The following pages provide an overview of the threats of targeted cyber intrusions, ransomware and external adversaries who destroy data and prevent computers/networks from functioning, as well as malicious insiders.
Implementation guidance for associated mitigation strategies is provided later in this publication, and a table summary of the associated mitigation strategies is provided in the complementary Strategies to Mitigate Cyber Security Incidents publication.
Targeted cyber intrusions involve external adversaries who steal data. This can damage the competitive advantages and reputation of affected organisations, damage a country’s economic wellbeing, influence public opinion, negatively impact citizens due to the release of their private data, and unnecessarily consume scarce financial and staff resources to respond to such intrusions.
Organisations need to identify the type and location of their sensitive data stored electronically, as part of a security risk assessment performed to identify the level of protection that their assets require from various threats. For the purpose of this publication, sensitive data refers to either unclassified or classified information identified as requiring protection. This protection is often focused on maintaining confidentiality of the data, although data integrity and availability are also important and are often overlooked. Such data might reside within organisations in various locations including government ministerial submissions and other documents detailing government intentions, strategic planning documents, business proposals, tenders, meeting minutes, financial and accounting reports, legal documents, and intellectual property holdings.
No set of mitigation strategies is guaranteed to prevent all targeted cyber intrusions. However, organisations should still implement mitigation strategies that address all three high level stages of targeted cyber intrusions.
Stage 1 – Malicious software (malware) delivery and execution:
Stage 2 – Network propagation:
Stage 3 – Data exfiltration:
The phrase ‘Most Likely Targets’ describes users who are most likely to be targeted as part of the first stage of a targeted cyber intrusion, and includes:
Understanding the goals of adversaries can provide insight into which other users are likely to be targeted based on their access to sensitive data. Targeting might occur just prior to a significant upcoming meeting or other business event of relevance to adversaries.
Ransomware denies access to data, typically by encrypting it, until a monetary ransom is paid within a specified time period. Ransomware can delete accessible backups, sometimes spreads to other computers, and encrypts all accessible data including data stored on local hard drives, network drives (file shares) and removable storage media such as USB drives. Ransomware can prevent computers from functioning, for example if operating system files or configuration data are encrypted. ‘Lockers’ are related malware that focus on preventing computers from functioning until a ransom is paid.
Some adversaries target specific organisations, for example hospitals are highly motivated to pay the ransom if lives are at risk, and educational institutions typically depend on access to their data. Such compromises might occur by adversaries sending spear phishing emails, by exploiting security vulnerabilities in internet-accessible computers such as websites and associated databases, or by using brute-force passphrase guessing to remotely access computers exposed to the internet via Remote Desktop Protocol (RDP). Additional techniques used by adversaries to motivate victims to pay the ransom include threatening to either delete files or publicly publish sensitive files on the internet.
A limited number of ransomware variants have cryptographic weaknesses or their master decryption key has been disclosed, enabling files to be decrypted in limited cases using free tools.
Paying a ransom has ethical implications and doesn’t guarantee that encrypted files will be decrypted. Adversaries might not be honest and trustworthy, the ransomware might not have the technical capability to decrypt data, or the data might be encrypted/deleted by multiple adversaries.
External adversaries who destroy data and prevent computers/networks from functioning, often motivated by political goals, could delete or corrupt data in a variety of ways. This includes deleting or corrupting user data, applications, operating system files, boot firmware accessed via BIOS/UEFI and other firmware, or configuration settings of computers and other network devices which prevent them from booting their operating system or otherwise operating normally.
Malicious insiders motivated by money or in some cases coercion, ideology, ego or excitement, might steal data such as customer details or intellectual property.
Malicious insiders motivated by revenge or disgruntlement due to reasons such as a negative job performance review, a denied promotion or involuntary termination of employment, might destroy data and prevent computers/networks from functioning.
For the purpose of this publication, the definition of the malicious insider threat excludes non-malicious employees who unintentionally and inadvertently facilitate a cyber security incident, for example by interacting with malicious emails sent by external adversaries – in this case the employee is not the threat, rather they are a weakness that the external threat is exploiting.
Some industry commentators suggest that malicious insiders who steal data can be mitigated using the same mitigation strategies, implemented in the same prioritised order, as for a targeted cyber intrusion that compromises an employee’s computer account to access and exfiltrate data. However, there is a difference in mitigating these two threats since malicious insiders who steal data usually already have an account and access to data, so therefore don’t need to use malware to obtain initial access. Also, malicious insiders have the option of using removable storage media such as USB drives to exfiltrate data. This typically isn’t a viable low-risk exfiltration option for a targeted cyber intrusion where adversaries are in a physically distant location such as a foreign country.
The complementary Strategies to Mitigate Cyber Security Incidents publication doesn’t explicitly provide mitigation guidance for the threat of ‘business email compromise’ or threats to industrial control systems. Nevertheless, non-exhaustive guidance is provided for these threats on the following pages to highlight how the existing mitigation strategies are relevant and can be leveraged as a baseline for mitigating these threats.
‘Business email compromise’ involves adversaries using social engineering or targeted cyber intrusion techniques to abuse the trust in the target organisation’s business processes with the typical goal of committing fraud. Examples include conducting unauthorised transfers of money or in some cases obtaining personnel details to commit tax fraud.
Sometimes adversaries compromise a legitimate email account or create an email account with a similar email address, which is then used to interact with the target. Adversaries might change bank account numbers and contact details on invoices so that the adversaries are inadvertently paid.
