How to verify a mitigation works
Verifying a mitigation means testing the control against the path the flaw needs, then re-checking it on a fixed interval. NIST SP 800-40 Rev. 4 makes these two steps. FedRAMP's 2026 rules set drift checks from every 7 days (Class D) to every 3 months (Class A), read 19 Nov 2026.
On this page
Two steps, and a record that shows both
NIST SP 800-40 Rev. 4, Section 2.2, lists step c, "Verify the risk response", and step d, "Continuously monitor the risk response". For controls, step c reads "For deploying additional security controls, ensure they are functioning as intended." Step d reads: "Make sure that the risk response continues to be in place: no one uninstalls the patch, deactivates the additional security controls, lets the cybersecurity insurance lapse, or restarts the decommissioned asset." A mitigation record with an install date and no check date has completed neither step.
The first check answers whether the control closes the path the flaw needs. It is not a check that the control exists. A rule that is installed in log-only mode exists and blocks nothing; a setting that was changed on one node of three exists too. So each check below reads the state that matters for that flaw, and each says who does it and how it usually fails.
The check for five real mitigations
These come from the CVE records in the site's data, read 19 Nov 2026. The mitigation is the vendor's or, where marked, a general control. The check reads configuration or logs on a system you own; none of it needs an exploit.
| CVE and mitigation | Verification step | Owner | How it usually fails |
|---|---|---|---|
| CVE-2026-104286, FortiMail: turn off the IBE service (Fortinet FG-IR-26-175) | In the GUI open Encryption > IBE and read IBE Service. In the CLI run show system encryption ibe, and confirm the status line reads disable. Then confirm from outside the trusted network that the web interface is not published, if that was the chosen option. | Messaging administrator | Status was changed on one appliance of a pair. A later configuration restore or upgrade brings back set status enable. Users who need IBE mail complain and someone re-enables it. |
| CVE-2026-0322, PAN-OS: limit the management interface to one jump box (Palo Alto) | Open Device > Setup > Interfaces > Management Interface Settings and read Permitted IP Addresses: one address, the jump box. Then open Network > Network Profiles > Interface Mgmt and list every profile that allows HTTPS or SSH, and which data-plane interfaces use it. | Firewall administrator | The management address is closed but an interface management profile on a data-plane interface still allows HTTPS. Panorama or a second firewall was not included. Palo Alto itself rates the jump-box case 8.5 instead of 8.6, so this control lowers risk only slightly. |
| CVE-2026-111497, Node.js and Brotli: reject inbound requests with Content-Encoding br at the proxy (researcher stopgap; the Node.js advisory lists no workaround) | From a test client you control, send a request with Content-Encoding br and an empty or harmless body to a staging endpoint, and confirm the proxy answers with an error before the application sees it. Check the proxy log for the rejected request. Confirm the rule is on every proxy node. | Platform or web operations | The rule covers inbound requests only. Brotli responses your own application fetches from other servers still reach the decoder. A node behind a different load balancer has no rule. |
| CVE-2026-19868, SonicOS SSLVPN: disable or delete every LDAP, RADIUS and TACACS+ server (SonicWall workaround of 9 Nov 2026) | Open Device > Users > Settings > Authentication, then Configure RADIUS, Configure LDAP and Configure TACACS+. In each Servers list, confirm none is listed and enabled. Then sign in once with each login type that should still work, such as the local admin. | Network security administrator | Someone re-adds a server for administrator logins and reopens the condition. Or the check stops at one of the three lists. |
| CVE-2026-21589, Atlassian Data Center: block parent-directory sequences with a WAF rule or the vendor's rewrite rule | For the rewrite option, check conf/server.xml for the RewriteValve line and WEB-INF/rewrite.config for the rule on every node (Bitbucket: the first rule in app/WEB-INF/urlrewrite.xml). Atlassian says to test that the rule blocks the encoded and double-encoded forms, so run that test on a staging copy and keep the test request and its forbidden or 404 response in the record. | Application administrator | One node of a cluster was not restarted after the edit, or a node was rebuilt from an older image. The rule matches only the plain form and not the encoded ones. |
A check that passes on a vulnerable build
For CVE-2026-111497 the obvious version check is wrong. ELLIO reports that process.versions.brotli shows 1.2.0 on both patched and vulnerable Node.js builds, because the fix was cherry-picked without changing the version string. The check that works is node --version on the host and inside each image, compared with 22.23.4, 24.22.1 or 26.12.1. A Dockerfile pinned to node:24.22.0 stays vulnerable until the pin moves. If Node.js links the system Brotli library, the check is the distribution package version instead. The lesson holds beyond this CVE: verify with the identifier the vendor says decides the outcome, and confirm it can tell a fixed system from an unfixed one.
