Detection rules › YARA-L

O365 Login Activity To Azure AD PowerShell App

Severity
medium
Type
hunt
Time window
5m
Match by
userid
Author
Google Cloud Security
Source
github.com/chronicle/detection-rules

Logins to Azure AD PowerShell app can have legitimate purposes, but are also abused to gain access to user information. Programmatic access to Entra ID (Azure AD) should generally be through apps, so reviewing these activities is needed.

MITRE ATT&CK coverage

TacticTechniques
Initial Access

Telemetry coverage

PlatformRecord / event type
Microsoft 365UserLoggedIn

Rules detecting the same action

These rules filter on the same operation.

Rule body

// Copyright 2025 Google LLC. Licensed under Apache-2.0.

rule o365_login_activity_to_azure_ad_powershell_app {

  meta:
    author = "Google Cloud Security"
    description = "Logins to Azure AD PowerShell app can have legitimate purposes, but are also abused to gain access to user information. Programmatic access to Entra ID (Azure AD) should generally be through apps, so reviewing these activities is needed."
    rule_id = "mr_0cc56eaf-4176-444d-874d-9780bbb29d4f"
    rule_name = "O365 Login Activity To Azure AD PowerShell App"
    tactic = "TA0001"
    technique = "T1078.004"
    type = "hunt"
    platform = "azure"
    data_source = "o365"
    severity = "Medium"
    priority = "Medium"

  events:
    $login.metadata.event_type = "USER_LOGIN"
    $login.metadata.product_event_type = "UserLoggedIn"
    $login.metadata.product_name = "Office 365"
    $login.metadata.vendor_name = "Microsoft"
    $login.target.resource.product_object_id = "1b730954-1685-4b74-9bfd-dac224a7b894"
    $login.security_result.action = "ALLOW"

    // Below filters ADFS Sync Account that runs every 30 minutes between AD and Entra ID (for federated environments); can use display name or email address
    //$login.target.user.userid != "insert your email address here" and
    //$login.target.user.user_display_name != "On-Premises Directory Synchronization Service Account"
    $login.target.user.userid = $userid

  match:
    $userid over 5m

  outcome:
    $risk_score = 65
    $event_count = count_distinct($login.metadata.id)
    $security_summary = array_distinct($login.security_result.summary)
    $user_agent = array_distinct($login.network.http.user_agent)
    $country_region_login_attempt = array_distinct($login.principal.ip_geo_artifact.location.country_or_region)
    //added to populate alert graph with additional context
    $principal_ip = array_distinct($login.principal.ip)
    $target_user_userid = array_distinct($login.target.user.userid)

  condition:
    $login
}

YARA-L rule structure

Condition

Fires when at least one $login event in the 5m window.

Events

$login USER_LOGIN

  • metadata.product_event_type = "UserLoggedIn"
  • target.resource.product_object_id = "1b730954-1685-4b74-9bfd-dac224a7b894"
  • security_result.action = "ALLOW"

Match and correlation

Match key
$userid (target.user.userid)
Within
5m

Outcome

Outcome variables add context to a detection. When the rule generates an alert, $risk_score is displayed with it.

FieldExpression
risk_score65
event_countcount_distinct($login.metadata.id)
security_summaryarray_distinct($login.security_result.summary)
user_agentarray_distinct($login.network.http.user_agent)
country_region_login_attemptarray_distinct($login.principal.ip_geo_artifact.location.country_or_region)
principal_iparray_distinct($login.principal.ip)
target_user_useridarray_distinct($login.target.user.userid)