Online Services

Category: Security

  • 9 WordPress Plugins Targeted in Coordinated 4.5-Year Spam Campaign

    On Tuesday last week we published a post that described how someone had released an update to the Display Widgets plugin which contained a backdoor that allowed them to publish content to any site using the plugin. We also described how they exploited that backdoor to publish spam.

    We then wrote a follow-up piece which we published on Wednesday identifying the person who modified Display Widgets as Mason Soiza.

    In today’s post, we are publishing research showing a coordinated effort by the same spammer that targeted WordPress plugins over a 4.5-year period. In some cases, site owners opted in to a vague agreement that didn’t make it clear that their sites would be serving spam; in other cases, plugins were simply “backdoored” to allow posting without a site owner’s permission.

    The content served from the ad network included ads for payday loans and escort services in the UK. The activity started in 2013 and ended this month. All nine plugins discussed today are linked to the same malicious actor, either through a financial trail or through the ad-serving domains used which share IP addresses.

    Which Plugins Were Targeted And Their Current Status

    We now have evidence that, in addition to the Display Widgets plugin, eight other plugins have been affected by spam in the past and all are linked in some way. These incidents range from May 2013 to September 2017.

    In four cases we have also been able to uncover financial transactions linked to Mason Soiza, either directly or indirectly. In other cases, we have no transaction, but we share the data we have which helps establish attribution.

    It is important to note that all plugins listed here no longer have the ability to inject spam into websites. They are either safe, or in two cases have been removed from the WordPress plugin repository.

    Here is the list of plugins we discuss today and their current status. When we list a plugin as “safe,” it means the plugin does not currently include any spam code that we are aware of.

    • 404 to 301: Safe
    • Display Widgets Plugin: Safe, but no longer maintained. Use Jetpack’s Widget Visibility Module instead.
    • WP Slimstat: Safe
    • WP Maintenance Mode: Safe
    • Menu Image: Safe
    • NewStatPress: Safe
    • Financial Calculator Plugin: Safe. Never included malicious code, but Soiza did have access for some time this year.
    • Weptile Image: Removed from repository
    • No Comment: Removed from repository

    Some plugin authors spoke with me anonymously for this post, but most went on the record. My thanks to all authors who helped us compile this data, because they have done the WordPress community a service in helping us understand the impact of these spam networks and who they belong to.

    Special thanks to the following plugin authors:

    • Display Widgets original author Steph Wells. We spoke via Skype voice and IM. She opened the floodgates.
    • 404 to 301 plugin author Joel James. We spoke via Skype IM.
    • WP Slimstat author who goes by the alias Jason Crouse. We spoke on the phone.
    • WP Maintenance Mode developer George and current owner Andrian. We chatted via email.
    • Menu Image plugin developer Alex. We chatted via email.
    • Financial Calculator plugin author Ciprian Popescu. We chatted via LinkedIn.

    Supply Chain Attacks and Avoiding Bad Partners

    What we are describing here today and in our previous posts is known as a supply chain attack. A few examples of supply chain attacks are:

    • A hacker breaks into a software company and infects its code with malware, which then gets distributed via the “supply chain” to that company’s customers.
    • A developer’s account is compromised and the hacker uses the account to release malicious code to customers.

    I have managed to verify that in most of these cases a malicious actor convinced a developer or company to add their code, usually through a financial arrangement, to a WordPress plugin. The malicious actor then uses that gateway to inject spam into customer websites.

    As developers, we have a giant target painted on our backs. We are constantly releasing code to customers, and if someone can inject their malicious code into our software, they can infect thousands or even millions of customers at once, rather than just a few at a time. It is very important for us to guard our login credentials and code and important for us to choose our partners wisely.

    One could argue that the plugin authors should have been more careful about who they partnered with. I think that is 20/20 hindsight because most partnerships require trust at some point in the relationship. If you put Google ads on your website, you trust Google won’t display offensive content. If another ad network pays you to include their code that displays their ads, you trust them not to include spam or malicious content. In this case, the trust was broken, and in some cases customers paid the price.

