Defensive runbook for active or suspected ransomware. You direct and document; the tech
executes every console action. Declare Critical on the first credible signal and keep a
timestamped log of every action — insurers, IR counsel and the postmortem run on it.
1. Isolate before investigating. Have the tech network-isolate affected endpoints from the
EDR or RMM console — host isolation, never power-off, which destroys volatile evidence.
Enumerate the devices, hand over a deep link for each, and set maintenance mode. No
isolation feature: physical or switch-level disconnection. Timestamp each.
2. Protect backups next, before scoping: disconnect or lock backup repositories from the
production network and disable backup credentials any domain account can reach —
operators target backups first. No restore yet.
3. Scope: hosts with encryption artifacts, shares touched, accounts active on those hosts,
earliest observed encryption timestamp. Check the preceding weeks of alerts — EDR
detections, odd sign-ins, disabled tooling; the intrusion predates the
encryption.
4. Engage per the client's documented IR plan and cyber-insurance carrier: most policies
require the carrier's approved IR firm to lead, and the wrong responder can void
coverage. Management makes that call — you package the facts.
5. Verify backups before restoring: test-restore a sample, confirm it predates the earliest
encryption timestamp, and scan it for the intrusion's artifacts — a backup taken after
initial compromise restores the attacker with the data.
6. Once the IR lead approves, sequence recovery: reset credentials, privileged first, then
every exposed account — assume theft; rebuild or clean-restore patient zero and
confirmed-compromised hosts rather than cleaning in place; restore from the verified-clean
point; re-admit hosts only after EDR confirms them clean. Preserve the isolated images as
evidence.
7. Client updates follow the defensive-writing-standard skill — factual, no blame or
speculation. Each: facts with timestamps, actions taken, what you need, next update time. Say "breach" only after confirmed system-level
findings and management sign-off; disclosure is management's, counsel's and the client's
policy's call. Hand the timeline to security-incident-postmortem; the ticket does not
close at containment.
Never contact or negotiate with the threat actor and never advise on paying — that is IR
counsel's and the insurer's alone; don't open ransom-note links or attacker portals. Nothing
is wiped, reimaged or cleaned up before the IR lead (or management, absent external IR)
approves — evidence outranks tidiness. No RMM: enumeration and maintenance mode are manual —
direct the tech and record it. No documentation system: ask management for the IR plan and
carrier.
Magic Library