New Microsoft Security and Compliance blog: How to Gain More from your Connection to an OT Network

How to Gain More from your Connection to an OT Network

One of the most productive and non-intrusive tools in the Cyber Security Engineer’s bag is passive Network Traffic Analysis (NTA).  Providing network maps, inventory, and firmware information among other benefits provides insights that are not generally known any other way.  Manual inventory collection methods are error-prone and expose this information to interception over corporate email networks, shared file folders, etc.  But how do we implement this kind of system without causing any bumps in the road for real-time processes?  What are the risks?  Which methods are best?  The best sensor does no good unconnected and is of little value connected in the wrong part of the network. 

 To discuss this, I will use a diagram that was developed for my last blog post Designing a Robust Defense for Operational Technology Using Azure Defender for IoT (microsoft.com).  This diagram (below) shows an example OT network monitored by Azure Defender for IoT. Defender for IoT is an agentless passive Network Traffic Analysis tool with strong roots in Operational Technology, now expanding to IoT. Defender for IoT discovers OT/IoT devices, identifies vulnerabilities, and provides continuous OT/IoT-aware monitoring of network traffic.  The recommended locations for Azure Defender for IoT  (AD4IoT) are shown in red color.  Why have these locations been chosen?  To explain this, we will break this network into pieces and address these issues for each type of traffic.

 

 

 Starting with the lower portion of this sketch, let’s look at traffic flows around the PLCs. 

 

  1. The first arrow shows traffic between a PLC and its ethernet-connected Input/Output (I/O) modules.  This traffic utilizes simplistic protocols and is very structured and periodic.  It can be leveraged as a threat to the overall OT system and is more vulnerable when I/O is remote from the PLCs in unsecured areas.  Malicious applications could perform inappropriate control actions and/or falsify data.  Firmware problems in I/O modules often go unpatched unless some form of undesirable behavior is experienced. In certain families of PLCs or controllers, the Defender for IoT can provide data on firmware levels and types of I/O modules if this data is requested by an HMI or historian. 

 The mechanism to monitor this traffic is to span switches used in the I/O subsystem as shown here.  If they are unmanaged switches, taps may be located at the connection to the PLC or controller.

 2. The second arrow identifies traffic from Variable Frequency Drives or similar equipment often interfaced with the PLCs or Controllers.  This communication may be Modbus, Rockwell Protocols, or CIP.  Equipment could be damaged or destroyed by inappropriate commands sent to such devices.  Good engineering practice would put bounds of reasonability around all potential setpoints, but this may not be the case.  These protocols are well understood and in the public domain.  A man-in-the-middle attack could affect this type of equipment.  Monitoring these communications can identify inappropriate function calls, program or firmware changes, and parameter updates. As above, switch span or taps are the mechanisms to monitor this traffic.

 3. Custom engineered systems may utilize well-known, open OT protocols such as Modbus, OPC, or others.  This traffic should be monitored even if it is not fully understood as the behavior patterns should be very predictable.  It is common for these systems to utilize unusual functions and atypical ranges for data.  This is the result of a developer reading a protocol spec with no actual field experience with the protocol.  Custom alerts can be configured and tuned based on the nature of the data.  Since such systems are engineered to order for a specific purpose, the damage could have long-term implications on plant production.

 4. Traffic crossing OT Access-level switches should always be monitored.  This is the primary point at which PLCs or controllers communicate with HMIs, engineering stations, and sometimes historians.  The problem here is that these switches carry the actual OT control traffic.  Any action that could compromise this traffic affects the reliability of the OT system.  Many switches at the I/O and access layers may be unmanaged devices.  By unmanaged, I mean that they are not configurable and therefore cannot support a SPAN (or mirror) session. 

 Unmanaged switches is not an insurmountable hurdle.  Two possible paths may be followed from this point.  The least intrusive is to install network taps.  The security engineer should consult with the OT engineer on the most valuable locations for taps.  Since a stand-alone tap monitors only one data stream, the most valuable assets (compromise targets) should be monitored. These would normally be at least the engineering station, historian and/or alarms server (if appropriate), and HMIs, particularly those with engineering tools installed.  If it is necessary to monitor all traffic, a tap aggregator may be used.