    My hope in publishing this story, and through the collaboration of these plugin authors who have kindly come forward, spoken up and shared their story, is that we as a community will learn how to avoid bad partnerships and how to form partnerships that do work.

    The rest of this post goes into detail about each plugin that we analyzed, how it is linked to the spammer we have been tracking and which versions were affected over what time frame.

    Display Widgets Plugin: An Update

    We wrote about Display Widgets in detail last week so I am going to summarize the status here and provide an update from the WP plugin team.

    I chatted with original plugin author Stephanie Wells via Skype voice and IM.  She sold the plugin to Soiza on May 19, 2017, and we’ve confirmed the transaction which came from pp@linkrocket.net.

    On June 30, 2017, a backdoor that allowed remote unauthenticated content updates was added to the plugin and remained there until the plugin was removed from the repository on September 8, 2017. The affected versions are version 2.6.1 to version 2.6.3.

    The domains used to serve spam were geoip2.io and stopspam.io. The WordPress.org plugin team have removed the historical versions that contained malicious code from the WP repository by the plugin team.

    The WordPress.org plugin team have removed the offending code from the latest version of the Display Widgets plugin and incremented the version number so that sites can update to the clean code. I chatted with Otto, one of the plugin repository maintainers, who had this to say on WordPress Slack:

    Version 2.7 is identical to the 2.05 version, codewise. This was done to remove the offending code from the plugin and to help eliminate it from sites that had updated it to the 2.6 series. The plugin is now closed and will receive no more updates. We recommend all users find an alternative plugin for their needs. Given what the intended functionality of the plugin is, I recommend using the Widget Visibility option in Jetpack instead.

    You can find out more about Jetpack’s Widget Visibility module on this page.

    404 to 301 Plugin

    We received data that Soiza was also involved in 404 to 301 and so we reached out to the author.

    We previously wrote about 404 to 301 distributing spam in August of last year. We also did a follow-up post with more technical detail and to counter criticism we received.

    I chatted with Joel James, the author of 404 to 301, for this post. He was kind enough to share transaction details with me. Based on a PayPal transaction dated May 9, 2016 from pp@linkrocket.net using Soiza’s name, Soiza also appears to be the perpetrator in this case.

    404 to 301 included Soiza’s code to distribute spam to websites from June 1, 2016 until July 12, 2016. The affected versions are 2.2.0 to 2.2.8.

    The domain used to fetch spam and inject it into the plugin was wpcdn.io. You can find the code here if you search for the domain.

    The plugin injected spam content onto sites, including a domain for a UK-based escort service which we verified belongs to Soiza in the Display Widgets follow-up post.

    The plugin asked users to agree to terms before injecting their content with spam, but the terms were lengthy and the ‘spam’ clause appeared right at the end.

    Joel James now controls the code to this plugin and has confirmed that the plugin is now clean. As a reminder, this happened over a year ago. At this point I have no hesitation in recommending that you use the 404 to 301 plugin.

    WP Slimstat Plugin

    We also received data that Soiza was involved with WP Slimstat. I chatted with the author of WP Slimstat on the phone. He goes by the alias Jason Crouse, and would prefer to remain anonymous. He was kind enough to provide all of the data we requested.

    He received a PayPal transaction dated November 22, 2015 from email jj@linkrocket.net and name “Garri Kiekbusch.” The email uses the same domain as two other emails that we have established belong to Soiza, but the transaction used a different name. This email was also displayed on an archived linkrocket.net home page along with the other two. The name may be a business partner or an alias.

    From October 22, 2013 until August 10, 2016 (2 years 10 months), WP Slimstat had code in it that is very similar to Soiza’s spam code. That period includes versions 3.4 to version 4.3.7.

    The domains used for Slimstat were wordpress.cloudapp.net, then wpcdn.io and later api.wp-stats.io. You can find a code sample on this page for 4.3.7.

    It’s important to note that wordpress.cloudapp.net and wpcdn.io share the same IP address at the time of writing. api.wp-stats.io historically also shared that IP address. This links the three domains and we will refer to this later in the post.

    Early versions of Slimstat that included Soiza’s code included a checkbox that you had to opt into, but the checkbox did not mention spam or ads. Later versions did mention that ads would be displayed.

