Agent-Learning-Hub

Why we included this project

This roadmap, curated by a Datawhale member, turns agent learning into a checklist you can work through instead of a pile of links. The staged todo list starts by having you distinguish chatbots, workflows, and agents, then moves to building a minimal agent loop, adding tool use and memory, and finishing with evaluation and safety. The maintainers hold clear opinions: they steer readers toward coding agents, harness engineering, and the MCP, A2A, and ACP protocols, and caution against role-play-style multi-agent frameworks that rarely ship. A project ladder gives runnable milestones for people who learn by building, and readers who already work with LLMs can jump in at later stages to fill specific gaps. Because each task is a checkbox, it is easy to tell whether you are actually making progress.