How the attack works
The Intune Wipe action restores a device to factory settings and removes its data, apps and configuration; Retire removes company data and takes the device out of management. Both are ordinary helpdesk actions for lost, stolen or returned devices. They are available to Intune Administrator, to the Intune roles Help Desk Operator and School Administrator, to any custom role with the Remote tasks Wipe or Retire permission, and to an application with DeviceManagementManagedDevices.PrivilegedOperations.All.
Used against the whole fleet, the same action is a destructive attack that needs no code on the endpoints: the devices carry out the instruction of the management service they are enrolled in. On 18 March 2026 CISA urged organizations to harden endpoint management systems after malicious activity that misused legitimate endpoint management software, and named device wiping among the actions that should need a second administrator's approval. Microsoft now documents daily tenant limits of 500 Wipe, 1,000 Retire and 1,000 Delete actions. The rule reads the Intune audit log, which reaches a SIEM only after an Intune diagnostic setting (category AuditLogs) sends it to a Log Analytics workspace or an event hub. One event is one device, so a mass wipe appears as many events from one actor. A threshold rule on this page counts those events and fires when one actor wipes or retires five or more devices within an hour.
What an attacker needs
- Control of an identity with the Remote tasks Wipe or Retire permission over the target devices, or an application with
DeviceManagementManagedDevices.PrivilegedOperations.All. - No Multi Admin Approval access policy for device actions, or a second compromised approver.
Impact
Loss of the data and configuration on every wiped device and of the working time needed to rebuild them. Wiped Windows devices have to be set up again, and the option that keeps wiping after power loss can leave a device unable to start. Retired devices lose company data and management. If administrators' devices are in scope, the people who would respond lose their tools at the same moment.
Detection rules
Intune Device Wipe or Retire Action
Detects a successful wipe or retire device action in the Intune audit log. Microsoft does not publish the list of OperationName values. Confirmed: the name wipe ManagedDevice is in the sample events and the bulk wipe detection of Splunk Threat Research. Inferred: retire ManagedDevice follows the same pattern (the Microsoft Graph action, then the resource type) but is not in a Microsoft or Splunk sample, so confirm it in your tenant. Each event is one device and both actions are routine for lost or returned devices, so use the rule as the base of an alert that counts events per Identity (for example five or more in an hour) and as an immediate alert for actors that never wipe devices.
title: Intune Device Wipe or Retire Action
id: 22ee1e3f-cd33-45c7-883a-3399c1819158
status: experimental
description: >-
Detects a successful wipe or retire device action in the Intune audit log.
Microsoft does not publish the list of OperationName values. Confirmed: the
name wipe ManagedDevice is in the sample events and the bulk wipe detection
of Splunk Threat Research. Inferred: retire ManagedDevice follows the same
pattern (the Microsoft Graph action, then the resource type) but is not in a
Microsoft or Splunk sample, so confirm it in your tenant. Each event is one
device and both actions are routine for lost or returned devices, so use the
rule as the base of an alert that counts events per Identity (for example
five or more in an hour) and as an immediate alert for actors that never
wipe devices.
references:
- https://attack.mitre.org/techniques/T1485/
- https://learn.microsoft.com/en-us/intune/device-management/actions/wipe
- https://learn.microsoft.com/en-us/graph/api/intune-devices-manageddevice-wipe
- https://research.splunk.com/cloud/c3f48aa9-878e-443f-8889-e42a11a9bea9/
author: Forestall
date: 2026-11-11
tags:
- attack.impact
- attack.t1485
logsource:
product: intune
service: audit
detection:
selection:
OperationName|endswith:
- wipe ManagedDevice
- retire ManagedDevice
ResultType: Success
condition: selection
falsepositives:
- Helpdesk wiping or retiring lost, stolen, returned or reassigned devices; expected from the service desk accounts and in small numbers.
- Offboarding runs and device refresh projects; announce them to the SOC and raise the count threshold for their duration.
level: medium
False positives
- Helpdesk wiping or retiring lost, stolen, returned or reassigned devices; expected from the service desk accounts and in small numbers.
