Tutorial: Reconnaissance playbook
The second tutorial in this four part series for Azure ATP security alerts is a reconnaissance playbook. The purpose of the Azure ATP security alert lab is to illustrate Azure ATP's capabilities in identifying and detecting suspicious activities and potential attacks against your network. The playbook explains how to test against some of Azure ATP's discrete detections, and focuses on Azure ATP’s signature-based capabilities. This playbook doesn't include alerts or detections based on advanced machine-learning, or user/entity based behavioral detections, as they require a learning period with real network traffic for up to 30 days. For more information about each tutorial in this series, see the ATP security alert lab overview.
This playbook illustrates the threat detections and security alerts services of Azure ATP for simulated attacks from common, real-world, publicly available hacking and attack tools.
In this tutorial you will:
- Simulate network mapping reconnaissance
- Simulate Directory Service reconnaissance
- Simulate user and IP address (SMB) reconnaissance
- Review the security alerts from the simulated reconnaissance in Azure ATP
- We recommend following the lab setup instructions as closely as possible. The closer your lab is to the suggested lab setup, the easier it will be to follow the Azure ATP testing procedures.
Simulate a Reconnaissance attack
Once an attacker gains presence in your environment, their reconnaissance campaign begins. At this phase, the attacker will typically spend time researching. They try to discover computers of interest, enumerate users and groups, gather important IPs, and map your organization's assets and weaknesses. Reconnaissance activities allow attackers to gain a thorough understanding and complete mapping of your environment for later use.
Reconnaissance attack testing methods:
- Network-mapping reconnaissance
- Directory Service reconnaissance
- User and IP Address (SMB) reconnaissance
Network-mapping reconnaissance (DNS)
One of the first things an attacker will attempt is to try to get a copy of all DNS information. When successful, the attacker gains extensive information about your environment that potentially includes similar information about your other environments or networks.
Azure ATP suppresses network-mapping reconnaissance activity from your suspicious activity timeline until an eight day learning period is completed. In the learning period, Azure ATP learns what is normal and abnormal for your network. After the eight day learning period, abnormal network-mapping reconnaissance events invoke the related security alert.
Run nslookup from VictimPC
To test DNS reconnaissance, we'll use the native command-line tool, nslookup, to initiate a DNS zone transfer. DNS servers with correct configuration will refuse queries of this type and won't allow the zone transfer attempt.
Sign into VictimPC, using the compromised JeffL credentials. Run the following command:
Type server then the FQDN or IP address of the DC where the ATP sensor is installed.
Let's try to transfer the domain.
ls -d contoso.azure
- Replace contosodc.contoso.azure and contoso.azure with the FQDN of your Azure ATP sensor and domain name respectively.
If ContsoDC is your first deployed sensor, wait 15 minutes to allow the database backend to finish deploying the necessary micro services.
Network-mapping reconnaissance (DNS) Detected in Azure ATP
Getting visibility of this type of attempt (failed or successful) is vital for domain threat protection. After recently installing the environment, you'll need to go to the Logical Activities timeline to see the detected activity.
In the Azure ATP Search, type VictimPC, then click on it to view the timeline.
Look for the "AXFR query" activity. Azure ATP detects this type of reconnaissance against your DNS.
- If you have a large number of activities, click Filter by and deselect all types except DNS query.
If your security analyst determined this activity originated from a security scanner, the specific device can be excluded from further detection alerts. In the top-right area of the alert, click on the three dots. Then, select Close and exclude MySecurityScanner. Ensuring this alert doesn't show up again when detected from "MySecurityScanner".
Detecting failures can be as insightful as detecting successful attacks against an environment. The Azure ATP portal allows us to see the exact result of the actions done by a possible attacker. In our simulated DNS reconnaissance attack story, we, acting as attackers, were stopped from dumping the DNS records of the domain. Your SecOps team became aware of our attack attempt and which machine we used in our attempt from the Azure ATP security alert.
Directory Service Reconnaissance
Acting as an attacker, the next reconnaissance goal is an attempt to enumerate all users and groups in the Forest. Azure ATP suppresses Directory Service enumeration activity from your Suspicious Activity timeline until a 30 day learning period is completed. In the learning period, Azure ATP learns what is normal and abnormal for your network. After the 30 day learning period, abnormal Directory Service enumeration events invoke a security alert. During the 30 day learning period, you can see Azure ATP detections of these activities using the activity timeline of an entity in your network. The Azure ATP detections of these activities are shown in this lab.
