Contact

What an ethical penetration test covers, and what to expect from the assessment

A richly furnished salon: a woman plays the piano while others listen and a dog sleeps on the carpet. (Mihály Munkácsy, The Music Room, 1878)
Mihály Munkácsy: The Music Room, 1878 · The Metropolitan Museum of Art · Public domain
Category
Security
Published
Reading time
~3 min read
Author

Summary

A penetration test is an authorised, controlled attack on your systems. It reveals what a real attacker could achieve before one actually tries.

What is a penetration test?

In a penetration test (or pentest), security specialists attempt to breach a system with the client's written authorisation and within strictly agreed boundaries, using the same techniques a real attacker would use. The goal is not disruption, but uncovering weaknesses and proving their real-world impact.

A test captures the state of a system at a specific point in time. It is not a certificate guaranteeing that a system is invulnerable, but an expert assessment of what could be compromised under the agreed scope.

How does it differ from a vulnerability scan?

Vulnerability scanning versus penetration testing
DimensionVulnerability scanPenetration test
MethodPrimarily automated scanning toolsHands-on expert testing supported by specialised tools
OutputA list of potential vulnerabilitiesVerified, exploitable flaws and their real business impact
DepthBroad but surface-levelTargeted and deep; chains multiple flaws together
Best used forRegular, frequent hygiene checksPeriodic in-depth assessments and after major changes

The two do not replace each other; they complement each other. Automated scans can run continuously, while a penetration test uncovers flaws no scanner can spot, such as broken business logic.

What can a test cover?

  • Web applications and APIs: authentication, access control, input handling, and session management.
  • Mobile applications: communication between the app and backend servers, plus data stored on the device.
  • External network: everything reachable from the public internet.
  • Internal network: what an attacker who has already gained a foothold (or an insider) could reach.
  • Cloud environment: configuration, IAM permissions, and publicly exposed storage buckets.
  • The human factor: social-engineering and phishing simulations, when explicitly requested by the client.

Depending on how much prior information the tester receives, engagements are classified as black-box (no internal knowledge), grey-box (partial knowledge, such as standard user accounts), or white-box (full visibility, including architecture docs and source code). The more context testers receive, the more efficiently they spend their time on high-value findings.

How the engagement unfolds

  1. Scoping: defining which systems are in scope, when testing takes place, which methods are allowed, and what is off-limits.
  2. Written authorisation: testing without the system owner's explicit consent is illegal.
  3. Reconnaissance: testers map out the target surface.
  4. Discovery and exploitation: identified weaknesses are safely exploited to demonstrate real impact.
  5. Reporting: every finding is documented with severity, evidence, and actionable remediation advice.
  6. Readout session: walking through the findings together with your developers and operations team.
  7. Retesting: verifying after remediation that the vulnerabilities have been properly closed.

What to expect from the report

A good pentest report serves two audiences. For executives, it provides a concise summary of overall risk exposure and top priorities. For the engineering team, it supplies enough technical detail to reproduce and fix every issue.

The value of a test lies in the fixes that follow. When commissioning a pentest, reserve engineering capacity for remediation and include retesting in the agreement. How often you should test depends on system criticality, release velocity, and regulatory requirements, but always test after major architectural changes.

Key takeaways

  • A penetration test is an authorised attack simulation conducted within strictly agreed boundaries.
  • Automated vulnerability scans are broad and frequent; penetration tests are deep and targeted; the two complement each other.
  • Never begin testing without written authorisation and a clearly defined scope.
  • A strong report gives leadership a clear view of risk and gives engineers reproducible steps and remediation guidance.
  • The real value of a test is realised through remediation and verification retesting.

Frequently Asked Questions

Can a penetration test disrupt live production systems?

That risk is manageable: during scoping you can rule out disruptive techniques, and testing can be performed in a staging environment or during low-traffic windows. Testers and your operations team stay in direct contact throughout the engagement.

When should we commission a penetration test?

Typically before launching a new system into production, after major architectural changes, and at regular intervals for business-critical systems or those handling sensitive data. Several industry standards and enterprise client contracts also mandate periodic testing.

If the test finds no vulnerabilities, does that mean our system is completely secure?

It means no exploitable flaws were found within the agreed scope and timeframe. That is a strong positive signal, not a permanent guarantee: both your software and the threat landscape evolve, so security remains an ongoing discipline.

Facing a similar decision? We are happy to talk it through.

Let's talk