Remediation vs mitigation in cybersecurity: a five-question tool that picks the response for one finding

Remediation vs mitigation in cybersecurity comes down to what is left afterward: remediation eliminates the vulnerability, mitigation leaves it in place and lowers exposure or impact. The tool below takes five answers about one finding and returns remediate, mitigate now and fix later, or accept with a dated exception, with the record to keep. Read 19 Nov 2026.

1Does a vendor fix exist today?
2Is the vulnerable component reachable from the internet?
3Is the CVE on the CISA Known Exploited Vulnerabilities catalog?
4Can the fix go in before its due date without breaking a critical system?
5Is a control already in front of the flaw, and has someone tested it?

Outcome for this finding

  1. Remediate
  2. Mitigate now, fix laterGuidance points here
  3. Accept with a dated exceptionAlso allowed

No answers yet, so each question reads the cautious way; the mapping table below shows how.

Rule 1: The vulnerability is exploited or severe and the vendor has not released a patch yet

NIST SP 800-40 Rev. 4 names emergency mitigation for this case and says to replace it with the permanent fix and remove the mitigation on a schedule. In the KEV feed, 791 of 1786 required actions mention mitigations, most often as 'apply mitigations per vendor instructions or discontinue use of the product if mitigations are unavailable'.

Record to keep: what the sources say (read 19 Nov 2026)

  • Section 3.4 maintenance scenarios, PDF page 20 SP 800-40r4 defines emergency mitigation as a temporary measure before a patch exists. Emergency mitigation. This is the emergency procedure in a crisis situation, like those described above for the emergency patching scenario, to temporarily mitigate vulnerabilities before a patch is available. Section 3.4 maintenance scenarios, PDF page 20
  • 3 more sources for this row
    • Section 3.5.3 The mitigation is meant to end: both dates are set and enforced. Once a permanent fix, such as a patch, is available, the patch will need to be deployed and the mitigation removed. Schedules should be set and enforced for both patch deployment and mitigation removal. Section 3.5.3
    • Section 2.1 SP 800-40r4 lists a missing patch as a reason patching is not immediately viable. A patch may not be available yet. Section 2.1
    • requiredAction field, catalog version 2026.11.18: 260 entries with this exact text KEV feed requiredAction text: the most common mitigation-or-discontinue wording, counted from the live feed. Apply mitigations per vendor instructions or discontinue use of the product if mitigations are unavailable. requiredAction field, catalog version 2026.11.18: 260 entries with this exact text

Record to keep: fields for mitigate now, fix later

The 13 situations, by where guidance points

solid: guidance points here light: also allowed empty: not named by the sources

The source behind each of the five questions

The tool takes one finding at a time. It runs in the browser and sends nothing. On load, with no answers, each question reads cautiously, Rule 1 matches, and the tool shows Mitigate now, fix later. The questions above are the ones published guidance asks, in the order that settles most findings fastest; each is listed here with its source.

  1. Question 1, vendor fix. NIST SP 800-40 Rev. 4 lists "A patch may not be available yet" as a reason a patch cannot go in immediately (NIST SP 800-40 Rev. 4).
  2. Question 2, reachability. Exposure is one of four inputs to CISA's BOD 26-04 Table 1, with KEV status, whether an adversary can automate exploitation, and technical impact (BOD 26-04).
  3. Question 3, KEV status. BOD 26-04 states that "Known exploited vulnerabilities are a frequent attack vector for malicious cyber actors", and makes KEV listing one of its four inputs; the clock starts when CISA adds the entry. The feed held 1,786 entries in catalog version 2026.11.18, read 19 Nov 2026 (BOD 26-04; KEV feed).
  4. Question 4, fix in time. SP 800-40 Rev. 4, Section 2.1, accepts waiting for an outage window, testing first, or updating interacting software (NIST SP 800-40 Rev. 4).
  5. Question 5, tested control. SP 800-40 Rev. 4, Section 2.2, step 3c, says to "ensure they are functioning as intended" when additional controls are deployed (NIST SP 800-40 Rev. 4).

