On this page
- Does SOC 2 Require Penetration Testing?
- What Do the Trust Services Criteria Say About Penetration Testing?
- Where Does a Penetration-Test Requirement Come From?
- When Does Penetration Testing Support a SOC 2 Audit?
- How Should You Scope a SOC 2 Penetration Test?
- Who Can Run the Test?
- How Do Penetration Tests Differ From Vulnerability Scans?
- What Should a Penetration-Test Report Contain?
- How Should You Document Remediation and Retesting?
- What Happens If You Skip the Penetration Test?
No. SOC 2 does not require a penetration test, and the criteria set no testing frequency. CC4.1 asks management to run ongoing or separate evaluations of its controls, and its points of focus list penetration testing among the example evaluations. A pentest becomes a requirement for you through other routes: a control you wrote into your system description, evidence your CPA firm needs for CC4.1, or a customer contract. Check each one, and get your CPA firm’s answer on scope and timing in writing.
A SOC 2 penetration test is an authorized attack simulation against systems inside the SOC 2 system boundary. NIST SP 800-115 describes penetration testing as security testing that mimics real-world attacks to identify ways around application, system, or network security features. The report is one evidence source your CPA firm weighs alongside the system description and the operating evidence for your other controls.
Does SOC 2 Require Penetration Testing?
No. The AICPA’s Trust Services Criteria do not require a penetration test or set a testing frequency. They ask management to evaluate its controls on a schedule it sets from risk, and they list penetration testing as one example.
Four questions settle it for your audit:
- Do our written controls or system description promise a test, and at what frequency?
- What does our CPA firm need to see for CC4.1, and have they said so in writing?
- Do customer contracts or security questionnaires name a test?
- If we skip the test, which other separate evaluations do we run instead?
What Do the Trust Services Criteria Say About Penetration Testing?
CC4.1 is the criterion most often tied to penetration testing, and its points of focus name the test. The criterion reads, in the AICPA’s words:
The entity selects, develops, and performs ongoing and/or separate evaluations to ascertain whether the components of internal control are present and functioning.
One point of focus says those evaluations may include “vulnerability scans, security assessments, penetration testing, and third-party assessments,” alongside internal audit and compliance assessments. A pentest is one item in a list, so the criterion asks for evaluations of your controls and leaves the method to you. An entity also does not need a control for every point of focus.
The other points of focus under CC4.1 decide how you run a test once you choose one. The middle column paraphrases the AICPA wording.
| What the point of focus asks | What it means for your test | |
|---|---|---|
| Adjusts Scope and Frequency | Management varies the scope and frequency of separate evaluations by risk. | Cadence is a documented risk decision, not a calendar rule. Write down why you chose it. |
| Considers Rate of Change | Management weighs how fast the business and its processes change. | A major architecture, identity, or data-flow change is a reason to retest or add a targeted assessment. |
| Objectively Evaluates | Separate evaluations are performed periodically to give objective feedback. | A tester who did not build the system is easier to defend. |
| Uses Knowledgeable Personnel | Evaluators know enough to understand what they evaluate. | Keep the tester’s qualifications with the report. |
| Considers Different Types of Evaluations | Management uses a variety of risk and control evaluations. | A pentest is one choice among scans, internal audit, and third-party assessments. If you skip it, document the others. |
CC7.1 is the criterion for detecting new vulnerabilities and configuration changes. Its point of focus is “Conducts Vulnerability Scans,” scans run periodically and after significant change, with timely remediation. It names scans, not pentests. A pentest can support CC7.1 evidence, but the criteria text does not require one there either.
Where Does a Penetration-Test Requirement Come From?
A test requirement comes from one of five places, and only the first is the AICPA.
| Does it require a pentest? | What it can require | What to keep | |
|---|---|---|---|
| AICPA criteria (CC4.1) | No | Ongoing or separate evaluations of your controls, scoped by risk. A pentest is one listed example. | The test, mapped to CC4.1 |
| Your system description and policies | Yes, once you state it | The scope and cadence you wrote, such as “annual, third party” | A test dated within your stated cadence and covering your stated scope |
| Your CPA firm | Varies by firm | The evidence it needs for CC4.1 and CC7.1 on your system | A written answer on scope, timing, and tester independence |
| Customer contract or questionnaire | Often, by name | A test, a frequency, or a copy of the report | The report, or a summary letter from the tester |
| Compliance platform rules | Platform practice, not SOC 2 | Its own checklist. Secureframe, for example, recommends an annual third-party test and says it is not necessarily a hard pass or fail control | The platform’s evidence mapping |
Checked October 2026 against the AICPA Trust Services Criteria (2017, with revised points of focus 2022) and Secureframe’s pen testing help article.
The second row is easy to overlook. A control that says “we perform an annual penetration test” binds you for the whole Type 2 period, and a skipped test becomes an exception. The third row is also easy to miss: a customer contract that requires a test makes it mandatory for that deal, whatever the criteria say.