- Offboarding runs and device refresh projects; announce them to the SOC and raise the count threshold for their duration.
One Actor Wiped or Retired Many Devices Through Intune
Counts the successful wipe and retire actions per actor (Identity) in the Intune audit log. One event is one device, so a destructive mass wipe shows as a run of events from one account or one application. The rule fires at five or more actions by the same actor within one hour. A service desk that handles lost and returned devices stays below that on a normal day, and Microsoft's daily tenant limits (500 Wipe and 1,000 Retire actions) still leave room for hundreds of devices. The operation names carry the caveat of the single-event rule: Microsoft does not publish the list of OperationName values, wipe ManagedDevice is confirmed by public sample events and retire ManagedDevice is inferred from the same pattern. The threshold and the window are starting points.
Fires when one Identity has at least 5 matching events within 1 hour.
title: Intune Device Wipe or Retire Action (Mass Wipe Base Event)
id: 9f73e62a-deb2-4f90-89ad-46fda3495849
name: intune_device_wipe_or_retire
status: experimental
description: >-
Base event of the threshold rule in this file. A successful wipe or retire
device action in the Intune audit log, with the same operation names as the
single-event rule of this attack.
logsource:
product: intune
service: audit
detection:
selection:
OperationName|endswith:
- wipe ManagedDevice
- retire ManagedDevice
ResultType: Success
condition: selection
level: low
---
title: One Actor Wiped or Retired Many Devices Through Intune
id: b78cce3d-8aa3-49f7-a9ef-c0c040b33531
status: experimental
description: >-
Counts the successful wipe and retire actions per actor (Identity) in the
Intune audit log. One event is one device, so a destructive mass wipe shows
as a run of events from one account or one application. The rule fires at
five or more actions by the same actor within one hour. A service desk that
handles lost and returned devices stays below that on a normal day, and
Microsoft's daily tenant limits (500 Wipe and 1,000 Retire actions) still
leave room for hundreds of devices. The operation names carry the caveat of
the single-event rule: Microsoft does not publish the list of OperationName
values, wipe ManagedDevice is confirmed by public sample events and retire
ManagedDevice is inferred from the same pattern. The threshold and the
window are starting points.
references:
- https://learn.microsoft.com/en-us/intune/device-management/actions/wipe
- https://learn.microsoft.com/en-us/intune/device-management/actions/#daily-tenant-limits
- https://learn.microsoft.com/en-us/azure/azure-monitor/reference/tables/intuneauditlogs
- https://research.splunk.com/cloud/c3f48aa9-878e-443f-8889-e42a11a9bea9/
- https://attack.mitre.org/techniques/T1485/
author: Forestall
date: 2026-11-18
tags:
- attack.impact
- attack.t1485
correlation:
type: event_count
rules:
- intune_device_wipe_or_retire
group-by:
- Identity
timespan: 1h
condition:
gte: 5
falsepositives:
- Offboarding runs, device refresh projects and the return of a batch of leased devices, in which one technician wipes many devices. Announce them to the SOC and raise the threshold for their duration.
- Automation that retires devices which have not checked in for a long time. Exclude its identity once it is known.
- The values of five actions and one hour are starting points, not tested limits.
level: high
False positives
- Offboarding runs, device refresh projects and the return of a batch of leased devices, in which one technician wipes many devices. Announce them to the SOC and raise the threshold for their duration.
- Automation that retires devices which have not checked in for a long time. Exclude its identity once it is known.
- The values of five actions and one hour are starting points, not tested limits.
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.
Wipe or retire by an actor that has not wiped a device before
Wipes come from a small and stable group, the service desk and the endpoint team. A destructive attack runs through an account that was taken over or given the permission for the purpose, and that account usually has no history of device actions. Alerting on the first wipe of a new actor gives the warning before the run of events reaches the threshold rule, while most devices have not yet checked in.
