Skip to content

Vulnerability Remediation: How to Prioritize, Fix and Prove It

A practical vulnerability remediation guide: how to prioritize with KEV, EPSS and exposure, set fix SLAs, run the workflow, and prove the fix with a retest.

Omar KhandakerBy , Senior Security ConsultantUpdated 13 min read

Vulnerability Management ServicesRecurring program from $1,500 per validated scanFixed scope, free re-scan, no hourly billing.

Get my fixed price

Vulnerability remediation is the process of prioritizing, fixing and verifying the security weaknesses found by a scan or a penetration test, so that each one is removed, reduced or formally accepted. In practice you have a report with more findings than you can fix at once, and remediation is the work that turns that list into a fix order, closes the issues that matter first, and proves they are actually closed. This guide covers how to prioritize when everything is labelled critical, how to set fix timelines you can defend to an auditor, the six-step process from finding to verified fix, how to remediate a single CVE, and a remediation plan you can copy. The last step, proving the fix, is where a free retest earns its place.

Remediation, mitigation and acceptance

Three responses close a finding, and the report should say which one applies:

  • Remediation removes the vulnerability: apply the patch, change the configuration, rewrite the code, or decommission the system. This is the goal.
  • Mitigation reduces the risk without removing the flaw: a firewall rule, a WAF signature, network segmentation, or disabling a feature. It buys time when a fix is not yet available.
  • Risk acceptance is a documented decision to live with a finding, with a named owner, a reason, a compensating control and a review date. It is a legitimate answer, but only in writing, never by silence.

A finding “closed” because everyone stopped looking at it is the one that comes back next quarter, usually reachable from a new feature and re-rated higher.

Why CVSS is the wrong sort order

Most reports arrive sorted by CVSS score, and most teams start at the top and work down. That wastes the first week. CVSS measures technical severity, not your risk. The specification is explicit: consumers “should enrich the Base metrics with Threat and Environmental metric values specific to their use of the vulnerable system to produce a score that provides a more comprehensive input to risk assessment” (FIRST, CVSS v4.0 specification). CVSS v4.0, released in 2023, keeps the familiar bands (Critical 9.0 to 10.0, High 7.0 to 8.9, Medium 4.0 to 6.9, Low 0.1 to 3.9) and asks that a score built from Base metrics alone be labelled CVSS-B, so readers can see it carries no threat or environmental context. Most published scores, including those in the National Vulnerability Database, are CVSS-B.

A CVSS 9.8 on an internal lab box with no data behind it can wait. A CVSS 7.5 on an internet-facing service that is being exploited right now cannot. Sorting by the base score alone puts them in the wrong order.

Prioritize on exploitation and exposure

Rank findings on four signals, in this order, and the real work list falls out of a long report:

  1. Active exploitation. Is the CVE on CISA’s Known Exploited Vulnerabilities catalog? A KEV entry means it is being exploited in the wild, which moves it to the top regardless of CVSS. CISA’s own guidance is that organizations “should use the KEV catalog as an input to their vulnerability management prioritization framework,” federal or not.
  2. Likelihood of exploitation. The Exploit Prediction Scoring System (EPSS) “estimates the probability that a published CVE will be exploited in the wild in the next 30 days” and publishes a fresh 0-to-1 score for every CVE daily. A high EPSS score on a not-yet-KEV vulnerability is an early warning.
  3. Exposure. Internet-facing systems before internal ones. An attacker reaches the perimeter first.
  4. Asset value. Systems that hold regulated data, credentials or payment data before the rest. An exploitable medium on the server that stores card data deserves a slot ahead of a high on an isolated test box.

Federal practice now formalizes exactly this logic. CISA’s Binding Operational Directive 26-04, issued June 10, 2026, sets remediation deadlines from four factors: whether the asset is publicly exposed, whether the CVE is on the KEV catalog, whether an adversary can automate the exploit, and whether exploitation gives partial or total control. It supersedes and revokes BOD 22-01 (and the older BOD 19-02), whose KEV deadlines applied regardless of where the asset sat. The directive binds only federal civilian agencies, but its table is a sound model for your own SLAs.

Set fix timelines you can defend

Put target fix times in writing, by priority, before the fixes start, so the retest can measure against them and an auditor can see the policy. If your organization is not bound by BOD 26-04, this simpler table, informed by the same risk signals, is a workable alternative to the federal model:

Priority Example fix target Trigger
Emergency 72 hours On the KEV catalog and internet-facing, or exploitation confirmed against you
Critical 15 days High CVSS with real exposure, or a high EPSS score
High 30 days Exploitable, limited exposure
Medium 90 days Real but low-exposure or low-impact
Low Next maintenance cycle Minimal risk, documented

