Offboarding a Windows Machine: The Decommission Checklist Your Fleet Actually Needs
Retire the identity before you wipe the disk: a six-step decommission checklist that keeps a forgotten laptop from becoming a live credential.
The machine nobody remembers to retire
Every fleet has one: the laptop in the storage closet, the desktop under a desk in a closed branch office, the VM nobody logged into for nine months. It still has a domain account. It still has cached credentials. It still has a VPN certificate. And because nobody ever formally retired it, every security control that depends on the machine being known quietly stops applying to it.
Onboarding gets a checklist. Offboarding a machine almost never does. That asymmetry is how a retired laptop becomes an attacker's quietest foothold — a device with valid credentials, no patching, no monitoring, and no owner.
Why a decommissioned machine is still a live asset
A machine stops being patched the day its owner leaves, but its identity does not expire on the same schedule. Consider what a forgotten machine still holds:
- Cached domain credentials — usable for offline attacks and pass-the-hash if the disk is recovered.
- Machine account and Kerberos tickets — valid until the account is disabled, not until the user leaves.
- VPN and Wi-Fi certificates — network access that outlives employment.
- BitLocker recovery keys in escrow — useful, but only if you can prove the disk was actually wiped.
- Local admin accounts — often with passwords set once, years ago, and shared across the fleet.
None of this is exotic. It is the ordinary residue of a machine that was set up carefully and retired casually.
The decommission checklist
Run this in order. Each step closes a door that the next step assumes is already shut.
- Freeze the machine first. Disable the computer account in Active Directory and revoke its certificates before you touch the hardware. This kills network access immediately, even if the wipe takes a week.
- Capture the evidence. Run a final audit and export the findings. If the machine is ever implicated in an incident later, this snapshot is your proof of what its posture was on its last day.
- Revoke, then remove. Revoke VPN and Wi-Fi certificates, remove the machine from your management and inventory tools, and delete its entries from any allowlists keyed to its hostname or MAC address.
- Wipe, don't just delete. A factory reset is not a wipe. Use a full-disk wipe or a BitLocker-encrypted drive reset followed by key destruction, and record the wipe certificate with the asset record.
- Close the asset record. Mark the asset retired in your inventory with a date, a reason, and the name of the person who signed off. An asset with no status is an asset that will be "found" again in six months.
- Verify from the outside. A week later, check that the machine account is still disabled, the certificates are still revoked, and no logon attempts have come from it. A retirement that nobody verifies is a hope, not a control.
Where this breaks at fleet scale
On one machine, the checklist above is an afternoon. On a hundred machines, it is a process — and processes fail in predictable ways. The two most common failures:
- The handoff gap. HR knows the employee left; IT learns about it two weeks later; the machine sits live in between. Tie the decommission trigger to the HR offboarding event, not to someone remembering to file a ticket.
- The partial retirement. The machine gets wiped, but the computer account stays enabled and the certificate stays valid. The hardware is clean; the identity is still armed. Step one exists precisely because step four tempts people to skip it.
The fix is boring and therefore effective: a single queue of machines pending retirement, reviewed weekly, where a machine leaves the queue only when every step is checked off. A machine that has been wiped but not revoked stays on the list. A machine that has been revoked but not wiped stays on the list. The list is the control.
How WinSentinel fits
A WinSentinel audit on the machine itself — free, full power on one machine — gives you the final posture snapshot for step two: cached credentials, local admin accounts, BitLocker state, certificate stores, and the rest of the 64 audit modules, all in one run before the machine is wiped. For a fleet, that per-machine snapshot is only useful if it rolls up somewhere: a central node that collects each machine's final audit, flags the machines still pending revocation, and keeps the retirement queue honest is fleet orchestration — the Pro side of the product. The audit is the evidence; the fleet view is what makes the queue closable at 100 machines instead of 10.
The one-line version
Retire the identity before you wipe the disk, verify the retirement a week later, and keep a queue until every step is done. A machine is not retired when it is gone — it is retired when nothing it ever held can still get in.