What you’ll learn
From basic terminology to advanced workflows for people who work with both agents in parallel.
Clear mental model
The real difference between CLAUDE.md e AGENTS.md, .claude/ e .codex/, no memorization required.
Anatomy of each runtime
Skills, agents, hooks, settings, MCP — where each piece lives in each ecosystem.
Conversion without losing anything
A prompt template for migrating a project from Claude Code to Codex (and vice versa) in minutes.
polyskill in practice
One canonical source, two optimized outputs. Full CLI: init, build, import, reconcile.
Side-by-side workflow
Two terminals, one project. How to pass context from one agent to the other without losing the thread.
Tool-agnostic
Don't get locked into one ecosystem. If one goes down, the other keeps going. If one gets stuck, the other gets it unstuck.
Course tracks
6 independent tracks — pick the one that makes the most sense for where you are right now.
🧱 Fundamentals
What Claude Code is, what Codex is, why use both, and never depend on just one.
Start track →🔷 Anatomy of Claude Code
CLAUDE.md, .claude/, skills with YAML, sub-agents, hooks, and settings.
🟣 Anatomy of Codex
AGENTS.md, .codex/, .agents/, agents in TOML and openai.yaml.
🟡 Converting your project
A prompt template, synchronized maintenance, and common pitfalls when running both.
Start track →🦜 polyskill — cross-runtime skills
One source definition, outputs optimized for each runtime. Claude ↔ Codex round-trip, drift policy, full CLI.
Start track →🌹 Advanced workflows
Session handoff, two terminals in the same project, shared MCP, and governance.
Start track →Cheat sheet — Claude Code vs. Codex
Same knowledge, different names. This is the chart that appears in almost every lesson.
💡The big insight
Project knowledge is shared — documents, scripts, references. Only the files in configuration change names and formats. Anyone who understands this map can migrate between runtimes in minutes, not days.
Who this course is for
✓ It’s for you if…
- ✓You already use Claude Code (or Codex) and want to learn the other without having to relearn everything
- ✓You want to run both agents in the same project without duplicating content
- ✓Creates skills/sub-agents and gets tired of maintaining two versions by hand
- ✓You want to be tool-agnostic and stop depending on just one ecosystem
✗ It’s not for you if…
- ✗Look for an introductory tutorial on “what is an LLM”
- ✗You want a model comparison (GPT vs Claude). This is about runtime, not models.
- ✗Expect the course to decide for you which tool is "better"
Ready to get started?
If you’ve never seen either one, start with Track 1. If you’re already fluent in both, go straight to Track 5 (polyskill) — that's the key trick.