MCADDF

[IMPACT-SERVICE-STOP-001]: Service Shutdown/Deletion (Multi-Env)

1. METADATA HEADER

Metadata

| Attribute | Details | |—|—| | Technique ID | IMPACT-SERVICE-STOP-001 | | Technique Name | Service Shutdown/Deletion | | MITRE ATT&CK v18.1 | Service Stop (T1489) & Account Access Removal (T1531) – https://attack.mitre.org/techniques/T1489/, https://attack.mitre.org/techniques/T1531/ | | Tactic | Impact | | Platforms | Windows Server / Windows Endpoint, Linux, Azure VMs, Hypervisors, SaaS admin portals | | Environment | Multi-Env (on‑prem, Azure, M365, other SaaS) | | Severity | High to Critical (depending on service criticality) | | CVE | N/A (abuse of system features; often used alongside other vulnerabilities) | | Technique Status | ACTIVE | | Last Verified | 2026-01-10 | | Affected Versions | Windows Server 2012 R2–2025, Windows 10/11, major Linux distros, Azure IaaS VMs, M365/Azure services | | Patched In | N/A – mitigated via hardening, RBAC, EDR, and monitoring; core stop/delete operations remain by design.[61][48][51] | | Author | SERVTEPArtur Pchelnikau |


2. EXECUTIVE SUMMARY

Operational Risk

Compliance Mappings

Framework Control / ID Description (Failure Mode)
CIS Benchmarks Windows CIS 18.x, Linux CIS 1.x Lack of restrictions on who can control system services and weak logging allow attackers to stop critical services undetected.
DISA STIG Windows, Azure Compute STIGs Non-enforcement of service hardening, unauthorized disabling of security controls.
CISA SCuBA Logging & Monitoring Inadequate monitoring of critical security service status and cloud service state.
NIST SP 800-53 Rev.5 AC-6, AU-12, SI-4, CP-10 Excessive privileges, insufficient logging, and inadequate failover procedures for critical services.[61][48]
GDPR Art. 32 Failure to ensure ongoing confidentiality, integrity, and availability of processing – especially if security services are disabled.
DORA Art. 11 Insufficient ICT resilience where attackers can disrupt core financial services by shutting down platforms.
NIS2 Art. 21 Missing operational safeguards and monitoring around essential service availability.
ISO 27001:2022 A.5.29, A.8.16 Poor monitoring of security controls and weak capacity management/availability for systems.
ISO 27005 Risk Scenario: “EDR/backup infrastructure disabled before destructive attack” Hinders detection and recovery, increases loss magnitude.

3. TECHNICAL PREREQUISITES


4. ENVIRONMENTAL RECONNAISSANCE

Windows – Service Inventory

Get-Service | Where-Object { $_.Status -eq 'Running' } |
  Select-Object Name,DisplayName,Status | Sort-Object Name

What to Look For (Attacker):

What to Look For (Defender):

Azure – VM Deallocation Monitoring

AzureActivity
| where OperationNameValue =~ "MICROSOFT.COMPUTE/VIRTUALMACHINES/DEALLOCATE/ACTION"
| where ActivityStatusValue =~ "Succeeded"
| summarize count() by Caller, ResourceGroup, bin(TimeGenerated, 1h)

What to Look For:


5. DETAILED EXECUTION METHODS AND THEIRS STEPS

METHOD 1 – Stopping Security & Backup Services on Windows

Supported Versions: Windows Server 2012 R2–2025, Windows 10/11.

Step 1: Stop Windows Defender and EDR Services

Objective: Disable preventive and detection capabilities.

sc stop WinDefend
sc config WinDefend start= disabled

Ransomware families such as Akira and RansomHub execute similar commands or use signed kernel drivers to kill security agents.[49][52][58]

Step 2: Stop Backup and Shadow Copy Services

net stop VSS /y
net stop SQLWriter /y

Combined with T1490 (Inhibit System Recovery), attackers may also remove shadow copies:

Get-WmiObject Win32_ShadowCopy | Remove-WmiObject

METHOD 2 – Stopping Services via PowerShell

$targets = 'WinDefend','VSS','SQLWriter','BackupExecVSSProvider'
foreach ($svc in $targets) {
  try { Stop-Service -Name $svc -Force -ErrorAction SilentlyContinue } catch {}
}

METHOD 3 – Azure VM Deallocation as Service Stop

Objective: Use Azure control plane to deallocate or stop key VMs, denying access to hosted services.[50]

az vm deallocate -g <RG_NAME> -n <VM_NAME>

Expected Impact:


6. ATTACK SIMULATION & VERIFICATION (Atomic Red Team)

Example (conceptual):

Invoke-AtomicTest T1489 -TestNumbers 1

Use a non-production service to verify detection rules and logging.


