What is threat modeling?
Threat modeling is a structured way to analyze a system, work out what could go wrong, and decide which mitigations matter most.
Definition
Threat modeling
Also called Threat model, STRIDE
Threat modeling is the practice of analyzing a representation of a system to identify what could go wrong with its security and privacy, and deciding which countermeasures to apply.
How it works
The Threat Modeling Manifesto defines threat modeling as analyzing representations of a system to highlight concerns about security and privacy, and organizes it around four questions: What are we working on? What can go wrong? What are we going to do about it? Did we do a good enough job? OWASP adds that a threat model is a structured view of an application and its environment through the lens of security, and that the work typically produces a prioritized list of security improvements.
Teams usually start by drawing the system: components, data stores, data flows, and the trust boundaries between them. They then walk each element and ask what an attacker could do. Microsoft's Threat Modeling Tool uses the STRIDE categories to prompt those questions: spoofing, tampering, repudiation, information disclosure, denial of service, and elevation of privilege. Each threat gets a mitigation, an accepted risk, or a follow-up. The Manifesto stresses that the practice works best early and often, iterating with design changes rather than as a one-time compliance exercise.
Example
A team adds a feature that lets customers upload files. The data flow diagram shows files crossing from the internet into a storage bucket, then being processed by a backend service. STRIDE questions surface threats: malicious files exploiting the parser (tampering, elevation of privilege), one customer reading another's uploads (information disclosure), and huge uploads exhausting resources (denial of service). Each becomes a design requirement before code is written.
Why it matters for a SOC
Threat modeling is traditionally an application security practice, but the same thinking tells a SOC where to focus: which assets matter most, which paths an attacker would take to reach them, and which of those paths have no detection. That makes it a natural input for detection engineering and hypothesis-driven threat hunting. A practical way to connect the two:
- List the handful of assets that would hurt most if compromised, such as the identity provider, source code, production data, and finance systems.
- For each, sketch the likely attack paths, for example a stolen SSO session leading to a code repository.
- Map each step to a MITRE ATT&CK technique and to the log source that would record it.
- Mark which steps have a tested detection, which have logs but no detection, and which have no logs at all.
The last two categories are the detection and logging backlog, ordered by what they protect. Cotool applies the idea to the whole environment: it maps systems, the identities that use them, and the data they hold, builds a threat model of what attackers could reach and how, and keeps it updated as the environment changes so hunts and detections follow it.
Related terms
- Threat intelligenceThreat intelligence is information about threats and threat actors that has been collected, analyzed, and enriched with context so defenders can make decisions, such as what to detect, hunt for, patch, or block.
- MITRE ATT&CKMITRE ATT&CK is a free, publicly maintained knowledge base that catalogs the tactics and techniques attackers use against real systems, based on observed intrusions, and gives defenders a shared vocabulary for detection and threat intelligence.
- Detection engineeringDetection engineering is the practice of designing, building, testing, deploying, and maintaining the rules and analytics that turn security telemetry into alerts, treated with the rigor of software engineering.
- AI red teamingAI red teaming is the practice of adversarially testing AI models and AI-enabled systems to find harmful or unintended behavior, security vulnerabilities such as prompt injection, and weaknesses in safeguards before attackers or users do.
Go deeper
- What is threat hunting?Threat hunting is a proactive, hypothesis-led search of security telemetry for attackers your detections missed. Hunt types, maturity models, and what a hunt produces.
- AI threat hunting: what agents can and can't hunt yetWhat AI agents can and can't do in a threat hunt, from BlueBench results on 26 models: where they fail, what they need, what a hunt costs, and where humans review.
Sources
- Threat Modeling | OWASP Foundation · accessed November 16, 2026
- Threat Modeling Manifesto · accessed November 16, 2026
- Threats - Microsoft Threat Modeling Tool - Azure · accessed November 16, 2026
- Security Log Storage and Federated Search for Agents | Cotool · accessed November 16, 2026
