Bret and Nirmal explore AI agents in Kubernetes with Eitan Yarmush, Senior Architect at . Eitan explains how AI agents work through three simple components (system prompts, LLMs, and tools), and demonstrates the kagent project, which provides a Kubernetes-native way to deploy and manage AI workflows. We cover the Model Context Protocol (MCP), which standardizes how agents communicate with external tools, along with practical use cases like incident response, debugging workflows, and CI/CD integration. Eitan shows how to create an agent through a web interface. We also address the challenges including security concerns (MCP lacks built-in security standards), cost considerations, and technical limitations like context window constraints. Listen to the audio version of this show: This edited version is from my live stream show June 5, 2025: 🙌 My next course is coming soon! I've opened the waitlist for those wanting to go deep in GitHub Actions for DevOps and AI automation in 2025. I'm so thrilled to announce this course. The waitlist allows you to quickly sign up for some content updates, discounts, and more as I finish building the course. 🍾 🗞️ Sign up for my weekly newsletter for the latest on upcoming guests and what I'm releasing: Show Links ========= Eitan Yarmush ============ LinkedIn: Nirmal Mehta =========== @nirmal Bret Fisher ========= Bluesky: LinkedIn: X: Website: Courses on Udemy: Join my Community 🤜🤛 ================ 💬 Join the discussion on our Discord chat server / discord 👨‍🏫 Coupons for my Docker and Kubernetes courses 🎙️ Podcast of this show Chapters ======== 00:19 Introduction 04:23 Hype vs Reality 06:15 Incident Response and Debugging 09:33 Defining Agents and MCP 15:20 Where Do You Run Agents? 18:24 What Problem does kagent Solve? 28:52 DEMO: kagent 33:52 Authentication and Tool Limitations