Another approach would be to replace the unmanaged switches with managed switches.  This may sound daunting but usually is not.  Most managed switches are configured to “wake up” in a basic mode which approximates an unmanaged switch.  So replacement, while requiring a system shutdown, can be accomplished rather quickly and have the system up and functioning again.  Once this is done, the configuration can be added to provide basic security and copy traffic to a SPAN or mirror port.  Make sure these configurations are saved as most switches make changes to operating memory which is not stored on power reset.  It is generally recommended to discuss this change with your OT support personnel and/or OEM service engineers.  They probably have some standard switch configurations that they apply when a customer requests managed switches.  Additionally, they should be able to provide you with approximate bus speeds needed to support OT traffic with mirroring.

What are the risks? In the case of switch SPAN (SwitchPort ANalyzer), or mirror sessions, the only concern of serious significance is the current traffic level on the switch.  If a SPAN session is added to a heavily loaded switch, the SPAN may drop packets because the SPAN session is a lower priority than actual switching traffic. This could mean that some packets might slip through unmonitored.  However, it does not affect the normal functioning of the switch for ICS traffic.  Some switches, if they are greatly overloaded can revert to ‘flood mode’ in which they act as a network hub.  This situation is extremely rare.  If switch SPANning is chosen as a method, it is wise to monitor network traffic on the switch prior to adding the session.  Assume that a full switch span will double the switch backbone traffic. 

 If network taps are installed, the risks are insignificant.  Passive taps should of course be chosen.  Passive means that the tap continues to pass control traffic even if it loses power.  Passive taps are simply inserted in-line with the existing traffic, see sketch below.  Installation needs to be coordinated with OT engineers to limit the impact on operating processes. 

 

 Next, we will discuss special equipment including analysis devices and robotics.  This portion of the overall diagram is shown below.  

    

Network traffic to analyzers typically looks like normal PC traffic using common IT protocols.  Most analyzers have some form of controller that is designed for a specific function.  Sometimes the PC is the controller, utilizing specialized I/O boards included in the machine. Some analyzers or groups of analyzers may be managed by mini computers.   In any case, from a network security perspective, these devices appear on the network as computers, not analyzers per se.  Patching of these customized machines often lags behind the upgrade strategies used for standard IT equipment.  Upgrades to analysis systems must be approved by, and often be implemented by the OEMs which may be expensive and involve downtime. Because of infrequent patching and/or OS upgrades, this equipment can become a security liability on a lab network. Ideally, lab equipment should be separated either physically onto separate networks or via VLANs, but such changes may require extensive planning and testing and still can be disruptive to ongoing lab processes.

Most major medical laboratories utilize either a LIMS (Laboratory Information Management System) or a middleware server to collect analytics data from these devices and forward that data to a patient information database managed either locally or in the cloud (see sketch below).  Hence, the traffic to/from the analyzer will be most easily recognized by the ultimate destination at the middleware or LIMS.  Since these potentially vulnerable machines may process interactions with users on the lab network for input data or maintenance functions, they should be monitored more closely than fully patched IT machines.  This presents a challenge to lab IT managers who may want to gain a handle on this type of OT equipment in their network but may not have good inventory information. 

Since medical testing facilities utilize normal switched networks, monitoring should be installed at an appropriate location to ‘see’ all the traffic from analyzers to the middleware or LIMS server.  This could be either core or distribution level switches depending on the network design.  Standard SPAN or mirror traffic can be used.

Dual-homed machines present special security challenges since they could be converted to active routers by malware.  It is common for expensive lab or analysis equipment to be leased.  OEM terms and conditions specify how this equipment may be used and what service it requires to achieve contracted performance.  This is often monitored via a ‘secure’ datalink to the manufacturer’s support site.  These may or may not be bi-directional.  These links are generally firewalled, either by the OEM, by the customer or by both.  Bi-directional links are inherently a threat.  Remote access to a computer on the lab network can put much more than that computer in jeopardy. 

 

 

In robotic applications, the primary issue is the speed of response.  The control systems are complex, utilizing high-level programming toolsets.  The low-level communication may not utilize standard ethernet framing.  Robot protocols vary widely and include Ethernet/IP, DeviceNet, Profibus-DP, Profinet, CC-Link, and EtherCat protocols.  Physical media may be Cat5/6, but RG-6 coaxial, twisted pair, RS-485, and fiber are also used.  Monitoring the low-level communication between controllers and robots requires careful coordination with the equipment designer and should not be attempted casually.  Network monitoring should utilize taps. Switch SPAN, or mirroring is not recommended. 

