Penetration Testing Scope: How to Scope a Pen Test (With Checklist)

Reviewed by Ali Aleali, CISSP, CCSP · Last reviewed November 21, 2026

Penetration testing scope is the written agreement on what a tester may attack, how, from where and when. A complete pen test scope sets the test type, the targets (URLs, IP ranges, APIs and user roles), the knowledge level (black, grey or white box), the environment, the rules of engagement, the exclusions, the testing window and the retest.

At a glance

  • Decide first: Which discipline: external network, internal network, web application, API, cloud configuration, mobile or social engineering
  • Targets: URLs, IP ranges, API endpoints, and test accounts for every user role
  • Knowledge level: Black box, grey box or white box. Grey box suits most SaaS web applications
  • Environment: Production for external network tests. A production-like staging copy with erasable data for most application tests
  • Paperwork: Statement of work with rules of engagement, exclusions, emergency contacts and a retest
  • Cadence: Annual as the common baseline, plus a test after major architecture changes
  • Framework fit: SOC 2 CC4.1 and CC7.1, ISO/IEC 27001:2022 A.8.8 and A.8.29

What goes into a penetration testing scope?

A penetration testing scope is a short set of decisions the buyer makes before any firm prices the work. Testers quote days of effort from these decisions. The elements of a pen test scope of work are:

  1. Discipline: the kind of test, for example web application or internal network.
  2. Targets: the exact URLs, IP ranges, endpoints and applications in scope.
  3. Accounts and roles: the test users the tester logs in as.
  4. Knowledge level: black, grey or white box.
  5. Environment: production, staging or an isolated copy.
  6. Rules of engagement: testing window, contacts, allowed and forbidden techniques.
  7. Exclusions: systems and techniques left out, with the reason.
  8. Timing: when the test happens relative to releases and audits.
  9. Retest: a second pass to confirm each fix.

Which type of penetration test does the scope need?

The type of penetration test follows from the question the company wants answered. Each discipline needs different inputs from the buyer:

Pen test disciplineWhat the tester attacksWhat the buyer provides
External networkInternet-facing hosts, VPN gateways, mail servers, exposed servicesPublic IP ranges and hostnames, proof of ownership
Internal networkLateral movement from a foothold inside the networkVPN access or a dropped device, in-scope subnets
Web applicationAuthentication, authorization, session handling, business logicURLs, test accounts per role, a short walkthrough of the application
APIEndpoint authorization, object-level access, input handlingAPI documentation or a collection file, keys or tokens per role
Cloud configurationIAM policies, storage exposure, network rules, loggingRead-only access to the cloud accounts in scope
Mobile applicationThe app binary, local storage, traffic to the backendApp builds for each platform, test accounts
Social engineeringStaff response to phishing, phone pretexting or physical entryApproved target list, pretexts and an executive sponsor

For on-premise and hybrid networks, the SOC 2 penetration testing guide for on-prem networks adds segmentation testing, wireless and out-of-band management interfaces such as iDRAC and iLO. Applications with LLM features need prompt injection and model abuse testing in scope as well, a discipline described on the AI penetration testing page.

A penetration test is human-led validation, and buyers often mix the term up with neighbouring disciplines:

  • Vulnerability scanning: automated detection of known CVEs, run continuously. See what a vulnerability scan is.
  • Vulnerability assessment: a scan plus human triage, often the right first engagement. See vulnerability assessment services.
  • Red team exercise: a goal-based attack with the defenders kept unaware.
  • Purple team exercise: attackers and defenders work together to check that detections fire.

Watch out

A vulnerability scan delivered as a pen test report gets rejected.

An auditor who reads a pen test report with no exploitation chain and no authorization findings often asks for a re-test. Human testers find IDOR (one user reading another user's records by changing an ID) and role boundary violations because scanners rarely reach authorization logic. Write manual authorization testing into the penetration testing scope by name.

How do you define penetration test targets, URLs, IP ranges and user roles?

Penetration test targets are the exact assets the tester is allowed to touch:

  • Web applications: every hostname, including admin portals and customer subdomains.
  • APIs: base URLs, versions and the full endpoint list.
  • IP ranges: CIDR blocks for external and internal testing, marked as company-owned or provider-owned.
  • Cloud accounts: account or subscription IDs and regions.
  • Third parties: SaaS tools and hosting providers the company relies on but does not own. These normally go in the exclusions. Check cloud workloads against the cloud provider's published penetration testing policy.

