Encryption is not the start of a ransomware incident — it's the finale. By the time files begin locking up or data shows up on a leak site, the operators behind the intrusion have typically been inside your environment for days or weeks. A recent analysis by Acronis, covered by BleepingComputer, reinforces what those of us who work ransomware incident response engagements already know from the trenches: encryption-based attacks follow a predictable pre-encryption playbook of credential theft, lateral movement, and deliberate sabotage of security controls and backups.
That dwell time is your window of opportunity. Every modern ransomware operation — whether it's a RaaS affiliate crew or a hands-on intrusion team — must perform a series of observable, high-signal actions before it can safely detonate. If your SOC is tuned to catch those five behaviors, you convert a catastrophic encryption event into a containable intrusion. If it isn't, your first alert will be the ransom note.
This post breaks down the five warning signs, why each one matters from an attacker's operational perspective, and gives you production-ready detections you can deploy today.
Why Pre-Encryption Detection Is the Real Ransomware Battleground
Two trends have converged to make pre-encryption detection the decisive control point. First, encryption-less extortion is on the rise — many groups now exfiltrate data and threaten publication without ever deploying a locker, meaning "we detected the ransomware binary" is no longer a complete strategy. Second, the average dwell time before detonation remains measurable in days, giving defenders a real but perishable window.
Across the ransomware IR engagements I've led, the failure pattern is almost never "we had no telemetry." It's that the telemetry existed — failed logons, a new service install, a shadow copy deletion — and nobody was alerting on it with appropriate severity. The five warning signs below are where your detection engineering effort should go.
The Five Warning Signs
1. Credential Theft and Abuse
Before an operator can move laterally or reach your crown jewels, they need valid credentials — ideally privileged ones. Expect LSASS memory dumping (via comsvcs.dll, procdump, or direct access from tools like Mimikatz), Kerberoasting (abnormal volumes of RC4-encrypted TGS requests), DCSync against domain controllers, and theft of cached credentials or password manager databases. Valid-account abuse is the number-one initial movement vector in the intrusions we investigate.
Defender's view: LSASS access from non-system processes, Event ID 4769 spikes for RC4 TGS tickets, and Event ID 4662 replication requests from non-DC hosts are all high-fidelity signals.
2. Lateral Movement
Once credentials are in hand, operators fan out. Classic patterns include PsExec-style service creation (Event ID 7045 with service binaries in ADMIN$ or user-writable paths), remote WMI/WinRM execution, RDP sessions outside administrative baselines, and SMB admin-share writes. In 2025–2026 intrusions we also see heavy abuse of legitimate RMM tooling (AnyDesk, ScreenConnect, Atera, Splashtop) — software your EDR may trust by default.
Defender's view: New service installations on servers, workstation-to-workstation SMB, RMM binaries installed outside your sanctioned toolset, and remote thread creation (Sysmon Event ID 8) into lsass.exe or Office processes.
3. Disabling Security Controls and Logging
This is one of the loudest pre-encryption signals. Operators routinely attempt to stop or delete EDR/AV services, clear event logs (wevtutil, Event ID 1102), disable Windows Defender via registry or PowerShell (Set-MpPreference -DisableRealtimeMonitoring $true), tamper with AMSI, and increasingly bring BYOVD (Bring Your Own Vulnerable Driver) payloads to blind kernel-level sensors. If your EDR agents start going quiet on a cohort of machines simultaneously, treat it as a critical incident, not a connectivity issue.
Defender's view: Log-clear events, Defender tampering registry writes, services stopped en masse, and sensor heartbeat loss correlated across hosts.
4. Backup Discovery and Destruction
Ransomware crews know backups are your leverage. Before encryption, they enumerate backup infrastructure (vssadmin list shadows, wbadmin, discovery of Veeam/Commvault/Acronis consoles), delete Volume Shadow Copies, corrupt or delete backup repositories, and in mature intrusions, compromise the backup admin accounts themselves. Acronis's guidance rightly emphasizes protecting backups as a first-class defense — an attacker who destroys your recovery path controls the negotiation.
Defender's view: vssadmin delete shadows, bcdedit recovery-option tampering, wbadmin delete catalog, and anomalous logons to backup consoles are immediate escalation triggers.
