← All posts
SOC 2 penetration testing requirementsdoes SOC 2 require a penetration testSOC 2 vulnerability scanning

Does SOC 2 Require a Penetration Test? A Startup Decision Guide

SOC 2 does not set one pen test schedule for every company. Decide what to test from your risks, controls, and promises, then keep the results and fixes reviewable.

Start with the system and risk you need to test, then keep the test scope, findings, fixes, and retest linked.
Start with the system and risk you need to test, then keep the test scope, findings, fixes, and retest linked.

SOC 2 does not give every company the same penetration test requirement or schedule. The AICPA Trust Services Criteria name penetration testing as one example of an evaluation under CC4.1. They do not say every startup must buy an annual test. A test may still be needed to support the controls you chose, a customer promise, a contract, or the CPA firm’s examination plan. Start with the system and risk in scope, then decide what the test must show.

What do the SOC 2 penetration testing requirements actually say?

CC4.1 asks an organization to perform ongoing or separate evaluations of its internal controls. The AICPA’s points of focus give examples of evaluations, including penetration testing. Those examples help explain a criterion; they are not a fixed shopping list of tests. CC7.1 addresses finding vulnerabilities and monitoring systems. Your control design and evidence need to address the applicable criteria for the service you describe.

This creates three questions for your founder or security owner:

  1. What does the risk assessment say about your exposed application, APIs, cloud setup, and customer data?
  2. What do your policy, control descriptions, customer contracts, and security answers promise about testing and timing?
  3. What will the CPA firm need to examine the controls in your chosen scope and period?

If you have already promised an annual outside test, the generic criteria do not erase that promise. If you have not, write a test plan that fits the risk and confirm the approach with the CPA firm. The CPA firm performs the SOC 2 examination and decides how it will test management’s controls.

When should a small team run a pen test?

Run one when it answers a security question your other checks cannot answer well. An internet-facing product with complex authorization rules is a common case: a tester can try to cross tenant boundaries and combine flaws in ways a routine scan may miss. A major rewrite of authentication or data access can also justify a new test. The test scope should name the service, environments, interfaces, exclusions, and dates. An old report for a retired architecture does little for today’s system.

Set the cadence from your risks and changes, then record the decision. Check the contract and your own policy before choosing the date. If a customer asks for a recent test, clarify its acceptable age and scope. Discuss your plan with the CPA firm early enough to fix a scope mismatch before the examination. An annual cycle may be sensible for your system, but it is a management choice or a specific commitment, not a universal SOC 2 interval.

For a team preparing its first report, the broader startup SOC 2 guide helps place this decision beside scope, ownership, and the CPA engagement.

Is a vulnerability scan enough for SOC 2?

A scan and a pen test answer different questions. NIST SP 800-115 covers both techniques. NIST describes a pen test as an attempt to get around security features under defined constraints, often by combining weaknesses. A scan looks for detectable weaknesses within its configured targets. Both need a known scope and a process to handle findings.

Question Vulnerability scan Penetration test
What is checked? Targets and checks configured in the scanner. Paths a tester can try within agreed rules.
What can it show? Known or detectable weaknesses at scan time. Whether weaknesses can be used together to reach an objective.
What should you retain? Scan scope, date, results, triage, and fixes. Authorization, scope, report, findings, fixes, and retest.

Neither result proves that every system is safe. A narrow pen test does not replace recurring scans or patch work, and a clean scan does not prove that a person could not break an access boundary. Match each method to the control and risk it is meant to check.

For the work after a scanner finds a weakness, use the vulnerability management guide. It connects scan coverage, triage, a production fix, and verification.

What should you keep for the CPA firm?

Treat the final report as one part of a trail. Keep the test authorization and rules, the tested assets and exclusions, test and report dates, tester, finding IDs, severity and owner, fix decision, change record, and retest result. If a finding remains open, record the reason, compensating action, due date, and person who accepted the decision. A bare PDF with no tie to fixes makes it hard to show what management did next.

For example, suppose a tester finds that one API route can read another customer’s record. Link the finding to the affected service and risk. Record the code change and deployment date, then attach the tester’s retest result or another documented verification. A later scan with no finding may help, but its scope must cover the same issue before you treat it as a retest.

Pen test reports may contain exploit steps, customer details, and sensitive system data. The evidence repository guide explains when to keep the report in a restricted store and retain only an approved reference in Git. The evidence examples guide shows how to connect source material to a control and review record. Ask the CPA firm for its specific request and use its approved delivery channel.

Where does FileGRC fit?

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. You can use it to connect the risk, control, evidence reference, and management’s review of a test and its fixes. Starter records are proposals, not compliance claims.

Your scanner, tester, ticketing and deployment systems still produce the source facts. FileGRC does not run a pen test, decide whether a finding is fixed, or decide whether evidence is sufficient for a SOC 2 examination. Keep the technical report in a restricted source when its contents do not belong in the repository, and put a safe reference in the program record.

Open source · MIT

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 a penetration test?

The AICPA Trust Services Criteria name penetration testing as an example of an evaluation under CC4.1, but do not set one universal penetration test requirement for every service organization. Your own controls, risk assessment, contracts, and CPA firm's examination plan may still make a test necessary for your scope.

Does SOC 2 require an annual penetration test?

The Trust Services Criteria do not set a universal annual cadence. Choose and document a schedule based on the systems and risks in scope, the rate of change, your written commitments, and the CPA firm's expectations. If your policy or customer contract says annual, operate to that commitment.

Must an outside company perform the pen test for SOC 2?

The Trust Services Criteria do not impose one universal outside-provider rule. Decide who can test the system competently and objectively, check customer and contractual requirements, and agree on suitable evidence with your CPA firm before relying on an internal test.

Is a vulnerability scan the same as a penetration test?

No. A scan looks for known or detectable weaknesses across its configured scope. A penetration test has testers try to exploit weaknesses under agreed rules to learn what an attacker could reach. NIST treats them as different testing techniques.

What SOC 2 evidence should we keep from a pen test?

Keep the authorization, scope, test dates, tester, final report or a restricted reference, finding IDs, owners, decisions, fixes, and retest results. Link them to the systems, risks, and controls they address. Ask the CPA firm which material it needs for its examination.

Does passing a pen test mean we will get a SOC 2 report?

No. A pen test covers a defined technical scope at a particular time. Management must operate the relevant controls, and an independent CPA firm examines those controls against the selected Trust Services Criteria. The firm decides what evidence is sufficient for its procedures.