Kyle and Jeremy from Dagger conduct a workshop on building AI agents using Dagger, a container runtime and workflow engine. They guide participants through installing Dagger, creating a CI pipeline, and adding an AI agent to an existing project. The agent is designed to develop new features for a Vue.js website, with capabilities to read and write files, list directory contents, and run tests. They demonstrate how to integrate the agent with GitHub, enabling automated feature requests via GitHub issues. Throughout the workshop, they address questions from the audience, discussing topics such as managing agent tools, implementing reflection agents, and the relationship between Dagger and MCP. The goal is to show how Dagger can be used to create reliable and efficient AI agents for software engineering workflows.
Sign in to continue reading, translating and more.
Continue