    I’d like to repeat that I have chatted with the author of this plugin on a fairly long phone call and he shared his real identity with me. He prefers to remain anonymous, but is in control of the plugin once again and has been for a year at this point. WP Slimstat no longer contains any spam code, and has not for some time.

    WP Maintenance Mode Plugin

    Once we identified several plugins linked with Soiza, we started looking through the repository to find plugins using the same domains. We were surprised to learn that WP Maintenance Mode was involved.  It is a very popular plugin with around 500,000 active installs. From June 20, 2013 to September 1, 2014, WP Maintenance Mode had code that allowed content to be remotely injected into a website without permission.

    The versions that were affected are 1.8.9 up until the code was removed in version 2.0.0. At that point, ownership of the plugin changed hands to the new owners who appeared to remove the code.

    I chatted to the new owners via email and they don’t have any information that can help us link this to other affected plugins. I also reached out to the previous owner and managed to connect with him via email but was not able to get any helpful data.

    The one link we do have is the domain that WP Maintenance Mode was loading ads from. It was wordpress.cloudapp.net, one of the domains used by WP Slimstat which we confirmed was paid by jj@linkrocket.net. So this appears to, at the very least, be linked to the same advertiser that used Slimstat.

    The difference with this plugin is that there was no opt-in required by the user. The advertiser could simply inject content into the website at will. You can find an example of the code in version 1.8.11 on this page.

    This page on a hosting company website shows a user in 2013 asking why WP Maintenance Mode is slow in loading. It turns out it was contacting wordpress.cloudapp.net to load content in the background.

    NewStatPress Plugin

    We found this plugin was involved when searching for the domain associated with the plugins above. We found this post from 2013 where a user was reporting the plugin slowing down when loading something from wordpress.cloudapp.net.

    NewStatPress would randomly select a WordPress hook like get_header, get_sidebar, wp_head or wp_footer and print out a decoded JSON blob which it fetched from the wordpress.cloudapp.net domain.

    This plugin, WP Maintenance Mode and WP Slimstat all use the same domain, which is linked to jj@linkrocket.net.

    This plugin ran this code from May 5, 2013 until June 29, 2013. The versions affected were 0.6.2 to version 0.6.7. You can see a sample of what the version 0.6.2 code looked like on this page.

    I posted a contact request on this plugin’s support forum but did not get a response in time for publication.

    Menu Image Plugin

    In our research, we found that Menu Image was also loading ads and injecting them into site content. This post, where users are discussing spam being injected into the plugin, is what initially caught our attention.

    The Menu Image plugin had opt-in ads that displayed if the user agreed to a long license that was vague in describing what would happen to your website. I chatted to the author via email and he was very helpful in helping us identify whether these ads are linked to the other plugins.

    It appears that the author was paid by someone with a different name, tied to a different company. The domain used to load the content was apistats.net.

    There is one tentative link: apistats.net at one point shared an IP address with the domain api.wp-stats.io which was a domain used by WP Slimstat which we linked to jj@linkrocket.net (that Soiza domain again). The api.wp-stats.io domain also shared an IP at one point with wordpress.cloudapp.net and wpcdn.io, which as we know are other domains used to serve content to these affected plugins.

    Menu Image contained this code from July 7, 2016 until November 20, 2016 – slightly less than a year ago. The versions affected were version 2.6.4 to versions 2.7.0.

    You can find a diff that shows when the code was added, in the file notice.php, on this page on github.

    Weptile Image Plugin

    This plugin caught our attention when we did a Google search for functions used by other plugins to inject content. This plugin also uses the notice.php file with the same function names as the Menu Image plugin. It also loads content from the apistats.net domain.

    Weptile included code to remotely load content on July 2, 2016. It appears to have been removed from the repository on the same day.

    According to this article, the plugin later changed ownership. The date the post was published is September 25, 2015, about 1 year before the injection code appeared.

    The domain used to inject content was apistats.net which, as we pointed out above, does have a tentative link to Soiza and linkrocket.net.

    I was not able to contact any authors associated with this plugin.

