# Threat Modeling

A structured method for identifying valuable assets, trust boundaries, plausible threats, and proportionate security responses before and during system change.

- Type: method
- Domain: Engineering & architecture
- Level: Intermediate
- Last reviewed: 2026-09-23
- Original: [Soroush’s Product Wiki](<https://wiki.soroush.uk/topic/threat-modeling/>)

## What it means

Threat modeling examines how a system could be abused or fail under adversarial conditions. Teams describe the system and its trust boundaries, identify threats and affected assets, assess risk, choose mitigations, and verify the resulting requirements. It is a recurring design activity, not a one-time diagram or penetration test.

## Why it matters

Security testing finds some implemented flaws, while threat modeling can reveal missing controls and unsafe architecture before implementation. It helps teams focus limited effort on plausible, consequential paths rather than generic checklists.

## When to use it

Threat-model new systems, high-risk features, identity and payment flows, major architecture changes, and newly exposed trust boundaries. Scale depth to consequences and uncertainty, then revisit when data, actors, dependencies, or deployment changes.

## How it works

1. Define scope, assets, security objectives, assumptions, and relevant attackers.
2. Map components, data flows, entry points, privileges, dependencies, and trust boundaries.
3. Identify threats systematically using misuse cases, attack trees, or a framework such as STRIDE.
4. Assess likelihood, impact, existing controls, uncertainty, and risk ownership.
5. Select design changes, preventive and detective controls, tests, and accepted residual risk.
6. Track actions and revisit the model as implementation and evidence evolve.

## A practical example

Before adding document sharing, a team maps invitation links, tenant boundaries, storage, previews, and email delivery. It identifies link leakage and cross-tenant access, then adds expiring scoped links, object-level authorization tests, and audit events.

## Where it breaks

A model reflects the participants' knowledge and can miss novel threats, implementation defects, or changing dependencies. Scoring can create false precision, and findings provide no protection unless owned, implemented, tested, and monitored.

## Common mistakes

- Starting from a generic threat list without mapping the actual system.
- Modeling only external attackers and ignoring compromised accounts, insiders, and dependencies.
- Treating every theoretical threat as equal instead of prioritizing consequence and plausibility.
- Producing a diagram and never tracking mitigations or revisiting assumptions.

## Visualization

[View the visualization on the website](<https://wiki.soroush.uk/topic/threat-modeling/#visualization-threat-modeling>)

## Sources

1. [Threat Modeling](<https://community.owasp.org/Threat_Modeling>) — OWASP Foundation · Security practice guidance
2. [Threat Modeling Manifesto](<https://www.threatmodelingmanifesto.org/>) — Threat Modeling Manifesto · Practitioner principles

## Related topics

- [Non-Functional Requirements](<https://wiki.soroush.uk/topic/non-functional-requirements/>)
- [Architecture Decision Record](<https://wiki.soroush.uk/topic/architecture-decision-record/>)
- [Incident Management](<https://wiki.soroush.uk/topic/incident-management/>)