When Does Penetration Testing Support a SOC 2 Audit?
A penetration test supports a SOC 2 audit when it covers the systems and attack paths of the service under examination and management can trace each material finding through remediation and retest. It is one evidence source for monitoring and vulnerability-management controls. It does not replace risk assessment, scanning, change management, or access reviews.
A SaaS company that tests only its marketing site hands the CPA firm a clean report that says nothing about the authenticated application where customer data lives. A test that checks whether one customer can read another customer’s records, or whether a cloud identity role reaches production data, answers the question the audit is asking.
The report counts as audit evidence only with the records around it:
Authorization, the systems tested and excluded, and the rules of engagement, so the CPA firm can judge relevance.
Dates, tester, method, findings, and stated limitations.
An owner and a decision per finding: fixed, accepted, or covered by a compensating control, with the change record.
Which findings were retested, when, by whom, and with what result.
The firm matches the evidence to the audit period and to the system as it ran.
Our SOC 2 controls library carries a sample penetration-test control mapped to CC4.1 and CC7.1. It asks for the same three artifacts, a report, remediation tickets, and a retest letter. Its common gap is a scope that leaves out the production application or its API, or findings filed with no remediation record or retest.
How Should You Scope a SOC 2 Penetration Test?
Scope the test from the SOC 2 system boundary, the service’s material attack paths, and the controls the test is meant to inform. The two boundaries do not have to match. Document what is included and excluded so management and the CPA firm can judge whether the test is relevant.
Which Systems Belong in the Test Scope?
Include the systems and interfaces that answer the test’s objective. For a SaaS service, that usually means the customer-facing application, externally exposed APIs, authentication flows, administrative interfaces, cloud configuration, and any path that reaches production data. Record why each candidate is in or out.
Third-party services need a boundary decision. Test the connection points you are authorized to assess, identify provider-managed components, and keep the dependency and exclusion notes. NIST SP 800-115 recommends planning technical security testing around objectives, scope, authorization, logistics, and rules of engagement.
How Should You Set the Testing Frequency?
SOC 2 does not prescribe a testing frequency. Annual testing is the common baseline in vendor and platform guidance, including Secureframe’s and Synack’s, and the criteria do not set it. Define your cadence in policy from the systems in scope, material changes, threat exposure, and the evidence you need during the audit period. NIST SP 800-53 Rev. 5, control CA-8 likewise calls for penetration testing at an organization-defined frequency on organization-defined systems or components.
A fast-changing SaaS product may need a targeted assessment after a material architecture, identity, or data-flow change. A stable service can run recurring vulnerability and configuration checks between deeper assessments. The policy should name the trigger, the decision-maker, the affected systems, and the evidence you keep.
Schedule the principal assessment early enough to fix material findings and get retest evidence before fieldwork. Lead time follows the scope, the testing window, the remediation work, and the audit timetable, so do not rely on a generic month-in-period rule. A report from before a major architecture change may not cover the current attack surface. Record whether the change calls for supplemental testing, and keep evidence that matches the system as it ran during the audit period.
How to Set a SOC 2 Penetration-Testing Cadence
These scenarios show how to write a cadence decision a CPA firm can review. They are not testing intervals.
| Cadence decision | Possible audit use | Evidence to retain | |
|---|---|---|---|
| First Type 2 examination | Schedule the first test, and any interim checks, in the evidence plan. | Monitoring and vulnerability-management evidence | A baseline report, time to fix material findings, and the record for the observation period |
| Material system change | Decide whether the change creates a new attack path or outdates the last test’s scope. | Change and risk-management evidence | The decision, affected systems, testing done, and any compensating evidence |
| Higher-risk service | Set a deeper or more frequent cadence when your documented risk assessment supports it. | Depends on the criteria and scope selected | The risk, systems, decision-maker, and evidence, not an industry label |
| Security incident or material finding | Decide whether a targeted retest is needed after remediation. | Incident and remediation evidence | The original finding, remediation decision, deployment evidence, and retest result |
Keep the policy, test plan, report, remediation records, and retest evidence together so the CPA firm can review the chain without reconstructing it from email.
Does the Test Have to Fall Inside the Audit Period?
The criteria set no date for the test, so the answer comes from your own control wording and your CPA firm. Two published sources describe the usual practice for each report type:
| Type 1 | Type 2 | |
|---|---|---|
| Secureframe, pen testing help article | The test can have taken place any time in the 12 months before the report date. | The test has to occur during the audit period. |
| Synack, SOC 2 penetration testing guide | A recent test is useful but not automatically mandatory. | Auditors may check that testing followed the stated cadence and fell within the review period. |
| The criteria | No date or frequency | No date or frequency |
These rows are platform and vendor practice, checked October 2026. They are not criteria text. Synack makes the point that a company can create its own audit problem by stating an annual test in its control and then not doing it, because the control language, the frequency, and the evidence all have to agree. Ask your CPA firm which timing it accepts before you book the test, and read the SOC 2 observation period guide to see where the period starts.