 As described above, most industrial robots are programmed using a computer workstation.  Downloading and selection of programs may be manual or automated using standard network protocols. So, monitoring should focus on the programming workstations and the source of robot program selections.  Robot program file downloads may be transferred from a central server.  These could occur over SFTP, FTP, SMB, or other methods. 

 Finally, we would like to address the OT interface to the business (Enterprise) network.  This can be a gateway for potential threats to OT systems.  Some vulnerabilities that may be unsuccessful in the IT network space may cause severe problems in the OT space because the machines may not be patched.  Out of date and unsupported operating systems may be in use.  As a result, traffic that enters from the Enterprise network and ultimately reaches the OT network should be monitored. 

 Generally, good practice prevents any direct traversal of the DMZ.  For instance, remote desktop sessions should be hosted by a RAS server in the DMZ which is then used to open a remote desktop session into an OT machine with different credentials. Elaborate credential systems with short password lives attempt to increase the challenge for attackers attempting to gain control.  Well designed implementations keep all machines in the DMZ patched up-to-date which should limit the effect of known vulnerabilities. 

 Zero day vulnerabilities will always be a threat prior to discovery.  So, monitoring sessions entering the DMZ from the Enterprise and those leaving the DMZ for the OT network are an important part of a security design.  Similarly, monitoring traffic from the OT network to a Historian server and Enterprise connections to that same server could uncover issues.  Since these sessions are often encrypted, efforts should focus on the legitimacy of the Enterprise hosts, times of access, data rates, and other indicators to validate these externally generated sessions.

 The DMZ is also used as a connection point for a variety of other facility systems such as IP phones; perimeter security systems; weather stations; contracted supply systems like water purification, compressed air supply and the like; wireless devices; etc.  In most cases, these various systems are assigned separate VLANs and subnets.  By monitoring all the VLANS in this zone, suspicious traffic can be identified and managed.  Traffic originating from any of these devices to the ICS network should not normally exist. 

Subnet-to-subnet traffic could be cause for concern.  This is another area where Defender for IoT can help.  By mapping the assets, assigning them to VLANs, subnets, and user assigned subsystems, communication between the various device groups can be easily seen greatly aiding efforts to perform or monitor network segregation.  

 The visual network map produced by Defender for IoT in conjunction with the filtering capabilities on the map make it easy to identify interconnections between various plant control systems. Having a powerful visual of group-to-group communication makes the effort of segmentation much easier.  This process is a long and tedious one using arp tables on switches.  Also, if this effort is underway, the map will show areas that may have been overlooked.

 Conclusions:

 Well-engineered connections to ICS networks can yield valuable results, including accurate inventories, network maps, and improved security with no risk to the reliability of the underlying OT systems.  This information can be combined, in Azure Sentinel or other SIEM/SOAR solutions, with agent-based Defender for endpoint data to produce a complete picture of OT networks.  Custom-designed playbooks can assist your analysts in responding to OT or IoT issues.  

Teamwork between OT engineers and IT security personnel can yield benefits for both groups while presenting a more challenging landscape to potential intruders.

 

New Azure Sentinel blog: What's new: Watchlists templates are now in public preview!

As we know, each organization is unique and have different use cases and scenarios in mind when it come to security operations. Nevertheless we've identified several use cases that are common across many SOC teams.

Azure Sentinel now provides built-in watchlist templates, which you can customize for your environment and use during investigations.

After those watchlists are populated with data, you can correlate that data with analytics rules, view it in the entity pages and investigation graphs as insights, create custom uses such as to track VIP or sensitive users, and more.

Watchlist templates currently include:

• VIP Users. A list of user accounts of employees that have high impact value in the organization.
• Terminated Employees. A list of user accounts of employees that have been, or are about to be, terminated.
• Service Accounts. A list of service accounts and their owners.
• Identity Correlation. A list of related user accounts that belong to the same person.
• High Value Assets. A list of devices, resources, or other assets that have critical value in the organization.
• Network Mapping. A list of IP subnets and their respective organizational contexts.

 

Watchlists templates insights in entity pages

 

We've created the watchlists templates schemas to be super easy and extensible, in order for you to populate it with the relevant data. more information about using the watchlists templates can be found here.

