A penetration test report should include an executive summary, the scope and testing dates, the methodology, every finding with evidence and reproduction steps, a risk rating for each finding, prioritized remediation guidance and retest results. Auditors accept reports that tie each finding to an in-scope system and show how its risk was rated and fixed.

That’s the short answer. The longer one matters because a report is the only thing that survives the engagement. Your auditor never sees the testing, only the document. Your engineers fix what the report tells them to fix, in the order it tells them. A weak report can turn a good test into a rejected audit artifact or a misdirected remediation sprint.

Below is a section-by-section walkthrough of a penetration test report, a plain-English guide to severity ratings using CVSS and CISA’s Known Exploited Vulnerabilities (KEV) catalog, and a checklist for evidence auditors accept.

What does each section of a penetration test report contain?

A complete report has ten parts. Some vendors merge or rename them, but if one is missing, ask why.

1. Cover page and document control

The report should identify the client, the tester, the engagement name, the report version, the delivery date and the classification (for example, Confidential). A distribution list and a version history matter more than they look: auditors check that the report they’re reading is the final version, not a draft.

2. Executive summary

This is written for leadership and the board, not engineers. In one or two pages it should say:

  • What was tested and why (for example, “annual external and web application test to support SOC 2”)
  • The overall risk posture in plain language
  • The number of findings by severity
  • The two or three issues that matter most, in business terms (“an attacker on the internet could read customer records”), not technical ones (“IDOR in /api/v2/orders”)
  • What happens next: remediation priorities and the retest plan

If you need to report results to leadership, this section should work on its own. A CISO should be able to forward it without editing.

3. Scope, rules of engagement and testing window

This section lists exactly what was tested: IP ranges, domains, applications, API endpoints, mobile builds, user roles and environments (production or staging). It also lists what was out of scope, the testing dates, the testing type (black box, gray box or white box) and any constraints, such as no denial-of-service testing.

Auditors read this section first. If your SOC 2 system description includes a customer portal and the report doesn’t, the test doesn’t cover the system being audited.

4. Methodology and standards

The report should name the frameworks the testers followed. Common ones are NIST SP 800-115 for technical security testing, the OWASP Web Security Testing Guide for web applications, OWASP MASVS for mobile apps, and the Penetration Testing Execution Standard (PTES). It should also explain the mix of manual testing and automated tooling. A methodology section that lists only scanners suggests a vulnerability scan was sold as a penetration test.

For a deeper look at how testing is structured, see our guide to penetration testing methodology.

5. Attack narrative

Good reports explain how findings connect. A narrative shows the path an attacker would take: a leaked credential leads to a VPN login, the VPN login reaches an unpatched server, and the server gives up domain admin. Individually, each finding might rate Medium. Chained, they’re a breach. This section is the heart of a red team report and should appear in any internal network or assumed-breach test.

6. Detailed findings

Each finding needs the same structure so it can be tracked to closure:

  • Finding ID and title (for example, CT-2026-014: Broken object-level authorization on order API)
  • Affected assets: hostnames, URLs, endpoints or app versions
  • Severity and score: CVSS score and vector string, plus any context that changed the rating (see the table below)
  • Description: what the weakness is and why it matters here
  • Evidence: screenshots, request/response pairs or command output, with sensitive data redacted
  • Reproduction steps: enough detail for your team to confirm the issue and validate the fix
  • References: CVE, CWE, or OWASP category where they apply
  • Remediation: specific fix guidance, not “apply vendor patches”

Evidence is what separates a pen test finding from a scanner alert. A scanner reports that a version is vulnerable; a tester shows the exploit worked.

7. Remediation roadmap and prioritized fix list

This answers the question buyers ask most often about deliverables: yes, you should get a risk rating for every finding and a prioritized list of what to fix first. A useful roadmap groups fixes by owner (infrastructure, application, identity) and by urgency, and flags quick wins that close several findings at once, such as enforcing MFA on a remote access gateway.

8. Compliance mapping

If the test supports an audit, the report should map findings or test coverage to the framework controls involved, such as SOC 2 Trust Services Criteria, PCI DSS Requirement 11.4, or ISO/IEC 27001 Annex A control 8.8 (management of technical vulnerabilities). This saves your compliance team from building the crosswalk by hand. See penetration testing requirements by framework for how each standard treats testing.

9. Appendices

Appendices hold supporting detail: the full list of tested hosts, tools and versions used, accounts created during testing (and confirmation they were removed), and raw output where it helps your engineers.

