This guide is for CTOs, product owners and engineering managers who have a penetration test booked for a customer request, an audit or a major release. It is written from the side of the team that builds and fixes the application, not the testing firm. If you need the testing itself, see application security testing.
In this guide
- What does a penetration test need from your team before day one?
- Preparation checklist
- Which environment should the testers use?
- How should findings be triaged, fixed and retested?
- When is a penetration test the wrong next step?
- Where AI fits
- What delivery record exists, and what does not?
- Alternatives: who prepares and fixes, and how to choose
- Limits of this guide
- Frequently asked questions
- Next step
What does a penetration test need from your team before day one?
A penetration test needs four things from the application team before it starts: a written scope, an environment the testers can reach, accounts for every role, and named people who answer questions during the test. Without them, testing days go to access problems.
List every asset by name: web applications, API base URLs, mobile builds, admin panels, background jobs that accept input, and the cloud resources behind them. Write the out-of-scope list just as carefully, especially third-party services you do not own, such as payment gateways, email providers and identity platforms, which you usually cannot authorise anyone to test.
Keep the scope in proportion to the testing days. A broad scope over a short engagement produces a shallow test of everything and a deep test of nothing. When the budget is fixed, rank assets by the data they hold and the money or permissions they move, and put the top of that list in scope first.
Choose the knowledge level on purpose. A black-box test shows what an outsider finds; a grey-box test, with accounts and API documentation, spends the time on authorisation and business logic, which is where business applications tend to fail. For most product teams, grey box gives more findings for the same days. The OWASP Web Security Testing Guide is a useful shared reference for what a web test covers.
Preparation checklist
-
Write the scope and the exclusions
Assets by name and URL, roles to test, out-of-scope systems, and any action the testers must not take, such as sending real emails or charging real cards.
-
Sign the authorisation and rules of engagement
Who approved the test, the testing window, source addresses, and the stop-and-call rule if the testers find live customer data or cause an outage.
-
Prepare the environment
A production-like build of the release being tested, with realistic but anonymised data, so features that depend on data can actually be exercised.
-
Create test accounts for every role
Administrator, standard user, read-only user and, in a multi-tenant product, two tenants, so testers can try to cross from one tenant into the other.
-
Clear the access path
Allowlist the testers' addresses in the firewall and web application firewall, issue VPN or MFA access, and test every login before day one.
-
Fix what a scanner would find
Run dependency and vulnerability scans first and close the obvious items, so paid testing time goes to logic flaws, not to outdated libraries or certificate warnings.
-
Brief the people involved
Tell operations and monitoring the dates and source addresses, set up one shared channel, and name a technical contact who answers within the working day.
-
Take a backup and freeze the build
Back up the environment's data and avoid deploying to it during the test, so findings can be reproduced against a known version.
-
Agree the retest now
Write into the order how many retest rounds are included, which severities they cover, and how long after the report the retest window stays open.
For the certificate and cipher part of step 6, the AdminKit SSL/TLS check from AdminKit, a Netbase brand, gives a quick certificate and cipher check for a website before the testers arrive.
Which environment should the testers use?
The environment decides what the test can prove. Production shows the real configuration but carries the risk of damaging live data; staging is safer but only useful if it matches production.
| Option | Strength | Risk | Choose it when |
|---|---|---|---|
| Production | Real configuration, real integrations, nothing to keep in sync | Test actions can change or expose live customer data; outages affect customers | The test targets the perimeter or configuration, and destructive actions are excluded in writing |
| Production-like staging | Same build and infrastructure code, safe to attack, test data under your control | Drift: a staging-only setting can hide or create findings | Most application tests; keep it built from the same infrastructure code as production |
| Short-lived copy for the test | Clean, isolated, destroyed afterwards; a second copy makes the retest easy | Effort to build; external integrations are often stubbed | Infrastructure is defined in code and the release is large enough to justify a dedicated copy |
Do not test production when the scope includes write actions on customer records, payment flows or bulk emails, unless those actions are disabled for the test accounts. Do not test an empty staging environment either: without data, the testers cannot check whether one user can read another user's records.
How should findings be triaged, fixed and retested?
Findings should be triaged within days of the report, fixed at the root cause with a regression test, and retested in the window agreed before the test. The retest proves the fix works; it does not prove the application is now secure, because it only checks the reported issues.
Start triage with one owner per finding and a severity your team agrees with. The CVSS v4.0 specification from FIRST says its scores measure technical severity, not risk, so adjust the order of work for the data and business process each finding touches. A medium finding in a payment flow may come before a high one in an internal tool.
If the testers reached another tenant's invoice through one endpoint, check every endpoint that loads records by identifier, not only the one in the report. Turning the vulnerable feature off is a valid emergency measure, but it is not a fix, and a good retest will say so.
Add an automated test for each fixed finding before closing it. The test keeps the weakness from returning in a later release, and it gives the retest a clear expected behaviour. Record the version, environment and commit of each fix, so the testers retest the right build.
Plan the retest window when you book the test, not after the report arrives. If an audit or customer deadline drives the test, count backwards from it: testing days, report, fixing time and retest must all fit before the date.
When is a penetration test the wrong next step?