The answers land on 6 of the 13 situations in the Decision Table: no fix yet, a KEV item that can be fixed in time, a non-KEV item that can, a control already in front, an unreachable flaw outside the KEV catalog, and a patch that breaks a critical system. The other seven are reached from the Decision Table directly: end-of-life software, SaaS that cannot be patched, a scheduled decommission, a PCI requirement that cannot be met, an OT or medical device, a federal agency that cannot meet a KEV date, and a false positive. The tool's usual answer is wrong in three cases. It does not ask whether exploitation can be automated, which BOD 26-04 treats as an input, so a non-KEV flaw with a public automated exploit needs a faster response than the tool returns. It does not ask whether you are a federal civilian agency, for whom the KEV date is binding. It does not ask what the product is, so end-of-life, SaaS and OT cases need your own reading of the matching rows in the Decision Table, listed at the foot of the rules table below. The rules table says where each rule assumes an answer. The method page explains how the answers pick a row.

One finding worked through

Four entries were added to the KEV catalog on 18 Nov 2026: Synology DiskStation Manager, BerriAI LiteLLM, Vanderbilt REDCap and Apache Axis. CVE-2024-10441, in Synology DiskStation Manager, has a due date of 21 Nov 2026, three days later, with forensic triage marked Yes. CVE-2026-40217, in BerriAI LiteLLM, has a due date of 2 Dec 2026, 14 days later, with forensic triage marked No. The REDCap and Axis entries are also due 21 Nov 2026 (KEV feed, catalog version 2026.11.18). The walk-through takes the Synology entry. The feed answers question 3; the other answers are assumptions that an asset owner replaces with facts.

Walk-through for CVE-2024-10441. Only the KEV facts come from the feed, read 19 Nov 2026.
QuestionAnswer usedBasis
1. Vendor fix exists?Yes (assumed)The entry's notes link Synology advisory Synology_SA_24_20; the owner confirms a fixed release exists
2. Reachable from the internet?Yes (assumed)The owner's architecture drawing
3. On the KEV catalog?YesAdded 18 Nov 2026, due 21 Nov 2026
4. Fix in before the due date?No (assumed)21 Nov 2026 is a Saturday; the next change window is assumed to be 24 Nov 2026
5. Tested control in front?Yes (assumed)A tested access restriction on the management interface

Rule 5 matches (fix yes, date missed, tested control yes), so the result is Mitigate now, fix later, on the row for a control already in front. The control goes into the record with its test date, the hosts it covers, the approver, and the date it comes off, which is after the fix lands on 24 Nov 2026. Had the control been untested, rule 7 would return the row for a patch that breaks a critical system, with the same response. The entry's required action is to apply mitigations per vendor instructions "or discontinue use of the product if mitigations are unavailable".

For a federal civilian agency the date stays 21 Nov 2026, and its KEV date row applies, not rule 5: CISA "does not issue waivers or exceptions for actions required in cyber directives" (implementation guidance). Outside that scope the KEV date is guidance, and FedRAMP providers have a separate rule, quoted in the section on the primary sources below. The dates page sets this entry's clocks side by side.

The three responses and the record to keep for each

Built from the evidence lists in the Compensating Controls Register and the fields of the Risk Acceptance Form. Read 19 Nov 2026.
ResponseWhat it meansRecord to keep
RemediateEliminate the vulnerability through patching, decommissioning the system, or another action (BOD 26-04 wording).Fixed version and date, the scan or test that no longer finds the flaw, the change ticket.
Mitigate now, fix laterA control reduces exposure or impact while the flaw stays in place.Control and mode (block or log-only), test date and tester, hosts and routes covered, approver, the date the control comes off and the date the fix lands.
Accept with a dated exceptionThe residual risk stays under a signed decision with an end date.Reason, residual risk against stated tolerance, controls in place, approver and date, expiry date, the event that reopens it.

The mitigate record is the one most often missing a date. NIST SP 800-40 Rev. 4, Section 3.5.3, says "Schedules should be set and enforced for both patch deployment and mitigation removal." The Compensating Controls Register lists the evidence to keep for each of 11 controls. The Risk Acceptance Form holds the 15 fields of an exception, and the exception register tracks them to expiry.