10. Retest results and attestation

After you fix the findings, the tester should retest and issue an updated report or letter showing each finding’s status: remediated, partially remediated, or open with accepted risk. Many auditors and enterprise customers ask for this attestation rather than the full report, because it shows the fix, not just the flaw. Canary Trap issues a verifiable Security Certificate once retesting confirms the findings are remediated.

How does CVSS scoring work, in plain English?

The Common Vulnerability Scoring System (CVSS), maintained by FIRST, scores a vulnerability from 0.0 to 10.0. The base score asks two questions. How hard is it to exploit: over the network or with physical access, with or without a login, with or without a user clicking something? And how bad is the result: can an attacker read data, change it or take the system down?

The base score describes the flaw in general, not in your environment. A 9.8 on a server nobody can reach is less urgent than a 7.5 on your internet-facing login page. That’s why a good report shows the CVSS vector string as well as the number, and explains when the tester adjusted priority for your context.

Many findings in a penetration test aren’t CVEs at all. Business logic flaws, misconfigured SSO, over-permissioned service accounts and weak password policies have no vendor advisory. For those, testers rate likelihood and impact directly, often using the OWASP Risk Rating Methodology, which scores likelihood factors (attacker skill, ease of exploit) and impact factors (data exposure, financial and compliance damage) separately and combines them.

What do severity ratings mean, and where does the CISA KEV catalog fit?

CVSS tells you how severe a flaw could be. CISA’s Known Exploited Vulnerabilities (KEV) catalog tells you whether attackers are using it right now. A report should give you both.

Severity CVSS score What it means in plain English If the CVE is in CISA’s KEV catalog
Critical 9.0–10.0 Exploitable remotely with little effort; leads to system takeover or major data exposure. Emergency. Attackers are already exploiting it. Fix or isolate now.
High 7.0–8.9 Serious impact, but exploitation needs one extra condition, such as a valid login or a user action. Treat like a Critical. KEV status outweighs the score.
Medium 4.0–6.9 A real weakness with limited impact or harder exploitation; often one link in an attack chain. Raise it to the top of the queue. Known exploitation beats a medium score.
Low 0.1–3.9 Minor exposure or a hardening gap with little direct impact. Rare, but if listed, fix it on the KEV schedule, not the low-severity one.
Informational 0.0 or not scored Observations and best-practice gaps without direct exploitability. Not applicable.

CISA itself has moved away from severity scores as the main driver. Its Vulnerability Review for fiscal years 2024 and 2025 argues that CVSS reflects theoretical severity, not real-world impact, and that most compromises in that period came from known, exposed vulnerabilities rather than zero-days.

How should you prioritize vulnerability findings?

Use exposure and exploitation first, and CVSS second. CISA’s Binding Operational Directive 26-04, issued June 10, 2026, sets this out for federal civilian agencies with four questions:

  1. Is the affected asset publicly exposed?
  2. Is the vulnerability in the KEV catalog (is there evidence of active exploitation)?
  3. Can exploitation be fully automated?
  4. Would exploitation give the attacker total control of the system?

The answers set the deadline. The directive applies only to federal agencies, but its timelines are a useful benchmark for any remediation plan:

Situation BOD 26-04 remediation timeline
Publicly exposed, in KEV, total control possible 3 days, plus forensic triage
Publicly exposed, in KEV, automatable, partial control 3 days
Publicly exposed, not in KEV, automatable, total control 3 days
Internal, in KEV, automatable, total control 3 days, plus forensic triage
Most other combinations involving exposure, KEV or automation 14 days
Publicly exposed, not in KEV, not automatable, partial control 60 days
Internal, not in KEV, automatable 60 days
Internal, not in KEV, not automatable At the next system upgrade

In practice, a prioritized fix list from your tester should work the same way:

  1. Fix KEV-listed and actively exploited issues on exposed systems first, whatever their CVSS score.
  2. Break attack chains. A Medium finding that is the first step to domain admin outranks an isolated High.
  3. Prioritize total-control outcomes: remote code execution, authentication bypass, credential exposure.
  4. Use CVSS to order what’s left, adjusted for whether the asset holds sensitive data.
  5. Schedule hardening items (Lows and Informationals) into normal maintenance.

For more on how vulnerability assessments and penetration tests produce different findings, see vulnerability assessment vs. penetration test.

What makes a pen test report acceptable to auditors?