    No Comments Plugin

    This plugin caught our attention because it used the wpl.io domain to inject content into websites. That domain was hosted on the same IP address as wpcdn.io which we linked directly to Soiza.

    The plugin contained code that injected content from April 25, 2014 until it was removed from the repository in October 2016. From version 1.1.5 the plugin contained code that could deliver spam.

    The code in this plugin checked to see if a user was signed in. If the user was not signed in, the plugin would fetch a blob of content from the wpl.io domain, decode it and inject it. The plugin would print meta-data in the affected site’s header and replace keywords in the content with links. The links were hidden from view using CSS. There was no option to disable this feature.

    I was not able to contact the authors of this plugin because it has been removed from the WordPress repository.

    Financial Calculator Plugin

    This plugin gets an honorable mention. It has never contained any spam code. However, the author, Ciprian Popescu, was kind enough to chat with me. It turns out that Soiza (pp@linkrocket.net) paid $600 to get access to this plugin but never used his access to add any code.

    The date on the transaction was April 24, 2017. After we chatted, Ciprian reverified that Soiza no longer has access, and assured us that the plugin is safe.

    A Coordinated Effort to Spam Via WordPress

    What this research reveals is a coordinated effort by a malicious actor to use WordPress plugins to serve spam that started in 2013 and lasted until this month.

    In some cases, the site owner had to opt in, but the opt-in message was either vague, misleading or obfuscated. In other cases, WordPress sites would serve ad content without the site owner being aware.

    The content from these ad servers includes ads for payday loans and ads for escort services based in the UK.

    How Plugin Authors Can Avoid Falling Victim

    WordPress is now the most popular content management system in the world by a significant margin. It is important that we, as plugin authors, recognize that we are attractive targets for bad people. If they can infect just a handful of plugins with malicious code, it can be worth millions.

    Stay secure and keep your code and customers secure. If you consider selling your plugin or a partnership, make sure you understand who you are selling to and what their history is. Also get a clear understanding of what their intentions are. Get references. You might also consider reaching out to the community, including Wordfence, to find out if anyone has any data on your buyer.

    I would also strongly recommend to plugin authors to never give ‘commit’ access to a partner. Have them send you a patch or feature request and review and commit the code yourself.

    Make sure your buyer or partner understands the plugin directory guidelines.

    How WordPress Site Owners Can Avoid Spam in Plugins

    As a WordPress site owner it is important that you keep track of which plugins are removed from the WordPress repository. In several cases, the plugins affected in this post were removed multiple times from the WordPress repository. If you are using Wordfence, it will notify you that a plugin you are using has been removed from the repository.

    When you see a plugin removed, go to the support forum and ask the developer what happened. If they aren’t clear and transparent in their answer, you might considering moving on and using a different plugin.

    Stay abreast of security developments by subscribing to our mailing list. When we discover that a plugin has been spamming, we immediately notify the community and take action.

    If a plugin does include “terms” that you need to agree to, read the entire T&C document included, especially the end. In several cases the plugins affected did have an ‘opt-in’ checkbox with terms that said ads would be injected into your website. The problem was that the language describing the ‘ads’ was at the very end of a very long document.

    If a plugin has an ‘opt-in’ checkbox for any kind of data sharing, make sure that it clearly describes what is being shared with or fetched from a remote server.

    A Final Note

    As a final note, I would like suggest that the WordPress.org maintainers enable two-factor authentication and enforce a strong password policy on the plugin repository. I’d like to suggest TOTP which integrates with password managers like 1Password.

    These supply chain attacks will only get worse.  This will probably be a large and complex project because it involves enabling two-factor on the code repository and other systems. So this will take time, because it risks breaking things.

    Mark Maunder – Wordfence Founder/CEO.

    Credits: Thank you to all the plugin authors who we interviewed and who shared data with us that helped us compile the connections described in this post. They are listed at the top of this post. Thanks to Brad Haas who compiled the data we gathered into a Maltego investigative graph (pictured at the top of this post) which was invaluable when understanding the complex connections in this story. Special thanks to Åsa Rosenberg, Ryan Britton, Matt Barry, Colette Chamberland, Dan Moen and other Wordfence team members who provided valuable assistance. Thanks to Andie La-Rosa for editing.

