to obtain the same data after the transaction is complete. Unfortunately, it is difficult to determine whether a site is safe without checking the entire codebase for any unauthorized entries.

Fake Tech Support Ads Now Indistinguishable from Real Counterparts

In the run-up to Google’s release of a verification program for third-party vendors to display ads, the company has been inundated with countless fake tech support advertisements that are nearly impossible to identify over a real vendor’s ads. The creators of these fake ads will go to almost any lengths to avoid detection, including creating entire companies to continue their illicit activities.

Unsecured Sites Leaving .git Repositories Easily Accessible

Nearly 400,000 websites have been found with exposed .git directories that could lead to major information exposure, if improperly accessed. These repositories contain everything from passwords and API keys for the site, to forgotten data stored on the sites. Fortunately for the website owners, the researcher who discovered the breach was not acting maliciously, and quickly began contacting them with information on how he found the leak and what they could do to resolve it.

If you saw a file called eicar.com on your computer, you might think it was malware. But, you would be wrong. Readers, if you haven’t yet met the EICAR test file, allow me to introduce you to it. If you have used the EICAR test file, let’s get a bit cozier with it.

If you ran this file through VirusTotal, 61 out of 62 antimalware scanners currently would detect the EICAR test file as if it were malicious. That’s because the EICAR file is actually a tool that was designed to help users verify their antimalware scanner is functioning properly. The EICAR test file is a harmless piece of code that most vendors have agreed to flag as if it was malicious. Essentially, it’s a false positive—by design—for your benefit. Some scanners detect it, some do not; neither outcome indicates that any scanner is better or worse than another.

If you have heard of EICAR, you may have seen it referred to as a “test virus,” but that’s inaccurate. Think of it more like the test button on a smoke detector in your home. The test button doesn’t simulate fire or smoke; it simply lets you know that the smoke detector is functional. The test button certainly doesn’t tell you anything about the quality of the smoke detector. Similarly, the EICAR test file does not simulate malware, it just causes a scanner to demonstrate how it would handle a threat it detected (assuming the vendor has chosen to recognize the file as malicious, that is.)

Using the EICAR Test File

Now that you know more about EICAR, let’s talk about why, how, and when you might want to use it.

1. Curiosity. The first time I used the test file, it was purely out of curiosity. What if I zipped the file up or changed its extension from .com to .xyz, and so on. Because the file itself is harmless, I could simulate any number of scenarios without risk to my computer or my data.
2. Smoke test. The intended purpose of the test file was always to verify that your scanner was properly installed and that the scan engine was functional. Any time you install a new antimalware product, you can give it a quick test with the EICAR file to make sure it is functioning as designed (if the vendor support the file, that is.)
3. Forensics. Malware writers often try to disable a scanner as soon as their malicious code gains a foothold on a given computer. If you periodically test your scanner and, one day, it fails to detect the test file, that could indicate of an infection. Keep in mind, it could also indicate that another layer of security blocked the file before it got to your scanner. The test itself is not conclusive and should only be considered as part of a bigger picture.
4. Behavioral information. Between 1997 and 2004, I worked at Microsoft, ensuring none of their software releases were infected. I used 11 different virus scanners on each of my test machines (don’t try this at home). The testing was not about the quality of the scanners, but rather how they’d react in different situations to help me make decisions and gain greater knowledge. For example, antivirus scanners have default configurations that I needed to test and potentially modify. Back then, not all scanners scanned all extension types by default. A directory with EICAR test files that each had different extensions would allow me to determine if my scanner’s default configuration for file types needed to be adjusted. Once I made modifications, I had to test those as well. There were a variety of tests I could run involving filenames with punctuation or foreign language characters, too. Basically, I could test virus handling without needing am actual virus.

Note: At the Virus Bulletin conference in 1999 I presented the paper, “Giving the EICAR Test File Some Teeth.” If you’re interested in the breadth of test scenarios I explored, you can read the paper on the Virus Bulletin website.

Where to Find EICAR

You’d think the easiest way to get your hands on this file would be to download it straight from www.eicar.org, except that your antimalware scanner might block the download. To get around that, you’d likely have to temporarily disable your web protection—WHICH I DO NOT RECOMMEND. Instead, I’ll show you how to create the file yourself.

Here are the step by step instructions.

1. Open Notepad.
2. Copy the following string and paste it into Notepad:
   X5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*
3. Save the file and cross your fingers that your scanner doesn’t detect it on close.

Note: You could create the file in Microsoft® Word, but you’d have to save it as plain text. The test file must begin with the test string, and Word includes additional information in .doc and .docx files.

The file eicar.com, will run on older operating systems, but not on a 64-bit OS. When you run it on a compatible OS, the file will display this text.

You can change the display message to anything you like. In the following example, I’ve replaced the word EICAR with my name.

However, if you change it as I did above, it will no longer be a valid test file and should not be detected by your antimalware program.

At the 1999 Virus Bulletin conference, I asked researchers for EICAR-like test files to test script and macro detection. Although we still don’t have that, the Anti-Malware Testing Standards Organization (AMTSO) provides a set of security feature checks at www.amtso.org/security-features-check. Just be sure to remember that the security feature checks, like the EICAR test file, don’t indicate the quality of the product, but they can be used to ensure that certain features are functioning.

Questions? Comments? Let’s talk on the Webroot community forum.