Microsoft Threat Intelligence has identified highly targeted social engineering attacks using credential theft phishing lures sent as Microsoft Teams chats by the threat actor that Microsoft tracks as Midnight Blizzard (previously tracked as NOBELIUM). This latest attack, combined with past activity, further demonstrates Midnight Blizzard’s ongoing execution of their objectives using both new and common techniques. In this latest activity, the threat actor uses previously compromised Microsoft 365 tenants owned by small businesses to create new domains that appear as technical support entities. Using these domains from compromised tenants, Midnight Blizzard leverages Teams messages to send lures that attempt to steal credentials from a targeted organization by engaging a user and eliciting approval of multifactor authentication (MFA) prompts. As with any social engineering lures, we encourage organizations to reinforce security best practices to all users and reinforce that any authentication requests not initiated by the user should be treated as malicious.
Our current investigation indicates this campaign has affected fewer than 40 unique global organizations. The organizations targeted in this activity likely indicate specific espionage objectives by Midnight Blizzard directed at government, non-government organizations (NGOs), IT services, technology, discrete manufacturing, and media sectors. Microsoft has mitigated the actor from using the domains and continues to investigate this activity and work to remediate the impact of the attack. As with any observed nation-state actor activity, Microsoft has directly notified targeted or compromised customers, providing them with important information needed to secure their environments.
Midnight Blizzard (NOBELIUM) is a Russia-based threat actor attributed by the US and UK governments as the Foreign Intelligence Service of the Russian Federation, also known as the SVR. This threat actor is known to primarily target governments, diplomatic entities, non-government organizations (NGOs), and IT service providers primarily in the US and Europe. Their focus is to collect intelligence through longstanding and dedicated espionage of foreign interests that can be traced to early 2018. Their operations often involve compromise of valid accounts and, in some highly targeted cases, advanced techniques to compromise authentication mechanisms within an organization to expand access and evade detection.
Midnight Blizzard is consistent and persistent in their operational targeting, and their objectives rarely change. They utilize diverse initial access methods ranging from stolen credentials to supply chain attacks, exploitation of on-premises environments to laterally move to the cloud, exploitation of service providers’ trust chain to gain access to downstream customers, as well as the Active Directory Federation Service (AD FS) malware known as FOGGYWEB and MAGICWEB. Midnight Blizzard (NOBELIUM) is tracked by partner security vendors as APT29, UNC2452, and Cozy Bear.
Midnight Blizzard regularly utilizes token theft techniques for initial access into targeted environments, in addition to authentication spear-phishing, password spray, brute force, and other credential attacks. The attack pattern observed in malicious activity since at least late May 2023 has been identified as a subset of broader credential attack campaigns that we attribute to Midnight Blizzard.
Use of security-themed domain names in lures
To facilitate their attack, the actor uses Microsoft 365 tenants owned by small businesses they have compromised in previous attacks to host and launch their social engineering attack. The actor renames the compromised tenant, adds a new onmicrosoft.com subdomain, then adds a new user associated with that domain from which to send the outbound message to the target tenant. The actor uses security-themed or product name-themed keywords to create a new subdomain and new tenant name to lend legitimacy to the messages. These precursory attacks to compromise legitimate Azure tenants and the use of homoglyph domain names in social engineering lures are part of our ongoing investigation. Microsoft has mitigated the actor from using the domains.
Social engineering attack chain
In this activity, Midnight Blizzard either has obtained valid account credentials for the users they are targeting, or they are targeting users with passwordless authentication configured on their account – both of which require the user to enter a code that is displayed during the authentication flow into the prompt on the Microsoft Authenticator app on their mobile device.
After attempting to authenticate to an account where this form of MFA is required, the actor is presented with a code that the user would need to enter in their authenticator app. The user receives the prompt for code entry on their device. The actor then sends a message to the targeted user over Microsoft Teams eliciting the user to enter the code into the prompt on their device.
Step 1: Teams request to chat
The target user may receive a Microsoft Teams message request from an external user masquerading as a technical support or security team.
Figure 1: Screenshot of a Microsoft Teams message request from a Midnight Blizzard-controlled account
Step 2: Request authentication app action
If the target user accepts the message request, the user then receives a Microsoft Teams message from the attacker attempting to convince them to enter a code into the Microsoft Authenticator app on their mobile device.
Figure 2: A Microsoft Teams prompt with a code and instructions.
Step 3: Successful MFA authentication
If the targeted user accepts the message request and enters the code into the Microsoft Authenticator app, the threat actor is granted a token to authenticate as the targeted user. The actor gains access to the user’s Microsoft 365 account, having completed the authentication flow.
