AnySec
Vulnerability Assessment Report: What It Must Contain — AnySec Insights
← InsightsPenetration Testing · 8 min read

Vulnerability Assessment Report: What It Must Contain

A scanner export is not a vulnerability assessment report. The sections a useful report needs, the coverage proof a reviewer can check, and the red flags to reject.

By AnySec engineering

The short answer

A useful vulnerability assessment report states what was actually scanned and what was not reachable, names the method, tools and credentials used, separates findings a person confirmed from raw scanner output, ranks findings with stated reasoning for your environment, and gives remediation per finding. A scanner export with a severity column is an input to that report, not the deliverable.

Who this is for

This is for a CISO, head of infrastructure or risk owner who has bought, or is about to buy, a vulnerability assessment and wants to know what the document at the end must prove. It is also for whoever will show that document to an auditor, a partner or a board. Before the report there is the purchase decision, covered in what a vulnerability assessment quote must say. After the report there is the work of turning findings into tickets, which scan results triage walks through. This article is about the middle: the document itself, and how to tell whether it describes your estate or only the scanner's view of it.

Why a scanner export is not a report

Disappointing assessment deliverables fail in the same few ways.

  1. They describe the tool, not your estate. A long list of findings sorted by the tool's own severity tells you what the rule set matched. It does not tell you which of your assets were reached, with which credentials, or on which day.
  2. They hide what was missed. Hosts that were offline, firewalled, rate-limited or out of reach during the window silently drop out of the output. The reader sees a short list and assumes a clean estate.
  3. They mix confirmed and unconfirmed findings. Version-based detections and banner matches are guesses until someone checks them. When they sit beside confirmed findings with the same formatting, the reader cannot tell which to fix first.
  4. They rank without reasoning. A critical label with no explanation of reachability, exposure or asset role forces you to redo the analysis yourself.

The report is the part of the engagement that outlives the people who ran it. That is why its contents belong in the quote, before any scan runs.

Two hands pass a thin stack of findings sheets with a single lime marker across a dark table, beside a much thicker stack of asset inventory pages
The thick stack is what was in scope; the thin one is what was found. A report must show how the second relates to the first.

The report anatomy matrix

Use this as a checklist when you write the quote or review a draft. It contains no scores or thresholds, because the right targets come from your own risk and your own asset inventory.

SectionWhat it must showWhat a weak version looks like
Scope as scannedThe in-scope asset list, the address ranges and applications actually covered, and the scan windowA sentence such as "the external environment"
Coverage tableEvery in-scope asset with a status: scanned, scanned with credentials, scanned without credentials, unreachable, or excluded with a reasonNo table; only the assets that produced findings
Method and toolsTool names and versions, rule or plugin set date, scan policy, whether authenticated checks were used"Industry-standard scanner"
ConstraintsRate limits, blackout periods, systems deliberately not touched and whyNone stated
Validated findingsFor each important finding, the affected asset and location, the evidence a person observed and why it is reachableScanner output pasted with no review
Unconfirmed findingsA separate list of detections not verified, labelled as suchMixed into the main table
Ranking and reasoningHow each finding was ranked for your environment: exposure, known exploitation, asset role, compensating controlsOnly the tool's own severity
Remediation guidanceA specific fix or mitigation per finding, and an owner field for you to fill"Apply vendor patches"
Duplicates and exclusionsHow duplicates were merged and false positives removedRaw counts only
Not coveredWhat the assessment does not claim: logic flaws, chained attacks, social engineering, unscanned assetsA generic disclaimer
Retest basisHow a later rescan can be compared with this one: same scope, same policy, same tool versionNo baseline kept

The coverage table is the first thing to read

Most readers go to the findings. Read the coverage table first. A finding list tells you about the assets that were reached; the coverage table tells you how much of your estate that is. Compare it against your own asset inventory, not the provider's. Where the two disagree, the difference is either an asset that was missed or an inventory that has gone stale, and both are worth knowing.

Ask in particular for the split between authenticated and unauthenticated checks. An unauthenticated scan sees what an outsider sees on the network. An authenticated scan sees installed packages and configuration, and typically finds more. Neither is better in the abstract. The report must say which was run on which assets, because the same host can look very different depending on the answer.

Validated, unconfirmed and false: three different piles

A good report keeps three groups apart.

  • Validated: a person checked the finding and recorded what they observed. These are the ones to fix first.
  • Unconfirmed: the scanner reported it, but nobody verified it. The report should say why it was not verified, and what check would settle it.
  • Removed as false or duplicate: listed by count and reason, so you can see the reviewer did the work.

If a provider reports a single list, ask how many findings were dropped before it reached you. A report in which nothing was ever removed was probably never reviewed. The follow-on question, how to rank what is left, is the subject of scan results triage; the report should already contain the reasoning, not leave it for you.