Compliance frameworks set their own floors. PCI DSS Requirement 6.3.3 requires patches for critical vulnerabilities to be installed “within one month of release,” and the clock starts at patch release, not when you opened the ticket (PCI Security Standards Council, PCI DSS v4.0.1). The CIS Controls ask for “a risk-based remediation strategy documented in a remediation process, with monthly, or more frequent, reviews” (Safeguard 7.2); our CIS Controls guide covers the rest of Control 7. CISA’s BOD 26-04 table is the more granular federal model: its deadlines run from 3 days plus forensic triage, for KEV-listed flaws that give total control and are publicly exposed or automatable, through 3, 14 and 60 days, down to “fix on system upgrade” for flaws that are not on KEV, sit on internal assets and cannot be exploited automatically, whatever their impact. Our guide to risk-based vulnerability management maps all 16 combinations. NIST frames the whole activity as routine preventive maintenance rather than firefighting (NIST SP 800-40 Rev. 4). Whatever numbers you choose, write them down and apply them the same way to every finding.

The vulnerability remediation process in six steps

Prioritization tells you what to fix. A repeatable process gets it fixed and keeps it fixed. The same six steps work for a scan report, a penetration test report or a single urgent advisory.

Flow diagram of the vulnerability remediation process in six boxes. First validate that the finding is real, reachable and exploitable, then deduplicate and group findings that share one fix, then give every finding an owner and a due date from the SLA table. Next fix, mitigate or formally accept the finding, verify the fix with a retest and record the result, and finally track remediation velocity and recurring findings every quarter. A closing note says a closed ticket is not a closed vulnerability.
  1. Validate before you assign. Confirm the finding is real, reachable and exploitable in your environment. Version-based scanner detections flag patched-but-backported packages and disabled features; a meaningful share of “critical” scan findings do not survive validation. Assigning false positives burns the goodwill you need for the real ones. Our network vulnerability assessment checklist covers how to validate raw output.
  2. Deduplicate and group. The same outdated OpenSSL build reported on every web server is one change to the base image, so raise it as one task with a host list, not a ticket per server. Then group into projects: patch cycles, configuration baselines, decommissioning and segmentation each close whole classes of findings. A list sorted by score hides that structure; grouping reveals that ten findings share one fix.
  3. Assign an owner and a due date. A finding with no owner is a finding that reappears. The due date comes from the SLA table.
  4. Fix, mitigate, or formally accept. Apply the fix through your normal change management process, put a compensating control in place while you wait, or document an accepted risk with a review date.
  5. Verify. Retest the fix and record the result. Unverified fixes are how “closed” findings survive into the next report.
  6. Trend it. Track remediation velocity and recurring findings quarter over quarter; our guide to vulnerability management metrics defines the numbers worth reporting. A finding that keeps returning is a process problem, not a patch problem, and it belongs in your work between penetration tests.

CVE remediation: fixing a single known vulnerability

When the finding is one CVE, the process above compresses into a short checklist:

  1. Read the vendor security advisory, not only the scanner text. The advisory names the affected versions, the fixed version and any conditions the flaw needs, such as a feature being enabled.
  2. Confirm you are affected. Check the installed version with an authenticated check. Linux distributions often backport a fix into a package that keeps the old upstream version number, so a banner can say vulnerable when the host is patched.
  3. Check exploitation status. Look the CVE up in the KEV catalog and check its EPSS score. That sets the deadline from your SLA table.
  4. Apply the fixed version where one exists, on a test system first if the service is critical.
  5. If no patch exists yet, mitigate. Apply the vendor’s workaround, disable the affected feature, restrict network access to the service, or put a virtual patch in front of it: a WAF or IPS rule that blocks the exploit request. Record it as a mitigation with a review date, because the flaw is still there.
  6. Verify and record. Re-scan or retest, attach the evidence, and close the entry.

CVE mitigation and CVE remediation are different outcomes. A workaround or virtual patch is mitigation: it buys time. The CVE is remediated only when the vulnerable code is updated or removed. If you want a decision model in place of a score, CISA’s Stakeholder-Specific Vulnerability Categorization (SSVC) sorts each vulnerability into one of four decisions (Track, Track with closer monitoring, Attend or Act) from its exploitation status and impact.

Remediating findings that have no patch

Many penetration test findings have no CVE and no vendor fix, because the flaw is in your own code or configuration:

  • Business logic and authorization flaws. The fix is a code change, such as a server-side check that the record belongs to the caller. Ask the tester for the exact request that reproduces the issue, add a regression test, and have the same request retested.
  • Misconfigurations. Open storage, default credentials, weak TLS settings and over-broad permissions are fixed in the configuration baseline or the infrastructure-as-code template, not on one host, or they return with the next deployment.
  • Design weaknesses. A flat network or a shared administrator password is a project, not a ticket. Put a compensating control in place now, then schedule the project with an owner and a date.

Each of these still gets an owner, a due date and a verified result in the plan below.

