Detect Email Addresses in Elastic Security
Adversaries may gather email addresses that can be used during targeting. Even if internal instances exist, organizations may have public-facing email infrastructure and addresses for employees. Adversaries may gather email addresses from publicly accessible sources such as social media, company websites, and leaked credential databases. Additionally, adversaries may actively enumerate valid email addresses by probing authentication services — for example, querying the Microsoft GetCredentialType API endpoint or Exchange Autodiscover to determine whether a given address is a valid account in Office 365 or on-premises Exchange environments. Gathered email addresses enable spearphishing campaigns, credential brute force attacks, business email compromise, and social engineering operations.
MITRE ATT&CK
- Tactic
- Reconnaissance
- Technique
- T1589 Gather Victim Identity Information
- Sub-technique
- T1589.002 Email Addresses
- Canonical reference
- https://attack.mitre.org/techniques/T1589/002/
Elastic Detection Query
sequence by source.ip with maxspan=1h
[authentication where event.dataset == "azure.signinlogs"
and azure.signinlogs.properties.status.error_code in (50034, 50053, 50055, 50056)
and source.ip != null
and azure.signinlogs.properties.user_principal_name != null] with runs=10Detects email address enumeration via Azure AD sign-in failure error code analysis in Elastic Security. Uses EQL sequence with maxspan=1h and runs=10 to identify source IPs generating 10 or more authentication failures matching error codes 50034 (UserAccountNotFound), 50053 (AccountLocked), 50055 (InvalidPassword), and 50056 (PasswordRequired) within a rolling hour window. For full distinct-username counting, deploy as an Elastic SIEM Threshold rule with unique_count on azure.signinlogs.properties.user_principal_name grouped by source.ip with a threshold of 10.
Data Sources
Required Tables
False Positives & Tuning
- Legitimate red team or penetration testing exercises targeting Azure AD authentication from authorized IP ranges that have not been allowlisted in detection thresholds
- Misconfigured enterprise SSO connectors or ADFS proxies retrying authentication with stale user lists following a directory migration or UPN domain rename
- Automated help desk tooling or IT provisioning scripts that iterate over large user lists to verify account status for bulk onboarding or offboarding operations
Other platforms for T1589.002
Testing Methodology
Validate this detection against 5 adversary techniques from Atomic Red Team. Each test below lists the behaviour to exercise and the telemetry you should expect to see. Executable commands and cleanup steps are available with Pro.
- Test 1Azure AD Email Enumeration via GetCredentialType API (PowerShell)
Expected signal: Sysmon Event ID 1: PowerShell.exe process with CommandLine containing 'GetCredentialType' and 'login.microsoftonline.com'. Sysmon Event ID 3: Outbound network connections from powershell.exe to login.microsoftonline.com (TCP 443). Proxy/web filter logs: HTTP POST to login.microsoftonline.com/common/GetCredentialType with PowerShell User-Agent. DeviceNetworkEvents in MDE: InitiatingProcessFileName=powershell.exe, RemoteUrl containing GetCredentialType.
- Test 2Azure AD Email Enumeration via AADInternals Module
Expected signal: Sysmon Event ID 1: powershell.exe with CommandLine containing 'AADInternals' and 'UserEnumerationAsOutsider'. Sysmon Event ID 3: Network connections to login.microsoftonline.com and related Microsoft AAD endpoints. PowerShell ScriptBlock Log Event ID 4104 capturing the full module invocation. Windows Security Event ID 4688: Process creation for powershell.exe with command line auditing enabled. Proxy logs: HTTP requests with AADInternals User-Agent string to Microsoft authentication endpoints.
- Test 3Email Enumeration via Sign-in Attempt with Non-Existent Accounts (Generates SigninLogs 50034)
Expected signal: Azure AD SigninLogs: 10 entries with ResultType=50034 (UserAccountNotFound) for each probed address, all from the same source IP. These entries are visible in the Azure Portal under Azure AD > Sign-in logs and in Microsoft Sentinel's SigninLogs table within 5-15 minutes. Field ClientAppUsed will show 'Other clients'.
- Test 4OSINT Email Harvesting Simulation via Python Requests (LinkedIn/Web Scraping Pattern)
Expected signal: Sysmon Event ID 1: python3.exe process creation with CommandLine containing 'theHarvester' User-Agent string. Sysmon Event ID 3: Network connections from python3.exe to target hosts over TCP 443/80. DeviceNetworkEvents in MDE: InitiatingProcessFileName=python3.exe, RemoteUrl matching target sites. Proxy logs: HTTP GET requests with theHarvester User-Agent string — this User-Agent string is well-known and often blocked/alerted by web proxies.
- Test 5Exchange Autodiscover Email Validation Probe (curl)
Expected signal: Sysmon Event ID 1: curl.exe process creation with CommandLine containing 'autodiscover' and target email addresses. Sysmon Event ID 3: Outbound network connections from curl.exe to autodiscover.targetdomain.com over TCP 443. Exchange IIS Access Logs (W3SVC): POST requests to /autodiscover/autodiscover.xml with 401 response codes, source IP, and User-Agent 'curl/*'. Windows Security Event ID 4625 (Logon Failure): if Basic Authentication is attempted against Exchange.
References (9)
- https://attack.mitre.org/techniques/T1589/002/
- https://o365blog.com/post/just-looking/
- https://github.com/gremwell/o365enum
- https://grimhacker.com/2017/07/24/office365-activesync-username-enumeration/
- https://www.hackers-arise.com/email-scraping-and-maltego
- https://www.cnet.com/news/massive-breach-leaks-773-million-emails-21-million-passwords/
- https://learn.microsoft.com/en-us/azure/active-directory/reports-monitoring/concept-sign-ins
- https://github.com/dafthack/MSOLSpray
- https://github.com/0xZDH/o365spray
Unlock Pro Content
Get the full detection package for T1589.002 including response playbook, investigation guide, and atomic red team tests.