Detection rules › Splunk

Windows AD Abnormal Object Access Activity

Status
production
Severity
low
Group by
user
Author
Steven Dick
Source
github.com/splunk/security_content

The following analytic identifies a statistically significant increase in access to Active Directory objects, which may indicate attacker enumeration. It leverages Windows Security Event Code 4662 to monitor and analyze access patterns, comparing them against historical averages to detect anomalies. This activity is significant for a SOC because abnormal access to AD objects can be an early indicator of reconnaissance efforts by an attacker. If confirmed malicious, this behavior could lead to unauthorized access, privilege escalation, or further compromise of the Active Directory environment.

Known false positives

  • Service accounts or applications that routinely query Active Directory for information.

MITRE ATT&CK coverage

Telemetry coverage

Rule body

name: Windows AD Abnormal Object Access Activity
id: 71b289db-5f2c-4c43-8256-8bf26ae7324a
version: 10
creation_date: '2023-08-18'
modification_date: '2026-05-13'
author: Steven Dick
status: production
type: Anomaly
description: The following analytic identifies a statistically significant increase in access to Active Directory objects, which may indicate attacker enumeration. It leverages Windows Security Event Code 4662 to monitor and analyze access patterns, comparing them against historical averages to detect anomalies. This activity is significant for a SOC because abnormal access to AD objects can be an early indicator of reconnaissance efforts by an attacker. If confirmed malicious, this behavior could lead to unauthorized access, privilege escalation, or further compromise of the Active Directory environment.
data_source:
    - Windows Event Log Security 4662
search: |-
    `wineventlog_security` EventCode=4662
      | stats min(_time) AS firstTime, max(_time) AS lastTime, dc(ObjectName) AS ObjectName_count, values(ObjectType) AS ObjectType, latest(Computer) AS dest count
        BY SubjectUserName
      | eventstats avg(ObjectName_count) AS average stdev(ObjectName_count) AS standarddev
      | eval limit = round((average+(standarddev*3)),0), user = SubjectUserName
      | where ObjectName_count > limit
      | `security_content_ctime(firstTime)`
      | `security_content_ctime(lastTime)`
      | `windows_ad_abnormal_object_access_activity_filter`
how_to_implement: Enable Audit Directory Service Access via GPO and collect event code 4662. The required SACLs need to be created for the relevant objects. Be aware Splunk filters this event by default on the Windows TA. Recommend pre-filtering any known service accounts that frequently query AD to make detection more accurate. Setting wide search window of 48~72hr may smooth out misfires.
known_false_positives: Service accounts or applications that routinely query Active Directory for information.
references:
    - https://medium.com/securonix-tech-blog/detecting-ldap-enumeration-and-bloodhound-s-sharphound-collector-using-active-directory-decoys-dfc840f2f644
    - https://learn.microsoft.com/en-us/windows/security/threat-protection/auditing/event-4662
    - https://attack.mitre.org/tactics/TA0007/
intermediate_findings:
    entities:
        - field: user
          type: user
          score: 20
          message: The account $user$ accessed an abnormal amount ($ObjectName_count$) of [$ObjectType$] AD object(s) between $firstTime$ and $lastTime$.
analytic_story:
    - Active Directory Discovery
    - BlackSuit Ransomware
asset_type: Endpoint
mitre_attack_id:
    - T1087.002
product:
    - Splunk Enterprise
    - Splunk Enterprise Security
    - Splunk Cloud
category: endpoint
security_domain: endpoint

Stages and Predicates

Stage 1: search

`wineventlog_security` EventCode=4662

Stage 2: stats

| stats min(_time) AS firstTime, max(_time) AS lastTime, dc(ObjectName) AS ObjectName_count, values(ObjectType) AS ObjectType, latest(Computer) AS dest count
    BY SubjectUserName

Stage 3: eventstats

| eventstats avg(ObjectName_count) AS average stdev(ObjectName_count) AS standarddev

Stage 4: eval

| eval limit = round((average+(standarddev*3)),0), user = SubjectUserName

Stage 5: where

| where ObjectName_count > limit

Stage 6: search

| `security_content_ctime(firstTime)`

Stage 7: search

| `security_content_ctime(lastTime)`

Stage 8: search

| `windows_ad_abnormal_object_access_activity_filter`

Indicators

These rows show field, operator, and value matches.

FieldKindValuesSearch
EventCodeeq
  • 4662 corpus 15 (splunk 7, kusto 4, elastic 3, chronicle 1)
field:"EventID" kind:eq value:"4662"
ObjectName_countcross_field_compare
  • limit transforms: op:gt
field:"ObjectName_count" kind:cross_field_compare value:"limit"