SOC 2 Vulnerability Management: From Scan to Verified Fix
Build a SOC 2 vulnerability management trail that shows scan coverage, triage, patch decisions, deployed fixes, and verification without inventing a universal SLA.

SOC 2 vulnerability management is the work of finding weaknesses in the systems that support your service, deciding what they mean, fixing or treating them, and checking the result. For a small SaaS team, start with the assets you operate, a repeatable scan and review schedule, and a record that follows each finding to a verified outcome. Keep the scanner’s dated output, including runs with no findings. A closed ticket alone cannot show whether a fix reached the affected production system.
What does SOC 2 say about vulnerability management?
The AICPA Trust Services Criteria address this in CC7.1. The criterion covers detecting configuration changes that introduce vulnerabilities and susceptibility to newly discovered ones. Its points of focus include infrastructure and software scans on a periodic basis and after significant changes, with timely action on identified gaps. Those points help explain the criterion; they do not name one scanning product, a universal monthly cadence, or a fixed remediation deadline.
Write a short operating rule for your actual service: which assets and software you cover, how you detect findings, who reviews them, how you set priority, when you act, who may accept a risk, and how you verify closure. Check customer contracts and the promises in your security answers before setting dates. Ask the CPA firm whether the proposed controls and records fit the agreed examination scope and period. The firm decides which evidence is sufficient for its work.
If you are still deciding whether to start the broader program, use the startup SOC 2 decision guide before setting control dates you cannot yet meet.
Set coverage before choosing a scan cadence
List the production assets and software that can affect the scoped service: public endpoints, cloud resources, hosts or containers you maintain, application dependencies, and relevant configuration. The asset inventory guide helps tie each component to the service and an owner. Then record the source that checks it, the expected interval or change trigger, and any part of the system the check cannot reach.
A dependency alert, cloud configuration check, and infrastructure scan see different things. Do not merge their counts into one vague claim that the service was scanned. For each source, retain its target list or filters, run time, result, and review. If a scheduled job fails, an agent is offline, or an asset is excluded, record the coverage gap and its repair. A report with zero findings is useful only after you know the intended assets and checks were in that run.
NIST’s Secure Software Development Framework recommends ongoing identification, confirmation, prioritization, and response to software vulnerabilities. It is engineering guidance, not an additional SOC 2 rule. Use it to test whether your discovery process still works when a new issue appears between planned scans.
Follow one finding from discovery to closure
Suppose a dependency check reports a vulnerable package used by the customer API. The alert is a lead, not yet proof that the deployed service is affected. An engineer checks the deployed version and reachable code path, then records the decision. If the package is affected, the work should have one reviewable trail:
| Record | What to preserve |
|---|---|
| Discovery | Source, scan or alert ID, run date, target, affected version, and original severity. |
| Triage | Service and asset, whether the issue applies, exposure, owner, priority, and reason. |
| Action | Fix or mitigation plan, due date, change or patch ID, and any approved exception. |
| Deployment | Build or patch version, production target, rollout date, and change record. |
| Verification | Follow-up scan or direct check of the same issue and target, result, reviewer, and date. |
A false positive still needs an explanation tied to the observed version or configuration. A mitigation such as disabling a route needs proof that the route changed and a plan to revisit the risk. A merged pull request does not prove that the deployed service changed. Link the production change evidence and verify the running target before closing the finding.
NIST’s patch management guide describes identifying, prioritizing, installing, and verifying updates. That last step matters when a patch needs a restart or a release did not reach every instance. Keep a retest result or another documented check that can show the risk response took effect.
Handle overdue work and exceptions honestly
Set remediation targets from exposure and capacity, then follow them. Do not copy a severity-to-days table from another company and call it an AICPA rule. If a fix will miss your own target, record when you knew, the affected assets, remaining exposure, interim safeguards actually in place, new owner and date, and the person authorized to accept that risk. The policy exception request guide shows how to keep a time-limited decision separate from an auditor’s testing result. A later approval cannot rewrite a missed deadline as timely work.
Review open findings and expiring exceptions on the cadence your policy sets. Preserve the list used for each review and the decisions made. For a proposed Type 2 period, keep the run history and finding population across the period, including failed or missed scans, so the CPA firm can select samples from a complete set. The evidence examples guide shows how to connect source artifacts to management’s operating records.
Keep the source facts and the program record separate
FileGRC is a Git-native GRC workspace for SOC 2 work. JSON holds structured records, Markdown holds long-form work, and Git supplies the change history. A team can link the scoped asset, control, risk decision, review, and fixed evidence reference so an agent can find an overdue item or missing verification in the records it can access. Starter records are proposals, not claims that scans or reviews already happened.
The scanner, package registry, cloud service, issue tracker, and deployment system remain the sources for detection and action. FileGRC does not scan, patch, or decide that a vulnerability is fixed. Keep exploit details, secrets, and personal data that may need erasure out of Git; use a restricted evidence location and a safe reference when a raw report is sensitive. Your team makes the risk and closure decisions. The independent CPA firm performs the SOC 2 examination and judges evidence sufficiency.
Run your SOC 2 program as files in Git.
Keep policies, controls, work, and evidence indexes in a repository your team and agents can inspect. Add optional FileGRC Autopilot email and Slack reminders to keep work moving.
Frequently asked questions
Does SOC 2 require vulnerability scans?
Not as a universal standalone rule. AICPA criterion CC7.1 requires detection and monitoring of configuration changes that introduce vulnerabilities and susceptibility to newly discovered ones. It lists periodic and post-change infrastructure and software scans as points of focus, but the AICPA says not every point must be addressed. Document how your controls meet the criterion for your scoped system and risks.
How often must we scan for SOC 2?
The Trust Services Criteria do not set one universal scan interval. Set and follow a cadence based on your systems, rate of change, risk, policy, and customer commitments. Record significant-change triggers and discuss your proposed controls with the independent CPA firm. A scheduled run that failed or skipped assets needs follow-up, not a clean-scan label.
Does SOC 2 set a remediation SLA for critical vulnerabilities?
The Trust Services Criteria do not prescribe one number of days for every finding. Management should set risk-based targets it can operate, check contractual promises, and document priority, owner, due date, treatment, and any overdue decision. An approved exception does not erase a missed control or guarantee the CPA firm will accept it.
What evidence shows that a vulnerability was fixed?
Keep the finding and affected asset, the risk decision, change or patch record, production deployment date, and a verification result that checks the same weakness on the affected system. A closed ticket or merged pull request alone does not prove that a fix reached production or worked there.
Is a clean scan enough vulnerability management evidence?
No. A clean scan supports one run when you can show its targets, date, checks, and review. To show the process operated over the relevant period, keep the scheduled run history, coverage gaps, findings from all detection sources, triage decisions, fixes, and verification. A zero-finding report says little if a production asset was missing or the run failed.
Does FileGRC run vulnerability scans or patch systems?
No. FileGRC is a Git-native GRC workspace for SOC 2 work. It can organize control, risk, review, and evidence records in JSON and Markdown with Git history. Scanners and source systems detect weaknesses and show patches or deployments; people make risk decisions, and a CPA firm performs the examination.