A remediation plan you can copy

Track every finding in one table so nothing drops. This is the kind of record an auditor or a customer’s security team expects to see as evidence of a remediation process:

Field Example
Finding ID VULN-014
Description Unauthenticated RCE in perimeter appliance (CVE listed on KEV)
Priority Emergency
Asset and exposure edge-gw-01, internet-facing
Owner Network team
Response Remediate (vendor patch)
Due date Within 72 hours of report
Status Fixed
Verified Yes, confirmed on retest (date recorded)
Evidence Retest screenshot, version confirmed

The two rows that matter most are Verified and Evidence. Without them, the plan records intentions, not outcomes. Our vulnerability assessment report template has a longer version with a worked example, and the penetration testing report template shows how the retest results attach to the original findings.

Proving the fix: the retest

The remediation is not done when the ticket is closed. It is done when someone confirms the vulnerability is gone and did not take a working feature or a new gap with it. A proper retest re-runs the original test against the fixed system, confirms the finding is closed, and checks for regressions the fix may have introduced. The output is the evidence auditors, customers and cyber insurers actually ask for: a report that shows each finding moving from open to fixed, with proof.

This is where a scan-and-report vendor and a testing firm diverge. A scanner can rescan, but it cannot tell you whether the authorization flaw it never found in the first place is now fixed. Every Invadel penetration test includes a free retest of remediated findings, and the report is reissued to show them closed, so the fix is proven, not assumed. You can request it any time within two months of report delivery. See the format on the sample report page, and for the scanning side of an ongoing program, our managed vulnerability scanning validates results and re-scans fixes from $1,500 per scan, with the device tiers on the vulnerability scanning pricing page.

Where remediation fits the bigger picture

Remediation is one phase of vulnerability management, which also covers discovery, assessment and continuous monitoring; our guide to building a vulnerability management program covers the whole cycle. If you would rather have the cycle run for you, our managed vulnerability management service repeats validated scanning on a schedule and tracks each finding to a verified fix. If you are choosing between a scan, an analyst-led vulnerability assessment and a penetration test, our guide to penetration testing vs vulnerability scanning explains which produces the findings you then remediate. The stronger the input, the more useful the remediation plan: a validated, prioritized report is a work order, while a raw scanner export is a to-do list nobody can trust.

Frequently asked questions

What is vulnerability remediation? The process of prioritizing, fixing and verifying the vulnerabilities found in a scan or a test. Fixing removes the flaw, mitigating reduces its risk, and accepting is a documented decision to live with it. The verification step proves the fix worked.

How should we prioritize which vulnerabilities to fix first? By exploitation and exposure, not CVSS alone. Put KEV-listed and actively exploited findings on internet-facing systems at the top, then use EPSS and asset value to order the rest. CVSS is one input, not the sort order.

What is a reasonable remediation SLA? Set targets by priority and write them down: emergency findings in days, critical in about two weeks, high in a month, medium in a quarter. PCI DSS requires critical patches within one month of release, so align to your strictest framework.

Do we have to retest after remediating? For assurance and for most compliance evidence, yes. PCI DSS Requirement 11.4.4 requires exploitable findings from a penetration test to be corrected and the penetration test repeated to verify the corrections. A retest confirms the fix and catches regressions, and ours is included in the fixed price. Our PCI DSS penetration testing page covers the requirement in full.

How is remediation different from mitigation? Remediation removes the vulnerability; mitigation reduces the risk while the flaw remains, such as a firewall rule or a disabled feature. Mitigation is a bridge to a fix, not a substitute for one.

What are the steps in the vulnerability remediation process? Six: validate the finding, deduplicate and group findings that share a fix, assign an owner and a due date, fix or mitigate or formally accept, verify with a retest, and track the trend over time.

How do you remediate a CVE when no patch is available? You mitigate until one is: apply the vendor’s workaround, disable the affected feature, restrict access to the service, or add a virtual patch such as a WAF or IPS rule. Log it as a mitigation with a review date, and apply the fixed version when it ships.

Who is responsible for vulnerability remediation? The team that owns the asset fixes it, and the security function sets the priority and verifies the result. If IT is outsourced to a managed service provider, the provider usually applies the patches, but the risk decisions, the SLA and the acceptance sign-off stay with you, so write the fix timelines into the contract.

The short version

Sort by exploitation and exposure, not by CVSS. Write down fix SLAs by priority and apply them consistently. Run the finding through validate, deduplicate, group, assign, fix and verify, and track every one in a plan whose most important columns are Verified and Evidence. Then prove the fix with a retest, because a closed ticket is not a closed vulnerability. If you want findings you can actually act on, delivered with a free retest that proves the fixes, scope an engagement.

Omar Khandaker

Written by

Omar Khandaker, Senior Security Consultant

11+ years in the field. Focus: PCI DSS, ISO 27001, ISO 9001 & CMA. Part of the team that scopes, tests, and reports for our clients. Full profile →

Find out what an attacker sees.

Tell us what to test and see your fixed price.

Prefer the full scoping questionnaire? 
Start the conversation

Get a quick proposal

We are online