Ranking that explains itself

A severity label is a starting point, not a decision. For each important finding the report should say why it matters in your environment: whether the asset is reachable from the internet or only from inside, whether the weakness is known to be exploited in the wild, what the asset does, and whether another control already limits the impact. Public references such as the CISA Known Exploited Vulnerabilities catalog can inform the exploitation question. The ranking should be written so that you can disagree with it, which means it has to show its inputs.

Remediation that someone can act on

For every finding the report needs a fix an engineer can start on: the specific update or configuration change, the version to move to or the setting to change, any caveats that could break a workload, and a way to confirm the fix. Where no fix exists yet, it should name a mitigation and say what it does and does not cover. Leave an owner and a due-date column for your own team; a report with no place for them is a report that will not be tracked.

Red flags when you review a draft

  • A cover page and a scanner export, nothing in between.
  • No coverage table, or a table listing only assets that had findings.
  • No named tool version or rule-set date.
  • No statement of whether checks were authenticated.
  • All findings in one list, with no split between confirmed and unconfirmed.
  • Severity copied from the tool with no reasoning.
  • Remediation lines that say only "apply patches".
  • No statement of what was not covered.
  • No way to repeat the same scan for comparison later.

Ask for the report before you sign

Treat the deliverable as a quote line. Before work starts, ask for a redacted sample report from a past engagement, the headings that will be used, the delivery date, whether a walkthrough is included, who keeps the raw scan data and for how long, and whether a rescan is priced in. If you cannot get a sample, treat that as information. An assessment report also does not replace testing the weaknesses it finds: for the question of what an attacker could chain together, see vulnerability assessment versus penetration testing and what a real pentest report looks like.

Limitations

This article is not legal or regulatory advice and does not state what any auditor, supervisor or licensing authority requires of an assessment report. It uses no external statistics and no incident case studies. The matrix is a review checklist to adapt to your own estate and contract, and it contains no thresholds or scores. An assessment describes the assets reached at one point in time with one tool set; it does not prove that the environment is free of vulnerabilities, does not find logic flaws or chained attacks, and is not a certification. New weaknesses are published daily, so a report ages the day it is written.

Decision and next step

If you can read the draft and, for each asset group, find what was reached, how it was checked, which findings were confirmed, why they were ranked as they were and what to do about them, the report is doing its job. If you cannot, ask for the evidence behind the summary or reject the deliverable. To have the report specified in the quote before any scan runs, get a fixed quote from AnySec and we will map your asset inventory to a scope, the credentials to agree and the sections you will receive.

Sources and review

The framing follows public guidance: NIST SP 800-115 Technical Guide to Information Security Testing and Assessment, which covers planning, executing and reporting on security assessments, and the CISA Known Exploited Vulnerabilities catalog, read 2026-11-21; check the current versions, and read your scanner vendor's documentation for its own coverage limits. AnySec prepared the report matrix from the public structure of assessment engagements; no AnySec-original statistics are used. Author: AnySec Engineering. Published 2026-11-21; reviewed by the AnySec security team.


Related reading

To have the deliverable specified up front, get a fixed quote.

Frequently asked questions

What should a vulnerability assessment report contain?
The scope as actually scanned, including the assets that could not be reached, the method and tools with versions, the dates and credentials used, findings that were validated by a person rather than only exported, a ranking that explains its reasoning, remediation guidance per finding, and a plain statement of what was not covered. A raw scanner export is an input to a report, not the report.
How do we know the scan covered everything in scope?
Ask for a coverage table that lists every in-scope asset with a status: scanned, scanned with credentials, scanned without credentials, unreachable, or excluded with a reason. Compare it against your own asset inventory. An assessment that cannot show what it did not reach cannot claim to describe your exposure.
Is a high severity count a sign of a good assessment?
No. Counts depend on the scanner's rule set, on whether duplicates were merged and on whether false positives were removed. A good report explains how each ranking was reached for your environment and which findings were confirmed, and it does not use volume as a quality signal.
Should the report include proof that findings are real?
Yes. For findings ranked as important, ask for the evidence that a person confirmed them: the observed version or response, the affected asset and where on it, and why it is reachable. Findings that were not validated should be labelled as unconfirmed rather than mixed in with confirmed ones.
Does an assessment report replace a penetration test report?
No. An assessment inventories and ranks known weaknesses across many assets; a penetration test tries to chain weaknesses into impact on a smaller scope. They answer different questions, and each report should say what it does not claim.

Rather not learn this in production.

Talk to the engineers behind these write-ups — no sales script, a straight read on where you stand.

Get a fixed quote