Adversaries might compromise the email account of the target’s CEO or senior executive, or send ‘spoofed’ emails that appear to come from a CEO or senior executive. These techniques are also referred to as ‘CEO fraud’, ‘senior executive impersonation’ and ‘business email spoofing’.
Mitigation guidance for ‘business email compromise’ includes:
Industrial control systems (ICS) leverage operational technology (OT) environments, which include components such as electronic sensors as well as systems such as networked computing hardware. This equipment is often used to monitor or control industrial equipment typically to support operational reliability and safety functions.
OT environments are special-purpose and are designed to be in production for decades. This extended asset lifecycle, characterised by infrequent upgrades and replacements, extends the period of time that OT assets are vulnerable to cyber threats and creates additional complexity over time with respect to applying mitigation strategies. Security solutions need to support the high reliability and availability requirements of OT environments and the infrequent opportunities for scheduled outages.
OT environments are distinct from the information technology (IT) environments common to many organisations, which are used for general purpose functions (e.g. email, writing documents and web browsing) and are designed to be in production typically for one to three years before being refreshed, with regular opportunities for scheduled outages.
Prioritise the protection of OT assets (including supporting computers) which are critical to the organisation’s ability to deliver essential services.
Mitigation guidance for OT environments includes:
Mitigation guidance for IT environments includes implementing the mitigation strategies listed in the Strategies to Mitigate Cyber Security Incidents for both targeted cyber intrusions as well as for ransomware and external adversaries with destructive intent, especially focusing on the computers that administer OT environments, develop software for OT environments, or otherwise can interact with OT environments.
The concept of allowing only approved applications or network communications is a key theme of the mitigation strategies. In such cases, activities such as application execution or network communication is denied by default and only activity explicitly approved of by system administrators and network administrators to meet business requirements is allowed to occur.
The traditional approach of blocking the limited subset of applications or network communication that is known to be malicious is a very reactive approach that provides limited security.
Vendor products increasingly advertise alternative approaches to determine whether applications, network communication, computer behaviour or associated logs exhibit indications of malicious activity. Example alternatives include leveraging threat intelligence consisting of more than just indicators of compromise, big data analytics, heuristics, machine learning, artificial intelligence and maths/statistics. Several of these alternative approaches assume that normal behaviour of users and computers can be accurately baselined to identify anomalies while avoiding false positives. Organisations need to critically assess the value of such approaches before purchasing such vendor products, noting that the value is likely to vary depending on each vendor’s implementation.
The effectiveness of network-based mitigation strategies continues to decrease due to evolutions in the architecture of IT infrastructure. For example, the network perimeter continues to be eroded due to the increasing use of external computer infrastructure such as cloud computing services as well as mobile computing devices used by employees. Also, it is increasingly infeasible to backhaul or otherwise steer network traffic to a single bottleneck location to implement network-based mitigation strategies such as ‘Network-based intrusion detection/prevention system’ and ‘Capture network traffic’. These evolutions also impact the ability to implement the mitigation strategy ‘Deny corporate computers direct internet connectivity’.
Paying for cyber insurance isn’t a substitute for investing in cyber security protection by implementing these mitigation strategies, although cyber insurance might encourage organisations to implement these mitigation strategies to reduce the cost of their cyber insurance premium.
Even if a cyber security incident is covered by the cyber insurance policy, an insurance payout might not be able to repair damage such as stolen intellectual property and the associated loss of long term competitive advantages, damage to the organisation’s reputation, lost customer loyalty, and citizens dealing with the consequences of their private data being compromised and used for malicious purposes such as identity theft and fraud.
Application control to prevent execution of unapproved/malicious programs including .exe, DLL, scripts (e.g. Windows Script Host, PowerShell and HTML Applications) and installers.
An appropriately configured implementation of application control helps to prevent the undesired execution of software regardless of whether the software was downloaded from a website, clicked on as an email attachment or introduced via CD/DVD/USB removable storage media.
Implementing application control on important servers such as Active Directory, email servers, and other servers handling user authentication can help prevent adversaries from running malware that obtains passphrase hashes or otherwise provides adversaries with additional privileges.
The following examples are not application control:
The ability of application control to provide a reasonable barrier for low to moderately sophisticated cyber security incidents depends on the solution chosen to implement application control, combined with its configuration settings, as well as the file permissions controlling which directories a user (and therefore malware) can write to and execute from.
Ensure that application control prevents unapproved programs running regardless of their file extension.
A very basic implementation to mitigate some unsophisticated malware from running involves using application control or filesystem permissions to block execution from user profile directories. Such directories include %AppData%, %LocalAppData%, their subdirectories, as well as %TEMP%. Additionally, to prevent malicious scripts from running when clicked on by users, the notepad program can be associated with script file extensions such as .hta, .js, .jse, .vbs, .vbe, .wsf and .ps1.
Organisations that don’t require the use of Windows Script Host are strongly advised to disable it, while other organisations should use application control to allow only approved scripts to run.
After performing testing to confirm that there is no significant business impact, deny typical low-privileged users the ability to run all script execution engines shipped with Microsoft Windows including Windows Script Host (cscript.exe and wscript.exe which run JScript and VBScript including Windows Script Files), powershell.exe, powershell_ise.exe, cmd.exe, wmic.exe and where possible Microsoft HTML Application Host (mshta.exe).