 

What’s next?

 

Beside surfacing the watchlists templates data inside the entity pages, we're working on embedding this information in the UEBA anomalies, and the entity risk score which is planned next. Understanding if a user is a VIP/Terminated or an asset is an HVA is important to provide both context and security value for the analyst while investigating.

 

New Azure Sentinel blog: What’s new: Fusion Detection for Ransomware

In collaboration with the Microsoft Threat Intelligence Center (MSTIC), we are excited to announce Fusion detection for ransomware is now publicly available!

 

These Fusion detections correlate alerts that are potentially associated with ransomware activities that are observed at defense evasion and execution stages during a specific timeframe. Once such ransomware activities are detected and correlated by the Fusion machine learning model, a high severity incident titled “Multiple alerts possibly related to Ransomware activity detected” will be triggered in your Azure Sentinel workspace.

 

In order to help your analyst quickly understand the possible attack, Fusion provides you with a complete picture for the suspicious activities happened on the same device/host by correlating signals from Microsoft products as well as signals in network and cloud. Supported data connectors include:

• Azure Security Center
• Microsoft Cloud App Security
• Microsoft Defender for Endpoint
• Microsoft Defender for Identity
• Azure Sentinel scheduled analytics rules. Fusion only considers scheduled analytics rules with tactics information.

 

The screenshot below shows a Fusion incident with 22 alerts. It correlates low severity signals that were detected around the same timeframe from the network and the host to show a possible ransomware attack and the different techniques used by attackers.

 

 

 

For more information, see Multiple alerts possibly related to Ransomware activity detected.

 

Why Fusion detection for ransomware?

Ransomware attack is a type of attack that involves using specific types of malicious software or malware to make network or system inaccessible for the purpose of extortion – ‘ransom’. There is no doubt that ransomware attacks have taken a massive turn in being the top priority as a threat to many organizations. A recent report released by PurpleSec revealed that the estimated cost of ransomware attacks was $20 billion in 2020 and with downtime increasing by over 200% and the cost being 23x higher than 2019.

 

Preventing such attacks in the first place would be the ideal solution but with the new trend of ‘ransomware as a service’ and human operated ransomware, the scope and the sophistication of attacks are increasing – attackers are using slow and stealth techniques to compromise network, which makes it harder to detect them in the first place.

 

With Fusion detection for ransomware that captures malicious activities at the defense evasion and execution stages of an attack, it gives security analysts an opportunity to quickly understand the suspicious activities happened around the same timeframe on the common entities, connect the dots and take immediate actions to disrupt the attack. When it comes to ransomware attacks, time more than anything else is the most important factor in preventing more machines or the entire network from getting compromised. The sooner such alerts are raised to security analysts with the details on various attacker activities, the faster the ransomware attacks can be contained and remediated. A detection like this will help analysts by giving the compilation of attacker activity around execution stage helping reduce MTTD (Mean Time to Detect) and MTTR (Mean Time to Respond).

 

Examples of the Fusion detection for ransomware

In the Incident 1 example, Fusion correlates alerts triggered within a short timeframe on the same device, indicating a possible chain of attacks from how the attackers got in through possible RDP brute-force attack, followed by the use of a ‘Cryptor’ malware and potential phishing activities using malicious document associated with the EUROPIUM activity group, to the detection of Petya and WannaCrypt ransomware in the network.

 

Incident 1

 

 

Incident 2 below is another example of the Fusion ransomware detection that was confirmed as true positive. This incident correlates alerts showing ransomware activities at defense evasion and execution stages on the same host, along with additional suspicious activities detected during the same timeframe to show you possible techniques used by attackers to compromise the host.

 

Incident 2

 

In these Fusion incidents, the alerts related to ransomware/malware detection might indicate that the ransomware/malware was stopped from delivering its payload but it is prudent to check the machine for signs of infection. Attackers may continue malicious activities after ransomware was prevented – it is also important that you investigate the entire network to understand the intrusion and identify other machines that might be impacted by this attack.

 

What’s next after receiving the Fusion detection?

