Techbanq

Cybersecurity · secure design

Ask what can go wrong while the design is still a drawing

Threat modelling is a structured way for the people building a system to ask what an attacker could do to it — before it exists, when the fix is a design change rather than a rebuild. Done as an engineering practice, its output is a prioritised mitigation backlog the team actually works through, and a model that evolves as the system does.

Serving Saudi Arabia from our Riyadh office and our engineering hubs in Lahore and Karachi

Software team working together at desks in an open office
  1. 01

    Market

    Saudi Arabia

  2. 02

    Written for

    CTO, CISO

  3. 03

    Decision

    Adopt threat modelling as a repeatable practice inside the engineering process

  4. 04

    Languages

    EN · AR

Enterprise delivery credentials

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

The working method

From a whiteboard drawing to a ranked threat backlog in one session

A threat modelling session is not a lecture; it is the team decomposing its own design with a checklist that makes sure nothing obvious is skipped.

A flow from drawing the system, through walking each boundary-crossing flow with the STRIDE questions, to ranking threats and writing mitigations into the backlog — with a loop back whenever the design changes.

  1. 01

    Draw the system honestly

    A data-flow diagram of what is actually planned — components, data stores, external parties — with trust boundaries marked wherever the level of trust changes.

  2. 02

    Walk each flow with the checklist

    For every flow crossing a trust boundary, the team asks the six STRIDE questions in turn — a discipline that surfaces the threats intuition misses.

  3. 03

    Rank what was found

    Each credible threat is scored by how hard it is to exploit and what it costs if it lands, because a list without an order is a list nobody starts.

  4. 04

    Write mitigations into the backlog

    The output lands in the team's own tracker as designed-in changes and control tickets with owners — not in a PDF that ends the conversation.

Always available

The design changes — the model follows

A new integration, a new data store or a moved trust boundary triggers a short delta session on the changed part, so the model tracks the system instead of describing its ancestor.

Timing decides the value

A threat model at design time costs a meeting; at audit time it costs a rebuild

The same finding has wildly different prices depending on when it is made. "The reporting service can read the payments database directly" is, at design time, one line moved on a diagram and an API put between two boxes. Eighteen months later, with the service live and a dozen jobs depending on that direct connection, the identical finding is a migration project competing for a quarter's roadmap — and most such findings lose that competition and are accepted as risk instead. Nothing about the flaw changed; only the moment of discovery did.

That asymmetry is the entire business case, and it dictates where the practice must live: inside the engineering process, at the design review, run by the team with a facilitator — not in the audit calendar, run annually by outsiders on systems already built. It also dictates the tone. A threat modelling session is not a test the design passes or fails; it is the design getting better in the room. Teams that experience it that way run it again without being told to, and that voluntary repetition, not any single model, is what changes an organisation's security posture.

The checklist

The six questions STRIDE asks of every flow that crosses a boundary

STRIDE is not a tool or a certification — it is a mnemonic that keeps a room from only finding the threats it already expected. Each letter is a plain question.

01

Spoofing — could someone pretend to be this?

Can a caller fake the identity of a user, a service or a device on this flow, and what actually verifies it — a signature, a token, or an assumption.

02

Tampering — could the data be altered in flight or at rest?

What stops a request, a file or a record being modified between the point it was written and the point it is trusted.

03

Repudiation — could someone deny having done it?

If this action were disputed — a transfer, an approval, a deletion — does a log exist that would settle the argument, and can that log itself be edited.

04

Information disclosure — who else could read this?

Where this data could leak — in transit, in storage, in logs, in error messages — and whether anyone who can see the channel should be able to see the content.

05

Denial of service — what happens when it is flooded?

What this component does under deliberate overload, what else fails with it, and whether anything limits how much work one caller can demand.

06

Elevation of privilege — could a small foothold become control?

If an attacker holds this component, what does its access let them reach next — and whether it holds any permission it does not strictly need.

Our role

How a threat modelling backlog stays alive after the workshop ends

  1. 01

    Facilitation that transfers the skill

    We run the first sessions on your real designs with your engineers doing the decomposing, because the aim is a team that models without us — a practice we must attend forever has failed.

  2. 02

    Threats turned into tracked work

    Every credible threat leaves the room as a ranked backlog item with an owner in your tracker — a design change, a control, or an explicitly accepted risk with a name against it.

  3. 03

    A model wired into change

    We define the triggers — a new integration, a new data category, a moved trust boundary — that call for a short delta session, so keeping the model current costs an hour occasionally rather than a yearly archaeology project.

Which risks are worth mitigating now and which the business accepts are product decisions; the model's job is to make sure they are decisions someone actually took, in writing, rather than defaults nobody noticed.

Common questions

What buyers ask

When in a project should the first threat model be run?+

When the architecture is drawn but not yet built — the moment there is a diagram with components, data flows and trust boundaries on it, and before the expensive commitments are poured into code. Earlier than that there is nothing concrete to decompose; later, every finding competes with sunk cost. For a system that already exists, the answer changes shape rather than disappearing: model the next significant change at its design stage, and model the riskiest existing flows retrospectively — the findings there feed hardening work rather than design changes, which moves less but still moves something.

Do engineers need to be security specialists to run STRIDE?+

No, and the method is designed on that premise. STRIDE is a checklist of six plain questions, and the knowledge that answers them — how the system fits together, where the data really flows, which shortcut was taken under deadline — belongs to the engineers who built it, not to a security specialist. What a facilitator adds at the start is discipline and calibration: keeping the walk systematic, pressing on flows a room glosses over, and helping rank findings realistically. After a few facilitated sessions, teams run their own, and the facilitator returns only for unusually sensitive designs.

How do we keep a threat model current as the system keeps changing?+

By updating on triggers, not on a calendar. An annual refresh guarantees the model is wrong for most of the year and makes each update a dreaded archaeology session. The sustainable pattern defines a short list of changes that reopen the model — a new external integration, a new category of data handled, a moved or removed trust boundary, a new deployment environment — and answers each with a delta session on the changed part only, usually under an hour. Wired into the existing design review process rather than run beside it, the model stays close to reality at a cost the team stops noticing.

Model your next design with us

A first threat modelling workshop on one real design of yours, leaving your team with a ranked mitigation backlog and the method to run the next one themselves.

Start the conversation