The ZMCIRT urges organisations to exercise caution when using publisher certificate rules to allow operating system files and other applications to execute. There is a security risk of inadvertently allowing applications that are digitally signed by the same publisher which can be used for legitimate purposes or malicious purposes such as network propagation and running malicious programs. To help mitigate this security risk, ensure that publisher certificate rules specify the ‘Product Name’ in addition to the ‘Publisher Name’.
Where possible, prevent users (and therefore malware running on the user’s behalf) from running system executables commonly used for malicious purposes as listed in mitigation strategy ‘Continuous incident detection and response’. Note the exception for regsvr32.exe and rundll32.exe – these are required for legitimate functionality but can be abused to circumvent application control, which can be mitigated by configuring rules in Microsoft’s Enhanced Mitigation Experience Toolkit (EMET).
It is advisable to deploy application control in phases, instead of trying to deploy it to an entire organisation at once. For example, after fully testing and understanding application control to avoid false positives, one approach is to deploy application control to the computers used by senior executives and their executive assistants. Such users are Most Likely Targets who usually run a limited number of software applications such as Microsoft Office, an email program and a web browser. An additional benefit is that, when these users are made aware that they clicked on a malicious email attachment or visited a malicious website and application control mitigated the compromise, they might provide additional support for the deployment of application control to more computers in the organisation.
Deploying application control is easier if the organisation has detailed visibility of what software is installed on computers. Such visibility can be obtained by using a Standard Operating Environment, maintaining an inventory of software installed and implementing a robust change management process. Initially testing application control in ‘audit’/’logging only’ mode helps organisations to develop an inventory of installed software, while taking care to avoid including existing malware in the inventory. Once an inventory has been established, application control can be properly configured in ‘enforce’ mode to prevent unapproved programs from running.
When installing new software, avoid creating hashes for added files that aren’t of an executable nature. Otherwise if every new file is hashed, the list of hashes is likely to become too large and if distributed via Group Policy, might unacceptably slow down users logging into their computers. Additionally, note that installing new software can create subdirectories in allowed paths that provide users (and therefore malware) with write and execute permissions, enabling arbitrary unapproved or malicious programs to run. Organisations need to verify the effectiveness of application control periodically and especially after installing new software.
Installers, or installation packages, can install, modify or remove programs. Common installer frameworks include Windows Installer and InstallShield. Installers often contain installation information as well as files to be installed all within one package. Windows Installer package files have an MSI/MSP filename extension and are commonly used to perform installation or modification of programs in Microsoft Windows environments.
Endpoint protection or anti-malware software from some vendors includes application control functionality. The ZMCIRT has witnessed application control conflict with anti-malware software from a different vendor that launched itself with a random filename in an attempt to hide from malware.
Windows Defender Application Control, introduced in Microsoft Windows 10 and Microsoft Windows Server 2016, is application control that uses virtualisation to help protect itself from being disabled either by malicious administrators or by malware that runs with administrative privileges which has already circumvented application control (somewhat negating the malware’s need to disable application control).
Further guidance, including applicability for operating systems other than Microsoft Windows, is available in the Implementing Application Control and Hardening Linux Workstations and Servers publications.
Information about Windows Defender Application Control and recommended block rules is available from Microsoft.
Patch applications especially Adobe Flash, web browsers and web browser plug-ins/add-ons/extensions, Microsoft Office, Java and PDF viewers. Also patch server applications such as databases that store important (sensitive or high-availability) data as well as web server software that is internet-accessible.
Patch or mitigate computers exposed to ‘extreme risk’ security vulnerabilities within 48 hours of the security vulnerability being identified. The ZMCIRT has developed guidance to facilitate a risk management approach to applying patches based on the severity and potential business impact of the associated security vulnerabilities.
Use the latest version of applications since they typically incorporate additional security technologies such as sandboxing and other anti-exploitation capabilities. For some vendor applications, upgrading to the latest version is the only way to patch a security vulnerability. Don’t use application versions that are no longer vendor-supported with patches for security vulnerabilities.
‘Extreme risk’ security vulnerabilities in software used by the organisation can enable adversaries to execute malicious code, which can result in significant consequences for the organisation. The level of security risk might also be affected by whether exploit code for a security vulnerability is available commercially or publicly, for example in an open source tool like the Metasploit Framework or in a cybercrime exploit kit.
There are a variety of approaches to deploying patches to applications and operating systems running on user computers, based on the organisation’s risk tolerance, as well as how many applications the organisation uses where the applications are legacy, unsupported, developed in-house or poorly designed.
A different approach involving more thorough testing is usually used for deploying patches to servers, as well as for deploying upgrades that introduce significant additional features and capabilities.
To obtain visibility of what software requires patching, maintain an inventory of software installed on every computer, especially laptops that might only occasionally connect to the organisation’s network, and include details about software versions and patching history.
Prioritise patching security vulnerabilities in software used to interact with content from the internet, as well as software which runs with elevated privileges such as anti-malware software and third party video drivers.
Use an automated mechanism to confirm and record that deployed patches have been installed, applied successfully and remain in place.
Don’t use software which is no longer vendor-supported with patches for security vulnerabilities. This is especially important for software that interacts with untrusted and potentially malicious data.
Don’t use Adobe Reader prior to version X, or unsupported Internet Explorer versions (currently version 10 and older) especially when accessing the internet.
Further guidance is available in the Assessing Security Vulnerabilities and Applying Patches publication.
