A vulnerability finding starts a decision process. Someone needs to establish whether it applies, judge its significance, decide what to do, and check the result. Keeping those decisions connected to the finding makes the work understandable to both the team fixing it and the people reviewing risk.
That process continues after the first rollout or audit. A new advisory can affect unchanged code. An application can become externally accessible. An exception can remain open after its original justification has disappeared. Current decisions matter as much as the original scan.
Give the initial backlog and new findings room to progress
Introducing scanners often reveals years of accumulated findings at once. Work through that backlog according to application impact and exposure, starting with the systems where a compromise would matter most. Confirm that the relevant scans completed and produced usable default-branch reports before treating a project's results as a basis for triage.
Within that context, critical findings deserve early attention, but severity alone doesn't establish the order of all work. An exposed credential or an actively exploited weakness may need immediate action while the wider backlog is still being reviewed.
Keep a route for newly reported vulnerabilities while bulk triage continues. Findings from a customer, penetration test, or public advisory can need attention before the next scheduled backlog review. Set first-triage targets that reflect risk and available response capacity, and examine whether the team meets them.
Make priority reflect the application
Read the finding and its supporting advisory, then examine how the affected code or component is used. Consider reachability, exploit availability and complexity, data sensitivity, service importance, and any compensating control. When applicability is uncertain, keep that uncertainty visible until it's investigated.
For example, the same vulnerable library can have different urgency in a publicly accessible service and a tool with no reachable path to the affected function. That difference needs supporting reasoning. A claim that a service is internal, or a low severity label, doesn't by itself establish that the risk is acceptable.
Give the work an owner and record why the chosen response fits. A dismissal needs an investigated reason; accepted risk needs someone authorized to accept it and a point for review. Where a mitigating control is part of that reasoning, revisit the decision when the control or the application's exposure changes.
Revoke or rotate exposed secrets. A source-code fix removes one copy of the credential, while revocation addresses the access it granted. Keep both actions visible when closing the finding.
Verify the fix in the software people use
Link remediation work to the finding, and check the resulting code and scan. GitLab's default-branch vulnerability view and merge request findings answer different questions: one describes the tracked code's findings, while the other helps review a proposed change.1 Confirm the branch or tracked ref being reviewed, especially when supporting multiple releases.
A fix merged into the default branch may still need deployment or a backport to a supported release. Decide which versions need remediation and how an urgent security release fits the normal release cadence. A finding that is no longer detected can also reflect changed scan coverage, so compare the fix and successful scan before concluding that the problem is resolved.
Look for variants of a confirmed weakness. A SQL injection fix may reveal other places that need parameterized queries, or a shared helper that would prevent the same mistake. Investigated false positives can guide scanner tuning. Both responses turn individual findings into improvements beyond the immediate ticket.
Vulnerability management from finding to closure explains how to put this into practice, including ownership, competing priorities, risk acceptance, and the difference between a merged fix and a protected release.
Keep the records needed for later reviews
Retain the owner, decision, linked fix, exception review, and relevant scan results so a later reviewer can follow what happened and why. Security policies can automate established rules,2 but automated changes to findings and exceptions still need review appropriate to their consequences.
Use each record for the question it answers. A compliance status report shows the latest checks against configured requirements;3 audit events record actions such as access changes.4 An SBOM describes dependencies, and provenance describes an artifact's build origin. Together with decision notes and deployment history, these records help investigate a release or assess continuing risk.
Check the projects, refs, dates, and successful scans behind a report. Child pipeline results, export limits, and artifact retention can affect what is available for later review. Keep or export the records required for that review before they become unavailable, and check that the exported data covers the intended period and scope.
Further Reading
Notes
-
The vulnerability report and integrated merge request security reports are available in Ultimate. Teams using scanner artifacts directly need another record of findings, decisions, and fixes. ↩
-
Security policies are available in Ultimate. They can apply common rules across projects; teams still need people to own findings and review exceptions. ↩
-
The compliance status report is available in Ultimate. It shows the latest results of checks against configured framework requirements. Check the scan time and requirements represented when using it in a review. ↩
-
Group and project audit events are available in Premium and Ultimate. Successful sign-in events are available on all tiers. ↩
- Triage and track findings in the vulnerability report explains status, filters, scope, and export.
- Remediate vulnerabilities connects findings with fixes and follow-up work.
- Configure vulnerability management policies describes automation for established handling rules.
- Interpret compliance status results explains scan timing, configured controls, and export limits.
- Use audit events in an investigation explains the available records and how to retrieve them.