After receiving Fusion detentions for possible ransomware activities, we recommend you to check with the machine owner if this is intended behavior. If the activity is unexpected, treat the machine as potentially compromised and take immediate actions to analyze different techniques used by attackers to compromise the host and to evade detection in this potential ransomware attack. Here are the recommended steps:

 

1. Isolate the machine from the network to prevent potential lateral movement.
2. Run a full antimalware scan on the machine, following any resulting remediation advice.
3. Review installed / running software on the machine, removing any unknown or unwanted packages.
4. Revert the machine to a known good state, reinstalling operating system only if required and restoring software from a verified malware-free source.
5. Resolve to recommendations from alert providers (e.g. Azure Security Center and Microsoft Defender) to prevent future breaches.
6. Investigate the entire network to understand the intrusion and identify other machines that might be impacted by this attack.

 

As you investigate and close the Fusion incidents, we encourage you to provide feedback on whether this incident was a True Positive, Benign Positive, or a False Positive, along with details in the comments. Your feedback is critical to help Microsoft deliver the highest quality detections!

 

New Microsoft Security and Compliance blog: Warn and Educate Users on Risky App Usage

 Recent reports show the high extent to which information workers are utilizing cloud apps while doing their everyday tasks. In an average enterprise, there are more than 1500 different cloud services used, where less than 12% of them are sanctioned or managed by the IT teams. Considering that more than 78GB of data is being uploaded monthly to risky apps we conclude that most organizations are exposed to potential data loss or risks coming out of these cloud applications.

 

Shadow IT usage of risky apps is usually mitigated by a strict approach of blocking any usage of these cloud apps that do not meet certain risk criteria, this approach is already enabled today by using Microsoft’s Cloud App Security Shadow IT Discovery capabilities with its native integration with Microsoft Defender for Endpoints and it’s native integration with other 3^rd party network appliances. 

But what about apps that are widely used by employees and enable their productivity (especially in the Work from home/Covid-19 era) and their risk is not conclusive enough for a strict block. To enable the delicate balance between employee’s productivity, and the need for risk and compliance awareness, organizations need to take a gradual approach:

• Warn users that this app is not recommended/allowed but allows users to bypass to enable productivity. 
• IT can monitor access and bypasses such apps and learn usage trends and importance.
• IT can offer sanctioned and managed alternatives for the users by creating a contextual company web page that provides sanctioned alternatives in the organization. 

We are pleased to announce the public preview for a new endpoint-based capability to allow management and control of Monitored cloud applications, manage these Monitored applications applying soft block experience for end-users when accessing these apps. Users will have an option to bypass the block.
IT admins will be able to add a dedicated custom redirect link so users can get more context on why they were blocked in the first place and what valid alternatives do they have for such apps in the organization.

Besides enabling the soft block experience, admins will be able to continuously monitor these apps and understand how many of the users adhered to the block and chose other alternatives, or, decided to bypass the block and continue using the app – this will serve as a strong indication, org-wide, whether this app is necessary and should be considered for deeper management by IT.

By adopting a more gradual and less strict approach for blocking cloud applications, IT organizations can reduce their overhead of handling exception requests, but in parallel drive employee awareness.

 

How does it work?

 

 In Cloud App security, tag the targeted app as Monitored.

 

The corresponding URL/Domains indicators will appear in the Microsoft Defender for Endpoints security portal as a new URL/Domain indicator with action type Warn.

 

When a user attempts to access the Monitored app, they will be blocked by Windows defender network protection but will allow the user to bypass the block or get more details on why he was blocked by redirecting him to a dedicated custom web page managed by the organization.

 

 

Over time, an IT admin can monitor the usage pattern of the app in Cloud App Security's discovered app page and monitor how many users have bypassed the warning message.

 

 

 

Get started

After you have verified that you have all the integration prerequisites listed in this article, follow the steps below to start warning on access to Monitored apps with Cloud App Security and Microsoft Defender for Endpoint.

Step 1

In Microsoft 365 Defender, go to settings > Endpoints > Advanced features and enable Microsoft Cloud App Security integration and Custom network indicators.

 

Step 2

In the Microsoft Cloud App Security portal, go to  Settings > Microsoft Defender for Endpoint:

• Mark the checkbox to enable blocking of endpoint access to cloud apps marked as unsanctioned in Cloud App Security
• [Optional] Set a custom redirect URL for a company coaching page.
• [Optional] Set the Bypass duration time after which the user will get warned once again on access to the app.