    The post 9 WordPress Plugins Targeted in Coordinated 4.5-Year Spam Campaign appeared first on Wordfence.

  • Staying Ahead of WordPress Attackers with the Real-Time IP Blacklist

    WordPress sites are under constant attack by criminals around the world. It is unnerving to see them at work, looking for security vulnerabilities to exploit and trying thousands of passwords. And when they are successful, they inflict pain in the form of lost revenue, damaged reputation and clean-up expenses. It’s no wonder that Wordfence users love our blocking features. There’s nothing more satisfying than taking direct action against an evil adversary.

    We like blocking the bad guys too. That’s why we added a real-time IP blacklist for our Premium customers in March. In today’s post we do a deep dive on how the blacklist works. We also find out whether keeping an eye on Live Traffic and Monthly Attack Reports and blocking all the bad guys yourself is a viable alternative to the real-time blacklist. Let’s take a look!

    Real-Time IP Blacklist Stats for August 2017

    For our deep dive we took a look at all of the changes to the blacklist for August 2017 and sliced and diced it for you. The blacklist is updated in real-time using algorithms we have developed that run on our Big Data platform. IPs can be added and removed multiple times during the month, so in the numbers below we included each add and remove in our counts.

    537,951 IPs Added in August

    During the month an average of 17,353 IPs were added to the blacklist each day. That’s over 12 per minute, so if you were to attempt to do this manually you would need to block an IP roughly every 5 seconds. And the IP Blacklist doesn’t sleep, you would need to find a few partners to work with you in shifts, as the updates are happening 7 days a week at all times of the day. And forget about holidays and vacations, the blacklist never stops working.

    This chart shows the number of IPs added by day for the month of August. As you can see, the number of additions dropped off sharply starting on August 27th. As we reported in the August WordPress Attack Report, we saw a dramatic drop-off in brute force attacks on the 27th and blacklist additions essentially mirrored that change.

    545,340 IPs Removed in August

    When we talk about blocking attackers, we almost always talk about adding new IPs to our list. If you’re managing a small list of IPs by hand, that’s probably not going to get you in much trouble. But when you’re managing a big list that’s used by thousands of site-owners worldwide you need to be really careful to remove IPs when they stop attacking. This is especially important because we know that attackers like to use compromised routers to attack WordPress sites, which we wrote about back in April. As soon as it’s clear that an IP has stopped attacking, we immediately remove it from the blacklist.

    In August an average of 17,591 IPs were removed from the blacklist each day. It’s no coincidence that is almost the same as the average number of additions, which was 17,353. Attackers are well aware that IP blacklists exist, so they cycle through IPs quickly in an attempt to evade them. That’s why IP blacklists need to be managed in real-time, blocking the bad guys as quickly as possible when they start attacking.

    The chart below shows the number of IPs removed by day for the month of August. There was a big drop-off in volume beginning on the 28th, the day after we saw volume drop for additions.

    How Long Do IPs Remain on the list?

    To answer this question we looked at IPs that were removed from the list during the month of August and calculated how many days they had been on the list at the time of removal. IPs that were on the list for less than 24 hours are in the “0” row.

    As you can see, a small percentage of IPs spend more than 24 hours on the list. It’s clear that attackers cycle through IPs quickly, most likely because they become much less effective after a day or two of use. The longest an IP was was on the list was 150 days. We also looked at time on the blacklist by number of hours and learned that the average time an IP address spends on the blacklist is just 10 hours.

    Do Attackers Reuse IPs?

    For this question we counted the total number of times IPs were added to the list during the month. It turns out that while attackers do re-use IPs, they don’t as often as you might think. Over two-thirds of the time IPs were added to the blacklist just once.

    Which Countries Dominate the Blacklist?

    Unlike the list of most active countries in our Monthly Attack Reports, which ranks countries based on total attacks, this list looks at the total number IPs that were added to the blacklist during the month for each country. The results might surprise you.

    Russia, the United States and Ukraine consistently dominate the top 3 in our Monthly Attack Reports. From this view, India and Brazil are at the top. In our monthly attack report, India and Brazil are generally towards the bottom of the top 10. Another surprise is that in this view, Ukraine isn’t even in the top 25.

