How to Write a Penetration Test Report That Gets Fixed

A penetration test is only as valuable as its report. Many technically excellent testers fail at this stage: they hand over a hundred pages of scanner output, and nothing gets fixed. A good report is clear to executives, actionable for engineers and defensible as evidence. This lesson shows you how to structure findings, rate risk fairly, write recommendations that developers can use and avoid the common mistakes that make reports ignored.

Who Reads the Report?

ReaderWhat they need
Executives and business ownersOverall risk in plain language, business impact and priorities
Security managersFindings ranked by risk, trends, compliance mapping, resourcing needs
Engineers and developersExact location, evidence, steps to reproduce and a concrete fix
Auditors and regulatorsScope, methodology, dates and proof of testing

One report serves all four by using layered structure: a short summary at the front and technical detail behind.

Recommended Report Structure

  1. Cover page and document control: title, client, version, dates, author, confidentiality classification.
  2. Executive summary: one page, no jargon. What was tested, the overall result, the top risks and the most important actions.
  3. Scope and methodology: what was included and excluded, the test type (black, grey or white box), tools and standards followed, testing dates and constraints.
  4. Risk rating method: how you scored findings (explain it once, clearly).
  5. Summary of findings: a table sorted by severity, with a count by rating.
  6. Detailed findings: one section per issue (template below).
  7. Positive observations: controls that worked well. This builds trust and tells the client what to keep.
  8. Remediation roadmap: grouped quick wins, medium-term fixes and strategic improvements.
  9. Appendices: evidence, tool output, attack path diagrams, retest results, glossary.

Anatomy of One Good Finding

FieldWhat to write
TitleSpecific and meaningful: “Customer order details can be viewed by other logged-in users”, not “Access control issue”
SeverityCritical, High, Medium, Low or Informational, with CVSS vector if used
Affected assetsHost names, URLs, parameters, versions
DescriptionWhat the weakness is, in two or three sentences
ImpactWhat an attacker could do and what that means for the business
EvidenceScreenshots, request and response excerpts, timestamps (with sensitive data redacted)
Steps to reproduceNumbered steps another engineer can follow
RecommendationA specific, achievable fix, with references to guidance
ReferencesCWE, OWASP, vendor advisories
StatusOpen, fixed or accepted risk (updated after retest)

A short example finding (fictional, for training)

Title: Order history page does not verify ownership of the requested order.
Severity: High (CVSS 3.1 vector reflects network access, low privileges, high confidentiality impact).
Affected asset: https://shop.example.test/orders/view?id=1043 (staging server).
Description: The application returns an order when given a numeric identifier but does not check that the order belongs to the logged-in user.
Impact: Any authenticated customer can read other customers’ names, addresses and purchases by changing the identifier. This risks breach of privacy obligations and customer trust.
Evidence: Using Test Account A, the order belonging to Test Account B was returned; screenshots 3 and 4 (personal data redacted).
Steps to reproduce: (1) Log in as Test Account A. (2) Open one of your own orders and note the identifier. (3) Change the identifier to the value belonging to Test Account B. (4) Observe that the order details are returned.
Recommendation: Enforce server-side authorisation on every order request, comparing the order’s owner with the session user; use unpredictable identifiers as defence in depth; add automated tests for access control.
References: OWASP Top 10 A01 Broken Access Control; CWE-639.

Rating Risk Fairly

Risk combines likelihood (how easy to exploit, how exposed) and impact (what is lost). A simple matrix:

Likelihood \ ImpactLowMediumHigh
HighMediumHighCritical
MediumLowMediumHigh
LowLowLowMedium

CVSS offers a standard score, but add context: an internal-only flaw on an isolated system differs from an internet-facing one. Explain your reasoning, so clients can challenge or accept it. Never inflate severity to impress, and never downplay a finding to please a client.

Writing the Executive Summary

Write it last, but make it first. Answer these questions in plain language:

  1. What did we test, when and why?
  2. What is the overall security posture, in one sentence?
  3. What are the three most important problems and their business impact?
  4. What should the organisation do first?
  5. What went well?

Example opening: “We tested the online shop’s staging environment between 5 and 9 March. The application has a sound login design, but we found a flaw that lets any customer view other customers’ orders. We recommend fixing this and two related access-control issues within 14 days.” Avoid acronyms and tool names in this section.

Writing Recommendations That Get Done

  • Be specific: “Parameterise queries in OrderRepository.find()” beats “sanitise inputs”.
  • Fix root causes: recommend the design change as well as the single patch.
  • Offer options: short-term mitigation and long-term fix.
  • Prioritise: tell readers what to do this week, this quarter and this year.
  • Assign owners and timelines in the remediation roadmap, where the client allows.
  • Link to authoritative guidance: OWASP cheat sheets, vendor documentation and standards.

Handling Evidence and Sensitive Data

  • Redact personal data, passwords and secrets in screenshots and logs.
  • Store evidence encrypted, share it through secure channels and destroy it as agreed.
  • Do not copy more real data than is needed to prove the issue.
  • Mark the report with its classification, and limit distribution.

Practice Lab: Write a Report From Your Own Lab

  1. Take the findings from your Nmap and vulnerability scans in earlier lessons; see Scanning and Enumeration With Nmap.
  2. Choose three findings. For each, write a full finding using the template above, with at least one piece of evidence.
  3. Rate each with the likelihood and impact matrix and explain why.
  4. Write a half-page executive summary for a non-technical reader.
  5. Ask a friend with no technical background to read the summary and tell you what they would do first. If they cannot answer, rewrite it.
  6. Fix one issue, retest, and add a “retest result” line to the report.

Common Mistakes

  • Pasting raw scanner output as the report.
  • Vague titles and recommendations that nobody can act on.
  • Inflating every finding to “Critical”, so priorities are lost.
  • Missing scope, dates and limitations, which weakens credibility.
  • Blaming people rather than describing systems and processes.
  • Leaving out the good news.

Practice Exercises

  1. Rewrite the vague finding “The server has weak security” into a complete finding with title, impact and recommendation.
  2. Rate three imaginary findings using the likelihood and impact matrix, and defend your ratings.
  3. Write a five-sentence executive summary for a test where nothing critical was found but several medium issues exist.
  4. List five items you would include in the “positive observations” section for a well-defended environment.

Mentor tip: The report is the product. Clients rarely remember how clever your technique was, but they always remember whether your report helped them fix things. Write for the person who has to open the ticket on Monday morning.

Educational use only: the sample finding is fictional and uses a reserved test domain.

Key takeaways
  • Layer the report: executive summary first, technical detail behind.
  • Every finding needs a clear title, impact, evidence, reproduction steps and a specific fix.
  • Rate risk using likelihood and impact, with transparent reasoning.
  • Protect evidence and redact sensitive data.
  • Include positive observations and a prioritised remediation roadmap, then retest.

Related Lessons