The actor then proceeds to conduct post-compromise activity, which typically involves information theft from the compromised Microsoft 365 tenant. In some cases, the actor attempts to add a device to the organization as a managed device via Microsoft Entra ID (formerly Azure Active Directory), likely an attempt to circumvent conditional access policies configured to restrict access to specific resources to managed devices only.
Recommendations
Microsoft recommends the following mitigations to reduce the risk of this threat.
Educate users about social engineering and credential phishing attacks, including refraining from entering MFA codes sent via any form of unsolicited messages.
Educate Microsoft Teams users to verify ‘External’ tagging on communication attempts from external entities, be cautious about what they share, and , and never share their account information or authorize sign-in requests over chat.
Educate users to review sign-in activity and mark suspicious sign-in attempts as “This wasn’t me”.
Customers hunting for related activity in their environment can identify users that were targeted with the phishing lure using content search in Microsoft Purview. A content search can be created for selected Exchange mailboxes (which include Teams messages) using the following keywords (remove the [] around the “.” before use):
msftprotection.onmicrosoft[.]com
identityVerification.onmicrosoft[.]com
accountsVerification.onmicrosoft[.]com
azuresecuritycenter.onmicrosoft[.]com
teamsprotection.onmicrosoft[.]com
We detected a recent change to your preferred Multi-Factor Authentication (MFA)
The search results will include the messages that match the criteria. The first result will appear to be from <threadid>@unq.gbl.spaces addressed to the target user and the threat actor (i.e., the request to chat as described in Step 1), followed by the message sent by the threat actor, as shown in the Microsoft Purview image below:
Figure 3: Message sent by the threat actor, as shown in Microsoft Purview
Microsoft Sentinel
Microsoft Sentinel customers can use the TI Mapping analytics (a series of analytics all prefixed with “TI map”) to automatically match indicators associated with Midnight Blizzard in Microsoft Defender Threat Intelligence with data in their workspace. If the TI Map analytics are not currently deployed, customers can install the Threat Intelligence solution from the Microsoft Sentinel Content Hub to have the Defender Threat Intelligence connector and analytics rule deployed in their Sentinel workspace. Learn more about the Content Hub.
Microsoft Sentinel also has a range of detection and threat hunting content that customers can use to detect activity related to the activity described in this blog:
Mozilla has released security updates to address vulnerabilities for Firefox 116, Firefox ESR 115.1, and Firefox ESR 102.14. An attacker could exploit some of these vulnerabilities to take control of an affected system.
CISA encourages users and administrators to review Mozilla’s security advisories for Firefox 116, Firefox ESR 115.1, and Firefox ESR 102.14 for more information and apply the necessary updates.
Empower your workforce with a new partner in productivity Read the e-book, Working Smarter with AI, to discover how AI-powered tools like Microsoft 365 Copilot can unlock new levels of productivity.
For years now, AI has played a crucial role behind the scenes, powering everything from online searches to personalized streaming recommendations. Now, we are entering a transformative phase where AI is no longer just an unseen assistant but a powerful partner in enhancing productivity. Download the Working Smarter with AI e-book to: Discover the untapped potential of language models and generative AI and explore innovative ways to leverage their capabilities.Gain valuable insights on how to ensure the secure and responsible use of AI, including how to build an ethical and transparent AI ecosystem.Learn how to maximize the intelligent features of Copilot and enable your workforce to achieve its potential.
Sensitive information on the Wi-Fi connection settings stored in the memories of inkjet printers (home and office/large format) may not be deleted by the usual initialization process.
Affected Products/Versions
Please check the affected inkjet printer models from here.
Mitigation/Remediation
When your printer may be in the hand of any third party, such as when repairing, lending or disposing the printer, take the following steps from the printer unit:
Reset all settings (Reset settings ‐> Reset all)
Enable the wireless LAN
Reset all settings one more time
For models that do not have the Reset all settings function, take the following steps:
Reset LAN settings
Enable the wireless LAN
Reset LAN settings one more time
Please refer to the operation manual of relevant model for specific Reset all or Reset LAN settings.
When investigating an incident, analysts follow certain steps – tasks – to ensure that the investigation is conducted effectively and efficiently. Standardizing the process is necessary for both generic steps and specific types of incidents, and their availability in the context of the incident allows for faster and more efficient management and remediation.