User roles set the depth of an application test. My approach is to give the tester two accounts for every role, for example two standard users and two administrators in separate tenants. With two accounts per role, the tester can check whether user A reads user B's data (horizontal access) and whether a standard user reaches admin functions (vertical access).

Black box, grey box or white box: which pen test approach fits?

The knowledge level sets how much the tester knows at the start. Each option spends the testing days differently:

Pen test approachWhat the tester starts withBest fit
Black boxA company name or a URL, no credentialsExternal network tests, simulating an outside attacker
Grey boxCredentials, role descriptions, API documentationMost web application and API tests
White boxSource code, architecture diagrams, configurationHigh-risk applications, code-assisted review

For most SaaS web applications, I recommend grey box. A penetration test is time-boxed, and a black box tester spends many of those days on discovery a short walkthrough would replace. Grey box puts the days into authorization and business logic.

Should a penetration test run against production or staging?

The test environment is a tradeoff between realism and risk. External network tests usually run against production, because the production perimeter is the thing under test. Application tests are a different case. Testers create records, submit forms and try to break workflows, and those actions leave data behind.

A staging environment works for an application pen test when staging matches production:

  • Same code version and same configuration
  • Same authentication and MFA settings
  • Same WAF and rate-limiting rules, or a documented difference
  • Realistic but non-sensitive data that the team can erase and reload

If production is the only realistic option, tighten the rules of engagement: a narrower testing window, a named on-call engineer, and destructive tests left out.

What belongs in the rules of engagement for a pen test?

The rules of engagement are the operating instructions for the pen test, written into the statement of work:

  • Testing window: dates, hours and time zone.
  • Source IP addresses: the tester's IPs, so the team knows which traffic is the test, plus a decision on whether the WAF allowlists those IPs or stays in place.
  • Contacts: a technical contact on each side and an escalation path for outages.
  • Critical finding protocol: the tester reports a critical vulnerability the same day instead of waiting for the final report.
  • Forbidden techniques: for example denial of service, social engineering, or physical access, unless the scope includes them.
  • Data handling: how the tester stores and deletes screenshots and any data accessed.
  • Stop conditions: what makes either side pause the test, for example signs of a real attacker in the logs.

What should a penetration test scope exclude?

Exclusions record what the tester leaves alone and why. Typical exclusions:

  • Third-party SaaS and hosting the company does not own
  • Systems outside the audit scope, such as a marketing website on a separate platform
  • Fragile production systems where an active test carries unacceptable risk
  • Techniques such as denial of service

Known issues belong in the scope document too. Flag the vulnerabilities the team already tracks, such as an outdated library with a planned upgrade, as expected findings. The tester confirms the known issues briefly and spends the remaining days on the unknown ones.

How does a legacy application change the penetration testing scope?

A legacy application needs a phased penetration testing scope, because framework-level vulnerabilities otherwise fill the report. One small software company I worked with needed a first pen test for an application in production for nearly two decades: an older web framework, a single developer, over 100,000 lines of code across thousands of pages, sensitive personal information, and no previous formal security testing.

The team split the test into two phases:

  1. Unauthenticated testing: authentication flows, MFA bypass attempts and the external attack surface.
  2. Authenticated testing: with valid credentials, authorization boundaries, business logic and data access controls.

Four decisions came out of the scoping conversation:

  • Sequential IDs: the application used numeric sequential IDs in URLs, so authorization testing focused on changing IDs to reach other users' records.
  • Expected findings: the known outdated UI libraries were flagged in advance, so the report separated framework risks the team already knew about from application-specific risks.
  • Isolated environment: the team built a separate test environment with an erasable copy of the database. The preparation took more effort than expected and kept production data out of the test.
  • API expansion: the API scope grew once the team found dozens of API functions with separate authentication, with a separate authorization check for each function.

Key insight

The scoping conversation documents the architecture.

Preparing the scope forced the team to write down the application architecture more thoroughly than the team had in years. For an older application, the scope document often becomes the most current architecture record the company has.

When a portfolio has dozens of applications and budget for a few tests, a maturity assessment first helps pick which applications to test. The 90 applications in 90 days assessment shows that approach.

When should a penetration test happen, and is the retest in scope?

Penetration test timing should leave room to fix findings before a customer or auditor needs the final report:

  • Test date: a few weeks before the report deadline, so the team can fix critical and high findings.
  • Release timing: after major releases, so the test reflects the code customers use.
  • Cadence: annual is the common baseline, with an extra test after a major architecture change or a significant incident.
  • Retest: included in the original statement of work, so the final report shows each finding as fixed or accepted.

