When Can You Close a Security Finding?
By Chuck Lerch, Founder and CEO, Cyberuptive ·
Close a security finding as fixed when a dated check shows that the original problem no longer exists across the agreed scope. If the problem remains and an authorized owner chooses to live with it, record risk acceptance instead. If the evidence is incomplete, keep the unverified part open.
Consider a provider report that says, “External sharing finding closed.” Did the provider verify that an outside user can no longer open the files? Did a partner approve continued access for a deal? Or did the provider stop monitoring the workspace? Ask for the record that distinguishes those answers before counting the finding as fixed.
What are the key takeaways?
- Write down what must be true before work begins, including which assets the finding covers.
- Match the retest to the original problem, not just to the change someone made.
- Give untested assets their own open record instead of treating them as successes.
- Keep verified fixes and accepted risks separate in the report that goes to leadership.
What is the difference between closing a finding and accepting risk?
Closing a finding as fixed is a statement about the system. Accepting risk is a decision about what the business will tolerate. A ticketing system may let you finish either workflow with a “closed” status. Ask for a separate field that says which decision was made.
The National Institute of Standards and Technology (NIST) published its Cybersecurity Framework (CSF) 2.0 on February 26, 2024. Outcome ID.RA-06 calls for choosing, prioritizing, planning, tracking, and communicating risk responses. GV.RM-02 addresses risk appetite and tolerance; GV.RR-02 addresses roles and authority. The framework describes outcomes, not a mandatory set of ticket labels or an approval form (NIST Cybersecurity Framework 2.0).
Use the following decision table as a proposed reporting convention. Apply it to the original finding, even if an implementation task or a provider support request is already complete.
| What the reviewer can show | Decision to record |
|---|---|
| The agreed condition holds across the affected scope. | Verified fixed. Attach the retest and identify the reviewer. |
| The condition still fails, and an authorized business owner has approved the remaining exposure. | Risk accepted. Link the decision; do not count it as a fix. |
| Some assets passed; others failed or could not be checked. | Partially remediated. Keep the remaining scope, owner, and next action visible. |
| The original observation was incorrect, or the record duplicates another finding. | Correct the record with evidence or link the controlling finding. Do not claim remediation. |
What evidence should a provider deliver before calling a fix complete?
Ask for a short record that connects the original observation to a new result. A change receipt answers “What did we do?” The retest should answer “What can we now show that we could not show before?”
- Original condition. Describe the unwanted behavior, the affected systems, and how it was observed. Keep identifiers that will still make sense after a host is renamed or a ticket is moved.
- Pass condition. State what result would contradict the original observation. Agree on acceptable coverage and name any boundaries before looking at the new results.
- Change evidence. Identify who changed what and when. Include a reference to the approved change rather than copying unrelated sensitive data into the ticket.
- Verification evidence. Identify the method, date, scope reached, result, and reviewer. Distinguish an asset that passed from an asset the check could not reach.
- Disposition. Explain the decision in one sentence. Link any remaining work or risk decision so a reader need not infer it from a status color.
NIST’s January 2022 control assessment publication provides procedures that organizations can tailor to their risk tolerance, along with guidance on planning assessments and analyzing results (NIST Special Publication 800-53A Revision 5). The checklist here is our recommendation, not a claim that NIST requires five fields. For a finding about permissions, choose a permissions check. A deployment receipt on its own cannot tell the reviewer what access the test user retained.
How would a U.S.-based SOC verify a partly fixed finding?
In a U.S.-based security operations center (SOC), we would start with the same access question that opened the finding, then account for every workspace in scope. Take a hypothetical finding: a test identity outside the firm could read a restricted sample document in each of three workspaces. No actual customer records are needed for this exercise.
Suppose the administrator disables public sharing at the organization level. In the first workspace, the outside test identity can no longer open the approved sample, while the intended internal user still can. Record the identities by role, the workspace, the date, and both results. The internal check helps distinguish a working access restriction from a document that nobody can open.
In the second workspace, a direct guest permission still lets the test identity read the sample. Record the failed check and assign the permission review. In the third, the test stops at an unavailable identity service. Record “not tested,” identify the dependency, and schedule another check. An error message is not a successful denial test.
Our summary to the buyer would be: one workspace verified, one still exposed, one unverified. If the business owner approves the guest access for a defined purpose, link that decision to the second workspace. It changes the disposition of that exposure, not the evidence for either of the others. Our guide to security exceptions covers what to record about that separate decision.
What if the finding disappears from the dashboard?
Ask whether the system passed the check or simply stopped returning results. Compare the current inventory and collection status with the original scope. If the team cannot account for an asset, keep it in the unverified group until someone can explain its absence.
Consider a retired server. Ask for the inventory identifier, evidence of decommissioning, and confirmation that a replacement did not inherit the same problem. If the original exposure ended through retirement, say so. Do not tell leadership that a patch was verified on a server that no longer exists. For a live asset, agree with the vulnerability scanning team on how to distinguish a clean result from a missed or incomplete check.
What should buyers ask to see at a service review?
Ask the provider to trace one recently closed finding from the first observation to the approval. Then choose one accepted risk from the same reporting period. If the executive report counts both as fixes, request separate totals and definitions before using the numbers to judge progress.
Have the provider show where unverified assets appear, who can change a disposition, and what happens when a later test contradicts a closure. Ask the business owner to explain the remaining exposure in plain English. If the decision requires authority the owner does not have, record the escalation rather than treating an unanswered message as consent.
Before signing, use the managed SOC pricing guide to frame a practical scope question: does the service include retesting and reporting exceptions, or only identifying findings? Specify who arranges the retest when another supplier makes the change. A useful handoff names the person who will bring back the result, not just the team that received the ticket.
What else should buyers and operators know?
Can an engineer approve their own fix?
Separate evidence of implementation from the decision to close. Set review authority according to the importance of the control and your policy. For a consequential finding, ask someone other than the implementer to reproduce or review the result.
Can a compensating control count as remediation?
Compare it with the agreed pass condition. If the alternative meets the control objective across the scope and the reviewer verifies it, describe exactly what was fixed. If it only reduces exposure while the original condition persists, document the remaining risk and obtain the appropriate decision.
Does NIST CSF 2.0 require a particular closure deadline?
The framework describes outcomes and does not prescribe how to achieve them. Set deadlines using your own obligations, exposure, and authority; do not treat a provider’s default ticket timer as a deadline supplied by the framework (NIST Cybersecurity Framework 2.0).
When should a closed finding be reopened?
Reopen it, or link a new finding, when credible evidence contradicts the recorded result. Preserve the earlier decision, its date, and its scope. A new failure on a different asset deserves investigation without rewriting what the earlier test established.
What is the next step for your organization?
Bring your scope, current evidence, and unresolved responsibilities to a focused discussion. Schedule a security review or Call 833-92-CYBER to clarify the next practical actions.