Tasks in Microsoft Sentinel can help security analysts streamline their workflow and improve their efficiency. Analysts can add tasks to specific incidents or alerts, enabling them to track the progress of investigations and remediation activities. While tasks can be added manually from within the Sentinel console, playbooks, and automation rules can be used to automatically create tasks based on certain conditions. Today we’re happy to announce the release of new playbooks, a workbook, and Log Analytics logs as well as an update to the SOC Process Framework. Along with a new integration with Microsoft 365 Defender SecOps playbooks, this will allow even more efficiency in managing incidents and the SOC’s tasks, with out-of-the-box content delivered by our security experts.
New playbooks with tasks for BEC, Ransomware, and Phishing investigation
Handling complex incidents can be a long, challenging task that requires lots of expertise in different fields. Microsoft 365 Defender SecOps workflows offer security analysts a detailed guided response playbook for investigating and responding to security incidents, including Phishing, Ransomware, and BEC. Developed by some of the world’s top security researchers and backed by Microsoft’s extensive experience in threat detection and response, these workflows provide unparalleled security value, significantly bolstering an organization’s defenses.. And now, with dedicated playbooks in Sentinel, these workflows can be easily transformed into tasks. Simply deploy the phishing, BEC, and ransomware playbooks in your workspace and apply them to the corresponding incidents. By integrating Defender workflows with Sentinel tasks, security teams can enjoy a more structured and efficient workflow, enabling them to respond to threats with greater speed and accuracy. The playbooks are now published in the “SOAR essentials” solution in Content Hub and are ready for use by your SOC.
New workbook to manage tasks in the SOC
Now that tasks are integrated into the SOC’s incidents, the new tasks workbook offers a way for security teams to analyze task progress and completion, providing a comprehensive overview of all the tasks that exist in the Sentinel workspace. This is particularly important for SOC managers, as it allows them to easily monitor and manage the security workflow of their team. With a centralized view of all tasks, managers can quickly identify any bottlenecks or areas for improvement, including tasks that take the longest to complete. Additionally, a dashboard that shows tasks per incident/incidents and owner can help managers gain greater visibility into the workload of their team and allows analysts to manage their tasks in a timely and organized manner. The new Workbook allows toggling between those perspectives. The Workbook is now published in the “SOAR essentials” solution in Content hub.
The new workbook is based on new information in Log Analytics’ SecurityIncident table. Please refer to Appendix 2 for documentation of the new tasks fields.
Tasks details in SecurityIncident table
Task details in Log Analytics can be used as dashboards to monitor task progress, investigate security incidents, and track compliance and auditing activities. The tasks details include the task name and status, task number, last completed time (in case the task was modified after it was closed), and more. If you wish to explore the new task details in the SecurityIncident table, please refer to documentation and some recommended queries provided in Appendix 2 at the end of this blog post.
Updates to the SOC Process Framework
The SOC Process Framework solution, which can be found in Content Hub, is also updated to support Tasks. With the new version, instead of writing tasks into incident comments, the SOC Process Framework will create tasks defined in the watchlist into Microsoft Sentinel Incident Tasks. As a reminder, the SOC Process Framework Solution is designed to easily integrate with Microsoft Sentinel and establish a standard SOC Process and Procedure Framework within your organization, including incident or alert tasks.
Summary
Automating task management in Microsoft Sentinel using playbooks and automation rules can help security analysts streamline their workflow and improve efficiency. Integrating Microsoft Defender workflows with Sentinel tasks provides security teams with a more structured and effective way to investigate and respond to security incidents, significantly improving an organization’s security posture. Additionally, Workbooks offer valuable insights into task progress and completion, enabling security teams to monitor their workflow and identify areas for improvement. The new SecurityIncident audits allow for full flexibility in querying tasks details and integrating them into more of the SOC’s tools.
Window notebook and All-in-one come with a Power setting set to balanced, but there is a Hidden way to expose to ChoosetheHigh performance mode powerplan.
After typing the above command, please confirm whether the High performance mode is available and selected in the power plan of Windows operating system, please refer to the following method.
Type and search [Choose a power plan] in the Windows search bar, then click [Open]. Make sure the [High performance] mode is available. Now you can configure addition setting including Process speed. If you have an Intel CPU Turbo boost will be able to reach the highest frequency.
PII compromised in a data breach could be used or sold for identity theft schemes. Dark web marketplaces, forums, and websites offer a spectrum of products and services that monetize stolen personal and financial data, corporate and social media account details, as well as counterfeit documents and money. The dark web also provides an arsenal of malicious tools and malware that, combined with this personal information, can allow threat actors to create official documents or identities to commit identity theft or launch cyberattacks. Threat actors may use compromised PII to launch cyberattacks in social engineering attempts via phishing emails, vishing, SMiShing, compromised websites, and social media scams to compromise accounts, steal funds, steal additional PII or financial information, open new fraudulent accounts, access computer networks and resources, and perform additional cyberattacks.