7. TOOLS & COMMANDS REFERENCE

Defenders must baseline legitimate usage and alert on anomalies.


8. SPLUNK DETECTION RULES

Rule 1: Suspicious Bulk Service Stop on Windows

index=wineventlog sourcetype="WinEventLog:Security" EventCode=7036
| search Message="*stopped*service*"
| stats count by host, servicename, user, bin(_time, 10m)
| where count >= 10

Adjust fields depending on your Windows logging configuration; 7036 may also appear in System logs (sourcetype WinEventLog:System).

Rule 2: Azure VM Deallocate Spikes

index=azure sourcetype="azure:monitor:activity"
| where operationName="Microsoft.Compute/virtualMachines/deallocate/action" AND ActivityStatus="Succeeded"
| bin _time span=15m
| stats count by _time, caller
| where count >= 5

9. MICROSOFT SENTINEL DETECTION

Query 1: Multiple Critical Services Stopped on Windows Hosts

SecurityEvent
| where EventID == 7036
| where Param1 has_any ("WinDefend","VSS","SQLWriter","BackupExecVSSProvider")
| summarize ServiceStops = count() by Computer, bin(TimeGenerated, 15m)
| where ServiceStops >= 5

Query 2: Azure VM Deallocate Operations

AzureActivity
| where OperationNameValue =~ "MICROSOFT.COMPUTE/VIRTUALMACHINES/DEALLOCATE/ACTION"
| where ActivityStatusValue =~ "Succeeded"
| summarize VMDeallocations = count() by Caller, bin(TimeGenerated, 15m)
| where VMDeallocations >= 3

10. WINDOWS EVENT LOG MONITORING

Key events:

Ensure Advanced Audit Policy and service control logging are enabled via Group Policy.


11. SYSMON DETECTION PATTERNS

Example Sysmon config to catch suspicious service control tools:

<ProcessCreate onmatch="include">
  <Image condition="end with">\sc.exe</Image>
  <Image condition="end with">\net.exe</Image>
  <Image condition="end with">\taskkill.exe</Image>
  <CommandLine condition="contains"> stop </CommandLine>
</ProcessCreate>

Forward these events to Sentinel/Splunk for correlation with service stop events.


12. MICROSOFT DEFENDER FOR CLOUD

Defender for Cloud and Defender for Endpoint can raise alerts when security agents are disabled, services are stopped, or tampering is detected on critical workloads.[52][82]

Ensure:


13. MICROSOFT PURVIEW (UNIFIED AUDIT LOG)

In M365, account access removal (T1531) can be seen in Purview/Unified Audit Logs (e.g., Remove-Mailbox, disabling user accounts, licensing changes).[51][54]

Example Purview query:

Search-UnifiedAuditLog -StartDate (Get-Date).AddDays(-1) -EndDate (Get-Date) \
  -Operations Remove-Mailbox,DisableUserAccount -ResultSize 5000

14. DEFENSIVE MITIGATIONS

Priority 1: CRITICAL

Priority 2: HIGH


15. DETECTION & INCIDENT RESPONSE

IOCs

Response

  1. Immediately restart critical security and backup services (or VMs) where possible.
  2. Investigate who issued the stop/deallocate commands (process lineage, Azure caller).
  3. Look for follow-on activity: ransomware, data destruction, account lockouts.
  4. Harden service configurations and rotate compromised credentials.

Step Phase Technique Description
1 Initial Access IA-VALID-001 / IA-PHISH-001 Attacker obtains privileged access.
2 Privilege Escalation PE-VALID-010 / PE-ACCTMGMT-011 Elevate to local admin, domain admin, or high Azure roles.
3 Current Step [IMPACT-SERVICE-STOP-001] Service Shutdown/Deletion Stop or disable security, backup, and business services.
4 Impact T1486, T1485 Deploy ransomware or destroy data with reduced chance of detection or recovery.[61][51]

17. REAL-WORLD EXAMPLES

Example 1: RansomHub Campaigns

Trend Micro documented RansomHub campaigns where attackers used scripts (disableAV.bat, tdsskiller.bat) and tools (STONESTOP, POORTRY) to kill AV/EDR processes and services, and to delete registry keys controlling security products.[49] They also removed backup-related services to inhibit recovery.

Example 2: Akira Ransomware Targeting Backups

Akira campaigns show systematic removal of shadow copies and stopping of backup services to prevent system recovery, often via PowerShell commands that remove Win32_Shadowcopy instances and stop Volume Shadow Copy and related services.[52]

These real-world incidents highlight Service Stop and Account Access Removal as standard components of modern ransomware playbooks.