Guidance for improving patch management practices is available from Microsoft.
Configure Microsoft Office macro settings to block macros from the internet, and only allow vetted macros either in ‘trusted locations’ with limited write access or digitally signed with a trusted certificate.
This mitigation strategy addresses adversaries using Microsoft Office macros in an attempt to run malicious code while evading basic email content filtering and application control.
When configuring the new security feature added to Microsoft Office to block macros from the internet, also configure the Microsoft Windows Attachment Manager to prevent users from removing zone information to circumvent this security feature.
For organisations with a business requirement to run Microsoft Office macros, configure Microsoft Office on a per-user and per-application basis to only run macros vetted as trustworthy and preferably placed in ‘trusted location’ directories which typical low-privileged users can’t write to, or less preferably digitally signed by trusted publishers. Note that adversaries might attempt to purchase or steal a code signing certificate issued by a trusted certificate authority, and use it to sign a malicious macro – even if the certificate is associated with an untrusted publisher, the user might undesirably be provided with the decision and ability to run the macro.
Enforce the macro security configuration settings via Group Policy to prevent users from changing them to run a malicious or otherwise unapproved macro.
Detailed guidance on implementing this mitigation strategy is available in the Microsoft Office Macro Security publication.
Further information about the new security feature in Microsoft Office to block macros from the internet is available from Microsoft.
User application hardening. Configure web browsers to block Flash (ideally uninstall it if possible), advertisements and untrusted Java code on the internet. Disable unneeded features in Microsoft Office (e.g. OLE), web browsers and PDF viewers.
This mitigation strategy significantly helps to reduce the attack surface of user computers. It also helps to mitigate adversaries using malicious content in an attempt to evade application control by either exploiting an application’s legitimate functionality, or exploiting a security vulnerability for which a vendor patch is unavailable.
Focus on hardening the configuration of applications used to interact with content from the internet. For web browsers, block Adobe Flash (ideally uninstall it), ActiveX, Java, Silverlight and QuickTime for Windows. Only allow trustworthy websites that require such web browser functionality for a specific business purpose, such as a legacy Flash application used on the organisation’s intranet. Note that some web browsers have an embedded version of Flash.
Ideally uninstall Flash, since simply disabling Flash in the web browser doesn’t mitigate all exploitation vectors such as via Microsoft Office or PDF viewers. Furthermore, web browser ‘click-to-play’ functionality provides limited mitigation since it relies on users to make correct security decisions. Some users might choose incorrectly, for example enabling a malicious Flash advertisement located on a legitimate website.
Block internet advertisements using web browser software (and web content filtering in the gateway), due to the prevalent threat of adversaries using malicious advertising (malvertising) to compromise the integrity of legitimate websites to compromise visitors to such websites. Some organisations might choose to support selected websites that rely on advertising for revenue by enabling just their ads and potentially risking compromise.
A variety of approaches can be used to mitigate running malicious Java code located on the internet, including:
Blocking JavaScript, except for approved websites, is ideal though challenging due to the large number of websites that require such functionality for legitimate purposes, and is difficult to implement in a large scale deployment.
Configure Microsoft Office to disable activation of object linking and embedding (OLE) packages.
Configure the Microsoft Office File Validation and Protected View features to inspect and validate Microsoft Office files for potentially malicious abnormalities.
Detailed guidance on configuring the Microsoft Office File Validation and Protected View features is available in the Hardening Microsoft 365, Office 2021, Office 2019 and Office 2016 publication.
Automated dynamic analysis of email and web content run in a sandbox, blocked if suspicious behaviour is identified (e.g. network traffic, new or modified files, or other system configuration changes).
Dynamic analysis uses behaviour-based detection capabilities instead of relying on the use of signatures, enabling the organisation to detect malware that has yet to be identified by the cyber security community.
Analysis could be performed in an instrumented sandbox located either in the organisation’s gateway, on a user’s computer, or in an external cloud computing environment subject to concerns about data sensitivity, privacy, and security of the communications channel.
Preferably use a vendor product that:
Use an implementation that is regularly updated by the vendor to mitigate evolving evasion techniques that challenge the effectiveness of this mitigation strategy. Avoid using implementations that are easily circumvented by adversaries using evasion techniques such as:
Email content filtering. Allow only approved attachment types (including in archives and nested archives). Analyse/sanitise hyperlinks, PDF and Microsoft Office attachments. Quarantine Microsoft Office macros.
Email content filtering helps to prevent the compromise of user computers via adversaries using malicious emails. Allowing only approved business-related attachment types is significantly more effective than attempting to identify and block a complete list of malicious file types and file extensions, including those increasingly leveraged by adversaries such as .lnk shortcut files, PowerShell and JScript files.
Block/quarantine content that can’t be inspected such as passphrase-protected archive files (e.g. zip or RAR). Inspect archive files in a controlled manner to avoid denial of service via resource exhaustion.
Reject incoming emails that have the organisation’s domain as the email sender but do not originate from email servers approved by the organisation.
One approach to sanitising approved business-related attachment types is to use ‘Content Disarm and Reconstruction’ software, which replaces an email attachment with a new file containing the same content but without potentially malicious code.
Preferably archive PDF and Microsoft Office attachments, and scan them again for malware every month for several months.
Preferably quarantine attachments and disable hyperlinks in emails from webmail providers that provide free email addresses to anonymous internet users, since adversaries often use such email addresses due to the lack of attribution.