For example, Charles Schwab Corp., the parent company of TD Ameritrade, Inc., recently disclosed a data breach resulting from exploited vulnerabilities discovered in the MOVEit file transfer software. The customer data stored on TD Ameritrade’s server was compromised and, therefore, impacted customers may be subjected to an increased risk of identity theft and other malicious activity. Other organizations affected by the MOVEit vulnerability to date include various financial institutions, British Airways, Shell PLC, and more. Furthermore, the Cl0p ransomware group published Shell’s data to the dark web after Shell refused to pay the ransom demand; therefore, TD Ameritrade’s data may similarly be at risk. Also, the Schwab/TD Ameritrade merger conversion is ongoing until Labor Day weekend of 2023, which may enable threat actors to target and exploit potential victims in social engineering schemes, such as fraudulent merger updates, security account alerts, and account updates.
CISA analyzed backdoor malware variants obtained from an organization that had been compromised by threat actors exploiting the vulnerability.
Barracuda Exploit Payload and Backdoor – The payload exploits CVE-2023-2868, leading to dropping and execution of a reverse shell backdoor on ESG appliance. The reverse shell establishes communication with the threat actor’s command and control (C2) server, from where it downloads the SEASPY backdoor to the ESG appliance. The actors delivered the payload to the victim via a phishing email with a malicious attachment.
SEASPY – SEASPY is a persistent and passive backdoor that masquerades as a legitimate Barracuda service. SEASPY monitors traffic from the actor’s C2 server. When the right packet sequence is captured, it establishes a Transmission Control Protocol (TCP) reverse shell to the C2 server. The shell allows the threat actors to execute arbitrary commands on the ESG appliance.
SUBMARINE – SUBMARINE is a novel persistent backdoor executed with root privileges that lives in a Structured Query Language (SQL) database on the ESG appliance. SUBMARINE comprises multiple artifacts—including a SQL trigger, shell scripts, and a loaded library for a Linux daemon—that together enable execution with root privileges, persistence, command and control, and cleanup. CISA also analyzed artifacts related to SUBMARINE that contained the contents of the compromised SQL database. This malware poses a severe threat for lateral movement.
For more information, including indicators of compromise and YARA rules for detection, on the exploit payload, SEASPY, and SUBMARINE backdoor, see the following Malware Analysis Reports:
The Securities and Exchange Commission today adopted rules requiring registrants to disclose material cybersecurity incidents they experience and to disclose on an annual basis material information regarding their cybersecurity risk management, strategy, and governance. The Commission also adopted rules requiring foreign private issuers to make comparable disclosures.
“Whether a company loses a factory in a fire — or millions of files in a cybersecurity incident — it may be material to investors,” said SEC Chair Gary Gensler. “Currently, many public companies provide cybersecurity disclosure to investors. I think companies and investors alike, however, would benefit if this disclosure were made in a more consistent, comparable, and decision-useful way. Through helping to ensure that companies disclose material cybersecurity information, today’s rules will benefit investors, companies, and the markets connecting them.”
The new rules will require registrants to disclose on the new Item 1.05 of Form 8-K any cybersecurity incident they determine to be material and to describe the material aspects of the incident’s nature, scope, and timing, as well as its material impact or reasonably likely material impact on the registrant. An Item 1.05 Form 8-K will generally be due four business days after a registrant determines that a cybersecurity incident is material. The disclosure may be delayed if the United States Attorney General determines that immediate disclosure would pose a substantial risk to national security or public safety and notifies the Commission of such determination in writing.
The new rules also add Regulation S-K Item 106, which will require registrants to describe their processes, if any, for assessing, identifying, and managing material risks from cybersecurity threats, as well as the material effects or reasonably likely material effects of risks from cybersecurity threats and previous cybersecurity incidents. Item 106 will also require registrants to describe the board of directors’ oversight of risks from cybersecurity threats and management’s role and expertise in assessing and managing material risks from cybersecurity threats. These disclosures will be required in a registrant’s annual report on Form 10-K.
The rules require comparable disclosures by foreign private issuers on Form 6-K for material cybersecurity incidents and on Form 20-F for cybersecurity risk management, strategy, and governance.
