Internal sysadmins · June 27, 2026
Caisey for Incident Response: How to Quickly Isolate and Investigate a Compromised Endpoint
When a workstation starts behaving suspiciously—unexpected network connections, strange processes, or file changes—the clock starts ticking. As a sysadmin, you need to investigate quickly without tipping off the attacker or destroying evidence. Traditional remote access tools like RDP can trigger alerts, leave obvious footprints, and lack the granular control needed for a clean investigation. Caisey offers a different approach: approval-gated actions, a full audit trail, and a remote troubleshooting console that lets you work quietly and methodically.
The Problem with Traditional Remote Access During an Incident
Using RDP or SSH to jump onto a potentially compromised machine is risky. The attacker might have keyloggers or session monitors that detect your connection. Even if you use a jump box, the act of logging in can modify system state—event logs, registry keys, and running processes. This can alert the adversary or contaminate evidence. Worse, if you need to run commands that require elevated privileges, you might have to enter credentials directly on the machine, exposing them to any malware present.
Standard remote support tools are built for convenience, not forensic integrity. They often lack a detailed record of every action taken, making it hard to prove what was done during an investigation. For internal IT teams handling security incidents, this gap is critical.
Why Caisey Fits Incident Response Workflows
Caisey is a remote troubleshooting console designed for MSPs and internal IT teams. It uses headless AI runtimes on enrolled devices, coordinated through a Cloudflare Worker control plane. For incident response, three features stand out:
- Approval-gated actions: Every sensitive command or script requires explicit approval from a designated team member. This prevents accidental or unauthorized changes during an investigation.
- Persistent audit trail: All sessions, commands, outputs, and approvals are recorded in SQLite Durable Objects. You get a tamper-evident log that can serve as chain of custody.
- Endpoint connectivity via bridges: Caisey connects to enrolled endpoints through secure bridges, not direct network access. This reduces the attack surface and avoids leaving obvious network traces.
These capabilities make Caisey a practical tool for investigating a compromised machine without alerting the attacker or losing evidence.
Step-by-Step: Investigating a Suspicious Workstation with Caisey
Let’s walk through a realistic scenario. You receive an alert from your EDR about unusual outbound traffic from a Windows workstation in accounting. The machine is enrolled in Caisey. Here’s how you proceed.
1. Initiate a Troubleshooting Session
From the Caisey cloud UI, you select the target endpoint and start a session. The session is established through a bridge—no direct network connection to the machine. The user on the endpoint sees a permission prompt (if configured) but you can also initiate silently if policy allows. For incident response, silent initiation is often preferred to avoid alerting a potential attacker who might be watching the user’s screen.
2. Run Initial Reconnaissance Commands
You need to see what’s running. Using Caisey’s remote terminal (PowerShell or bash), you run:
Get-Process | Where-Object {$_.CPU -gt 10} | Select-Object Name, CPU, StartTimeThis command lists processes with high CPU usage. Because this is a potentially invasive action, Caisey’s approval gate triggers. A notification goes to a second team member (e.g., the security lead) who reviews the command and approves it. The attacker never sees this approval workflow—it happens in the cloud UI. Once approved, the command runs on the endpoint and the output streams back to your console.
3. Check Network Connections
Next, you want to see active connections. You run:
Get-NetTCPConnection | Where-Object {$_.State -eq 'Established'} | Format-Table LocalAddress, RemoteAddress, RemotePortAgain, approval is required. The audit trail logs the command, the approver’s identity, the timestamp, and the full output. This creates a verifiable record of your investigation steps.
4. View Key Logs
You need to check recent security events. You run:
Get-WinEvent -LogName Security -MaxEvents 50 | Where-Object {$_.Id -in (4624,4625,4634)} | Format-Table TimeCreated, Id, MessageThis pulls logon events. Caisey’s session history preserves the exact command and output, so you can later correlate findings with other evidence.
5. Isolate the Machine (If Needed)
If you confirm compromise, you might want to isolate the endpoint from the network. Caisey can run a script to disable the network adapter or modify firewall rules—again, only with approval. The audit trail captures the isolation action, providing proof that it was authorized and executed at a specific time.
Preserving Chain of Custody with Audit Trails
One of Caisey’s strongest features for incident response is the full audit trail. Every action taken during a session is recorded: who initiated the session, who approved each command, the exact command text, the output, and the timestamp. This log is stored in durable, replicated storage (SQLite Durable Objects) and can be exported for forensic analysis.
For internal investigations, this audit trail serves as chain of custody. If you need to present evidence to management or legal, you can show exactly what was done and who authorized it. No more relying on screen recordings or manual notes.
Approval Gates Prevent Unauthorized Actions
In a high-stress incident, it’s easy to make a mistake—run a command that deletes evidence, or change a setting that breaks the machine. Caisey’s approval-based remote support adds a safety net. By requiring a second person to approve each action, you reduce the risk of unintended damage. This is especially valuable when multiple team members are involved in the response.
The approval workflow is flexible. You can set policies that require approval for all commands, or only for specific categories (e.g., registry edits, service changes). During an incident, you might temporarily tighten the policy to require approval for everything, then relax it after containment.
Conclusion
Investigating a compromised endpoint requires speed, precision, and careful evidence handling. Caisey gives internal sysadmins a remote troubleshooting console that respects those needs. With approval-gated actions, a persistent audit trail, and secure bridge connectivity, you can isolate and investigate without tipping off the attacker or losing forensic integrity. Next time you face a suspicious workstation, consider using Caisey—not just for remote support, but as a first-response tool for incident response.
For more on how Caisey handles remote diagnostics, see our guide on remote endpoint diagnostics.