Containers transformed modern application deployment, enabling faster development with portable and scalable systems. They also introduce new security risks that are difficult to navigate, particularly when development teams don't understand fundamental infrastructure security principles. Having a threat model of containerized applications is critical for developers, security engineers, and policymakers alike. This talk will break down the key security risks at each layer of the container ecosystem while providing actionable insights for assessing and mitigating threats.

We'll open with **how containers work** to understand our risks across the application's lifecycle. Misconceptions about their security properties lead to dangerous assumptions. While offering process-level isolation, they are not virtual machines. Container isolation is not as strong as assumed, meaning applications inside containers can still be exposed to host-level threats.

Next, we’ll dive into **host (or node) OS risks**, where a shared kernel and a broad attack surface can expose the entire system. We’ll discuss how improper user access rights and file system tampering can lead to privilege escalation. We'll then demonstrate a common container escape to gain persistence and lateral movement on a node's filesystem.

From there, we’ll examine **container runtime risks**, such as vulnerabilities in the runtime software itself and misconfigurations that allow attackers to break out. Application-level security flaws, such as injection attacks and mismanaged secrets, can also be exploited here.

Containers rarely exist on a single system, so next we’ll dive into **orchestrator risks**. Poorly managed administrative access and improper segmentation can lead to unintended data exposure. We'll show a few default Kubernetes configurations that are more risky than they seem and dive into why and how to mitigate those risks.

All of these containers come from a **container registry**. These become a security liability if improperly managed. We’ll discuss threats such as insecure connections to registries, stale or vulnerable images lingering in repositories, and insufficient controls - and ways to mitigate each of them. We'll leave with patterns for commercial software factories that work in the real world.

Next, we'll dive into **container image security risks**, including vulnerabilities within base images, misconfigurations, and the presence of embedded malware or cleartext secrets. The use of third-party images without validation can introduce serious supply chain risks, emphasizing the need for strong image provenance and validation practices - using attacks from the field as examples.

To conclude, we’ll examine **how these risks play out in the real world**, drawing from industry case studies and best practices. Attendees will leave with practical guidance on prioritizing security fixes, assessing risk within their own containerized environments, and speaking about container security within the context of regulatory frameworks like **NIST 800-190** without falling asleep in the process.

By the end of this session, attendees will:
- Understand the full security landscape of containerized applications, including threats across the stack.
- Learn how to assess and triage security risks effectively, prioritizing fixes based on real-world impact.
- Gain the vocabulary to discuss container security within legal and regulatory frameworks, ensuring compliance while maintaining agility.

Whether you’re a developer, security engineer, or policy professional, this session will help you get squared away on container security with confidence.