Security testing

Choose, scope and commission a security assessment, interpret its findings and verify that the fixes work.

A useful security test starts with a decision you need to make. You might need to know whether two customer accounts are properly separated, whether a new service is ready for release, or whether your team can detect a realistic intrusion. Those questions call for different work.

This section takes you from choosing an assessment to closing its findings. It is written for service owners, technical leads and people buying testing for a UK organisation. The examples are fictional and the working documents are editorial tools, not certificates of assurance.

Choose your starting point

Your immediate task Read What you will leave with
Decide which assessment answers your question Assessment types A comparison and three explained choices
Agree the assets, permissions and limits Scoping and authorisation A worked scope and a list of decisions to resolve
Compare testing proposals Choosing a provider Questions and an evidence-based comparison sheet
Turn a report into work your team can do Reading a report An annotated finding and report-review checklist
Decide what to fix first Prioritising remediation A sample backlog with reasons, owners and review points
Establish whether a fix is effective Retesting A verification plan and a completed closure record

A practical commissioning sequence

For a fictional customer portal, the owner first writes: “We need evidence that one customer cannot retrieve another customer's documents.” That question identifies the relevant user roles and business functions. A scope then identifies the test environment, representative data and allowed actions. The provider's proposal should explain how its work will answer the question. The report should show the coverage and its limits; the service team should own the resulting improvements.

Reserve time and funding for that final stage. A report left in an inbox does not change the service. Use the scope worksheet to capture reporting and retest expectations before work starts.

Understand the limits of a test

The NCSC's commissioning guidance treats penetration testing as one part of vulnerability management. Its conclusions are bounded by the systems, access and period examined. Routine maintenance, secure development and monitoring continue between assessments.

An assessment that excludes customer administration cannot answer questions about that function. A test of a representative environment also needs an explanation of any differences from production. Record those limitations with the decision the report informs.

Historical context

The 2018 conference archive includes a documented red-teaming presentation. Historical material can introduce useful questions, while these guides provide a separate, newly written route through commissioning and follow-up. The historical speakers are not the authors or endorsers of this guidance.

Sources and further reading

  1. NCSC: Penetration testing — checked 2 November 2026

In this collection

6 pages

Read a security-testing report

Review coverage, evidence, impact and limitations, then turn an annotated fictional security finding into an actionable remediation task.

Prioritise security findings

Build a remediation backlog that considers exposure, exploitation evidence, business impact, dependencies and the proof needed to close each issue.

Security-test scope worksheet

Prepare a clear test scope with blank fields, a fictional worked example and editable downloads covering permission, boundaries and reporting.

CRESTCon and IISP Congress 2018

Explore the documented 2018 technical programme and a surviving Netscylla red-teaming presentation, with clear source and availability notes.