The final rules will become effective 30 days following publication of the adopting release in the Federal Register. The Form 10-K and Form 20-F disclosures will be due beginning with annual reports for fiscal years ending on or after December 15, 2023. The Form 8-K and Form 6-K disclosures will be due beginning the later of 90 days after the date of publication in the Federal Register or December 18, 2023. Smaller reporting companies will have an additional 180 days before they must begin providing the Form 8-K disclosure. With respect to compliance with the structured data requirements, all registrants must tag disclosures required under the final rules in Inline XBRL beginning one year after initial compliance with the related disclosure requirement
In cloud environments, cryptojacking – a type of cyberattack that uses computing power to mine cryptocurrency – takes the form of cloud compute resource abuse, which involves a threat actor compromising legitimate tenants. Cloud compute resource abuse could result in financial loss to targeted organizations due to the compute fees that can be incurred from the abuse. In attacks observed by Microsoft, targeted organizations incurred more than $300,000 in compute fees due to cryptojacking attacks.
While there are fundamental differences in how cloud providers handle authentication, permissions, and resource creation, a cloud cryptojacking attack could unfold in any environment where a threat actor can compromise an identity and create compute, and the attack lifecycle is largely the same. Microsoft security experts have surfaced tell-tale deployment patterns to help defenders determine, identify, and mitigate cloud cryptojacking attacks.
To perform cloud cryptojacking, threat actors must typically have access to compromised credentials obtained through various means, highlighting the need to implement common best practices like credential hygiene and cloud hardening. If the credentials do not have the threat actors’ desired permissions, privilege escalation techniques are used to obtain additional permissions. In some cases, threat actors hijack existing subscriptions to further obfuscate their operations.
Once access to the tenant is gained, threat actors create large amounts of compute, preferring core types that allow them to mine more currency faster. Threat actors use these deployed resources to start mining cryptocurrency by installing cryptomining software in the newly created virtual machines (VMs) and joining them to mining pools.
In this blog post, we present insights from our research on how attackers launch cryptojacking attacks in cloud environments. These insights deepen our understanding of these threats, which in turn inform the protections that we continuously build into our cloud security solutions. We share patterns that administrators and defenders can look out for to identify if a cryptojacking attack is occurring within their cloud environment. We also provide information on how Microsoft Defender for Cloud, Microsoft Defender for Cloud Apps, and other solutions can detect cryptocurrency mining threats and related malicious activity.
While this blog covers mitigation and protections against cloud cryptojacking, in general, strengthening cloud security posture, protecting cloud workloads from threats, and better control of cloud app access can help organizations defend against a wide range of cloud-based threats and risks.
Cryptocurrency mining in cloud environments
In incident response investigations and proactive research in the past year, we observed threat actors abusing administrative features to deploy and manage cryptocurrency mining resources in compromised tenants. Many of these attacks take advantage of automation, which increases the potential threat to cloud environments.
Cryptocurrency mining using central processing unit (CPU) or graphics processing unit (GPU) compute in cloud environments is not financially viable if one is paying for the compute used. In order to profit, threat actors use malicious methods to avoid paying for the resources, such as abusing free trials or compromising legitimate tenants to conduct cryptojacking attacks.
Unlike free trial abuse, which the cloud provider may be able to detect, cryptojacking in compromised tenants is more challenging to identify since it involves the threat actor having access to a legitimate user account. This complex method impacts the user more directly, as it allows the threat actor to make more intrusive changes in the target environment:
Utilize available compute quota from compromised tenants, and provision significantly more compute and other additional resources.
Mask resource provisioning activity as legitimate when operating within a compromised tenant.
Use access to the compromised tenant to do further lateral movement, achieve persistence, and conduct information theft.
Successful cloud cryptojacking attacks could result in significant unexpected charges to the compromised tenant and depletion of resources that the tenant might need for business continuity, potentially resulting in service interruption, highlighting the need to prevent, detect and mitigate cloud cryptojacking attacks.
Attack lifecycle
Cryptojacking requires the threat actor to reach a certain level of access to the cloud environment, which we explain in more detail in the next sections. The diagram below shows the stages of a typical cloud cryptojacking attack.
Figure 1. Diagram of cryptojacking attack on a compromised cloud tenant
In the above example, the attacker generally keeps their operational infrastructure separate from the compromised infrastructure used for mining.