Further guidance on malicious email mitigation strategies is available in the Malicious Email Mitigation Strategies publication.
Web content filtering. Allow only approved types of web content and websites with good reputation ratings. Block access to malicious domains and IP addresses, ads, anonymity networks and free domains.
An effective web content filter reduces the security risk of malware being accessed, as well as making it more difficult for adversaries to communicate with their malware. Defining a list of approved types of web content will assist in removing one of the most common malware delivery techniques.
Preferably block all executable content by default and use a process to enable selected users to access specific executable content if a business justification exists.
Preferably block access to websites that the web content filter considers to be ‘uncategorised’ or in a category that is not required for business purposes.
Ideally block Flash, ActiveX and Java, except for approved websites that require such functionality for legitimate purposes. However, the administrative resources required to analyse legitimate business requirements in larger organisations could be significant.
Implement a solution that inspects HTTPS traffic for malicious content, especially HTTPS communications with unfamiliar websites, noting that encrypted network traffic has become pervasive.
If the web content filter has the capability to inspect Microsoft Office files, quarantine such files if they contain macros, especially if they are downloaded from the internet rather than from the organisation’s intranet.
Block internet advertisements using web content filtering in the gateway (and web browser software), due to the prevalent threat of adversaries using malicious advertising (malvertising) to compromise the integrity of legitimate websites to compromise visitors to such websites. Some organisations might choose to support selected websites that rely on advertising for revenue by enabling just their ads and potentially risking compromise.
Block outbound network connections to anonymity networks such as Tor, Tor2web and I2P, to help mitigate malware that uses such networks for command and control as well as for data exfiltration. Some organisations might choose to support inbound network connections from anonymity networks to the organisation’s public internet-accessible websites, to cater to website visitors who wish to remain anonymous for privacy reasons.
Cyber security incidents often involve the use of ‘dynamic’ domains and other domains provided free to anonymous internet users, due to the lack of attribution. Block access to such domains after confirming that the organisation does not access any legitimate websites using these domains.
Where possible, block attempts to access websites by their IP address instead of by their domain name, to force adversaries to obtain a domain name which can contribute to an audit trail that can assist with identifying related cyber security incidents.
The effectiveness of this mitigation strategy is reduced by adversaries using legitimate websites, which are required for business purposes, for malware delivery, command and control, and exfiltration. Such websites include web forums, social networking websites, cloud computing services, legitimate but temporarily compromised websites and a range of other web infrastructure.
Deny corporate computers direct internet connectivity. Use a gateway firewall to require use of a split DNS server, an email server and an authenticated web proxy server for outbound web connections.
A gateway firewall limits external adversaries from accessing corporate computers running vulnerable network services, and serves as a logging and choke point for incoming and outgoing network traffic.
Malware of lower sophistication might fail to exfiltrate data and operate correctly if it expects direct internet connectivity and is unable to traverse the organisation’s internet gateway, resulting in the internet gateway detecting and blocking such unauthorised network communication.
The firewall should be configured to only allow approved networking ports and protocols required for business functionality, and should be capable of handling IPv6 traffic.
Implement a web proxy that decrypts and inspects encrypted HTTPS traffic for malicious content, especially HTTPS communications with unfamiliar websites.
Preferably configure computers with a non-routing network capture device as the default route to help detect malware attempting to directly communicate with the internet, noting that some legitimate applications or operating system functionality might generate false positives.
Servers should have a very restricted ability, and ideally no ability, to browse websites and access emails from the internet.
This mitigation strategy should not be interpreted that internet users visiting the organisation’s public internet-accessible websites need to be authenticated by a web proxy.
Operating system generic exploit mitigation e.g. Data Execution Prevention (DEP), Address Space Layout Randomisation (ASLR) and Enhanced Mitigation Experience Toolkit (EMET).
Security-Enhanced Linux (SELinux) and grsecurity are examples of exploit mitigation mechanisms for Linux operating systems.
These technologies provide system-wide measures to help mitigate techniques used to exploit security vulnerabilities, including for applications which EMET is specifically configured to protect, even in cases where the existence and details of security vulnerabilities are not publicly known.
Configure DEP hardware and software mechanisms to apply to all operating system programs and other software applications that support DEP.
Configure ASLR for all operating system programs and other software applications that support ASLR.
In addition to configuring system-wide EMET rules, configure EMET rules for applications that interact with potentially untrusted content, for example web browsers, Microsoft Office and PDF viewers.
Configure EMET rules to mitigate the legitimate Microsoft Windows operating system files regsvr32.exe and rundll32.exe being abused to circumvent application control.
Use a 64-bit version of Microsoft Windows instead of a 32-bit version, since the 64-bit version contains additional security technologies.
Microsoft note that their Microsoft Windows 10 operating system and Edge web browser natively implement many of EMET’s features and mitigations, making EMET less relevant for Microsoft Windows 10. EMET is most useful to help protect previous operating system versions, legacy applications and third party software.
Server application hardening especially internet-accessible web applications (sanitise input and use TLS not SSL) and databases, as well as other server applications that access important (sensitive or high-availability) data (e.g. customer, finance, human resources and other data storage systems).
Server application hardening helps the organisation to conduct its business with a reduced security risk of malicious data access, theft, exposure, corruption and loss.
