Agentic System Course
A 22-chapter open course teaching how to design, build and operate production AI agents, meant to be studied alongside your own AI assistant rather than as a stand-alone tutorial, with a companion skill that reviews agent designs against the course.
This course covers an agentic system's load-bearing topics and architectural trade-offs, planning, tool use, adaptation, memory, rather than a single walked-through project or one specific framework, so it stays useful as framework details change. It is written as a pair of files, the course itself and an AGENTS.md, meant to be read by an AI as much as by a person, and the intended way to study it is to open it in an editor while pointing an AI agent such as Claude Code or Codex at the same project and asking it to quiz, translate a pattern into a smaller working version, or compare it against how a leading coding agent solves the same problem.
It ships a bundled skill, agentic-system-reviewer, that reviews a product requirements document, design document, implementation plan or agent's own code against the course's material, usable on any agentic system, not only ones built while following the course.
It is released under the MIT license.