1. Resources
  2. Detection rules
  3. Microsoft Entra ID Protection Policy Changed

Microsoft Entra ID Protection Policy Changed detection rules

An attacker with a security administrator role disables or loosens an ID Protection policy (user risk, sign-in risk or MFA registration), so that risky sign-ins are no longer challenged and new accounts stay without MFA.

Platform
Microsoft Entra ID
Severity
Medium
Log source
Microsoft Entra ID audit logs
Last updated
Rule formats
Sigmaand queries for: SplunkMicrosoft SentinelIBM QRadar
What an attacker needs
Microsoft Entra ID P2 in the tenant and a compromised identity with the Security Administrator or Global Administrator role.
Impact
Risky sign-ins and users are no longer challenged, blocked or made to change their password, and new accounts are not made to register for MFA.
Response
Restore the policy, contain the identity that changed it and review every risky sign-in and risky user recorded while the policy was weakened.
Mitigation
Manage risk-based access in Conditional Access with protected actions, and limit the Security Administrator role to a few PIM-managed accounts.
Audit activities
Update Sign-In Risk PolicyUpdate User Risk and MFA Registration PolicyUpdate IdentityProtectionPolicyConfirmAccountSafeConfirmSafeDismissRiskDismissUserDismissServicePrincipal

How the attack works

Microsoft Entra ID Protection rates each sign-in and each user for risk, and policies act on that rating. The sign-in risk policy asks for MFA or blocks a risky sign-in. The user risk policy forces a secure password change or blocks a user who is probably compromised. The MFA registration policy makes users register for MFA when they sign in and gives them 14 days to complete it. Microsoft names Security Administrator as the least privileged role that can edit these policies.

An attacker who reaches that role turns a policy off, raises the risk level it reacts to, or excludes the accounts they use. Nothing looks broken afterwards. Risk is still detected and reported, but nothing acts on it, so a stolen password that ID Protection flags keeps working. With the MFA registration policy off, new accounts are no longer made to register.

Microsoft lists three audit activities for these changes: Update Sign-In Risk Policy and Update User Risk and MFA Registration Policy in the Policy category, and Update IdentityProtectionPolicy in the IdentityProtection category. Microsoft announced the retirement of the legacy user risk and sign-in risk policies of ID Protection for 1 October 2026 and moved risk-based access control to Conditional Access. In a tenant that has migrated, a change to a risk policy is a Conditional Access policy change and is covered by the page Conditional Access Policy Weakened or Deleted. The rule on this page keeps watching the MFA registration policy and any legacy policy that still exists. A second rule reads the administrator actions that clear risk without touching a policy, such as dismissing the risk of a user or confirming a user as safe. An attacker who holds a security role can use them to take a compromised account out of reach of the risk policies.

What an attacker needs

  • Microsoft Entra ID P2 or Microsoft Entra Suite, which the policies need.
  • A compromised identity with the Security Administrator or Global Administrator role.

Impact

Sign-ins and users that ID Protection rates as risky are no longer challenged, blocked or made to change their password, so stolen credentials stay usable for longer. With the MFA registration policy disabled, new and unregistered accounts are not made to register for MFA. The change affects every user the policy covered.

Detection rules

ID Protection Risk or MFA Registration Policy Updated

Detects a change to a Microsoft Entra ID Protection policy: the audit activities "Update Sign-In Risk Policy" and "Update User Risk and MFA Registration Policy" (Policy category) and "Update IdentityProtectionPolicy" (IdentityProtection category). Disabling one of these policies, raising its risk level or excluding accounts stops the automatic response to risky sign-ins and risky users, or stops new accounts from being made to register for MFA. Microsoft announced the retirement of the legacy user risk and sign-in risk policies of ID Protection for 1 October 2026. Risk policies that were recreated in Conditional Access are logged as Conditional Access policy changes and need the rule for those. The event does not say what changed, so compare the old and the new values in TargetResources.

