O365 Admin Consent Bypassed by Service Principal

SPLUNK detection rule by Mauricio Velazco, Splunk · Apache-2.0

The following analytic identifies instances where a service principal in Office 365 Azure Active Directory assigns app roles without standard admin consent. It leverages `o365_management_activity` logs, specifically focusing on the 'Add app role assignment to service principal' operation. This activity is significant for SOCs as it may indicate a bypass of critical administrative controls, potentially leading to unauthorized access or privilege escalation. If confirmed malicious, this could allow an attacker to misuse automated processes to assign sensitive permissions, compromising the security of the environment.

name: O365 Admin Consent Bypassed by Service Principal
id: 8a1b22eb-50ce-4e26-a691-97ff52349569
version: 9
creation_date: '2024-02-14'
modification_date: '2026-05-13'
author: Mauricio Velazco, Splunk
status: production
type: TTP
description: The following analytic identifies instances where a service principal in Office 365 Azure Active Directory assigns app roles without standard admin consent. It leverages `o365_management_activity` logs, specifically focusing on the 'Add app role assignment to service principal' operation. This activity is significant for SOCs as it may indicate a bypass of critical administrative controls, potentially leading to unauthorized access or privilege escalation. If confirmed malicious, this could allow an attacker to misuse automated processes to assign sensitive permissions, compromising the security of the environment.
data_source:
    - O365 Add app role assignment to service principal.
search: "`o365_management_activity` Workload=AzureActiveDirectory Operation=\"Add app role assignment to service principal.\" | eval len=mvcount('Actor{}.ID') | eval userType = mvindex('Actor{}.ID',len-1) | eval roleId = mvindex('ModifiedProperties{}.NewValue', 0) | eval roleValue = mvindex('ModifiedProperties{}.NewValue', 1) | eval roleDescription = mvindex('ModifiedProperties{}.NewValue', 2) | eval dest_user = mvindex('Target{}.ID', 0) | search userType = \"ServicePrincipal\" | eval src_user = user | fillnull | stats count min(_time) as firstTime max(_time) as lastTime by signature dest user src vendor_account vendor_product dest_user roleId roleValue roleDescription | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)` | `o365_admin_consent_bypassed_by_service_principal_filter`"
how_to_implement: You must install the Splunk Microsoft Office 365 Add-on and ingest Office 365 management activity events.
known_false_positives: Service Principals are sometimes configured to legitimately bypass the consent process for purposes of automation. Filter as needed.
references:
    - https://attack.mitre.org/techniques/T1098/003/
    - https://msrc.microsoft.com/blog/2024/01/microsoft-actions-following-attack-by-nation-state-actor-midnight-blizzard/
    - https://www.microsoft.com/en-us/security/blog/2024/01/25/midnight-blizzard-guidance-for-responders-on-nation-state-attack/
    - https://attack.mitre.org/techniques/T1098/002/
    - https://www.microsoft.com/en-us/security/blog/2024/01/25/midnight-blizzard-guidance-for-responders-on-nation-state-attack/
    - https://winsmarts.com/how-to-grant-admin-consent-to-an-api-programmatically-e32f4a100e9d
drilldown_searches:
    - name: View the detection results for - "$dest_user$"
      search: '%original_detection_search% | search  dest_user = "$dest_user$"'
      earliest_offset: $info_min_time$
      latest_offset: $info_max_time$
    - name: View risk events for the last 7 days for - "$dest_user$"
      search: '| from datamodel Risk.All_Risk | search normalized_risk_object IN ("$dest_user$") | stats count min(_time) as firstTime max(_time) as lastTime values(search_name) as "Search Name" values(risk_message) as "Risk Message" values(analyticstories) as "Analytic Stories" values(annotations._all) as "Annotations" values(annotations.mitre_attack.mitre_tactic) as "ATT&CK Tactics" by normalized_risk_object | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)`'
      earliest_offset: 7d
      latest_offset: "0"
finding:
    title: Service principal $user$ bypassed the admin consent process and granted permissions to $dest_user$
    entity:
        field: dest_user
        type: user
        score: 50
analytic_story:
    - Office 365 Persistence Mechanisms
asset_type: O365 Tenant
mitre_attack_id:
    - T1098.003
product:
    - Splunk Enterprise
    - Splunk Enterprise Security
    - Splunk Cloud
category: cloud
security_domain: identity
tests:
    - name: True Positive Test
      attack_data:
        - data: https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1098.003/o365_bypass_admin_consent/o365_bypass_admin_consent.log
          source: o365
          sourcetype: o365:management:activity
      test_type: unit

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

46735599-74bc-4410-94ec-06bee8d11970

Propose edits, view related rules, and engage with the community through comments.

Executable code, not a declarative rule

This submission is arbitrary [[ currentRule.format ? currentRule.format.toUpperCase() : '' ]] code, not a declarative detection rule — review it as you would any script before running it, and note that it cannot be safely aggregated or auto-executed alongside YARA/Sigma/Suricata/CRS rules.

Rejected pattern present

[[ reason ]]

What does this mean?

Format & documentation suggestions

Not a security issue — this rule is valid and was accepted as-is. These are optional fields/conventions its own format recommends, mainly affecting its quality score and how well it documents itself for other reviewers.

[[ reason ]]

What do these messages mean?

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

[[ currentRule.title ]]

Community Discussion