A penetration test (pentest) is a check in which specialists act like attackers: they look for ways around the defences and find out how far an attack could go. NIST defines it as security testing in which evaluators mimic real-world attacks to identify ways to circumvent the security features of an application, system or network.
The main difference from an automated check: a pentest does not just list vulnerabilities, it shows which of them can really be exploited and what an attacker would gain.
Pentest, vulnerability scanning, red team and bug bounty
| What it is | Outcome | When it fits | |
|---|---|---|---|
| Vulnerability scanning | An automated check of systems for known vulnerabilities | A list of possible vulnerabilities, not checked for being real | Regularly, between pentests |
| Pentest | A manual check by specialists within an agreed scope | Confirmed vulnerabilities, attack chains, evidence, recommendations | After major changes and on a schedule set by the risks |
| Red team | A simulated targeted attack on the whole organisation that only a few people know about | An assessment of whether monitoring and the team notice the attack and how they respond | When basic protection and monitoring are already in place |
| Bug bounty | A reward programme for external researchers who find vulnerabilities | Individual researcher reports | For public products and services, as an addition |
A scanner answers “which known vulnerabilities do our systems have”, a pentest — “what can an attacker really do”, a red team — “will we notice the attack and how will we respond”.
Types of penetration test
- External perimeter. Everything reachable from the internet: VPN gateways, email, websites, remote access services, firewalls. The scenario is an attack from outside with no access at all.
- Internal network. The “attacker is already inside” scenario: an infected laptop, an employee’s or contractor’s account. Testers check Active Directory, network segmentation, servers and workstations: can critical systems be reached from a single foothold.
- Web applications and APIs. Websites, customer portals, APIs. For such checks there is an open methodology, the OWASP Web Security Testing Guide (current stable version 4.2).
- Social engineering. Simulated phishing campaigns and other checks of employees. The topic is pressing: according to CERT-UA, the team handled 1,727 phishing cases in 2025 against 843 a year earlier.
There are also penetration tests of mobile apps, wireless networks and cloud environments — the scope always depends on what needs to be checked.
Black, grey and white box
- Black box. Testers get no information — like an external attacker. Realistic, but part of the time goes on reconnaissance.
- Grey box. Partial information: an ordinary user account, a list of addresses, a basic diagram. Allows a deeper check in the same time.
- White box. Full access to documentation, configurations and sometimes code. The deepest check, closer to an audit.
Stages of a penetration test
NIST SP 800-115 describes four phases of penetration testing:
- Planning. Scope, goals, rules and timing are agreed, and management approves the test in writing. No testing happens at this stage.
- Discovery. Information gathering, scanning, finding and analysing vulnerabilities.
- Attack. Attempts to exploit vulnerabilities within the agreed limits. If an attack opens new possibilities, testers go back to discovery — and so on.
- Reporting. A description of what was found, the evidence and recommendations.
In practice one more step is added — a retest after the company has fixed what was found.
What the report must contain
- A management summary: which business risks were found, without technical details.
- A list of vulnerabilities with a severity rating (for example, on the CVSS scale), a description of the attack scenario and evidence.
- Remediation recommendations and priorities: what to fix first.
- Retest results: what has been fixed and what has not.
A report without evidence and priorities is just a list that is hard for an IT team to work with.
How often to run a pentest
- After major changes: launching a new application, migrating to the cloud, rebuilding the network, connecting a contractor to your systems.
- Regularly — the schedule is set by risks and industry requirements. For example, for companies that process payment cards, PCI DSS v4.0 (requirement 11.4) requires internal and external penetration testing at least once every 12 months and after significant changes to the environment.
- Between pentests — regular vulnerability scanning and security event monitoring.
How to prepare a company
- Define the goal. A client or partner requirement, compliance with a standard, checking a new application, checking whether your monitoring notices an attack.
- Agree the scope. Addresses, domains, applications — and what must not be touched.
- Set the rules. Timing, allowed methods (for example, no denial-of-service attacks), contacts in case of a failure.
- Get written approval from the system owner. If systems are hosted in the cloud or by a provider, check its rules on testing.
- Decide who knows about the test. If the IT team and monitoring are not warned, the pentest also checks whether they notice the attack.
- Check backups of critical systems before starting.
- Plan time for fixes and a retest — without them, a pentest only records the problems.
Pentest and monitoring
A pentest shows the state of protection on a given date. A new vulnerability or a configuration mistake may appear the next day. That is why regular scanning and security event monitoring are needed between tests, to notice an attack that is already under way. More on the SOC as a Service page.
How Bereg-IT can help
Penetration tests are carried out by Bereg-IT’s own team. We test the external perimeter, the internal network, web applications and APIs, and run social engineering checks. Scope and rules are agreed before work starts.
As a result, you receive a report with a list of vulnerabilities, severity ratings and evidence, remediation recommendations, a management summary — and a retest after the fixes. Penetration testing complements the other areas of information security, including vulnerability management.
The first consultation is free. Send a request and an information security specialist will get in touch.
Sources
- NIST CSRC Glossary, penetration testing (definition from NIST SP 800-115): https://csrc.nist.gov/glossary/term/penetration_testing
- NIST SP 800-115, Technical Guide to Information Security Testing and Assessment (September 2008), section 5.2: https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-115.pdf
- OWASP Web Security Testing Guide: https://owasp.org/www-project-web-security-testing-guide/
- SSSCIP, “CERT-UA handled almost 6,000 cyber incidents in 2025: the number of enemy attacks rose by 37%” (07.01.2026, in Ukrainian): https://cip.gov.ua/ua/news/cert-ua-u-2025-roci-opracyuvala-maizhe-6000-kiberincidentiv-kilkist-vorozhikh-atak-zrosla-na-37
- FIRST, Common Vulnerability Scoring System (CVSS): https://www.first.org/cvss/
- PCI SSC, Self-Assessment Questionnaire D for Merchants, PCI DSS v4.0 (April 2022) — requirements 11.4.2, 11.4.3: https://listings.pcisecuritystandards.org/documents/PCI-DSS-v4-0-SAQ-D-Merchant.pdf