Worked example: FortiMail IBE switched off on 1 Oct 2026, drift found on 8 Oct
Fortinet published FG-IR-26-175 on 1 Oct 2026 and said the flaw was exploited. CISA added CVE-2026-104286 to the KEV catalog the same day with a due date of 4 Oct. Fortinet's solution update of 5 Oct 2026 lists FortiMail 7.4.9 as the fixed 7.4 build. The record below is an illustration: the vendor dates are real, the organization's dates and the cause of the drift are assumed. The organization is not a federal agency, so the 4 Oct date is a target for it, not a directive. It runs FortiMail 7.4.8 and picks the weekly interval from the FedRAMP Class D ladder.
| Date | Event | Record entry |
|---|---|---|
| 1 Oct 2026 | IBE Service set to off on the appliance. Step c: status read as disable, and the web interface confirmed closed to the internet. | Change ticket, configuration export, name of the person who read the status, next check 8 Oct |
| 5 Oct 2026 | Fortinet's solution update lists 7.4.9 as the fixed 7.4 build. The upgrade is booked for the 10 Oct window and the mitigation stays until then. | Advisory revision read, upgrade booking, mitigation kept with removal tied to the upgrade |
| 6 Oct 2026 (assumed) | A configuration is restored from a backup taken before 1 Oct. Nobody reads the status. | None: the setting is silently enable again |
| 8 Oct 2026 | Interval check finds set status enable. The finding is opened as drift, IBE is turned off again, the restore procedure is amended, and the appliance logs are checked against the indicators Fortinet lists, because the flaw was exploitable while the setting was on. | Check date, result, drift finding, logs reviewed, next check 10 Oct |
| 10 Oct 2026 | Appliance upgraded to 7.4.9. Version read as 7.4.9 on the appliance. The IBE setting is no longer needed as a control, so it is returned to the state the business wants, and the mitigation is marked removed. | Upgrade date, version read, mitigation removal date 10 Oct, closure |
The 8 Oct row is the case the record exists for. Without the weekly check, the restore would have stood until the 10 Oct upgrade or the next audit, and the register would still have said "mitigated". The residual figure on the residual risk page is only as old as the last check, so the next check date belongs in the entry.
How often to re-check
FedRAMP's rule on drift is the only published ladder of intervals. It defines drift as "Changes to information resources that cause deviations from the intended and assessed state; common forms of drift include changes to configurations, deployed software, privileges, running processes, and availability." The table is for detection on resources likely to drift. Non-machine-based resources must be verified and validated at least every 3 months (VDR-TFR-NMV).
| FedRAMP class | Drift detection (VDR-TFR-PDD, SHOULD) | Machine-based status check (VDR-TFR-MVF) |
|---|---|---|
| Class A | At least once every 3 months | Not listed |
| Class B | At least once every month | SHOULD, at least monthly |
| Class C | At least once every 14 days | MUST, at least monthly |
| Class D | At least once every 7 days | MUST, at least monthly |
These are provider timeframes. An organization with no rule of its own can adopt the ladder and write the interval on the exception register. A KEV-listed flaw on an internet-facing system belongs at the weekly end.
Which exposure result settles the question
When the mitigation is meant to change exposure, the method matters. FedRAMP says prevention that intercepts, inspects or filters triggering payloads means that, "once this prevention is in place the vulnerability should no longer be considered an internet-reachable vulnerability". CISA's BOD 26-04 is stricter about proof: agencies "may use multiple overlapping systems to determine whether an asset or set of assets are publicly exposed", and "If one such method finds that a system is publicly exposed, the value for the decision point should be set to Yes." One method showing the asset closed does not outvote another showing it open. The control itself is covered under removing internet exposure.
Check without weakening the system
- FedRAMP rule VDR-CSO-MSP says providers "SHOULD NOT weaken the security of information resources to facilitate vulnerability scanning, detection, or assessment activities." Do not switch a rule to log-only or open a port to make it easier to test.
- NIST SP 800-115 says "Because of the risk of operational impact, it is recommended that an established incident response plan be in place during testing."
- For OT, SP 800-82 Rev. 3 recommends "A complete verification" to ensure the controls "are not affecting or degrading the performance and safety of the OT system".
Sources
- NIST SP 800-40 Rev. 4, Section 2.2, read 19 Nov 2026
- FedRAMP 2026, Vulnerability Detection and Response: VDR-TFR-PDD, MVF, NMV, CSO-MSP, read 19 Nov 2026
- FedRAMP 2026 definitions: Drift, read 19 Nov 2026
- FedRAMP 2026, Vulnerability Evaluation and Reporting, internet-reachable note, read 19 Nov 2026
- CISA BOD 26-04, Appendix A, read 19 Nov 2026
- NIST SP 800-115, Section 7, read 19 Nov 2026
- NIST SP 800-82 Rev. 3, Task I-1, read 19 Nov 2026
- Fortinet PSIRT FG-IR-26-175 (FortiMail), read 19 Nov 2026
- Palo Alto Networks advisory CVE-2026-0322, read 19 Nov 2026
- Node.js October 2026 security releases, read 19 Nov 2026
- ELLIO analysis of CVE-2026-111497, read 19 Nov 2026
- SonicWall PSIRT SNWLID-2026-0014, read 19 Nov 2026
- Atlassian advisory CVE-2026-21589, read 19 Nov 2026
- CISA KEV entry, CVE-2026-104286, read 19 Nov 2026
Last reviewed