End-user consent stopped due to risk-based consent

Hunt Hypothesis

Adversaries may attempt to bypass multi-factor authentication by exploiting OAuth consent flows, leveraging risky applications to gain unauthorized access. SOC teams should proactively hunt for this behavior in Azure Sentinel to identify potential credential compromise or lateral movement tactics.

KQL Query

AuditLogs
  | where OperationName has "Consent to application"
  | where Result =~ "failure"
  | extend InitiatingUserPrincipalName = tostring(InitiatedBy.user.userPrincipalName)
  | extend InitiatingAadUserId = tostring(InitiatedBy.user.id)
  | extend InitiatingIPAddress = tostring(InitiatedBy.user.ipAddress)
  | extend userAgent = iif(AdditionalDetails[0].key == "User-Agent", tostring(AdditionalDetails[0].value), tostring(AdditionalDetails[1].value))
  | where isnotempty(TargetResources)
  | extend TargetAppName = tostring(TargetResources[0].displayName)
  | extend TargetAppId = tostring(TargetResources[0].id)
  | mv-expand TargetResources[0].modifiedProperties
  | extend TargetResources_0_modifiedProperties = columnifexists("TargetResources_0_modifiedProperties", '')
  | where isnotempty(TargetResources_0_modifiedProperties)
  | where TargetResources_0_modifiedProperties.displayName =~ "MethodExecutionResult."
  | extend TargetPropertyDisplayName = tostring(TargetResources_0_modifiedProperties.displayName)
  | extend FailureReason = tostring(parse_json(tostring(TargetResources_0_modifiedProperties.newValue)))
  | where FailureReason contains "Risky"
  | extend InitiatingAccountName = tostring(split(InitiatingUserPrincipalName, "@")[0]), InitiatingAccountUPNSuffix = tostring(split(InitiatingUserPrincipalName, "@")[1])
  | project-reorder TimeGenerated, OperationName, Result, TargetAppName, TargetAppId, FailureReason, InitiatingUserPrincipalName, InitiatingAadUserId, InitiatingIPAddress, userAgent

Analytic Rule Definition

id: 009b9bae-23dd-43c4-bcb9-11c4ba7c784a
name: End-user consent stopped due to risk-based consent
description: |
  'Detects a user's consent to an OAuth application being blocked due to it being too risky.
    These events should be investigated to understand why the user attempted to consent to the applicaiton and what other applicaitons they may have consented to.
    Ref: https://docs.microsoft.com/azure/active-directory/fundamentals/security-operations-applications#end-user-stopped-due-to-risk-based-consent'
severity: Medium
requiredDataConnectors:
  - connectorId: AzureActiveDirectory
    dataTypes:
      - AuditLogs
queryFrequency: 1d
queryPeriod: 1d
triggerOperator: gt
triggerThreshold: 0
tactics:
  - Persistence
  - PrivilegeEscalation
relevantTechniques:
  - T1078.004
tags:
  - AADSecOpsGuide
query: |
  AuditLogs
    | where OperationName has "Consent to application"
    | where Result =~ "failure"
    | extend InitiatingUserPrincipalName = tostring(InitiatedBy.user.userPrincipalName)
    | extend InitiatingAadUserId = tostring(InitiatedBy.user.id)
    | extend InitiatingIPAddress = tostring(InitiatedBy.user.ipAddress)
    | extend userAgent = iif(AdditionalDetails[0].key == "User-Agent", tostring(AdditionalDetails[0].value), tostring(AdditionalDetails[1].value))
    | where isnotempty(TargetResources)
    | extend TargetAppName = tostring(TargetResources[0].displayName)
    | extend TargetAppId = tostring(TargetResources[0].id)
    | mv-expand TargetResources[0].modifiedProperties
    | extend TargetResources_0_modifiedProperties = columnifexists("TargetResources_0_modifiedProperties", '')
    | where isnotempty(TargetResources_0_modifiedProperties)
    | where TargetResources_0_modifiedProperties.displayName =~ "MethodExecutionResult."
    | extend TargetPropertyDisplayName = tostring(TargetResources_0_modifiedProperties.displayName)
    | extend FailureReason = tostring(parse_json(tostring(TargetResources_0_modifiedProperties.newValue)))
    | where FailureReason contains "Risky"
    | extend InitiatingAccountName = tostring(split(InitiatingUserPrincipalName, "@")[0]), InitiatingAccountUPNSuffix = tostring(split(InitiatingUserPrincipalName, "@")[1])
    | project-reorder TimeGenerated, OperationName, Result, TargetAppName, TargetAppId, FailureReason, InitiatingUserPrincipalName, InitiatingAadUserId, InitiatingIPAddress, userAgent
entityMappings:
  - entityType: Account
    fieldMappings:
      - identifier: FullName
        columnName: InitiatingUserPrincipalName
      - identifier: Name
        columnName: InitiatingAccountName
      - identifier: UPNSuffix
        columnName: InitiatingAccountUPNSuffix
  - entityType: Account
    fieldMappings:
      - identifier: AadUserId
        columnName: InitiatingAadUserId
  - entityType: IP
    fieldMappings:
      - identifier: Address
        columnName: InitiatingIPAddress
  - entityType: CloudApplication
    fieldMappings:
      - identifier: AppId
        columnName: TargetAppId
      -

Required Data Sources

MITRE ATT&CK Context

• Tactic: Persistence
• Tactic: PrivilegeEscalation
• Technique: T1078.004 — T1078.004

References

• Source Query

False Positive Guidance

• Scenario: A system administrator is testing a new OAuth application in a development environment and accidentally triggers the consent flow.
  Filter/Exclusion: Exclude events where the requesting IP address is from the internal network or matches a known admin IP range (e.g., 10.0.0.0/8).

• Scenario: A scheduled job or automation tool (e.g., Jenkins, Ansible, or Terraform) is attempting to authenticate with an OAuth service to provision resources.
  Filter/Exclusion: Exclude events where the user agent string matches known automation tools (e.g., Jenkins/2.301.1, Ansible, or Terraform).

• Scenario: A legitimate enterprise application (e.g., Microsoft Teams, Salesforce, or ServiceNow) is requesting consent during a routine integration setup.
  Filter/Exclusion: Exclude events where the client ID matches known internal or partner applications (e.g., teams.microsoft.com, salesforce.com, or servicenow.com).

• Scenario: A user is attempting to consent to a third-party application that is part of a known business partnership or vendor integration.
  Filter/Exclusion: Exclude events where the application domain is whitelisted (e.g., partner.example.com or trusted.vendor.net).

• Scenario: A user is testing multi-factor authentication (e.g., using Okta or Azure AD) and the consent flow is triggered as part of the authentication process.
  Filter/Exclusion: Exclude events where the consent request is associated with a known MFA flow (e.g., mfa.microsoft.com, okta.com, or azure.com).