LevelMediumStatusExperimentalLog sourceMicrosoft Entra ID audit logs
title: ID Protection Risk or MFA Registration Policy Updated
id: 35732c4a-f69b-4837-8015-af4149887832
status: experimental
description: >-
  Detects a change to a Microsoft Entra ID Protection policy: the audit
  activities "Update Sign-In Risk Policy" and "Update User Risk and MFA
  Registration Policy" (Policy category) and "Update IdentityProtectionPolicy"
  (IdentityProtection category). Disabling one of these policies, raising its
  risk level or excluding accounts stops the automatic response to risky
  sign-ins and risky users, or stops new accounts from being made to register
  for MFA. Microsoft announced the retirement of the legacy user risk and
  sign-in risk policies of ID Protection for 1 October 2026. Risk policies
  that were recreated in Conditional Access are logged as Conditional Access
  policy changes and need the rule for those. The event does not say what
  changed, so compare the old and the new values in TargetResources.
references:
  - https://learn.microsoft.com/en-us/entra/identity/monitoring-health/reference-audit-activities
  - https://learn.microsoft.com/en-us/entra/id-protection/concept-identity-protection-policies
  - https://learn.microsoft.com/en-us/entra/id-protection/howto-identity-protection-configure-mfa-policy
  - https://attack.mitre.org/techniques/T1556/009/
author: Forestall
date: 2026-11-16
tags:
  - attack.defense-impairment
  - attack.persistence
  - attack.t1556.009
logsource:
  product: azure
  service: auditlogs
detection:
  selection:
    OperationName:
      - Update Sign-In Risk Policy
      - Update User Risk and MFA Registration Policy
      - Update IdentityProtectionPolicy
  filter_failed:
    Result: failure
  condition: selection and not filter_failed
falsepositives:
  - The planned migration of the legacy risk policies to Conditional Access, which disables them in ID Protection. Confirm that equivalent Conditional Access policies are enabled before closing the alert.
  - Rollouts that widen the MFA registration policy to more users, and routine tuning of risk levels by the identity team. Match each change to a change record.
level: medium

False positives

  • The planned migration of the legacy risk policies to Conditional Access, which disables them in ID Protection. Confirm that equivalent Conditional Access policies are enabled before closing the alert.
  • Rollouts that widen the MFA registration policy to more users, and routine tuning of risk levels by the identity team. Match each change to a change record.

ID Protection Risk Dismissed or Confirmed Safe by an Administrator

Detects an administrator clearing risk in Microsoft Entra ID Protection: a user risk is dismissed, a user or a sign-in is confirmed safe, or the risk of a service principal is dismissed. Microsoft describes "Confirm user safe" as removing the risk and the detections on the user, and "Dismiss user risk" as the action for a risk that is real but not malicious. Both are daily work for analysts. In the hands of an attacker who holds a security role they are a way around the risk policies without touching a policy: once the risk of the taken-over account is cleared, a policy that blocks or challenges risky users no longer applies to it. Microsoft lists the activities ConfirmAccountSafe, ConfirmSafe, DismissRisk, DismissUser and DismissServicePrincipal for ID Protection in the audit category Other and does not describe them one by one, so the rule matches the names. The actor is in InitiatedBy.

LevelLowStatusExperimentalLog sourceMicrosoft Entra ID audit logs
title: ID Protection Risk Dismissed or Confirmed Safe by an Administrator
id: 6b0256e9-4d97-4de1-bf79-ceeaca5ae112
status: experimental
description: >-
  Detects an administrator clearing risk in Microsoft Entra ID Protection: a
  user risk is dismissed, a user or a sign-in is confirmed safe, or the risk
  of a service principal is dismissed. Microsoft describes "Confirm user safe"
  as removing the risk and the detections on the user, and "Dismiss user risk"
  as the action for a risk that is real but not malicious. Both are daily work
  for analysts. In the hands of an attacker who holds a security role they are
  a way around the risk policies without touching a policy: once the risk of
  the taken-over account is cleared, a policy that blocks or challenges risky
  users no longer applies to it. Microsoft lists the activities
  ConfirmAccountSafe, ConfirmSafe, DismissRisk, DismissUser and
  DismissServicePrincipal for ID Protection in the audit category Other and
  does not describe them one by one, so the rule matches the names. The actor
  is in InitiatedBy.
references:
  - https://learn.microsoft.com/en-us/entra/identity/monitoring-health/reference-audit-activities
  - https://learn.microsoft.com/en-us/entra/id-protection/howto-identity-protection-risk-feedback
  - https://attack.mitre.org/techniques/T1556/009/
author: Forestall
date: 2026-11-18
tags:
  - attack.defense-impairment
  - attack.persistence
  - attack.t1556.009
logsource:
  product: azure
  service: auditlogs
detection:
  selection:
    OperationName:
      - ConfirmAccountSafe
      - ConfirmSafe
      - DismissRisk
      - DismissUser
      - DismissServicePrincipal
  filter_failed:
    Result: failure
  condition: selection and not filter_failed
falsepositives:
  - Security analysts who close investigated risk detections. This is the expected use, so alert on actors outside the team that handles ID Protection, on an actor who clears a risk on their own account and on many dismissals in a short time.
  - Automation that dismisses risk after a ticket is closed. Exclude its identity once it is known.
level: low

False positives

  • Security analysts who close investigated risk detections. This is the expected use, so alert on actors outside the team that handles ID Protection, on an actor who clears a risk on their own account and on many dismissals in a short time.
  • Automation that dismisses risk after a ticket is closed. Exclude its identity once it is known.