- Intune audit log: the successful wipe and retire records that the rules on this page match (
ResultTypeSuccess). KeepIdentityand the time. - Build the list of
Identityvalues with such records in the last 90 days. - Alert on the first wipe or retire by an
Identitythat is not on the list, whatever the count. - Check the actor's sign-in before the action in the Microsoft Entra sign-in logs (
UserPrincipalName,IPAddress,RiskLevelDuringSignIn), and look for changes to its Intune role or admin group in the days before.
LimitsNew service desk staff and a new automation account appear once. The baseline is empty until 90 days of records exist, and the 90 days are a starting point. It does not see an attacker who uses an account that wipes devices as part of its daily work. Microsoft does not describe the Identity column, so confirm in your data that it names the account or application that acted.
Triage
- Count the events per actor (Identity) in the last hour; a burst from one account or from an application is an incident until proven otherwise.
- Identify the devices (TargetObjectIds in Properties, and the device ID in the modified properties) and whether they include administrators' or executives' devices.
- Confirm the action with the actor through a separate channel, and check the sign-in that preceded it for a new location, device or token.
- If unauthorized, disable the actor and revoke its sessions at once; a device receives pending actions when it next checks in, so minutes matter.
Response
Once the attack is confirmed:
- Remove the actor's ability to wipe, not only its session. Take away the Intune role assignment or custom role, or remove
DeviceManagementManagedDevices.PrivilegedOperations.Alland the credentials from the application. - Tell the users of devices that have not carried out the action yet to keep them offline until data that exists only on the device is saved. A device receives a pending action when it next checks in.
- Check the Intune audit log for the actor's other actions in the same period, such as retire, delete, role changes and changes to Multi Admin Approval policies, and reverse what can be reversed.
- Rebuild in order of need. Start with the devices of administrators and responders, then business-critical users, on your tested recovery path.
- Restore user data from backups or cloud copies, and record which devices held data that existed nowhere else.
- Count the wipe, retire and delete events against your device inventory to confirm the full extent before you close the incident.
Mitigation
- Create a Multi Admin Approval access policy for Device actions, which covers wipe, retire and delete.
- Give the Wipe and Retire permissions through a custom role with narrow scope groups instead of assigning Help Desk Operator over all devices.
- Remove
DeviceManagementManagedDevices.PrivilegedOperations.Allfrom applications that do not need it. - Alert on every wipe by an unusual actor and on more than a handful of wipes by one actor within an hour.
- Keep a tested recovery path for devices, including Autopilot registration and the recovery of data that exists only on endpoints.
Frequently asked questions
Why does a mass wipe not appear as one event?
In the Intune audit log one event is one device. A mass wipe therefore appears as many events from one actor. A threshold rule on this page counts the events per actor and fires at five or more within an hour, which is a starting point. The single-event rule serves as an immediate alert for actors that never wipe devices.
Do Microsoft's daily limits on device actions prevent a destructive wipe?
They cap it, they do not prevent it. Microsoft documents daily tenant limits of 500 Wipe, 1,000 Retire and 1,000 Delete actions. Hundreds of devices can still be reset in a day, so a Multi Admin Approval access policy for device actions is still needed to require a second administrator.
How does the impact of Retire differ from Wipe?
Wipe restores a device to factory settings and removes its data, apps and configuration. Retire removes company data and takes the device out of management. Both are ordinary helpdesk actions for lost, stolen or returned devices, which is why the count per actor matters more than a single event.
Is the operation name for Retire confirmed?
Not from a public sample. The name wipe ManagedDevice appears in the sample events and the bulk wipe detection of Splunk Threat Research. The name retire ManagedDevice is inferred from the same pattern, so confirm it with a test retire in your tenant.
References
- MITRE ATT&CK T1485: Data Destruction
- Microsoft Learn: Wipe devices with Microsoft Intune
- CISA: CISA Urges Endpoint Management System Hardening After Cyberattack Against US Organization
- Microsoft Intune Customer Success blog: Best practices for securing Microsoft Intune
- Microsoft Learn: Use access policies to require multi admin approval
- Microsoft Graph: managedDevice wipe action
- Microsoft Learn: Device actions in Intune, daily tenant limits
- Azure Monitor Logs reference: IntuneAuditLogs
- Azure Monitor Logs reference: SigninLogs