5. Data Staging and Exfiltration Preparation
Double extortion means the theft happens before — or instead of — the encryption. Watch for bulk archive creation (rar.exe/7z.exe with password flags against large directory trees), compression tools appearing on servers, installation of exfil clients (rclone, MEGASync, FileZilla), and sustained high-volume outbound transfers to cloud storage or unfamiliar ASNs, often scheduled for off-hours.
Defender's view: Command lines with archive flags plus passwords, rclone-style remotes in configs, and egress anomalies measured in gigabytes per host per night.
Detection Content
Sigma Rules
The following rules target three of the highest-fidelity pre-encryption behaviors: backup/shadow-copy destruction, security control tampering, and data staging with archiving tools. Tune allowlists for your environment before deploying at high level.
---
title: Shadow Copy and Backup Catalog Deletion
description: Detects deletion of Volume Shadow Copies or backup catalogs, a hallmark pre-encryption ransomware behavior intended to destroy recovery options.
references:
- https://www.bleepingcomputer.com/news/security/five-warning-signs-of-ransomware-breach-before-encryption-or-data-theft/
- https://attack.mitre.org/techniques/T1490/
author: Security Arsenal
date: 2026/02/10
status: experimental
logsource:
category: process_creation
product: windows
detection:
selection_img:
Image|endswith:
- '\vssadmin.exe'
- '\wbadmin.exe'
- '\bcdedit.exe'
- '\wmic.exe'
selection_cmd:
CommandLine|contains:
- 'delete shadows'
- 'delete catalog'
- 'shadowcopy delete'
- 'recoveryenabled no'
- 'ignoreallfailures'
condition: selection_img and selection_cmd
falsepositives:
- Legitimate backup maintenance scripts (rare; validate against change windows)
level: critical
---
title: Security Tool Tampering and Log Clearing
description: Detects attempts to disable real-time protection, stop security services, or clear Windows event logs — common attacker steps before ransomware deployment.
references:
- https://attack.mitre.org/techniques/T1562.001/
- https://attack.mitre.org/techniques/T1070.001/
author: Security Arsenal
date: 2026/02/10
status: experimental
logsource:
category: process_creation
product: windows
detection:
selection_mp:
Image|endswith: '\powershell.exe'
CommandLine|contains:
- 'DisableRealtimeMonitoring'
- 'DisableBehaviorMonitoring'
- 'DisableIOAVProtection'
- 'Add-MpPreference -ExclusionPath'
selection_logclear:
Image|endswith:
- '\wevtutil.exe'
- '\powershell.exe'
CommandLine|contains:
- 'wevtutil cl'
- 'Clear-EventLog'
- 'Remove-EventLog'
selection_svc:
Image|endswith:
- '\sc.exe'
- '\net.exe'
- '\net1.exe'
CommandLine|contains:
- 'stop windefend'
- 'stop sense'
- 'stop sentinel'
- 'stop carbonblack'
- 'stop cb'
- 'stop crowd'
- 'stop sophos'
condition: 1 of selection_*
falsepositives:
- Administrative Defender exclusions during software deployment
- Legitimate log rotation tooling (rare on endpoints)
level: high
---
title: Bulk Data Staging with Archiving Utilities
description: Detects password-protected archive creation by common compression tools, a frequent data staging behavior prior to exfiltration and extortion.
references:
- https://attack.mitre.org/techniques/T1560.001/
- https://attack.mitre.org/techniques/T1048/
author: Security Arsenal
date: 2026/02/10
status: experimental
logsource:
category: process_creation
product: windows
detection:
selection_img:
Image|endswith:
- '\rar.exe'
- '\7z.exe'
- '\7za.exe'
- '\winzip.exe'
- '\winrar.exe'
selection_cmd:
CommandLine|contains:
- ' -p'
- ' -hp'
- ' -v'
filter_users:
Image|contains:
- '\7-Zip\7z.exe'
CommandLine|contains: ' -p0'
condition: selection_img and selection_cmd and not 1 of filter_*
falsepositives:
- Legitimate backup packaging by IT staff
- Software packaging pipelines
level: medium
KQL — Microsoft Sentinel / Defender Hunt
This hunt surfaces hosts exhibiting a pre-encryption behavioral chain — security-control tampering or shadow-copy deletion followed by archive staging — within a 24-hour window on the same device. Correlating behaviors on a single host sharply reduces noise versus alerting on each signal independently.