Auditors don’t grade the findings. They check that the test was real, independent, in scope and acted on. Before you submit a report, confirm it shows:

  • Independence and qualifications. The testers are separate from the team that built or runs the system, and the report names the firm and lead tester.
  • Scope that matches the audit boundary. The SOC 2 system description, the PCI DSS cardholder data environment or the ISO 27001 scope statement is covered.
  • Dates inside the audit period. A test from 14 months ago won’t support a 12-month SOC 2 Type II window.
  • A documented methodology. PCI DSS v4.0 Requirement 11.4.1 requires an industry-accepted approach that covers network and application layers and tests from inside and outside the network.
  • Evidence for each finding and a stated method for rating risk.
  • Remediation and retest proof. PCI DSS Requirement 11.4.4 requires exploitable vulnerabilities found in testing to be corrected and the corrections verified by repeat testing.
  • A final, versioned, signed document, with an attestation letter or certificate for sharing with customers.

SOC 2 doesn’t name penetration testing as a hard requirement, but the Trust Services Criteria list it as a type of evaluation under CC4.1, and most auditors expect one for systems that face the internet. Our SOC 2 penetration testing guide and ISO 27001 penetration testing guide cover each framework in detail.

Which vendors are known for good reporting?

Don’t take a vendor’s word for it. Ask every firm you’re evaluating for:

  1. A redacted sample report from a real engagement, not a template
  2. An explanation of how they rate risk beyond the raw CVSS score
  3. Their quality assurance process before a report is delivered
  4. Whether a retest and an attestation letter are included
  5. Whether they map findings to SOC 2, PCI DSS or ISO 27001 controls

Canary Trap delivers a Report of Findings with an executive summary for business stakeholders, detailed technical findings with severity ratings, and remediation guidance tailored to the environment tested.

How do vulnerability assessment, pen test and red team reports differ?

Report type Main question it answers What the findings look like Retest
Vulnerability assessment What known weaknesses exist across these assets? Broad list of validated vulnerabilities, rated and prioritized Often a rescan
Penetration test Can an attacker exploit these weaknesses, and how far can they get? Exploited findings with evidence, attack narrative, remediation Retest of findings, with attestation
Red team exercise Would we detect and stop a realistic attacker pursuing a goal? Objective-based narrative, detection and response observations Not included; findings feed the next exercise

Red team reports focus on whether your people, processes and tools noticed the activity. See red team engagement phases for what to expect from the engagement itself.

Frequently asked questions

What should a penetration test report include?

An executive summary, scope and testing dates, methodology, an attack narrative, detailed findings with evidence and reproduction steps, a severity rating for each finding, a prioritized remediation roadmap, compliance mapping where relevant, appendices and retest results.

Do I get a risk rating and a prioritized fix list with a vulnerability assessment?

You should. Every finding needs a severity rating, and the report should rank fixes by real-world risk (exposure, known exploitation and impact), not just by CVSS score. If a vendor delivers raw scanner output, it isn’t a complete assessment.

What is the difference between CVSS and the CISA KEV catalog?

CVSS scores how severe a vulnerability could be on a 0–10 scale. The KEV catalog lists vulnerabilities CISA has confirmed attackers are exploiting. A KEV-listed vulnerability should be fixed urgently even if its CVSS score is Medium.

Can one web app pentest satisfy SOC 2 and give both an executive summary and technical findings?

Yes. A single report can serve both audiences: an executive summary for leadership and the auditor, and technical findings for engineers. Make sure the scope matches your SOC 2 system description and the test falls inside your audit period.

Will auditors accept a pen test report with open findings?

Often, yes, if open findings have a documented remediation plan or formal risk acceptance. Auditors look for evidence that you acted on the results. Retest results showing closed findings make the strongest evidence.

What should a mobile app pentest report include?

Everything above, plus the app versions and platforms tested (iOS, Android), whether the backend APIs were in scope, and mapping to OWASP MASVS categories such as storage, cryptography, authentication and network communication.

What should an LLM red team report include?

The models, prompts and integrations tested; findings such as prompt injection, sensitive data disclosure and unsafe tool use, with the exact inputs that triggered them; and mapping to the OWASP Top 10 for LLM Applications. See our guide to LLM security testing.

Get a report your auditor and your engineers can both use

Canary Trap’s penetration tests end with a Report of Findings built for two readers: leadership and auditors who need the risk in plain terms, and engineers who need evidence and fix steps. Once retesting confirms remediation, we issue a verifiable Security Certificate you can share with customers and auditors.

Start your scoping questionnaire –>