Initial access: Compromised credentials
To perform this attack, the threat actor must have access to credentials that can be used to access the tenant. These credentials need to have the virtual machine contributor role, or provide a path to a user account that does. Threat actors abusing tenants in this way utilize multiple methods to gain account credentials such as phishing, using leaked credentials, and on-premises device compromise. Microsoft Incident Response investigations found that in nearly all cases observed, the accounts did not have multi-factor authentication (MFA) enabled, and no evidence of password spray or brute force was present, suggesting leaked credentials might be the most common vector.
After gaining access, some threat actors use attacker-controlled virtual machines within legitimate tenants as their operational infrastructure. By using living-off-the-land techniques, threat actors can operate without any infrastructure external to the cloud environment. This attack cycle is shown in the diagram below.
Figure 2. Initial access attack cycle
In the above example, the attacker generally keeps their operational infrastructure separate from the compromised infrastructure used for mining.
Privilege escalation: Elevating access
In some observed cases, threat actors compromise the global administrator account. By design, global administrator accounts might not have access to all subscriptions and management groups within the directory; the elevate access option needs to be elevated for the account to have permissions over all resources. Access to global administrator accounts must therefore be adequately secured to prevent threat actors from elevating their access or granting roles that allow the creation of compute resources.
Subscription hijacking is an evasion technique that allows the threat actor to hide some of their activities from the tenant administrator and security teams. Migrating a subscription directory requires the threat actor to have sufficient privileges in the target subscription. In cases observed by Microsoft, the destination tenant may be attacker-controlled or another affected tenant that the threat actor has access to.
Additionally, subscription hijacking is disruptive forensically. Microsoft Incident Response has observed instances where a threat actor compromised accounts in customer environments that were over-privileged. Abusing over-privileged accounts allowed the threat actor to migrate the subscription to a separate tenant (often attacker-controlled) to spin up additional resources. While activity logs at the subscription level remain with the subscription, anything recorded at the tenant role-based access control (RBAC) level is recorded in the new tenant, making forensic analysis, understanding the full timeline, or incident response by or for the customer, more challenging.
Impact: Increasing core quotas
Once a threat actor has access to a tenant, they can either create compute using existing core quota, or they may choose to increase core quotas within the tenant. Increasing core quotas is potentially risky for the actor as quota increases undergo review. Some quotas can’t be immediately adjusted and require a support ticket to increase.
Threat actors without permission to increase quotas use whatever is available. This often leads to them exhausting available core counts across multiple regions. Quota increases have occurred up to a month before resources are deployed by the threat actor.
GPU compute offerings are often targeted by threat actors. GPU compute provides access to high performance NVIDIA and AMD GPU cores, allowing cryptocurrency mining magnitudes more effective than any CPU compute offering. A complete overview of GPU compute types can be found in GPU optimized virtual machine sizes.
The NVIDIA T4, V100, and A100 GPU compute options are most abused by threat actors. At time of writing, the NVIDIA A100 is the best mining card available that is not a dedicated application-specific integrated circuit (ASIC). When comparing NVIDIA GPU performance for cryptomining, the number of Compute Unified Device Architecture (CUDA) cores can be used as a rough representation of the card’s performance. CUDA is designed specifically for high performance parallel computing, which allows more computations to take place at once. For NVIDIA GPUs, more CUDA cores generally means more mining potential. The table below shows the comparative hash rate for the top three most abused GPU compute cards within cloud environments based on mining Ethereum Proof of Work (ETHW).
Azure VM versions
GPU
CUDA cores
ETHW*
NC T4 v3
NVIDIA T4
2,560
25.1MH/s
NCv3
NVIDIA V100
5,120
89.5MH/s
ND A100 v4
NVIDIA A100 (40GB)
6,192
175MH/s
* Mining rates based on the Ethereum Proof of Work complexity in February 2023
As the table above shows, threat actors who can provision NVIDIA GPU cores can mine a meaningful amount of currency in a relatively short period of time. In attacks observed by Microsoft, cryptojacking activities were seen to incur compute fees more than $300,000, illustrating how unprofitable mining is within cloud environments without committing resource theft.
Impact: Deploying compute
There are several ways to deploy compute, and threat actors have adapted to abusing features to speed up deployment. As resource hijacking is an attack of scale, the threat actor needs a way to rapidly spin up and manage multiple devices. In observed cases, threat actors have employed VM scale sets, Azure Machine Learning compute instances, Azure Batch, and Azure Container Instances. Each of these systems allows compute to be deployed quickly and centrally managed.
Malicious provisioning behavior of compute using the above methods generally does not match existing compute provisioning patterns within the tenant. The graph below shows an attacker deploying NVIDIA compute cores within a target environment using VM scale sets. The Y axis shows the capacity of the VM whilst the X axis represents time, this activity spans a three-hour period. Each color represents a single region, with the attacker iterating the various regions to create compute.