let Lookback = 7d;
let Window = 1d;
let Tamper =
DeviceProcessEvents
| where TimeGenerated > ago(Lookback)
| where ProcessCommandLine has_any (
"vssadmin delete shadows", "wbadmin delete catalog",
"DisableRealtimeMonitoring", "wevtutil cl",
"bcdedit", "recoveryenabled no")
| project TamperTime = TimeGenerated, DeviceName, TamperCmd = ProcessCommandLine, AccountName;
let Staging =
DeviceProcessEvents
| where TimeGenerated > ago(Lookback)
| where FileName in~ ("rar.exe", "7z.exe", "7za.exe", "winrar.exe", "rclone.exe", "megasync.exe")
| project StageTime = TimeGenerated, DeviceName, StageProc = FileName, StageCmd = ProcessCommandLine;
Tamper
| join kind=inner Staging on DeviceName
| where StageTime between (TamperTime .. (TamperTime + Window))
| project DeviceName, AccountName, TamperTime, TamperCmd, StageTime, StageProc, StageCmd
| summarize arg_min(TamperTime, *) by DeviceName, AccountName
| sort by TamperTime desc;
A secondary, simpler Sentinel query worth running continuously against classic Windows event telemetry is log-clear detection: SecurityEvent | where EventID == 1102 | summarize count() by Computer, Account, bin(TimeGenerated, 1h). Any occurrence outside a documented maintenance window warrants immediate triage.
Velociraptor VQL — Endpoint Hunt
This artifact hunts across a fleet for processes matching pre-encryption behaviors and for suspicious service installations indicative of lateral movement tooling.
-- Hunt: Pre-encryption ransomware behaviors (tampering, staging, shadow deletion)
LET suspicious = '''(?i)(vssadmin.*delete\s+shadows|wbadmin.*delete\s+catalog|DisableRealtimeMonitoring|wevtutil\s+cl|bcdedit.*recoveryenabled|rclone.*--config|7z.*-p|rar.*-hp)'''
SELECT Pid, Ppid, Name, Exe, CommandLine, Username, CreateTime
FROM pslist()
WHERE CommandLine =~ suspicious
OR Exe =~ '(?i)(rclone|megasync|filezilla)\.exe$'
ORDER BY CreateTime DESC
-- Hunt: Recently installed services with binaries in user-writable or ADMIN$ paths (PsExec-style lateral movement)
SELECT Name, DisplayName, PathName, StartName, StartMode
FROM wmi_query(namespace='root/cimv2',
query='SELECT Name, DisplayName, PathName, StartName, StartMode FROM Win32_Service')
WHERE PathName =~ '(?i)(\\users\\|\\temp\\|\\appdata\\|\\programdata\\|admin\$)'
AND StartMode =~ '(?i)auto'
Hardening and Verification Script
Use this PowerShell script (run elevated, via GPO startup script or your RMM) to verify and enforce the controls that blunt pre-encryption attacker behavior: tamper protection, shadow copy accessibility, audit policy, and RMM tool inventory.