A penetration test is the wrong next step when the application has never been scanned, has no access-control design, or is about to change substantially. In those cases, a vulnerability scan, a code review or a threat-modelling session finds the same basic issues for less effort, and the test is then booked for a stable release.
- No scanning yet. Automated dependency and vulnerability scans in CI find outdated components quickly; paying testers to report them wastes the engagement.
- No authorisation model. If roles and permissions were never designed, the test will report symptoms of one missing design. Design the roles first, as in our security review gate for AI-generated code.
- A rebuild is weeks away. Testing code that is about to be replaced produces findings nobody will fix. Test the new release instead.
- You need assurance about a vendor, not an application. Ask the security questions for a development partner first.
Where AI fits
AI helps most with the paperwork around a test. A model can draft the asset inventory from route files and API definitions, group duplicate findings from scanners and the report, and draft regression tests for each fix. Engineers still confirm the scope, decide severity and review every generated test, because a wrong inventory or a test that passes for the wrong reason gives false comfort. Netbase works with the major commercial and open-source AI tools and models, chosen per project.
- Available capability
AI-assisted engineering under human review
AI-assisted changes are held to the same secure code review, version control and scanning as any other change; not yet tied to a published penetration-test case.
What delivery record exists, and what does not?
Netbase's published security practices are secure code review and version control, TLS in transit and AES at rest, role-based access control, MFA for admin dashboards, vulnerability scanning and penetration testing, and disaster recovery; NDAs, data processing agreements and SLAs are available on request. Security by design is part of how Netbase delivers.
Netbase holds ISO 27001 certification and a SOC 2 Type II attestation for its own operations. These cover how Netbase works, not the security of a client's application, and they are not test results.
Delivery follows a six-step lifecycle from discovery to ongoing support, with weekly reviews and a named project manager, which is where a test, its fixes and the retest are scheduled. Netbase also works as offshore development partner on a multi-tenant cloud ERP for a US client, a SaaS product for small and mid-sized businesses and the kind of product where tenant isolation is the first thing a tester tries to break.
Most Netbase projects are delivered on fixed-price contracts. No Netbase record publishes penetration test results, finding counts, retest outcomes or response times, and none is offered here as evidence that a particular control works.
Alternatives: who prepares and fixes, and how to choose
The test is usually bought from an independent firm; the question is who prepares the application and fixes findings.
- Your in-house team. Best when the engineers who built the code are still there and have time in the next release to fix findings. Check that someone owns the retest date.
- The development partner that built the application. Best when the partner holds the code and deployment; ask for the fix plan and regression tests in writing. See the security questions for a development partner and the SaaS security checklist.
- A separate maintenance team. Best when the original team has left; give it time to learn the code before the test, through application maintenance and support, and keep patching and scanning between tests.
Choose by three criteria: who can reproduce a finding fastest, who can ship a fix without breaking the release, and who will still be responsible when the next test comes. For testing that runs alongside each release, see quality engineering and testing.
Limits of this guide
- It covers application tests (web, API and mobile back ends), not physical, social-engineering or red-team exercises.
- It is not legal or compliance advice; an auditor or the standard you report against decides what evidence is required.
Plan the next step with a Netbase consultant
Frequently asked questions
Start when the test is booked. The scope, authorisation and test accounts depend on decisions by people outside engineering, and the environment often needs data and access changes, so leaving them to the last week moves the problem into the paid testing days.
Most application tests should run against a production-like staging environment built from the same infrastructure code as production. Test production only for perimeter and configuration checks, with destructive actions excluded in writing.
Fix whatever an automated scanner already reports, such as outdated dependencies, missing security headers and certificate problems. Testers will report those too, and the time is better spent on authorisation and business logic.
It depends on the order, so agree it before the test starts. Write down how many retest rounds are included, which severities they cover and how long the retest window stays open after the report.
Next step
If a penetration test is coming up and you want the scope, environment and fix plan prepared by the team that will fix the findings, book a solution review.
Related services and solutions
Application security assurance for web, SaaS and AI features Netbase provides application security testing for product and engineering leaders to find and fix vulnerabilities before release, including in AI features. We review code with AI-assisted triage, scan dependencies and test running applications, then rank findings and retest fixes. Netbase's own alignment with ISO 27001 and SOC 2 Type II covers Netbase's operations, never your application.
Learn More
AI-assisted QA and software testing that run inside delivery Netbase provides software testing and QA services for product and commerce teams to release often without breaking what customers use. QA runs inside delivery, not after it: acceptance criteria, automated regression with AI-generated test drafts, performance budgets and release readiness checks. Page speed shows the result: USticker's page load fell 21%, and PrintLeo's improved 35%.
Learn More
Application maintenance and support for live web, mobile and business software Netbase maintains and supports software that is already live: web platforms, online stores, mobile apps, ERP and CRM systems and internal tools, including software another vendor built. We take it over after a code and environment review, then fix defects, keep frameworks current, monitor it, make small changes and report monthly, under service levels agreed in your contract.
Learn More
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.
Learn More
Work with Netbase remotely
Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.
Discuss a project
Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.project@netbasejsc.com
+84 937 869 689
91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam
Get in touch
Tell us what you want to build, modernize, or operate.