Who Can Run the Test?
Any qualified, objective tester can run it. The criteria ask for objective evaluation by knowledgeable people, and they name no certification or outside firm. Your engineers can test between assessments, but a report from a party independent of the team that built and runs the system is easier to defend. Synack’s guide notes that auditors may look at the tester’s competence and independence, so keep the statement of work, tester credentials, and rules of engagement with the report.
The firm that issues your SOC 2 report is a separate question. The AICPA Code defines a self-review threat as the risk that a CPA will not appropriately evaluate the results of a service it performed itself and will rely on it in an attest judgment. Sensiba, a CPA firm, lists penetration testing among the nonattest services that raise this threat when the auditing firm also delivers them. A different firm is the cleaner arrangement. If your CPA firm also offers penetration testing, ask it to explain in writing how it addresses the self-review threat before you buy both.
If a CPA firm sells testing through a partner, ask who performs the test and expect the report to say so. A partner’s test is the partner’s work, and the report should name who did it.
How Do Penetration Tests Differ From Vulnerability Scans?
A scan identifies potential known weaknesses across a defined set of targets. A penetration test uses authorized attack techniques to find out whether selected weaknesses or attack paths can be exploited. Your risk assessment and control design decide whether you need one or both. CC7.1 names scans, and CC4.1 lists both.
A scan is generally automated and repeatable. A penetration test combines tools with analyst judgment under agreed rules of engagement. Neither label guarantees scope, depth, or evidence quality, so read the procedures and limitations in the delivered report.
Penetration Testing vs Vulnerability Scanning for SOC 2
Compare the evidence each activity produces before treating one report as a substitute for the other.
| Vulnerability Scanning | Penetration Testing | |
|---|---|---|
| Methodology | Automated checks for known vulnerabilities and common misconfigurations | Tool-assisted, analyst-led testing under agreed rules of engagement |
| Primary goal | Identify potential weaknesses across the defined target set | Validate selected attack paths and their service-specific impact |
| Possible SOC 2 use | Evidence that management ran a defined vulnerability-identification and response process | Evidence from an authorized adversarial assessment of selected systems and attack paths |
| Review questions | Was the target population complete? How were findings validated, prioritized, and closed? | Did scope match the objective? What was excluded? How were material findings treated and, where appropriate, retested? |
| Key output | Potential findings that need validation and disposition | Tested attack paths, findings, limitations, and remediation guidance |
What Should a Penetration-Test Report Contain?
A useful report tells management and the CPA firm what was tested, what was not, what the assessor did, what it found, and what happened next. No single template or scoring version is required for SOC 2.
Keep these report elements:
- Authorization, dates, and scope: the legal parties, test window, targets, exclusions, and rules of engagement.
- Methodology and limitations: the testing approach, access level, constraints, and assumptions.
- Findings and evidence: affected assets, reproduction information, impact, and the basis for severity.
- Management response: each accepted finding linked to an owner, decision, and remediation record.
- Retest status: which findings were retested, when, by whom, and with what result.
Risk scores help prioritize findings, but name the scoring system and version instead of assuming one, and pair the score with the service-specific impact and the affected system. For evidence organization, see SOC 2 documentation and our SOC 2 evidence request list.
How Should You Document Remediation and Retesting?
Record a disposition for each validated finding: remediation, accepted risk, a compensating control, or another approved response. The evidence should link the finding to an owner, a decision, an implementation record, and a current status.
- Validate the finding and assign an owner.
- Record severity and service-specific impact.
- Document the response decision and due date.
- Link implementation evidence, such as a change record or configuration update.
- Decide whether independent retesting or another validation method fits.
- Keep the result and any remaining limitation.
Retesting terms vary by provider and contract. Confirm the included window, eligible findings, deliverable, and extra fees before signing, and see the SOC 2 penetration testing cost guide for published starting prices. A closed ticket records workflow status. The underlying evidence shows what changed and how the result was validated. Findings that need fixing before fieldwork are work a SOC 2 readiness firm can take on.
What Happens If You Skip the Penetration Test?
A SOC 2 report can be issued without one, provided your controls do not promise a test and your CPA firm accepts the other evidence you rely on for CC4.1 and CC7.1. Synack’s guide says the same: a company may complete a SOC 2 examination without a pentest but still needs sufficient evidence that its security controls are suitably designed and, for Type 2, operated effectively.
If you skip it, document the other separate evaluations you run, such as vulnerability scans with remediation records, internal audit assessments, and third-party assessments, and get your CPA firm’s acceptance in writing. Customers may still ask for a test in a questionnaire or contract, and the report will not answer that request for you. A pentest also does not guarantee a clean opinion when you do run one.
If you decide to commission a test, compare SOC 2-scoped options on the SOC 2 penetration testing firms page. If you already have the report and need a CPA firm to evaluate it in the wider examination, use the SOC 2 auditor directory.
More in Audit Preparation