 

 

 

New Microsoft Blog -Trend-spotting email techniques: How modern phishing emails hide in plain sight

 With the massive volume of emails sent each day, coupled with the many methods that attackers use to blend in, identifying the unusual and malicious is more challenging than ever. An obscure Unicode character in a few emails is innocuous enough, but when a pattern of emails containing this obscure character accompanied by other HTML quirks, strange links, and phishing pages or malware is observed, it becomes an emerging attacker trend to investigate. We closely monitor these kinds of trends to gain insight into how best to protect customers.

This blog shines a light on techniques that are prominently used in many recent email-based attacks. We’ve chosen to highlight these techniques based on their observed impact to organizations, their relevance to active email campaigns, and because they are intentionally designed to be difficult to detect. They hide text from users, masquerade as the logos of trusted companies, and evade detection by using common web practices that are usually benign:

• Brand impersonation with procedurally-generated graphics
• Text padding with invisible characters
• Zero-point font obfuscation
• Victim-specific URI

We’ve observed attackers employ these tricks to gain initial access to networks. Although the examples we present were primarily seen in credential theft attacks, any of these techniques can be easily adapted to deliver malware.

By spotting trends in the threat landscape, we can swiftly respond to potentially malicious behavior. We use the knowledge we gain from our investigations to improve customer security and build comprehensive protections. Through security solutions such as Microsoft Defender for Office 365 and the broader Microsoft 365 Defender, we deliver durable and comprehensive protection against the latest attacker trends.

Brand impersonation with procedurally-generated graphics

We have observed attackers using HTML tables to imitate the logos and branding of trusted organizations. In one recent case, an attacker created a graphic resembling the Microsoft logo by using a 2×2 HTML table and CSS styling to closely match the official branding.

Spoofed logos created with HTML tables allow attackers to bypass brand impersonation protections. Malicious content arrives in users’ inboxes, appearing to recipients as if it were a legitimate message from the company. While Microsoft Defender for Office 365 data shows a decline in the usage of this technique over the last few months, we continue to monitor for new ways that attackers will use procedurally-generated graphics in attacks.

Figure 1. Tracking data for small 2×2 HTML tables

How it works

A graphic resembling a trusted organization’s official logo is procedurally generated from HTML and CSS markup. It’s a fileless way of impersonating a logo, because there are no image files for security solutions to detect. Instead, the graphic is constructed out a specially styled HTML table that is embedded directly in the email.

Of course, inserting an HTML table into an email is not malicious on its own. The malicious pattern emerges when we view this technique in context with the attacker’s goals.

Two campaigns that we have been tracking since April 2021 sent targets emails that recreated the Microsoft logo. They impersonated messages from Office 365 and SharePoint. We observed the following email subjects:

• Action Required: Expiration Notice On <Email Address>
• Action Required: 3 Pending Messages sent <date>
• New 1 page incoming eFax© message for “<Email Alias>”

Figure 2. Sample emails that use HTML code to embed a table designed to mimic the Microsoft logo

Upon extracting the HTML used in these emails, Microsoft analysts determined that the operators used the HTML table tag to create a 2×2 table resembling the Microsoft logo. The background color of each of the four cells corresponded to the colors of the quadrants of the official logo.

Figure 3. Page source of the isolated HTML mimicking the Microsoft logo

HTML and CSS allow for colors to be referenced in several different ways. Many colors can be referenced in code via English language color names, such as “red” or “green”. Colors can also be represented using six-digit hexadecimal values (i.e., #ffffff for white and #000000 for black), or by sets of three numbers, with each number signifying the amount of red, green, or blue (RGB) to combine. These methods allow for greater precision and variance, as the designer can tweak the numbers or values to customize the color’s appearance.

Figure 4. Color values used to replicate the Microsoft logo

As seen in the above screenshot, attackers often obscure the color references to the Microsoft brand by using color names, hexadecimal, and RGB to color in the table. By switching up the method they use to reference the color, or slightly changing the color values, the attacker can further evade detection by increasing variance between emails.

Text padding with invisible characters

In several observed campaigns, attackers inserted invisible Unicode characters to break up keywords in an email body or subject line in an attempt to bypass detection and automated security analysis. Certain characters in Unicode indicate extremely narrow areas of whitespace, or are not glyphs at all and are not intended to render on screen.