OWASP guidance helps to mitigate web application security vulnerabilities such as SQL injection, and covers code review, data validation and sanitisation, user and session management, protection of data in transit and storage, error handling, user authentication, logging and auditing.
The ZMCIRT has developed guidance for securing content management systems running on web servers, as part of the ZMCIRT responding to cyber security incidents involving adversaries compromising internet-accessible web servers and using ‘web shells’ which can facilitate remote access, administration and pivoting to the organisation’s internal systems.
Further guidance on protecting web applications is available in the Protecting Web Applications and Users publication.
Further guidance on securing content management systems is available in the Securing Content Management Systems publication.
Operating system hardening (including for network devices) based on a Standard Operating Environment (SOE), disabling unneeded functionality (e.g. RDP, AutoRun, LanMan, SMB/NetBIOS, Link-Local Multicast Name Resolution (LLMNR) and Web Proxy Auto-Discovery (WPAD)).
Benefits of computers and network devices having a consistent managed SOE configuration include:
Harden file and Windows Registry permissions, for example where possible, prevent users (and therefore malware running on the user’s behalf) from running system executables commonly used for malicious purposes as listed in mitigation strategies ‘Application control’ and ‘Continuous incident detection and response’.
Configure the Windows Task Scheduler service to prevent user computers from creating scheduled tasks (especially on servers) to execute malicious programs.
Configure the DLL search path algorithm to help mitigate malicious DLL files being loaded via DLL search order hijacking techniques.
Disable Server Message Block (SMB) and NetBIOS services running on computers where possible, especially to help mitigate internal reconnaissance and network propagation.
Disabling LLMNR and associated name resolution services such as NetBIOS Name Service where possible, helps to mitigate adversaries on the organisation’s network from responding to name queries performed by the organisation’s other computers and collecting their authentication credentials.
Organisations should create a WPAD DNS record in their internal DNS server and/or in the ‘hosts’ file of user computers. Organisations that don’t use Proxy Auto-Configuration should disable this feature in web browsers.
Configuring file extensions to be displayed assists users to understand a file’s type, otherwise an email attachment called ‘file.txt.exe’ could appear as ‘file.txt’ making the user think it is a harmless text file.
The scarcity of unused and available publicly routable IPv4 address results in an increasing need for IPv6 to be used by computers that directly connect to the internet. However, IPv6 might not be needed by computers on an organisation’s internal network which use IPv4 addresses in the reserved range.
Antivirus software using heuristics and reputation ratings to check a file’s prevalence and digital signature prior to execution. Use antivirus software from different vendors for gateways versus computers.
Specifically, this includes checking the prevalence of a questionable file among the vendor’s user base, and ideally also checking whether a digitally signed file uses a reputable vendor certificate that hasn’t been revoked and wasn’t expired when the digital signature was added to the file.
Antivirus software helps to detect malware that includes computer viruses, worms, Trojans, spyware and adware.
Configure the heuristic behaviour analysis capability to achieve an acceptable balance between identifying malware, while avoiding negatively impacting users and the organisation’s incident response team due to false positives.
Scan files when they are accessed and on a scheduled basis.
Endpoint protection or anti-malware software from some vendors includes heuristics and reputation rating functionality.
Control removable storage media and connected devices. Block unapproved CD/DVD/USB storage media. Block connectivity with unapproved smartphones, tablets and Bluetooth/Wi-Fi/3G/4G/5G devices.
Using removable storage media and connected devices in a controlled and accountable manner reduces the security risk of malware execution and unauthorised data exposure.
USB flash storage devices infected with malware might be deliberately provided to targeted users as a gift, and have been inadvertently distributed by major vendors at several Zambian cyber security conferences. Additionally, adversaries might scatter USB flash storage devices, CDs and DVDs containing malicious content in the car park of targeted users.
Follow a robust storage media transfer policy and process when using removable storage media to transfer data between computers, especially if they are located on different networks or in different security domains. Ideally, an alternative corporately approved method of data transfer should be established which avoids the need to use removable storage media.
Computers without a need to use removable storage media or connected devices can be configured to help prevent such connectivity by removing associated drivers from the operating system, using third party solutions to allow and block access to specific classes of devices, configuring computer BIOS/UEFI settings to disable access to associated hardware, and physically removing or disabling associated hardware used for external data storage or external device connectivity.
Block spoofed emails. Use Sender Policy Framework (SPF) or Sender ID to check incoming emails. Use ‘hard fail’ SPF TXT and DMARC DNS records to mitigate emails that spoof the organisation’s domain.
SPF, or alternative implementations such as Sender ID, reduce the likelihood of spoofed emails being delivered to the targeted user.
Configure ‘hard fail’ SPF TXT DNS records for the organisation’s domains and subdomains, and configure a wildcard SPF TXT DNS record to match non-existent subdomains.
Sender ID is an alternative version of SPF that checks the legitimacy of the sender’s email address that is displayed to the email recipient. Additional implementations include DomainKeys Identified Mail (DKIM).
Domain-based Message Authentication, Reporting and Conformance (DMARC) enables a domain owner to specify a policy stating what action the recipient’s email server should take if it receives an email that has failed an SPF check and/or a DKIM check. DMARC also contains a reporting feature which enables a domain owner to obtain some visibility of whether their domain is being spoofed in emails sent by adversaries.
