Running LLM Agents Safely: Hands-On with Docker Sandboxes

## Why LLM Agent Security Matters

As AI agents become more capable, they are increasingly being trusted to take actions on behalf of users—calling tools, writing files, executing code, accessing APIs, and interacting with external systems. While this unlocks powerful automation, it also introduces serious security risks. This makes LLM agent security a critical concern for anyone building or deploying autonomous AI systems.

## Hands-On LLM Agent Security with Docker Sandboxes

In this hands-on session, we explore practical approaches to LLM agent security by running LLM agents inside isolated Docker Sandboxes. You’ll learn how to safely execute agent workflows while keeping full control over permissions, system access, and runtime behavior.

## Understanding Risks in LLM Agent Security

We will start by understanding the core risks of autonomous agents and why traditional application security approaches are not enough for LLM-based systems. From there, we introduce Docker Sandboxes (SBX) as a lightweight isolation layer that enables secure execution of AI agents.

## Building Secure AI Agents with LLM Agent Security Practices

Throughout the session, we will build a sandboxed LLM agent, assign it real tasks, and observe how its actions are constrained by the environment. We will also progressively apply security controls to strengthen LLM agent security and demonstrate how to balance agent autonomy with safe execution.

## How Docker Sandboxes Improve LLM Agent Security

By the end of the session, you will have a working understanding of how to design and run secure LLM-powered agents using container-based isolation techniques.

## What You Will Learn

• Why LLM agent security matters in modern AI systems
• Risks of autonomous agents, tool use, and code execution
• How Docker Sandboxes (SBX) enable secure agent execution
• How to build and run sandboxed LLM agents
• How to manage permissions, file access, and tool usage safely
• Best practices for production-ready LLM agent security

## Who Should Attend

• AI engineers and developers
• Software engineers working with LLMs and agents
• DevOps and platform engineers
• Security engineers exploring AI systems
• Anyone building or deploying autonomous AI agents

## Join the Session

Join us to learn how to implement practical LLM agent security and build safer, production-ready AI agent systems using Docker Sandboxes.