c3.15 Perform threat modeling
Threat modeling identifies who might attack each workload, the paths they could take, and the impact if they succeed — so you invest in the controls that matter most. In the cloud, focus on what you can do to strengthen security, not just on everything that could go wrong.
Implementation guidance
Model threats per workload
For each workload, analyze which actors would want to harm its confidentiality, integrity, or availability, the attack paths and techniques they might use, and the impact if a threat succeeds. That analysis tells you where to invest. The cloud lets you reach a stronger security posture than on-premises, so frame the exercise around what you can do to strengthen defenses — shifting security from the "department of no" to an enabler of safe innovation. The AWS blog How to approach threat modeling is a solid method to follow.
Tools and reference material
- Threat Composer is a lightweight tool that reduces time-to-value when threat modeling.
- The Threat Technique Catalog for AWS maps MITRE ATT&CK cloud techniques to AWS-specific techniques observed by the AWS Customer Incident Response Team.
- The Threat Modeling for Builders workshop teaches a repeatable practice.
Automate it with AWS Security Agent
Beyond manual modeling, AWS Security Agent (part of AWS Continuum) can generate a threat model automatically from your design documents, your source code, or both. It produces a system overview describing how the system works and a set of threats — each with a severity level, a STRIDE classification, and actionable recommendations. Because Security Agent builds this application context once and reuses it, the same model feeds its design and code review and its agentic penetration testing. Use it to scale threat modeling across many workloads and to keep models current as the code changes, while your teams still reason about the highest-risk designs.
Related
Threat modeling is a core skill for your Security Champions and a natural input to DevSecOps.
How to validate
- Do you perform threat modeling for your critical workloads?
- Does it identify threat actors, attack paths, and impact to prioritize controls?
- Do you use a repeatable method and tooling (for example Threat Composer), and involve builders?
Threats addressed
- Without threat modeling, security effort is spread evenly or by intuition rather than concentrated on the most likely, highest-impact attack paths, leaving critical gaps unaddressed.