Configure a DMARC DNS record for the organisation’s domain, specifying that emails from the organisation’s domain and subdomains should be rejected if they fail SPF checks (and/or DKIM checks if DKIM is configured for the organisation’s domain). In the absence of a DMARC DNS record, the ZMCIRT responded to a cyber security incident involving a major free webmail provider that delivered a spoofed email to the recipient’s inbox even though the email failed SPF checks.
Organisations can conservatively deploy DMARC if they are concerned about legitimate emails sent from their domain being incorrectly rejected.
Reject incoming emails that have the organisation’s domain as the email sender but do not originate from email servers approved by the organisation.
Further guidance on spoofed email mitigation strategies is available in the How to Combat Fake Emails publication.
User education. Avoid phishing emails (e.g. with links to login to fake websites), weak passphrases, passphrase reuse, as well as corporately unapproved removable storage media, connected devices and external IT services such as cloud computing including webmail.
Educate users, especially Most Likely Targets, about internet threats such as identifying spear phishing emails or unexpected duplicate emails, and reporting such emails to the organisation’s IT security team. Users should also report potential cyber security incidents, including suspicious phone calls such as unidentified callers attempting to solicit details about the organisation’s IT environment. Finally, users should avoid using weak passphrases, reusing passphrases, using unapproved removable storage media and connected devices, and exposing their email addresses for example via social networking.
User education should focus on influencing user behaviour.
User education can complement technical mitigation strategies. Users can notice and report unexpected behaviour such as a suspicious email, or a blank document or irrelevant document content being displayed when an email attachment is opened. This can assist in detecting spear phishing emails as an intrusion vector. However, to prevent and automatically detect an attempted compromise, implementing a technical mitigation strategy (such as application control configured to log and report violations) is preferable to relying on user education.
Putting users in the position of making a security-related decision and hoping that they are all educated to always choose correctly, is likely to result in some users choosing incorrectly resulting in a compromise.
The ZMCIRT is aware of some spear phishing emails that use clever tradecraft and are believable such that no amount of user education would have helped to prevent or detect a compromise.
User education won’t prevent a user from visiting a legitimate website that has been temporarily compromised to serve malicious content as part of a ‘drive by download’, ‘watering hole’ or ‘strategic web compromise’, including where malvertising runs malicious software without requiring user interaction. Visiting such a website might compromise the user’s computer without any obvious indications of compromise for the user to detect.
Educate users to avoid:
Educate users as to why following cyber security policies helps them to protect and appropriately handle the sensitive data they have been entrusted to handle. Share with users the anecdotal details of previous cyber security incidents affecting the organisation and similar organisations, highlighting the impact that such incidents have to the organisation and to the user. Such education might reduce the level of user resistance to the implementation of mitigation strategies. For example, users might be less likely to resist the removal of their unnecessary administrative privileges if they understand why the mitigation strategy is required.
User education needs to be tailored to the job role of the user. Additional specialised education is useful for users with specific roles, for example:
The success of educating users needs to be measured using evidence such as whether user education contributed to:
Further guidance for users on detecting socially engineered emails is available in the Detecting Socially Engineered Messages publication.
Antivirus software with up-to-date signatures to identify malware, from a vendor that rapidly adds signatures for new malware. Use antivirus software from different vendors for gateways versus computers.
Antivirus software helps to detect malware that includes computer viruses, worms, Trojans, spyware and adware. However, signature-based antivirus software is a reactive approach that has difficulty protecting against targeted malware that is not yet known to the antivirus vendor.
Scan files when they are accessed and on a scheduled basis.
TLS encryption between email servers to help prevent legitimate emails being intercepted and subsequently leveraged for social engineering. Perform content scanning after email traffic is decrypted.
Enabling TLS encryption on both the originating and accepting email servers helps to prevent legitimate emails being intercepted in transit and subsequently being leveraged for social engineering.
Perform content scanning after email traffic is decrypted.
Restrict administrative privileges to operating systems and applications based on user duties. Validate the requirement for users to be granted administrative privileges, and revalidate this requirement at least annually and preferably monthly.
Privileged users should use a separate unprivileged account, and preferably a separate physical computer, for activities that are non-administrative or risky such as reading email, web browsing and obtaining files via internet services such as instant messaging or social networking – technical controls should be implemented to block all privileged user accounts from performing such activities.
The consequences of a compromise are reduced if users (and therefore malware running on the user’s behalf) have low privileges instead of administrative privileges.
This mitigation strategy applies to:
Further guidance is available in the Restricting Administrative Privileges publication.
Patch operating systems. Patch or mitigate computers (including network devices) exposed to ‘extreme risk’ security vulnerabilities within 48 hours of the security vulnerability being identified. The ZMCIRT has developed guidance to facilitate a risk management approach to applying patches based on the severity and potential business impact of the associated security vulnerabilities.
Use the latest version of operating systems since they typically incorporate additional security technologies such as anti-exploitation capabilities. Don’t use operating system versions that are no longer vendor-supported with patches for security vulnerabilities.
‘Extreme risk’ security vulnerabilities in operating systems used by the organisation can enable adversaries to perform actions such as elevating their privileges, which can result in significant consequences for the organisation. The level of security risk might also be affected by whether exploit code for a security vulnerability is available commercially or publicly, for example in an open source tool like the Metasploit Framework or in a cybercrime exploit kit.
Refer to the implementation guidance provided for mitigation strategy ‘Patch applications’.
Apply firmware patches, including for network devices such as routers, switches and firewalls, and especially for those devices that are internet-accessible.