To demonstrate a common Directory Service reconnaissance method, we'll use the native Microsoft binary, net. After our attempt, examining the Activity timeline of JeffL, our compromised user, will show Azure ATP detecting this activity.
Directory Service Enumeration via net from VictimPC
Any authenticated user or computer can potentially enumerate other users and groups in a domain. This enumeration ability is required for most applications to function properly. Our compromised user, JeffL, is an unprivileged domain account. In this simulated attack, we'll see exactly how even an unprivileged domain account can still provide valuable data points to an attacker.
From VictimPC, execute the following command:
net user /domain
The output shows all users in the Contoso.Azure domain.
Let's try to enumerate all groups in the domain. Execute the following command:
net group /domain
The output shows all groups in the Contoso.Azure domain. Notice the one Security Group that isn't a default group: Helpdesk.
Now, let's try to enumerate only the Domain Admins group. Execute the following command:
net group "Domain Admins" /domain
Acting as an attacker, we've learned there are two members of the Domain Admins group: SamiraA and ContosoAdmin (built-in Administrator for the Domain Controller). Knowing no security boundary exists between our Domain and Forest, our next leap is to try to enumerate the Enterprise Admins.
To attempt to enumerate the Enterprise Admins, execute the following command:
net group "Enterprise Admins" /domain
We learned that there's only one Enterprise Admin, ContosoAdmin. This information wasn't important since we already knew there isn't a security boundary between our Domain and the Forest.
With the information gathered in our reconnaissance, we now know about the Helpdesk Security Group. Although that information isn't interesting yet. We also know that SamiraA is a member of the Domain Admins group. If we can harvest SamiraA's credential, we can gain access the Domain Controller itself.
Directory Service Enumeration Detected in Azure ATP
If our lab had real live activity for 30 days with Azure ATP installed, the activity we just did as JeffL would potentially be classified as abnormal. Abnormal activity would show up in the Suspicious Activity timeline. However, since we just installed the environment, we'll need to go to the Logical Activities timeline.
In the Azure ATP Search, let's see what JeffL's Logical Activity timeline looks like:
We can see when JeffL signed onto the VictimPC, using the Kerberos protocol. Additionally, we see that JeffL, from VictimPC, enumerated all the users in the domain.
Many activities are logged in the Logical Activity timeline making it a major capability to performing Digital Forensics and Incident Response (DFIR). You can even see activities when the initial detection wasn't from Azure ATP but from Windows Defender ATP, Office 365, and others.
Taking a look at ContosoDC's page, we can also see the computers JeffL logged into.
We can also get Directory Data, including JeffL's Memberships and Access Controls, all from within Azure ATP.
Now, our attention will be shift towards SMB Session Enumeration.
User and IP Address reconnaissance (SMB)
Active Directory’s sysvol folder is one of the, if not the, most important network share in the environment. Every computer and user must be able to access this particular network share to pull down Group Policies. An attacker can get a goldmine of information from enumerating who has active sessions with the sysvol folder.
Our next step is SMB Session Enumeration against the ContosoDC resource. We want to learn who else has sessions with the SMB share, and from what IP.
Use JoeWare’s NetSess.exe from VictimPC
Run JoeWare’s NetSess tool against ContosoDC in context of an authenticated user, in this case, ContosoDC:
We already know that SamiraA is a Domain Admin. This attack gave us SamiraA's IP address as 10.0.24.6. As an attacker, we learned exactly who we need to compromise. We also got the network location where that credential is logged in.
User and IP Address reconnaissance (SMB) Detected in Azure ATP
Now we can see what Azure ATP detected for us:
Not only are we alerted on this activity, we're also alerted on the exposed accounts and their respective IP addresses at that point in time. As the Security Operations Center (SOC), we don't just have the attempt and its status, but also what was sent back to the attacker. This information aids our investigation.
The next phase in the attack kill chain is typically an attempt at lateral movement.
Join the Community
Have more questions, or an interest in discussing Azure ATP and related security with others? Join the Azure ATP Community today!