Figure 3. Attacker compute deployment pattern
In the graph above, the actor followed a predictable and anomalous deployment pattern across several hijacked subscriptions. Microsoft Threat Intelligence analysis shows that this deployment pattern is unique to a specific threat actor. While this specific pattern may change, the automated nature of malicious compute deployments means that an unusual pattern almost always emerges.
Some staggering of deployment is used, but the threat actor ultimately needs to provision compute very quickly to make the attack profitable. This time restriction means that patterns in provisioning generally emerge over relatively short periods of time. In the above case, the entire provisioning stage of the attack took place over a three-hour period.
In addition to the pattern of deployment, in this case, the following additional anomalies were also observed:
The user accounts used to provision compute had never provisioned compute before.
The compromised user provisioned GPU compute, when no GPU compute had been provisioned in this environment before.
Compute was deployed to regions anomalous for the environment.
Other cases observed by Microsoft showed the following deployment anomalies:
A user with a recent Azure AD anomaly creating large volumes of compute.
A user suddenly causing multiple deployment failures spanning multiple core types due to a core quota unavailability.
Once compute resources are deployed, the actor may need to install GPU drivers to take full advantage of the graphics card, especially on N-series VMs. Actors have been observed abusing Azure Virtual Machine extensions such as an NVIDIA GPU Driver Extension for Windows or Linux, or an AMD GPU Driver Extension for Windows, to facilitate driver installation. These extensions allow for the mass-deployment of drivers, reducing the threat actors’ setup time before mining.
The following anomalies have been observed when actors use these extensions:
Sudden or unusual high-volume provisioning of GPU drivers using a GPU Driver Extension.
A user account suddenly deploying GPU extensions, especially where that user account has no history of deploying VM extensions.
With compute prepared, the threat actor can begin mining cryptocurrency by deploying mining software to the newly created VMs. The installed mining software joins the VM to a mining pool, which allows the threat actor to pool their stolen processing power from multiple compromised tenants.
Data from Microsoft Defender for Cloud shows some of the most recent pools in use by threat actors using already-compromised Azure tenants. Below is the list of the top 10 mining domains observed being used:
nanopool[.]org
nicehash[.]com
supportxmr[.]com
hashvault[.]pro
zpool[.]ca
herominers[.]com
f2pool[.]com
minexmr[.]com
moneroocean[.]stream
miner[.]rocks
Seeing connections to any mining pool from a VM within an environment is a strong indication of compromise. Microsoft Defender for Cloud has multiple detections for this behavior.
Recommendations to identify and mitigate cryptojacking attacks
Security teams should monitor and regularly review alerts specific to these scenarios. In environments where the creation of compute or increases in quota are uncommon, additional alerts should be built to monitor associated operations within your SIEM tool like Microsoft Sentinel. These are highly environmentally specific.
While every situation is unique to the customer and their environment, Microsoft Incident Response has identified several recommendations that are broadly applicable to help identify and mitigate cryptojacking attacks, alongside specific product detections. These recommendations are based on observations from responding to multiple resource abuse engagements.
Separation of privileged roles: Keep administrator and normal user accounts separate. Non-administrator users who require privileged roles in the environment for specific functions should utilize Privileged Identity Management to access the roles on an as-needed basis in a way that can be audited and tracked, or also have separate accounts created. In most resource abuse cases Microsoft Incident Response has investigated, the initially compromised user is over privileged in some way. Thus, it is good practice to limit the number of accounts that have the virtual machine contributor role. In addition, accounts with this role should be protected by MFA and Conditional Access where possible. Also, since a global admin must enable the elevate access option to have permissions over all Azure resources, it should be considered a very sensitive activity that should be monitored and reviewed.
Multifactor authentication: Tenant administrators should ensure that MFA is in use comprehensively across all accounts. This is especially important if the account has virtual machine contributor privileges. Users should also be discouraged from reusing passwords across services. Microsoft Defender for Cloud provides a range of recommendations to secure cloud environments. A full list can be found in Security recommendations – a reference guide.
