How to Remove Sticky Keys Backdoors on Windows Login Screens
Sticky Keys is a legitimate Windows accessibility feature that lets users press modifier keys one at a time. It can also create a serious security weakness when an attacker changes the way Windows launches its accessibility tools. Pressing Shift five times at the sign-in screen may then open a command shell instead of Sticky Keys, potentially with SYSTEM-level privileges and without requiring a normal account password.
Removing this abuse requires an incident-response approach rather than a simple settings change. An attacker capable of modifying protected registry keys may also have installed other persistence mechanisms, created privileged accounts or accessed credentials. The response must therefore combine containment, evidence collection, removal, recovery and verification.
How the sticky keys backdoor works
The usual technique targets sethc.exe, the executable responsible for the Sticky Keys interface. Attackers create or modify an Image File Execution Options registry entry under HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Image File Execution Options\sethc.exe. Its Debugger value is then redirected to cmd.exe, PowerShell or another executable.
When Windows starts the accessibility tool, it also starts the configured debugger. Because the modified tool runs during the sign-in process, the resulting command shell may appear before a user signs in. This differs from a normal Sticky Keys configuration: the shortcut is not merely available, but its underlying executable has been redirected.
Related attacks may target utilman.exe, osk.exe or magnify.exe. They can also combine the registry modification with scheduled tasks, services, startup programs, malicious user profiles or remote administration software. A visible shortcut is therefore one warning sign, while its absence does not prove that the device is clean.
Check the sign-in screen without triggering the shortcut
Do not test the suspected behaviour by repeatedly pressing Shift on a potentially compromised device. Doing so could execute an attacker-controlled command, generate useful artefacts or interrupt the Windows environment. Instead, isolate the endpoint from untrusted networks while retaining access to EDR, management and forensic tools. Preserve volatile evidence where practical, and record who placed the device in isolation.
On a suspected device, an administrator can inspect the relevant registry location with reg query "HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Image File Execution Options\sethc.exe" /s. A legitimate system should not have a Debugger value pointing to a shell under that entry. Check the corresponding locations for the other Windows accessibility executables, then review local administrators, new services, remote-access tools, unusual logons and recent security events.
In Australian security conversations, this may be called a “dodgy sign-in shortcut”, but it should be handled as suspected system-level compromise. Some employees legitimately need Sticky Keys, magnification or on-screen keyboards, so their documented accessibility requirements must be considered. The distinction to investigate is not accessibility use itself, but unauthorised redirection of the executable and changes made outside the approved support process.
Remove the malicious registry hook
Begin from a known administrator account or the Windows Recovery Environment if normal sign-in is unavailable. Before editing the registry, export the relevant Image File Execution Options keys and collect other evidence required by the organisation’s incident process. If the system is evidential, work from a copy or follow the approved forensic procedure rather than modifying the original unnecessarily.
Locate the sethc.exe entry and remove the malicious Debugger value. If the key exists only to support the abuse, remove the complete entry. Also inspect utilman.exe, osk.exe and magnify.exe for equivalent redirection. When the running Windows installation cannot be trusted, use Windows Recovery Environment to load the offline SOFTWARE hive and examine it without starting the affected operating system.
Registry removal alone is not enough. Review Run keys, startup folders, scheduled tasks, services, WMI event subscriptions, browser extensions, user profiles and files across local drives. Run a Microsoft Defender full scan and an offline scan, then use approved EDR and endpoint-forensics tools to identify anything missed. When confidence in the device’s integrity is low, rebuilding it from a known-good image is safer than trying to prove that every altered component has been found.
Finally, restore accessibility settings according to the user’s approved needs and organisational policy. Validate the endpoint from a controlled administrative session, monitoring registry and process activity. Do not reactivate the shortcut on the sign-in screen until the environment has been cleaned and the investigation has established how the attacker obtained sufficient privileges.
Recover accounts and meet Australian obligations
Resetting passwords should follow containment, not precede it. If the attacker had SYSTEM access, assume that credentials, tokens and secrets accessible to that operating system may have been exposed. Revoke active sessions, rotate privileged and domain credentials, review service accounts and examine authentication logs. Resetting an affected user password without revoking existing tokens can leave the attacker signed in.
Australian operational conditions add important complications. Many organisations in Sydney, Melbourne, Brisbane and Perth support hybrid workers across different time zones, while smaller businesses often depend on managed service providers for endpoint monitoring. Isolation must preserve the provider’s visibility, and responsibility for registry changes, evidence handling and credential recovery must be clear across client, provider and internal teams.
Regulatory context also matters. Entities subject to the Australian Prudential Regulation Authority’s cyber security assessment obligations should align recovery evidence with the APRS process, while ASD’s Essential Eight provides a useful control baseline for registry auditing, application control and incident recovery. If personal information may have been accessed, assess the Australian Privacy Act and Notifiable Data Breaches scheme with legal and privacy specialists. The Joint Cyber Security Centre and CERT Australia can provide relevant coordination and advice during significant incidents.
CARM Security’s guidance on sensitive personal data breaches can help teams connect endpoint containment with notification and legal assessment. The link between technical evidence and business decisions is essential because a registry artefact may establish system access, while affected records determine whether notification duties are triggered.
Make the response repeatable
A repeatable runbook gives Australian security teams a consistent way to remove the hook without overlooking related persistence:
- Isolate the endpoint while preserving EDR, management and forensic access.
- Export relevant registry keys and record evidence before making changes.
- Inspect
sethc.exe,utilman.exe,osk.exeandmagnify.exeImage File Execution Options entries. - Review accounts, services, scheduled tasks, startup items and remote-access software.
- Rebuild the endpoint when its integrity cannot be confidently established.
- Revoke sessions and rotate exposed credentials after containment.
- Document the removal and restore accessibility settings through an approved process.
The following comparison separates expected accessibility use from signs of a deeper compromise:
| Observation | Likely interpretation | Required action | Evidence to preserve |
|---|---|---|---|
| Sticky Keys is available and the user has a documented need | Legitimate accessibility use | Retain it under normal policy | Approval and configuration record |
| Sticky Keys is enabled without an approved reason | Configuration drift or misuse | Disable or restore it and investigate | Registry export and user history |
sethc.exe opens a command shell |
Image File Execution Options abuse | Isolate, remove the hook and rebuild if necessary | Registry key, process and Defender records |
| Another accessibility executable behaves abnormally | Related persistence technique | Inspect and remove the corresponding hook | Registry keys, logs and EDR alerts |
| Unknown administrator accounts or remote sessions exist | Possible deeper compromise | Revoke access and rebuild the device | Identity, network and authentication logs |
The practical takeaway is to treat any unexpected command shell at the Windows sign-in screen as possible system-level compromise: isolate the device, preserve registry and identity evidence, remove malicious Image File Execution Options entries, hunt for related persistence, and rebuild when endpoint integrity remains uncertain.