cagent - Build Declarative AI Agents with Docker - AI Guide to the Galaxy Episode 5

Learn to build and deploy declarative AI agents using Docker's open-source cagent framework in this 38-minute technical demonstration. Explore how to define AI agents using YAML configuration files, covering essential components like instructions, model selection, tools integration, and chat loop management. Discover how to run agents locally with Docker Model Runner (DMR) or connect to external cloud providers such as OpenAI and Gemini. Master the integration of Model Context Protocol (MCP) servers through the Docker MCP Toolkit for context-aware document lookup and utilize built-in toolsets including shell access, filesystem operations, to-do management, and thinking capabilities. Understand how to architect multi-agent systems where sub-agents function as tool calls within larger workflows. Learn to package and distribute agents as OCI artifacts on Docker Hub for seamless sharing across teams and projects. Explore the bootstrapping feature that enables agents to create other agents from command-line prompts, and get practical guidance on obtaining binaries for Linux, Windows, and macOS platforms through GitHub Releases.