Blog

Pentest, audit and automated scan: what each one answers

Compliance audits, vulnerability scans and pentests answer different questions. See what each delivers and when your company needs it.

Five lock picks, tools for opening locks, side by side

Audits, vulnerability scans and pentests often sit on the same budget spreadsheet, sometimes under the same name. They are different jobs, done by different people, answering different questions. When your company buys one expecting the answer of another, the report arrives correct and still does not help with the decision that had to be made.

Three different questions

  • Compliance audit: do the controls the company says it has exist, are they documented, and are they followed?
  • Automated vulnerability scan: which known flaws show up in the systems today?
  • Pentest: how far would someone with intent and skill get, starting from what exists today?

All three questions are legitimate. The trouble starts when one of them is treated as the answer to the others.

What an audit checks

An audit compares the company with a framework: ISO/IEC 27001, PCI DSS for anyone who stores, processes or transmits card data, the CIS controls, a sector regulator’s requirements, or internal policy. The auditor reads policies, procedures and records, interviews the teams and asks for evidence by sampling: last quarter’s access review, the change log, proof that the backup was tested.

The result says whether a control exists and works as described. It is what a certifier, a large customer or a regulator wants to see, and it puts the house in order: who owns each process, where the records are, what happens when someone leaves the company.

The limit lies in the method. An audit checks what was declared, by sampling. A password policy can be written, approved and applied to almost every account, and there can still be an old service account, outside the inventory, whose password never changed. The audit was not designed to look for that account.

What an automated scan finds

A vulnerability scanner is a tool that walks through addresses, ports and applications and compares what it finds with a database of known flaws: software versions with a published CVE, insecure settings, expired certificates, services that should not be exposed. Tools like this run fast, cover many assets and can run every week.

It is a low-cost way to keep basic controls in place across many assets: to know when a server missed an update or someone opened a port they should not have. Mature vulnerability management programs scan continuously and track time to fix by severity.

Its limits come from the method too. A scan mostly flags what it recognizes, and does not always confirm whether a flaw is exploitable in that environment. That produces false positives, which eat the team’s time, and false negatives, because a scanner understands little about context. As a rule it does not know that an ordinary user can see another customer’s data by changing a number in the URL, nor that three low-severity flaws together give access to the file server. A tool’s severity is not a business risk.

What a pentest shows

In a pentest, specialists attack the systems within an agreed scope and rules, as an attacker would. Public methodologies such as PTES and the OWASP WSTG describe the phases: information gathering, analysis, controlled exploitation and reporting. What differs from a scan is human judgment: the specialist confirms whether a flaw is real, understands the system’s logic and chains small findings into a concrete impact.

The result answers the question a board usually asks: what could an attacker achieve, and with how much effort? Each finding comes with evidence of the exploitation, its business impact and the way to fix it.

A pentest has limits too. It is a snapshot of a period, covers the agreed scope and depends on the experience of the testers. It does not replace continuous scanning, because the environment changes the following week, nor the audit, because it does not assess compliance and processes comprehensively.

An illustrative example

This scenario is a composite of situations common in security assessments, not a real customer’s case.

A retailer passed its annual PCI DSS audit. The policies were approved, the segmentation of the payment network was documented and the quarterly scans were up to date: the external ones passed, the internal ones showing a few open medium-severity flaws and a remediation plan.

In the internal network pentest, run with a low-privilege user, the testing team found a monitoring server installed years earlier by a supplier. It did not appear in the inventory used for the audit, had access to both networks because the tool needed it, and stored credentials for several machines. The scan had listed the server with a medium-severity alert, which sat in the queue. Combining that alert with a permissive share setting, the test reached a device inside the payment network that the documentation said was isolated.

What failed was the scope. The server outside the inventory stayed outside the assessment, and the segmentation test that PCI DSS requires did not cover the path it opened. The scan found the alert, and the pentest showed what the alert meant in that environment. The company fixed the server, reviewed its inventory and the scope of its PCI DSS assessment, and began prioritizing the scan queue by the path each flaw opens, not only by the tool’s score.

How to combine the three

  • Continuous scanning for hygiene: new assets, overdue updates, settings that changed.
  • Periodic audits to prove and organize controls, in the cycles your framework requires.
  • Pentests before significant changes, after major fixes, and in cycles set by business risk and your sector’s requirements.

The results feed each other. The pentest shows which scan alerts deserve priority and which audit controls exist only on paper. The audit shows where the inventory is out of date, which improves the scope of the next test.

What to ask a supplier

  1. How much of the work is done by hand, and how much is tool output? Ask for a sample report.
  2. Does each finding come with evidence of exploitation and its business impact, or only the tool’s score?
  3. Does the test chain flaws to show a path to a critical asset?
  4. Which methodology is followed, and how does it show in the report?
  5. Is the retest of the fixes included, and how is it recorded?
  6. Who will run the test, and what certifications do they hold?

How Leonnes does pentests

In a Leonnes pentest, specialists with certifications such as OSCP, OSWE and OSCE test applications, APIs, mobile and infrastructure by hand, chaining flaws like a real attacker. You follow progress through a weekly status, receive the technical, executive and financial risk reports and, after the fixes, the retest report. See Leonnes’ pentest.

See what an attacker would see first

We reply within one business day.