Some invisible Unicode characters that we have observed being used maliciously include:

• Soft hyphen (U+00AD)
• Word joiner (U+2060)

Both of these are control characters that affect how other characters are formatted. They are not glyphs and would not even be visible to readers, in most cases. As seen in the following graph, the use of the soft hyphen and word joiner characters has seen a steady increase over time. These invisible characters are not inherently malicious, but seeing an otherwise unexplained rise of their use in emails indicates a potential shift in attacker techniques.

Figure 5. Tracking data for the invisible character obfuscation technique

How it works

When a recipient views a malicious email containing invisible Unicode characters, the text content may appear indistinguishable from any other email. Although not visible to readers, the extra characters are still included in the body of the email and are “visible” to filters or other security mechanisms. If attackers insert extra, invisible characters into a word they don’t want security products to “see,” the word might be treated as qualitatively different from the same word without the extra characters. This allows the keyword to evade detection even if filters are set to catch the visible part of the text.

Invisible characters do have legitimate uses. They are, for the most part, intended for formatting purposes: for instance, to indicate where to split a word when the whole word can’t fit on a single line. However, an unintended consequence of these characters not displaying like ordinary text is that malicious email campaign operators can insert the characters to evade security.

The animated GIF below shows how the soft hyphen characters are typically used in a malicious email. The soft hyphen is placed between each letter in the red heading to break up several key words. It’s worth noting that the soft hyphens are completely invisible to the reader until the text window is narrowed and the heading is forced to break across multiple lines.

Figure 6. Animation showing the use of the invisible soft hyphen characters

In the following example, a phishing email has had invisible characters inserted into the email body: specifically, in the “Keep current Password” text that links the victim to a phishing page.

Figure 7. Microsoft Office 365 phishing email using invisible characters to obfuscate the URL text.

The email appears by all means “normal” to the recipient, however, attackers have slyly added invisible characters in between the text “Keep current Password.” Clicking the URL directs the user to a phishing page impersonating the Microsoft single sign-on (SSO) page.

In some campaigns, we have seen the invisible characters applied to every word, especially any word referencing Microsoft or Microsoft products and services.

Zero-point font obfuscation

This technique involves inserting hidden words with a font size of zero into the body of an email. It is intended to throw off machine learning detections, by adding irrelevant sections of text to the HTML source making up the email body. Attackers can successfully obfuscate keywords and evade detection because recipients can’t see the inserted text—but security solutions can.

Microsoft Defender for Office 365 has been blocking malicious emails with zero-point font obfuscation for many years now. However, we continue to observe its usage regularly.

Figure 8. Tracking data for emails containing zero-point fonts experienced surges in June and July 2021

How it works

Similar to how there are many ways to represent colors in HTML and CSS, there are also many ways to indicate font size. We have observed attackers using the following styling to insert hidden text via this technique:

• font-size: 0px
• font-size: 0.0000em
• font-size: 0vw
• font-size: 0%
• font: italic bold 0.0px Georgia, serif
• font: italic bold 0em Georgia, serif
• font: italic bold 0vw Georgia, serif
• font: italic bold 0% Georgia, serif

Being able to add zero-width text to a page is a quirk of HTML and CSS. It is sometimes used legitimately for adding meta data to an email or to adjust whitespace on a page. Attackers repurpose this quirk to break up words and phrases a defender might want to track, whether to raise an alert or block the content entirely. As with the invisible Unicode character technique, certain kinds of security solutions might treat text containing these extra characters as distinct from the same text without the zero-width characters. This allows the visible keyword text to slip past security.

In a July 2021 phishing campaign blocked by Microsoft Defender for Office 365, the attacker used a voicemail lure to entice recipients into opening an email attachment. Hidden, zero-width letters were added to break up keywords that might otherwise have been caught by a content filter. The following screenshot shows how the email appeared to targeted users.

Figure 9. Sample email that uses the zero-point font technique

Those with sharp eyes might be able to spot the awkward spaces where the attacker inserted letters that are fully visible only within the HTML source code. In this campaign, the obfuscation technique was also used in the malicious email attachment, to evade file-hash based detections.

Figure 10. The HTML code of the email body, exposing the use of the zero-point font technique

Victim-specific URI

Victim-specific URI is a way of transmitting information about the target and creating dynamic content based upon it. In this technique, a custom URI crafted by the attacker passes information about the target to an attacker-controlled website. This aides in spear-phishing by personalizing content seen by the intended victim. This is often used by the attacker to create legitimate-seeming pages that impersonate the Single Sign On (SSO) experience.

