- 01
Market
Saudi Arabia
- 02
Written for
CISO, CTO
- 03
Decision
Scope and commission a web application penetration test
- 04
Languages
EN · AR
Cybersecurity credentials
NCA registration
Registered cybersecurity service provider

ISO/IEC 27001
ISO/IEC 27001 certified
Scoping
What actually determines the size of the job
We do not have an office in Saudi Arabia
We serve organisations there from our Riyadh office and our engineering hubs in Lahore and Karachi, travelling for the parts of an engagement that need to happen in person.
Suppliers commonly scope a web test on page count, and page count is close to irrelevant. Fifty screens that all render the same list for the same user are one test repeated fifty times. Twenty screens exercised as an administrator, a branch manager, a clerk, an auditor, a partner and an unauthenticated caller is a matrix of a hundred and twenty checks, and each cell can fail differently.
The second input is whether administrative and back-office interfaces are included. They are frequently excluded to reduce cost, and they are where the highest-impact findings sit — an admin panel reachable from the public internet, a support tool that can impersonate any user, a bulk-export function with no rate limit. Excluding them produces a cheaper test and a false sense of the estate.
Third is the state of the environment. A staging system with relaxed authentication, stubbed payment integrations or a fraction of production data produces findings that do not apply and misses findings that do. We ask for parity on the authorisation configuration specifically, and where it does not exist we say so at scoping rather than discovering it on day two.
Coverage
What the test covers, in the order it runs
Authorisation, horizontal and vertical
Every function attempted as every role, and every object identifier substituted for one belonging to a second account in the same role. This is the bulk of the engagement and the source of most findings that reach a board.
Session and authentication handling
Password reset flows, session fixation, concurrent session behaviour, whether logout invalidates server-side, and whether multi-factor can be skipped by replaying an intermediate step.
Business logic
Workflow steps taken out of order, validation skipped, a discount applied twice, a negative quantity, a state transition reversed after it should be final. No tool finds these; a tester who understands the process does.
Injection and input handling
Still present in the places a framework does not cover: raw queries written for reporting, search and filter parameters, template rendering, file paths. The older endpoints are tested first because they predate the conventions the newer ones follow.
Data exposure
What the response contains against what the interface renders. Filtering in the front end is the recurring pattern: the object carries an identity number or another user's details and the page simply does not display them.
Retest
Fixes verified, and — more usefully — checked for whether the fix was applied everywhere the same pattern occurs. A gap fixed on four endpoints and left on twenty is the most common outcome of a report that lists instances without generalising.
Estimating
What moves the day count, and by how much
Two proposals for the same application can differ by a factor of three, and it is usually not margin. These are the inputs that actually change the number, so a quotation can be read against them rather than accepted.
Distinct permission levels
The list of roles, and two accounts within at least one of them.
Effect on the estimate: The largest single driver. The authorisation matrix grows with roles, not with screens.
Administrative interfaces
Whether back-office and support tooling is in scope.
Effect on the estimate: Adds days and returns the highest-impact findings. Excluding it is the most common false economy.
Integration surface
Payment, identity, partner and internal systems the application calls.
Effect on the estimate: Each integration is a trust boundary, and each boundary is a place authorisation can be assumed rather than checked.
Environment parity
Whether staging matches production on authorisation specifically.
Effect on the estimate: Where it does not, days go to working around it and coverage drops. Worth fixing before the window, not during it.
Rate limiting and protective controls
Whether a gateway will throttle testing traffic, and whether we test through it.
Effect on the estimate: Testing through aggressive throttling costs days for coverage a test account could have avoided.
Retest
Whether a retest pass is included and when it runs.
Effect on the estimate: A quotation without one is not cheaper; it has moved a cost to a month nobody budgeted.
Scope boundary
What a test establishes
What we do
- Produce findings with the exact request, the account they were sent as, and the response that demonstrates the problem.
- Identify the pattern behind a finding, not only the instances, so the fix reaches every place the shape occurs.
- Rate exploitability against what the affected object actually contains, rather than a generic scale.
Where a test is commissioned to satisfy a framework, the assessor decides whether the evidence is acceptable. Establish the scope, tester independence and report format they will accept before the engagement is scoped, not after it is finished.
Primary sourceCommon questions
What buyers ask
Can you test against production?+
Where staging does not mirror production authorisation, sometimes that is the only honest option — but with conditions: a defined change freeze, a named contact who can stop the test within minutes, read-only scope on destructive endpoints, and agreement on what happens if a finding is discovered that requires immediate action. Testing production without those is how a security exercise becomes an incident.
How many roles do we need to provide credentials for?+
Every distinct permission level, plus two accounts within at least one of them — the second account is what makes horizontal object-level testing possible at all, and its absence is the most common reason a test misses the finding class that matters most. Also useful: an authenticated user with no role assigned, which is a state most applications have and few have thought about.
What if the application changes during the test window?+
Findings become unreproducible and the report loses its value. We agree a change freeze on the tested environment for the duration, and where a release genuinely cannot wait we pause, take the new build and restate the window rather than producing a report against a version that no longer exists.
In this cluster
Next step
Send your role count and scope
Send the application count, the role count and whether admin interfaces are in scope, and you get a day estimate and a fixed quotation rather than a range.
Send your role count and scope
What happens next
- 1We reply within one business day.
- 2A 30-minute scoping call with a senior consultant.
- 3A written proposal with scope, team, and timeline.
Or reach us directly