    What this indicates is that India and Brazil see significantly more IPs being added and removed from our blacklist throughout the month. This indicates malicious activity that uses a larger number of IP addresses and that switches IP addresses frequently. This kind of activity is usually seen when attackers are using a botnet – a large number of machines that they control and use to launch a coordinated attack.

    The algorithm that we use to identify malicious IPs and add them to the blacklist includes logic that detects IPs that are part of a botnet and are engaging in a coordinated attack. These can be very hard to detect manually because each individual IP address only generates a small number of attacks on target websites.

    Which Organizations Do The IPs Belong To?

    By looking at the top 25 organizations that our newly blacklisted IPs belonged to in August we see a lot of the same telecom companies that we referenced in our post in April where we talked about the home router botnet attacking WordPress sites. Attackers clearly view vulnerable home routers as an attractive source of new IP addresses to launch attacks from.

    Conclusion

    While the blocking features in Wordfence are very popular and useful, the Real-time IP Blacklist feature included with Wordfence Premium is much more effective at protecting your site. Attackers are aware that IPs are eventually blocked so they cycle through them very quickly. For just $99 per year, which works out to just 27 cents a day, we’ll do the IP blocking for you.

    The post Staying Ahead of WordPress Attackers with the Real-Time IP Blacklist appeared first on Wordfence.

  • WordPress Security Update 4.8.2 – Update Immediately

    WordPress Core version 4.8.2 has just been released. This is a minor update and a security release which means that your sites will update automatically within the next 24 hours unless you have disabled auto updates.

    The update includes a fix to $wpdb->prepare() to help protect against SQLi injection attacks. WordPress core is not vulnerable to SQLi injection attacks directly, but certain plugins and themes may be vulnerable depending on how they use the $wpdb->prepare() function in their code. This fix alone is reason to update immediately to 4.8.2.

    The release fixes five cross site scripting vulnerabilities. These are in:

    • oEmbed discovery
    • The visual editor
    • The plugin editor
    • In template names

    Two path traversal vulnerabilities were fixed. These are:

    • In the file unzipping code
    • In the customizer

    An open redirect was also fixed on the user and term editing screens. 4.8.2 also includes 6 maintenance fixes.

    Now that the existence of these vulnerabilities is public, it becomes much more likely that they will be exploited. It is very important that you update as soon as possible to 4.8.2.

    To update manually now you can sign into your WordPress site, mouse over the Dashboard on the top left and click ‘Updates’ and complete the update process.

    Please share this information with the rest of the community to ensure everyone updates in a timely fashion. Thanks.

    Resources:

    The post WordPress Security Update 4.8.2 – Update Immediately appeared first on Wordfence.

  • The August 2017 WordPress Attack Report

    This is the ninth edition of the WordPress Attack Report series we’ve been publishing since December 2016. You can find reports from the previous months here:

    This report contains the top 25 attacking IPs for the month of August and their details. It also includes charts of brute force and complex attack activity for the same period. We also include the top themes and plugins that were attacked and which countries generated the most attacks for this period.

    The Top 25 Attacking IPs

    The next section is our standard explanation of how the table below works. If you are familiar with our attack reports, you can skip down to the table below this section, which contains the data for August along with some commentary.

    Brief Introduction (If You Are New to Viewing These Reports)

    In the table below, we’ve listed the most active attack IPs for August 2017. Note that the “Attacks” column is in millions, and is the total of all attacks that originated from each IP. Farther right in the table (you may have to scroll right) we break down the attacks into “brute force” attacks and “complex” attacks.

    Brute force attacks are login-guessing attacks. You can learn more about how brute force attacks work in our Learning Center article about them. What we refer to as “complex attacks” are attacks blocked by a rule in the Wordfence firewall.

    We have also included the netblock owner, which is the organization (usually a company) that owns the block of IP addresses that the attack IP belongs to. You can Google the name of each owner for more information. A Google search for any of these IP addresses frequently shows reports of attacks.