Use a 64-bit version of Microsoft Windows instead of a 32-bit version, since the 64-bit version contains additional security technologies.
Further guidance is available in the Assessing Security Vulnerabilities and Applying Patches publication.
Microsoft’s guidance for improving patch management practices is available in the If you do only one thing to reduce your cybersecurity risk… blog post.
Multi-factor authentication especially for Most Likely Targets, VPNs, RDP, SSH and other remote access capabilities, and for all users when they perform a privileged action (including system administration) or access an important (sensitive or high-availability) data repository.
Multi-factor authentication involves users verifying their identity by using at least any two of the following three mechanisms:
If implemented correctly, multi-factor authentication can make it significantly more difficult for adversaries to use stolen user credentials to facilitate further malicious activities against the organisation, including establishing their own VPN or other remote access connection to the organisation’s network.
Different multi-factor authentication mechanisms provide varying levels of security. Examples include:
Servers that store user authentication data and perform user authentication are frequently targeted by adversaries, therefore additional effort needs to be invested to secure such servers.
The use of multi-factor authentication for remote access does not fully mitigate users entering their passphrase on a compromised computing device. Adversaries who have obtained a user’s passphrase could gain physical access to a corporate computer and simply log in as the user. Mitigations for this include using multi-factor authentication for all user logins including corporate computers in the office, or ensuring that user passphrases for remote access are different to passphrases used for corporate computers in the office. Furthermore, adversaries could use a stolen passphrase to access the user’s network drives once any other user who has access to the organisation’s corporate network has been remotely compromised.
Ensure that administrative service accounts, and other accounts that are unable to use multi-factor authentication, use a strong passphrase.
Multi-step authentication using a single factor is not multi-factor authentication, for example, a user accessing the organisation’s remote access VPN by authenticating using just a single factor, and then accessing the organisation’s internal email or other internal server application by authenticating using just a single factor, even if the first factor is different to the second factor (e.g. two different passphrases).
Further guidance on multi-factor authentication is available in the Implementing Multi-factor Authentication publication.
Disable local administrator accounts or assign passphrases that are random and unique for each computer’s local administrator account to prevent adversaries from easily propagating throughout the organisation’s network using compromised local administrator credentials that are shared by several computers.
Disabling local administrator accounts or assigning random unique passphrases helps to prevent adversaries from propagating throughout the organisation’s network.
In cases where it is not feasible to disable the local administrator account on servers such as the Active Directory authentication server, ensure that the local administrator account has a strong passphrase. Appropriately protect records of the passphrases used for such servers.
Microsoft developed a free tool called ‘Local Administrator Password Solution’ (LAPS) to periodically change the passphrase of the local administrator account on every Microsoft Windows computer in the domain to a random value.
Further information about the LAPS tool is available from Microsoft.
Network segmentation. Deny traffic between computers unless required. Constrain devices with low assurance (e.g. ‘Bring Your Own Device’ (BYOD) and ‘Internet of Things’ (IoT)). Restrict user access to network drives and data repositories based on user duties.
Network segmentation helps to prevent adversaries from propagating throughout the organisation’s network. If implemented correctly, it can make it significantly more difficult for adversaries to locate and gain access to the organisation’s important (sensitive or high-availability) data.
Restrict access based on the connectivity required, user job role, business function, trust boundaries and the extent to which data is important.
Develop and enforce a ruleset controlling which computers are allowed to communicate with other computers. For example, on most corporate networks, direct network communication between user computers should not be required or allowed.
Network controls that can assist with restricting network access include switches, virtual LANs, enclaves, data diodes, firewalls, routers and Network Access Control.
Permissions on files and network drives (file shares) can be used to limit access to data.
Constrain VPN and other remote access, wireless connections, IoT devices, as well as user-owned laptops, smartphones and tablets which are part of a BYOD implementation.
Organisations using operating system virtualisation, (especially third party) cloud computing infrastructure, or providing users with BYOD or remote access to the organisation’s network, might require controls that are less dependent on the physical architecture of the network. Such controls include ‘micro-segmentation’ firewalling implemented by the virtualisation platform layer, software-based firewalling implemented in individual computers and virtual machines, and ‘IPsec Server and Domain Isolation’.
The use of IPsec authentication can ensure that a specific network port or ports on a sensitive server can only be accessed by specific computers such as those computers belonging to administrators.
Important servers such as Active Directory and other authentication servers should only be able to be administered from a limited number of intermediary servers referred to as ‘jump servers’, ‘jump hosts’ or ‘jump boxes’. Jump servers should be closely monitored, be well secured, limit which users and network devices are able to connect to them, and typically have no internet access. Some jump servers might require limited internet access if they are used to administer defined computers located outside of the organisation’s local network.
Organisations with critically important data might choose to store and access it using air-gapped computers that are not accessible from the internet. Security patches and other data can be transferred to and from such air gapped computers in accordance with a robust media transfer policy and processes.
Adversaries could propagate throughout the network by leveraging the organisation’s existing systems used to distribute software such as patches for security vulnerabilities, login programs or scheduled tasks configured via Group Policy Objects, updated anti-malware detection engine software, or the computer Standard Operating Environment master image. Alternatively, adversaries could turn the organisation’s intranet website into a watering hole to compromise users when they visit. Therefore, protect software distribution systems from modifications which are malicious or otherwise unauthorised, combined with implementing a robust change management process.
Copyright @2023 ZAMBIA CIRT