The following graph shows cyclic surges in email content, specifically links that have an email address included as part of the URI. Since custom URIs are such a common web design practice, their usage always returns to a steady baseline in between peaks. The surges appear to be related to malicious activity, since attackers will often send out large numbers of spam emails over the course of a campaign.

Figure 11. Tracking data for emails containing URLs with email address in the PHP parameter

In a campaign Microsoft analysts observed in early May 2021, operators generated tens of thousands of subdomains from Google’s Appspot, creating unique phishing sites and victim identifiable URIs for each recipient. The technique allowed the operators to host seemingly legitimate Microsoft-themed phishing sites on third-party infrastructure.

How it works

The attacker sends the target an email, and within the body of the email is a link that includes special parameters as part of the web address, or URI. The custom URI parameters contain information about the target. These parameters often utilize PHP, as PHP is a programming language frequently used to build websites with dynamic content—especially on large platforms such as Appspot.

Details such as the target’s email address, alias, or domain, are sent via the URI to an attacker-controlled web page when the user visits the link. The attacker’s web page pulls the details from the parameters and use that to present the target with personalized content. This can help the attacker make malicious websites more convincing, especially if they are trying to mimic a user logon page, as the target will be greeted by their own account name.

Custom URIs containing user-specific parameters are not always, or even often, malicious. They are commonly used by all kinds of web developers to transmit pertinent information about a request. A query to a typical search engine will contain numerous parameters concerning the nature of the search as well as information about the user, so that the search engine can provide users with tailored results.

However, in the victim identifiable URI technique, attackers repurpose a common web design practice to malicious ends. The tailored results seen by the target are intended to trick them into handing over sensitive information to an attacker.

In the Compact phishing campaign described by WMC Global and tracked by Microsoft, this technique allowed the operators to host Microsoft-themed phishing sites on any cloud infrastructure, including third-party platforms such as Google’s Appspot. Microsoft’s own research into the campaign in May noted that not only tens of thousands of individual sites were created, but that URIs were crafted for each recipient, and the recipient’s email address was included as a parameter in the URI.

Newer variants of the May campaign started to include links in the email, which routed users through a compromised website, to ultimately redirect them to the Appspot-hosted phishing page. Each hyperlink in the email template used in this version of the campaign was structured to be unique to the recipient.

The recipient-specific information passed along in the URI was used to render their email account name on a custom phishing page, attempting to mimic the Microsoft Single Sign On (SSO) experience. Once on the phishing page, the user was prompted to enter their Microsoft account credentials. Entering that information would send it to the attacker.

Microsoft Defender for Office 365 delivers protection powered by threat intelligence

As the phishing techniques we discussed in this blog show, attackers use common or standard aspects of emails to hide in plain sight and make attacks very difficult to detect or block. With our trend tracking in place, we can make sense of suspicious patterns, and notice repeated combinations of techniques that are highly likely to indicate an attack. This enables us to ensure we protect customers from the latest evasive email campaigns through Microsoft Defender for Office 365. We train machine learning models to keep an eye on activity from potentially malicious domains or IP addresses. Knowing what to look out for, we can rule out false positives and focus on the bad actors.

This has already paid off. Microsoft Defender for Office 365 detected and protected customers from sophisticated phishing campaigns, including the Compact campaign. We also employed our knowledge of prevalent trends to hunt for a ransomware campaign that might have otherwise escaped notice. We swiftly opened an investigation to protect customers from what seemed at first like a set of innocuous emails.

Trend tracking helps us to expand our understanding about prevalent attacker tactics and to improve existing protections. We’ve already set up rules to detect the techniques described in this blog. Our understanding of the threat landscape has led to better response times to critical threats. Meanwhile, deep within Microsoft Defender for Office 365, rules for raising alerts are weighted so that detecting a preponderance of suspicious techniques triggers a response, while legitimate emails are allowed to travel to their intended inboxes.

Threat intelligence also drives what new features are developed, and which rules are added. In this way, generalized trend tracking leads to concrete results. Microsoft is committed to using our knowledge of the threat landscape to continue to track trends, build better protections for our products, and share intelligence with the greater online community.