Web Applications, APIs and Thick Clients

Application Penetration Testing

Experienced testers test your web applications, APIs and thick clients the way an attacker with an account would, across every role and the boundaries between them. Findings reach your portal as they are approved, so your team can start fixing while testing is still under way.

What's in Scope

Everything your users and integrations touch

Most application tests cover more than one of these: a web app, the API behind it, and the roles and customer organizations that share them. We scope around how your application is actually built and used.

Web applications

Customer-facing apps, admin consoles and internal tools, tested signed in as each kind of user, not only from the login page.

APIs

REST, GraphQL and SOAP, including the endpoints your own front end never calls but an attacker can still reach.

Thick clients

Installed desktop applications, along with the servers and services they talk to and what they trust on the user's machine.

Multi-role and multi-tenant

Authenticated testing across every role you have, and between customer organizations, so tenant isolation is tested rather than assumed.

What every application test covers

  • ▸OWASP Top 10 as the floor, not the checklist
  • ▸Authentication and authorization testing
  • ▸Business logic and workflow abuse
  • ▸Session management and token security
  • ▸File upload and server-side vulnerabilities
  • ▸Source code assisted (white-box) testing when you want it

How We Test It

Signed in, as every kind of user

The issues that hurt most in an application are rarely visible from the login page. They show up once you are signed in and start asking what this account should, and should not, be able to reach.

Map every role, then test each boundary in both directions

Can a standard user reach an admin function? Can one customer read another customer's records? We work out who should be allowed to do what, then try every crossing.

Work the business logic a scanner has no model for

Skipped steps, replayed requests, negative amounts, approvals granted to yourself. The flaws that live in how your workflow is meant to run.

Chain small issues into what they add up to

A verbose error, a predictable ID and a missing check are each minor. Together they can be an account takeover or a data exposure, and that is how we report them.

Show the exact request that proves it

Findings include the exact request that demonstrates the issue, so your developers can reproduce it, fix it, and check the fix themselves.

Testing a multi-tenant app in production. Our preferred setup is two test customer organizations, each with one account per role. That lets us test tenant isolation in both directions between accounts you control, without touching a real customer's data. If the app acts on connected systems such as payments or email, we agree how to handle those before testing starts.

From our testing platform: every URL and API operation in scope, tagged with the role that reaches it and tracked until it has been tested. Paths that are new since the last engagement are flagged. Hostname blurred.

From Our Own Engagement Data

What we find most often

These are the High and Critical issues we reported most often in application tests in 2025 and 2026, ranked by how frequently they came up. Not a generic vulnerability list: this is what our own findings data says, and a fair preview of what a test of your application is likely to look for first.

In web applications and APIs

Most frequent critical and high findings

  1. 01Improper authorization (broken access control between users and roles)
  2. 02Cross-site scripting
  3. 03Insecure direct object references
  4. 04SQL injection
  5. 05Server-side request forgery
  6. 06Missing authentication on sensitive functions
  7. 07Path traversal

Authorization leads by a wide margin

Improper authorization and insecure direct object references are the same underlying problem: a user reaching data or actions that belong to another user, role or tenant. Together they lead the list by a wide margin.

They are also exactly what automated scanners cannot see. A request for another customer's invoice looks perfectly normal to a scanner, because the scanner does not know who should be allowed to do what. Finding these takes a tester who has learned your roles and then tries to cross them.

What You Get

Findings you can act on, and evidence you can show

Everything arrives through your portal, not a PDF emailed at the end. See how it works on the platform page, or read a full sample application report first.

Findings as they are approved

Each finding is reviewed and approved by a second consultant, then released to your portal during testing, so a serious issue found early can be fixed early.

A thread with the tester on every finding

Ask how it was found or whether your fix covers it, and get the answer from the person who did the work.

Retests from the finding

Request a retest from the finding itself within the retest window. The result is recorded against the same finding.

Jira export

Send findings to your own Jira as tickets, with severity mapped to priority and the evidence attached.

The report and a letter of attestation

A full report for your team and auditors, plus a one-page letter of attestation you can hand to customers who ask.

Why Teams Commission It

Usually there is a reason, and a date

SOC 2 and security questionnaires

Your auditor expects a recent test, and your customers' security reviews ask for one by name. See SOC 2 penetration testing for what auditors look for.

A launch or major release

A new product, a new API, or a rebuilt permissions model is the moment to find out what it lets people do before your customers do.

A customer or partner requires it

Enterprise customers and integration partners often require a third-party test before they sign or connect. The report and letter of attestation are built to answer that request.

Shipping every week? An annual test soon describes an application that has moved on. With continuous penetration testing, testers come back to new or changed features, using testing time agreed up front.

FAQ

Application penetration testing questions

How long does an application penetration test take?

It depends on what there is to test: how many user roles the application has, how large the API is, and how much business logic sits behind it. An app with two roles is a much shorter engagement than a multi-tenant platform with an admin console and a public API. We scope it from a short call, test accounts, or a few screenshots, and the quote gives you a delivery range before you sign. Findings reach your portal during testing, so you can start fixing before the report is done.

Do you test in production?

Yes, with care, and often production is the only environment that matches what your customers use. For a multi-tenant production app we prefer to set up two test customer organizations, each with one account per role, so tenant isolation is tested between accounts you control and real customers are never touched. Anything that could disrupt the application, such as high-volume requests or actions that trigger connected systems like payments or email, is agreed with you first.

Do you need our source code?

No. Most application tests run from the outside with test accounts, the way an attacker with a login would approach it. Source code helps when the logic is complex or parts of the application are hard to reach from the interface, and source code assisted testing is available when you want it.

Is retesting included?

Yes. Retesting is part of the quote, used within the retest window. When you fix something, request the retest from the finding in your portal, and the result is recorded against that same finding.

Will the report satisfy our SOC 2 auditor or our customer?

That is what it is built for. The report covers scope, methodology, and each finding with severity, evidence and remediation guidance. A letter of attestation is also available for customers and auditors who want a one-page confirmation of what was tested and when. Our SOC 2 penetration testing page covers what auditors typically expect.

How is this different from an automated scan?

A scanner can spot known patterns, but it cannot judge authorization: it does not know that one customer should never see another customer's invoices, or that a support role should not be able to change billing. Those are the serious issues we report most often in application tests. Experienced testers work out what each role should be allowed to do, test every boundary, and decide what is real, so what reaches you is worth fixing.

Find out what one account can really reach

Tell us about the application, its roles and what is driving the test. You get a fixed price per assessment, with scope, methodology, deliverables and retest terms in writing, scoped by the people who test.

Budgeting first? Read our penetration testing cost guide.