Teaching Agent — Development skill for Claude Code

Teaching-Agent 🎓

A BMad-Method-powered, multi-agent AI assistant for structured course development with [LiaScript](https://liascript.github.io)

**Teaching-Agent** helps educators design, structure, and author complete courses through guided, iterative workflows — from initial concept to published, interactive materials. It coordinates four specialized agents (teaching, visual design, learner review, and publishing) around a single project file, `journal.md`, that holds the entire course state.

It is **editor-agnostic**: the same specs build into configurations for Claude Code, GitHub Copilot, OpenAI Codex, Cursor, Windsurf, and any web chat.

---

Purpose & Audience of This Repository

**Read this before the technical details below.** This repository is a worked example: it shows the LiaScript toolchain in practice, using a NIS2 compliance course as the concrete case.

The repository has **two layers**:

1. A sample course — "NIS2 Ready." A complete, self-paced course (six units) on the EU NIS2 Directive that works as a real training resource on its own. NIS2 was chosen as a fitting example because it concerns public administration and critical infrastructure, so the subject matter is directly relevant to the sectors LiaScript courses are often written for.
2. The tooling behind it. A separate LiaScript document — `about-liascript.md` ("Behind the Scenes") — walks through the tools used to build the course — LiaScript itself, the Teaching-Agent multi-agent system, and the liaex exporter — and refers to concrete features in the course as examples. Written from scratch for readers new to all three, it documents how an interactive, accessible, standards-based Open Educational Resource is produced.

In short: the course is the example, and the tooling document explains how it was made — with open, standards-based, vendor-independent tools.

Who the course material is for

Taken verbatim from the course de