Detect Browser Fingerprint in Microsoft Sentinel
Adversaries may attempt to blend in with legitimate traffic by spoofing browser and system attributes like operating system, system language, platform, user-agent string, resolution, time zone, etc. The HTTP User-Agent request header is a string that lets servers and network peers identify the application, operating system, vendor, and/or version of the requesting user agent. Adversaries may gather this information through System Information Discovery or by users navigating to adversary-controlled websites, and then use that information to craft their web traffic to evade defenses. This technique was documented in the FatDuke malware, which mimics a compromised user's traffic by using the same user agent as the installed browser.
MITRE ATT&CK
- Tactic
- Defense Evasion
- Technique
- T1036 Masquerading
- Sub-technique
- T1036.012 Browser Fingerprint
- Canonical reference
- https://attack.mitre.org/techniques/T1036/012/
KQL Detection Query
let KnownMalwareUA = dynamic([
"Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; Trident/6.0)",
"Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1)",
"Mozilla/5.0 (Windows NT 6.1; WOW64; rv:40.0) Gecko/20100101 Firefox/40.1"
]);
DeviceNetworkEvents
| where Timestamp > ago(24h)
| where RemoteIPType == "Public"
| where InitiatingProcessFileName !in~ ("chrome.exe", "msedge.exe", "firefox.exe", "iexplore.exe", "opera.exe", "brave.exe", "vivaldi.exe", "MicrosoftEdgeUpdate.exe", "msedgewebview2.exe")
| where isnotempty(InitiatingProcessCommandLine)
| extend HasUserAgent = InitiatingProcessCommandLine has "User-Agent" or InitiatingProcessCommandLine has "useragent" or InitiatingProcessCommandLine has "-H \"Mozilla"
| where HasUserAgent
| project Timestamp, DeviceName, InitiatingProcessFileName, InitiatingProcessCommandLine,
RemoteIP, RemotePort, RemoteUrl, AccountName
| sort by Timestamp desc Detects non-browser processes making HTTP requests with explicit User-Agent strings, which may indicate C2 malware or tools spoofing browser fingerprints to blend with legitimate traffic. Excludes known browser executables and focuses on processes like PowerShell, cmd, rundll32, or custom binaries that explicitly set User-Agent headers in their command lines. Also maintains a list of known malware User-Agent strings for direct matching.
Data Sources
Required Tables
False Positives & Tuning
- Legitimate software updaters that use HTTP with browser-like User-Agent strings (e.g., Windows Update, Adobe updaters, application auto-update mechanisms)
- System administration tools like curl, wget, Invoke-WebRequest used in legitimate scripts that set custom User-Agent strings for API compatibility
- Monitoring and health-check agents that use HTTP requests with User-Agent strings to verify web service availability
- Development and testing tools (Postman, Selenium, Playwright) that set User-Agent headers as part of web application testing workflows
Other platforms for T1036.012
Testing Methodology
Validate this detection against 3 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 1HTTP Request with Spoofed Browser User-Agent via curl
Expected signal: Sysmon Event ID 1: Process Create for curl.exe with '-A' and the full User-Agent string in CommandLine. Sysmon Event ID 3: Network Connection from curl.exe to httpbin.org IP on port 80. DeviceNetworkEvents with InitiatingProcessFileName=curl.exe connecting to external IP. Proxy logs show HTTP request with Chrome User-Agent from a non-browser process.
- Test 2PowerShell HTTP Request with Spoofed User-Agent
Expected signal: Sysmon Event ID 1: Process Create for powershell.exe with 'User-Agent' and 'MSIE 10.0' in CommandLine. Sysmon Event ID 3: Network Connection from powershell.exe to httpbin.org. PowerShell ScriptBlock Log Event ID 4104 with the full script content. Proxy logs show IE10 User-Agent from PowerShell process.
- Test 3Python HTTP Request Mimicking Browser Fingerprint
Expected signal: Auditd EXECVE record for python3 with the HTTP request script. Network connection from python3 to httpbin.org on port 80. Proxy logs show full browser-like header set (User-Agent, Accept, Accept-Language, Accept-Encoding) originating from a Python process rather than a browser.
References (7)
- https://attack.mitre.org/techniques/T1036/012/
- https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Headers/User-Agent
- https://arxiv.org/pdf/2110.10129
- https://www.welivesecurity.com/wp-content/uploads/2019/10/ESET_Operation_Ghost_Dukes.pdf
- https://engineering.salesforce.com/tls-fingerprinting-with-ja3-and-ja3s-247362855967/
- https://learn.microsoft.com/en-us/defender-endpoint/advanced-hunting-devicenetworkevents-table
- https://github.com/SigmaHQ/sigma/tree/master/rules/network
Unlock Pro Content
Get the full detection package for T1036.012 including response playbook, investigation guide, and atomic red team tests.