RMM identity security · 12 minute read
RMM MFA recovery code security checklist
Recovery codes preserve access when an authenticator is lost. They can also become a durable alternate credential if scope, storage, factor replacement, detection, and revocation are not deliberately controlled.
Start with the right threat model
A recovery code is not automatically an MFA bypass.
NIST treats a saved recovery code as a fallback authenticator and says it should be stored securely offline. The risk comes from the complete flow: weak generation, storage beside the password, unlimited attempts, reusable codes, silent factor replacement, or recovery that survives offboarding.
Judge the design and the operator workflow together. A strong code stored in the same compromised account as the password does not provide useful separation.
Eight controls to test before privileged rollout
1. Model recovery codes as authenticators
A saved recovery code is a fallback authenticator, not an administrative note. Document exactly what it can do: sign in, replace another factor, reset MFA, create new codes, or regain a privileged session.
- Map each recovery action to the roles that can use it and the privileges obtained afterward.
- Confirm whether a code satisfies one factor or completes the entire recovery flow.
- Keep recovery capability out of routine help-desk and shared-administrator workflows.
2. Separate the code from the password and daily session
The fallback loses value when it is stored beside the password, synced into the same compromised account, or left in a technician's downloads folder. Treat storage location and access path as part of the control.
- Store break-glass material offline or in a separately controlled vault with audited access.
- Do not paste codes into tickets, chat, documentation, scripts, or browser notes.
- Test whether endpoint backup and file-sync policies accidentally copy downloaded code files.
3. Verify generation, hashing, throttling, and one-time use
Recovery secrets need adequate randomness, protected storage, attempt throttling, and clear invalidation. A code that can be reused or brute-forced becomes a durable alternate credential.
- Ask the provider how codes are generated, stored, compared, and rate limited.
- Use one code in a lab account and prove it cannot be used again.
- Regenerate the set and prove every older unused code stops working.
4. Protect factor replacement as a high-risk action
Signing in and replacing an enrolled factor are different risks. Require fresh verification before changing MFA, adding a passkey, disabling a factor, or generating a new recovery set.
- Require reauthentication rather than trusting only an existing browser session.
- Apply stronger review to privileged technicians, billing owners, and platform administrators.
- Record the old and new factor state without logging the secret itself.
5. Give administrators policy and revocation controls
An MSP needs to know whether recovery codes are required, optional, disableable, or enforceable by role. The control plane should make current exposure visible without revealing the codes.
- Inventory which technicians have recovery enabled and when their set was last replaced.
- Confirm an administrator can revoke recovery access after loss, departure, or suspected compromise.
- Make policy changes auditable and test them on a noncritical account before broad rollout.
6. Notify and investigate every recovery event
Recovery is rare and consequential. Successful use, failed attempts, regeneration, factor replacement, and administrative reset should create security evidence and an independent notification.
- Send notifications through a channel that does not depend only on the recovered session.
- Alert on unusual source, device, time, geography, repeated failure, and privileged-account recovery.
- Retain actor, account, source, event, result, and investigation outcome without retaining secrets.
7. Drill the break-glass path without weakening it
A recovery plan that has never been tested may fail during an outage. A plan tested carelessly may expose the fallback. Use a dedicated lab account and preserve only the result evidence.
- Run a scheduled drill with two authorized participants and a written stop condition.
- Verify sign-in, notification, factor replacement, code invalidation, and session revocation.
- Record the drill outcome, not the code, screenshot, QR image, or full secret-bearing file.
8. Revoke during offboarding and incident response
Disabling a user is incomplete if recovery material, active sessions, alternate authenticators, or delegated reset paths remain usable. Include recovery in both offboarding and compromise containment.
- Revoke sessions, recovery codes, passkeys, authenticator bindings, API tokens, and delegated reset rights.
- Confirm the old account cannot recover itself after technician removal.
- Reconcile every privileged identity and document exceptions with an owner and expiry.
Free worksheet
Record evidence, not the recovery secret.
The 24-control CSV records owner, evidence, result, exception, remediation, due date, and retest status across the full recovery lifecycle. It intentionally contains no field for a code, password, QR image, or other secret.
Relationship disclosure: Nizlo maintains this vendor-neutral, ungated worksheet. It can be used to review Nizlo or another RMM and does not certify a product or establish compliance.
Download the 24-control CSVSeven questions for a vendor review
| Control | Evidence question |
|---|---|
| Scope | What does one recovery code authorize after it is accepted? |
| Storage | Can the code remain outside the password, email, and endpoint-sync path? |
| Lifecycle | Is the code single-use, revocable, replaceable, and throttled? |
| Factor change | What fresh proof is required before another factor is added or removed? |
| Administration | Can policy and exposure be reviewed without revealing the codes? |
| Detection | Who is notified, and what evidence reaches the incident queue? |
| Exit | Does offboarding invalidate every recovery and session path? |
Test the identity boundary
Run the checklist on one controlled account.
Nizlo offers a 14-day trial without a credit card. Use an authorized noncritical account and endpoint to test MFA, session controls, least privilege, audit evidence, recovery, and offboarding before broader deployment.