Findings then go into the same remediation process as scanner results, with the deadlines set in the vulnerability management policy.

Pen test scoping questionnaire: what information will a tester ask for?

A pen test scoping questionnaire is the form a testing firm sends before pricing the work. Expect questions like these:

  1. What is the reason for the test: an audit, a customer request, a release or an incident?
  2. Which disciplines are in scope?
  3. How many web applications, and roughly how many pages or screens in each?
  4. How many API endpoints, and is documentation available?
  5. How many user roles, and can the company provide two accounts per role?
  6. How many external IP addresses and internal subnets?
  7. Which cloud providers and accounts are in scope?
  8. Production or staging, and does staging match production?
  9. Which known issues should the tester treat as expected findings?
  10. What dates and hours are allowed for testing?
  11. Who receives the report, and does an auditor or customer need a letter of attestation?

Penetration testing scope checklist

Use this checklist before signing a penetration test statement of work:

  • Disciplines chosen and named in the statement of work
  • Every URL, endpoint, IP range and cloud account listed
  • Two test accounts per user role created
  • Knowledge level agreed (black, grey or white box)
  • Environment chosen, and staging checked against production
  • Rules of engagement signed, including the critical finding protocol
  • Exclusions listed with reasons
  • Manual authorization and business logic testing named in scope
  • Methodology named, for example NIST SP 800-115 or the OWASP Web Security Testing Guide
  • Retest included in the price
  • Testing window set with time to fix findings before the report deadline

Free download · Google Doc

Penetration Test Scoping Checklist and Questionnaire

The main pen test scoping decisions in one document: disciplines, targets, test accounts, rules of engagement, exclusions and answers to the questions testers usually ask, plus a checklist to use before you sign. Enter your email to get it.

Does SOC 2 require a penetration test? The evidence a well-scoped test produces

No SOC 2 criterion makes penetration testing mandatory, but auditors commonly expect an annual pen test. CC4.1 lists penetration testing among the evaluation types an organization may use, and a pen test report is the separate evaluation auditors most often ask for. A well-scoped test produces evidence for several controls at once:

ControlWhat the control asks forEvidence from a well-scoped pen test
SOC 2 CC4.1Ongoing and separate evaluations of whether controls workStatement of work, report from an independent tester, retest results
SOC 2 CC7.1Detection and monitoring procedures to identify vulnerabilitiesValidated findings tracked to closure in the remediation process
ISO/IEC 27001:2022 A.8.8Management of technical vulnerabilitiesFindings, remediation records, risk acceptances
ISO/IEC 27001:2022 A.8.29Security testing in development and acceptanceApplication test results before or after major releases

A clear target list, documented exclusions and a retest show an auditor that the test matched the systems in the audit.

Who can help scope a penetration test?

Truvo Cyber scopes penetration tests as part of building the security testing capability, one of the roughly 19 security capabilities in an effective security program. The penetration testing advisory service covers scoping, choosing a testing firm and turning the report into a remediation plan.

Scope the pen test before the quotes

Get a penetration test scoped to the systems your effective security program depends on.

Frequently asked questions

What is included in a penetration testing scope?

A penetration testing scope includes the test type, the exact targets (URLs, IP ranges, API endpoints, cloud accounts), the user roles and test accounts, the knowledge level, the test environment, the rules of engagement, the exclusions with reasons, the testing window and a retest. All of these go in the statement of work before testing starts.

What is the difference between a pen test scope and rules of engagement?

The pen test scope defines what the tester may test: targets, roles, environment and exclusions. The rules of engagement define how the test proceeds: dates and hours, tester source IPs, contacts, forbidden techniques, data handling, stop conditions and how critical findings are reported. Both belong in the same statement of work.

Should a penetration tester get credentials?

For most web application and API tests, yes. Grey box testing with two accounts per user role lets the tester check whether one user can read another user's data and whether a standard user can reach admin functions. Credentials put the testing days into authorization and business logic.

Does SOC 2 require a penetration test?

No SOC 2 criterion makes penetration testing mandatory. CC4.1 lists penetration testing among the evaluation types an organization may use, and auditors commonly expect an annual test from an independent firm. The same report supports ISO/IEC 27001:2022 A.8.8 and A.8.29.

Ready to Start Your Compliance Journey?

Get a clear, actionable roadmap with our readiness assessment.

Contact Us

Share this article:

Sample SOC 2 Control List

Input your email to download the list.

About the Author

Former security architect for Bank of Canada and Payments Canada. 20+ years building compliance programs for critical infrastructure.