Detection logic

Patterns that need more than one event, a baseline or knowledge of your own environment. Each is a hypothesis to build and tune in your SIEM, not a ready rule.

Risk cleared by an account that is itself flagged, or on its own account

Detection hypothesisJoins two logs

Analysts clear the risk of other people after an investigation. An actor who dismisses the risk on their own account, or who has an open risk detection while clearing the risk of others, is more likely an attacker removing an obstacle than an analyst doing triage.

  1. Audit log: activity DismissUser, DismissRisk, ConfirmSafe or ConfirmAccountSafe. Take the actor from InitiatedBy and the affected account from TargetResources.
  2. Alert when the actor and the affected account are the same.
  3. Risk detection log, 7 days before the action: detections whose UserPrincipalName is the actor. Alert when one exists.

EventsDismissUserDismissRiskConfirmSafeConfirmAccountSafe

LimitsMicrosoft documents the names of these activities and not the layout of their events, so check in a real event where the affected account is recorded before you build the comparison. Analysts are flagged at times too, for example while travelling. The 7 days are a starting point.

Triage

  • Read the old and the new values in TargetResources. Was the policy disabled, did the risk level change, or were users and groups excluded?
  • Identify the actor (InitiatedBy) and look for a change record. A planned migration to Conditional Access disables the legacy policies, so confirm that the replacement policies are on and not in report-only mode.
  • Check the actor's sign-ins before the change for unfamiliar IP addresses or risk.
  • Review the risky sign-ins and risky users since the change, starting with the actor and with any account that was excluded.
  • Look for the activity Update NotificationSettings in the IdentityProtection category by the same actor. It changes who receives the emails about users at risk.

Response

Once the attack is confirmed:

  1. Restore the policy. Enable it, put the risk level back and remove the new exclusions.
  2. Revoke the sessions and reset the credentials of the identity that made the change, and remove its Security Administrator role until the investigation is closed.
  3. Go through the risky sign-ins and risky users recorded while the policy was weakened. For each account that signed in at risk without being challenged, revoke the sessions and reset the password.
  4. List the accounts that were created or stayed unregistered while the MFA registration policy was off, and have them register under supervision.
  5. Check the Conditional Access policies and the ID Protection notification settings for changes by the same actor.

Mitigation

  • Move risk-based access control to Conditional Access, where a change can be protected as a protected action and tried in report-only mode, and disable the legacy policies once the new ones are on.
  • Limit Security Administrator and Global Administrator to a few accounts managed with Privileged Identity Management.
  • Keep policy exclusions in dedicated, monitored groups that hold only the emergency access accounts.
  • Record the intended state of each policy and compare it regularly, so that drift is found even when an alert is missed.

Frequently asked questions

Were the ID Protection risk policies not retired?

Microsoft announced the retirement of the legacy user risk and sign-in risk policies configured in ID Protection for 1 October 2026 and asks tenants to recreate them in Conditional Access. The MFA registration policy remains in ID Protection. The rule therefore matters most for that policy and for tenants where a legacy policy still exists. Risk policies in Conditional Access are watched through the Conditional Access audit activities.

Does ID Protection stop detecting risk when a policy is disabled?

No. The policies only decide what happens when risk is detected. Risky sign-ins and risky users still appear in the reports and in the risk logs. What is lost is the automatic response, such as the MFA challenge, the block or the forced password change.

How do we tell a planned migration from an attack?

A migration has a change record, is carried out by the identity team and leaves an enabled Conditional Access policy with the same risk condition behind. Microsoft's migration steps create that policy first, confirm it in report-only mode and only then disable the old one. A policy that was disabled without a replacement, or that gained new exclusions, needs an explanation.

References

We respect your privacy

We use cookies to keep this site secure and working properly. With your permission, we also use optional cookies to understand usage and improve the experience. Cookie Policy

You can change your choice at any time.