Isolate the host immediately, then preserve process, registry, file and network telemetry before remediation begins. With stealers, the priority is to stop further credential harvesting and exfiltration while retaining enough evidence to confirm what was collected, where persistence was created and whether any sessions or secrets were already reused.
Why the first move is containment, not cleanup
When a stealer is found on a Windows endpoint, the first objective is to stop the blast radius. MITRE ATT&CK Enterprise Matrix is useful here because credential theft, local discovery, and follow-on lateral movement are common post-compromise paths. Isolating the host cuts off additional theft and reduces the chance that harvested material is immediately reused elsewhere.
That priority is especially important on endpoints that have active browser sessions, cached tokens, password managers, or administrative access. A stealer often needs only minutes to copy what it can reach, so containment should happen before any broad remediation, reboots, or interactive investigation that could destroy volatile evidence or trigger more exfiltration.
On a practical level, “isolate” means removing the endpoint’s ability to reach the internet and sensitive internal services while preserving the local state as found. If remote containment is available through EDR or network controls, use it immediately rather than waiting for a hands-on response team to arrive.
What evidence to preserve before remediation starts
Once the endpoint is contained, teams should preserve the telemetry needed to answer three questions: what ran, what persisted, and what left the machine. That usually means process trees, command lines, autoruns, registry changes, dropped files, browser artifacts, and network connections. The point is to keep enough context to determine whether the malware only attempted theft or also established persistence.
For Windows systems, registry and file artifacts can show startup entries, scheduled task abuse, or secondary payload staging. Process and network evidence can show whether the stealer launched from a user context, injected into another process, or reached out to a known command-and-control or upload destination. Preserving those details before cleanup makes later scoping much more reliable.
If available, capture memory or at least the highest-value volatile telemetry before the system is altered. Even when a full forensic image is not feasible, defenders still need enough preserved state to support credential reset decisions, hunt other affected hosts, and determine whether sessions or tokens were already stolen and reused.
How to sequence response so you do not lose the case
The response sequence should be containment, evidence preservation, triage, then remediation. Triage should determine which accounts used the endpoint, whether the user had privileged access, whether browser or cloud sessions were active, and whether the stealer reached credential stores, mail, chat, or source-control tools. Those answers decide the urgency of password reset, token revocation, and wider incident scoping.
Because stealers often target reusable secrets, teams should assume any exposed secret may already have been copied. A good response therefore treats identity reset as part of the containment workflow, not as a later hygiene task. If the endpoint held administrative credentials, incident scope expands quickly beyond the host itself.
When a machine has been isolated and its telemetry preserved, remediation can proceed with less risk of erasing the evidence needed for root cause analysis and post-incident hunting. That is why the first move is not malware removal, it is preserving control over the endpoint and the information it may have exposed.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
MITRE ATT&CK addresses the attack and risk surface, while NIST SP 800-53 Rev 5 and CIS Controls v8 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| MITRE ATT&CK | T1003 — OS Credential Dumping | Stealer incidents often harvest credentials and tokens before lateral movement. |
| Recommendation — Map the compromise to credential-access techniques and hunt for adjacent lateral-movement activity. | ||
| NIST SP 800-53 Rev 5 | AU-6 — Audit Record Review, Analysis, and Reporting | Incident triage depends on preserved logs and telemetry to reconstruct execution and exfiltration. |
| IR-4 — Incident Handling | Immediate isolation and evidence preservation are core incident-handling actions for malware on endpoints. | |
| Recommendation — Review preserved audit data quickly to scope execution, persistence, and outbound activity. Contain the endpoint first, preserve evidence, then execute remediation under incident-handling procedures. | ||
| CIS Controls v8 | CIS-17 — Incident Response Management | The question is about first-response sequencing after malware detection on an endpoint. |
| CIS-8 — Audit Log Management | Process, registry, file, and network telemetry are the evidence base for stealer investigations. | |
| Recommendation — Use your response plan to isolate, preserve evidence, and coordinate follow-on containment. Retain endpoint and network logs long enough to support scoping and recovery decisions. | ||
Practitioner Guidance
What to prioritize: Cut network reachability first, then preserve the smallest set of artifacts that still tells the story of execution, persistence, and outbound activity. If the host is a jump box, admin workstation, or browser-heavy user endpoint, escalate the response because the likely credential impact is materially higher.
What to verify: Confirm whether the machine had active sessions, saved passwords, tokens, or credentials in use at the time of detection. If you cannot verify that data was untouched, assume reuse risk and widen the response to nearby accounts and services.
Decision rule: If containment would destroy volatile evidence, use the least destructive isolation method available first, then collect telemetry immediately after. If the endpoint is already unavailable, prioritize remote logs, EDR telemetry, and adjacent authentication records to reconstruct what happened.
Practitioner takeaway: With stealers, speed matters twice: the first minutes are for stopping further collection, and the next minutes are for preserving enough evidence to scope identity exposure accurately.
Related resources from NHI Mgmt Group
- What should security teams do first when a cloud-hosted Windows endpoint must be remediated at scale after a sensor-related outage?
- Why are NHIs a critical concern for security teams?
- How should teams reduce the risk of exposed AI credentials being abused?
- What steps should security teams take to prevent Shadow AI risks?
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on October 12, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org