Risk-based sign-in behaviors and conditional access policies: In cases investigated, attackers who have signed in using compromised credentials have triggered high Azure Active Directory (Azure AD) risk scores. Monitoring risky user alerts and tuning detections that take advantage of this security information help prevent these attacks. In addition to analyzing Azure AD risk scores, correlating risky Azure AD behavior with follow-on activity can help produce additional true positive detections. Risk-based conditional access policies can be designed to require multifactor reauthentication, enforce device compliance, force the user to update their password, or outright block the authentication. In many cases, policies such as these can be disruptive enough to provide security teams with enough time and signal to respond or alert the legitimate user to an issue before the resource abuse begins. Standard login anomaly detections were also found applicable in cases investigated by Microsoft Incident Response, with threat actors commonly using proxy services, signing in from anomalous locations, and accessing accounts using anomalous user agents. One group of activity tracked by Microsoft Threat Intelligence used Python requests and the default user agent (python-requests/2.26.0) for all operations. Microsoft 365 Defender uses detections such as Access elevation by risky user and Risky user performed suspicious Azure activities, which correlate users marked as risky by Azure AD with anomalous actions to raise the severity of alerts in Microsoft 365 Defender. Lastly, authentication to a tenant from an IP that is outside of that tenant should be anomalous. Defenders can identify which IP addresses are allocated within a tenant using the az vm list-ip-addresses command.
Limit unused quotaand monitor for unexpected quota increases: Looking for multiple unexpected quota increases occurring in a short period of time, quota increases across multiple regions, or quota increases within regions that the environment does not normally use might allow for early detection of a resource abuse attack. Quota increases are one of the first signals Microsoft Incident Response looks for when investigating suspected resource abuse attack. Quota increase detections can potentially be refined by looking for increases to commonly abused core types, especially if their usage is otherwise rare in an environment.
Monitor for external Azure IP addresses authenticated with your tenant: Threat actors performing these attacks also use Azure compute resources to conduct their operations. Monitoring for successful sign in activity from Azure IP addresses that are not owned by your tenant is often a strong indicator of suspicious activity. Seeing multiple authentication attempts from Azure IP addresses using the same browser user agent is another strong indicator of potential password guessing.
Detection details
Microsoft 365 Defender
Microsoft 365 Defender uses its cross-workloads detection capabilities to provide enhanced protection against cryptocurrency mining attacks. Microsoft 365 Defender customers who have enabled their Azure connector in Microsoft Defender for Cloud Applications can benefit from the following alerts:
Access elevation by risky user
Suspicious Azure activities related to possible cryptocurrency mining
Mass provisioning of GPU virtual machines for possible cryptocurrency mining
Suspicious creation of multiple Azure ML clusters and workspaces
Suspicious role assignment in Azure subscription
VM quota modified after risky user signed in
Microsoft Defender for Cloud Applications
The following Microsoft Defender for Cloud Application alerts indicate threat activity related to the attack discussed in this post:
Multiple delete VM activities
Multiple VM creation activities
Microsoft Defender for Cloud
Microsoft Defender for Cloud detects threat components associated with the activities outlined in this article with the following alerts:
Azure Resource Manager operation from suspicious proxy IP address
Crypto-mining activity
Digital currency mining activity (Preview)
Fileless attack toolkit detected
Possible Cryptocoinminer download detected
Process associated with digital currency mining detected
Potential crypto coin miner started
Suspicious Azure role assignment detected (Preview)
Suspicious creation of compute resources detected (Preview)
Suspicious installation of a GPU extension was detected in your virtual machine (Preview)
Suspicious invocation of a high-risk ‘Execution’ operation by a service principal detected (Preview)
Suspicious invocation of a high-risk ‘Execution’ operation detected (Preview)
Suspicious invocation of a high-risk ‘Impact’ operation by a service principal detected (Preview)
Suspicious invocation of a high-risk ‘Impact’ operation detected (Preview)
Suspicious subscription transfer to external tenant was detected (Preview)
Microsoft Defender for Endpoint
The following Microsoft Defender for Endpoint alert can indicate associated threat activity:
Possible cryptocurrency miner
Hunting queries
Microsoft Sentinel
Microsoft Sentinel customers can use the TI Mapping analytics (a series of analytics all prefixed with ‘TI map’) to automatically match the malicious domain indicators mentioned in this blog post with data in their workspace. If the TI Map analytics are not currently deployed, customers can install the Threat Intelligence solution from the Microsoft Sentinel Content Hub to have the analytics rule deployed in their Sentinel workspace. More details on the Content Hub can be found here:
In addition, Microsoft Sentinel customers can leverage the following content to hunt for and detect related activity in their environments:
For the latest security research from the Microsoft Threat Intelligence community, check out the Microsoft Threat Intelligence Blog: https://aka.ms/threatintelblog.
To get notified about new publications and to join discussions on social media, follow us on Twitter at https://twitter.com/MsftSecIntel.