# Pre-encryption hardening verification — run elevated
$report = @()
# 1. Verify Defender Tamper Protection status
$mp = Get-MpPreference
$report += [pscustomobject]@{
Check = 'Defender Realtime Monitoring'
Status = if ($mp.DisableRealtimeMonitoring -eq $false) { 'ENABLED (good)' } else { 'DISABLED (bad)' }
}
try {
$tamper = Get-MpComputerStatus
$report += [pscustomobject]@{
Check = 'Defender Tamper Protection'
Status = if ($tamper.IsTamperProtected) { 'ENABLED (good)' } else { 'DISABLED — enable via Intune/MDE portal' }
}
} catch { $report += [pscustomobject]@{ Check = 'Tamper Protection'; Status = 'Query failed — verify MDE onboarding' } }
# 2. Verify VSS is operational and shadow copies exist on the system volume
$shadows = Get-CimInstance Win32_ShadowCopy -ErrorAction SilentlyContinue
$report += [pscustomobject]@{
Check = 'Volume Shadow Copies'
Status = if ($shadows) { "$($shadows.Count) shadow copies present" } else { 'NONE — verify VSS schedule and backup solution' }
}
# 3. Verify key audit policies are enabled (process creation + cmdline, log clearing visibility)
$audit = auditpol /get /subcategory:"Process Creation" 2>$null
$report += [pscustomobject]@{
Check = 'Audit: Process Creation'
Status = if ($audit -match 'Success') { 'Auditing enabled' } else { 'NOT AUDITING — run: auditpol /set /subcategory:"Process Creation" /success:enable' }
}
$cmdline = Get-ItemProperty 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System\Audit' -Name ProcessCreationIncludeCmdLine_Enabled -ErrorAction SilentlyContinue
$report += [pscustomobject]@{
Check = 'Cmdline in Process Audit Events'
Status = if ($cmdline.ProcessCreationIncludeCmdLine_Enabled -eq 1) { 'Enabled' } else { 'DISABLED — set ProcessCreationIncludeCmdLine_Enabled = 1' }
}
# 4. Inventory unsanctioned RMM / remote access tooling (common lateral movement vector)
$rmm = @('anydesk','screenconnect','connectwise','atera','splashtop','teamviewer','ninjarmm','level.io','action1')
$found = Get-CimInstance Win32_Service | Where-Object {
$svc = $_.Name + ' ' + $_.PathName
($rmm | Where-Object { $svc -match $_ })
} | Select-Object Name, PathName, State
if ($found) {
$report += [pscustomobject]@{ Check = 'RMM Tools Installed'; Status = "REVIEW: $(($found | ForEach-Object Name) -join ', ')" }
} else {
$report += [pscustomobject]@{ Check = 'RMM Tools Installed'; Status = 'None detected' }
}
$report | Format-Table -AutoSize
Remediation and Hardening Recommendations
This news item describes attacker tradecraft rather than a single patchable vulnerability — there is no CVE to remediate. The correct response is control hardening against the five pre-encryption behaviors:
- Protect credentials at the source. Enable Credential Guard on supported Windows 10/11 and Server 2022+ systems, block LSASS access via Attack Surface Reduction rule "Block credential stealing from the Windows local security authority subsystem," and enforce tiered administration so Tier-0 credentials never touch workstations.
- Constrain lateral movement. Disable PsExec-style remote service creation where not required, restrict workstation-to-workstation SMB/RDP via host firewall rules, and maintain an allowlist of sanctioned RMM tools — alert on everything else.
- Harden security controls against tampering. Enable Tamper Protection across your fleet, forward Event ID 1102 (log cleared) to your SIEM as a high-severity alert, and alert on EDR sensor heartbeat loss correlated across multiple hosts — not single-host disconnects.
- Make backups attacker-proof. Follow the 3-2-1-1-0 model with at least one immutable or offline copy, isolate backup infrastructure with dedicated credentials (never domain-joined to production AD), and monitor backup consoles for anomalous logons and job deletions.
- Detect staging and egress. Alert on archiving tools executing with password flags, maintain an egress policy that flags new or high-volume destinations (personal cloud storage, unfamiliar VPS ASNs), and baseline normal per-host outbound volume so 50 GB overnight from a file server is unmistakable.
- Rehearse the window. Run tabletop exercises scoped to the 24–72 hour pre-encryption window: if your SOC detects shadow-copy deletion at 02:00 on a Saturday, who gets paged, what gets isolated, and how fast?
The Bottom Line
Ransomware groups are running the same pre-encryption playbook because defenders keep ignoring its loudest steps. Credential theft, lateral movement, control tampering, backup destruction, and data staging are not subtle — they are noisy, observable, and detectable with the telemetry most organizations already collect. The gap is almost always detection engineering and alert prioritization, not tooling. Tune your stack to the five warning signs above, and the ransom note becomes an incident you already contained, not a catastrophe you discovered.
Related Resources
Security Arsenal Incident Response Services AlertMonitor Platform Book a SOC Assessment incident-response Intel Hub
Is your security operations ready?
Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.