    The hostname included is the PTR record (reverse DNS record) that the IP address owner created for the IP, so this is not reliable data, but we still include it for interest. For example, we have seen PTR records that claim an IP is a Tor exit node, when, based on traffic, it is clearly not.

    We also include the city and country, if available. To the far right of the report, we show the date in August when we started logging attacks and the date the attacks stopped.

    The Top Attacking IPs

    The total attacks from the top 25 attacking IPs increased by almost 9% from August.

    Brute force attacks made up 91% of total attacks for August, up from 87% in July, which we had thought was an amazing number. Complex attacks accounted for 9% of the volume.

    The United States dominated the list this month, with 13 of the top 25 IPs. Ukraine had the second most IPs with 7. Turkey was notably absent from the top 25 this month.

    Brute Force Attacks on WordPress in August 2017

    In the chart below, we show the number of daily brute force attacks on the sites we monitor for the month of August.

     

    The average number of daily brute force attacks was down 9% from last month after growing the previous two months. Daily attack volumes were stable for most of the month, but plummeted on the 27th, dropping by roughly half for the final five days.

    Complex Attacks on WordPress in August 2017

    In the graph below, we show the daily complex attacks (attacks that attempt to exploit a security vulnerability) for August.

    Average daily attack volume for August was up 1% from August for the sites that we protect at 7.8 million. Daily volume was stable throughout the month, similar to what we saw in July.

    Attacks on Themes in August 2017

    The table below shows the total number of attacks on WordPress themes. We identify each theme using its slug, which is the directory where it is installed in WordPress.

    As usual we saw a lot of movement in the top 25 in August compared to July. The biggest move  on the list was the ‘twentyseventeen’ theme, moving up 95 spots to #2. Another big mover, the ‘sketch’ theme, moved up 34 spots to #6. Looking at the attacks involving both, they appear to be probe requests, looking for vulnerabilities that can be exploited in a subsequent request. We do not believe the attackers were attempting to compromise a vulnerability in these themes directly; rather, they were referencing very common themes in the probe requests.

    Another big mover on the list is the ‘nemesis’ theme, a premium WordPress theme that has been around since 2012. All of those attacks attempted to exploit a vulnerability in TimThumb versions <= 1.33, so we assume that the the theme included a vulnerable version of TimThumb some time in the past. The attacks appear to be coming from at least one large botnet as we saw attacks from 1637 unique IP addresses.

    Attacks on Plugins in August 2017

    The table below shows the total number of attacks on WordPress plugins. As with themes, we identify each plugin by its unique slug, which is the unique installation directory where the plugin is installed.

    As usual the top 25 list for plugins was pretty stable at the top with quite a bit of movement toward the bottom. There were 8 plugins on the list that weren’t in the top 25 in July. We looked into the details behind the first big mover toward the top of the list, ‘formcraft’. Almost all of the attacks were malicious file upload attacks, likely attempting to exploit a vulnerability that was made public in February of 2016.

    We took a look at the attacks on the next highest big mover, ‘easyrotator-for-wordpress’, and found that almost all of the requests appear to be attempting to probe for and in some cases send commands to back door files. Over 30 thousand unique sites were hit with these attacks, so we assume that the attacker was looking for backdoors that another attacker had already installed versus communicating with back doors they had installed themselves.

    The next big mover on the list, ‘hb-audio-gallery-lite’, moved up as a result of attempts to exploit a vulnerability that has been public since March of 2016. Over 89% of the attacks originated from just two IP addresses.

    Another big mover, ‘rb-agency’, saw attempts to exploit a vulnerability made public in September of 2016. Interestingly, over 91% of the attacks originated from the same two IP addresses responsible for the majority of the ‘hb-audio-gallery-lite’ attacks.

    Attacks by Country for August 2017

    The table below shows the top 25 countries from which attacks originated in the month of August on WordPress sites that we monitor.

    The top of the list was pretty stable, with the United States and Russia trading places at the top and the Ukraine holding at number 3.

    Conclusion

    That concludes our August 2017 WordPress attack report. It was nice to see attack volumes drop off at the end of the month, and we hope that trend continues through the month of September.

    The post The August 2017 WordPress Attack Report appeared first on Wordfence.