Remediation and mitigation in the primary sources

NIST's glossary gives mitigation as "A decision, action, or practice intended to reduce the level of risk associated with one or more threat events, threat scenarios, or vulnerabilities" (SP 800-160 Vol. 2 Rev. 1) and as "The temporary reduction or lessening of the impact of a vulnerability or the likelihood of its exploitation" (SP 800-216) (NIST glossary, mitigation). Mitigation lowers risk and leaves the flaw in place; remediation removes it. The glossary has both wordings.

CISA's BOD 26-04 revoked BOD 22-01 on 10 Jun 2026. It sets 3, 14 or 60 calendar days, or "fix on system upgrade", from exposure, KEV status, automation and technical impact, and defines remediation as the vulnerability being "eliminated through patching, decommissioning the system, or another action" (BOD 26-04).

FedRAMP's 2026 rules go further on vocabulary: "A fully mitigated vulnerability will still exist (with negligible risk) until it has been remediated." Providers should remediate KEV entries by the CISA due date even when fully mitigated, and the rules are required for Rev5 certification from 7 Dec 2026 (FedRAMP VDR). The difference guide sets the definitions of each source side by side.

What the tool cannot settle

Some decisions belong to a named party. In PCI DSS, "QSAs are responsible to determine whether a compensating control is sufficient to meet the intent of a requirement" (PCI SSC FAQ 1046). PCI SSC also separates two words: critical vulnerabilities under Requirement 6.3.3 "must be resolved within one month of release", while a vulnerability that is "addressed" may be handled with a compensating control or by disabling a vulnerable service (PCI SSC FAQ 1597). In federal systems, "the authorizing official is the only person who can accept risk" (NIST SP 800-37 Rev. 2, Task R-3). How long an acceptance may run is on the maximum durations page, and who may sign is on the who may sign page.

How the five answers pick a row

How the five answers pick a row of the Decision Table. Rules are tried in order and the first match wins. Vendor fix exists: not sure is read as no; Internet-reachable: not sure is read as yes; On the KEV catalog: not sure is read as yes; Fix in before due date: not sure is read as no; Tested control in front: not sure is read as no.
RuleVendor fix existsInternet-reachableOn the KEV catalogFix in before due dateTested control in frontDecision Table rowGuidance points toAssumes
1NoAnyAnyAnyAnykev-no-patch-yetMitigate now, fix laterNothing else is asked once there is no fix. The row is written for an exploited or severe flaw (its KEV example). For a finding outside the KEV catalog the response, mitigate now and fix later, is the same, but the row says nothing about a deadline: set the interim control and a dated plan from the finding's own exposure and impact, and record that it is not a KEV entry.
2YesAnyYesYesAnykev-patch-available-exposedRemediateThe row is written for an asset reachable from the internet. Not reachable gives the same response, remediate; BOD 26-04 Table 1 sets that clock.
3YesNoNoYesAnyinternal-asset-not-kev-moderateRemediateAssumes exploitation is automatable, which the tool does not ask. If it is not, Table 1 gives fix on system upgrade.
4YesYesNoYesAnykev-patch-available-exposedRemediateNo row covers a reachable flaw outside the KEV catalog. The nearest row is written for a KEV entry; the response is the same, remediate, and Table 1 sets the clock.
5YesAnyAnyNoYescompensating-control-in-frontMitigate now, fix laterMatches whatever the KEV and reachability answers are.
6YesNoNoNoAnynot-reachable-low-riskAccept with a dated exceptionAssumes an adversary cannot automate exploitation, which the tool does not ask.
7YesAnyAnyNoAnypatch-breaks-critical-systemMitigate now, fix laterCatches a missed change window with no tested control: the row lists "an outage window is not available yet" as a signal, so the closest row is the patch-breaks-a-critical-system row, which has the same response, mitigate now and fix later. A federal civilian agency with a KEV date it cannot meet goes to the KEV date row in the Decision Table; the tool does not ask who you are.

Rows the five questions do not reach: end-of-life-software; saas-cannot-patch; decommission-scheduled; pci-cannot-meet-requirement; ot-medical-cannot-patch; kev-due-date-cannot-be-met; false-positive.

Sources

Last reviewed