Generating a plan about Feature Envy is the easy half; the half that makes a session land is finding Feature Envy in the repository the team ships every day, which is what the repository analysis does — it returns actual files and line ranges to refactor. The board tool then turns the plan into a Miro workshop, so preparation ends where facilitation starts. Two of the five tools stay dark until you add the matching token.
A content generator for Technical Coaches running Learning Hours: structured 60-minute practice sessions built on the 4C model — Connect, Concept, Concrete, Conclusion. It writes the session, the code to practise on, and the board to run it from.
- A full session plan for a named topic, following the 4C model from the opening connection activity through concept, hands-on exercise and closing reflection
- Before and after code examples for a topic in the language you name, so the exercise has something concrete to refactor
- A Miro board created from the session content, turning the plan into a workshop surface the team can work on together
- Real examples of a named code smell — Feature Envy, Long Method, Duplicate Code — located in a GitHub repository you point at, with file and line references
- A read of a repository's tech stack, so the generated content uses the frameworks and test tools the team already works in
An Anthropic API key in ANTHROPIC_API_KEY — it is the only required one, and the generated content bills to it. Two optional tokens each unlock a tool: MIRO_ACCESS_TOKEN from a Miro app for board creation, and GITHUB_TOKEN with repo read access for the repository and tech-stack analysis. It starts over stdio as npx -y learning-hour-mcp. MIT.
One command plus a key — npx -y learning-hour-mcp, then supply credentials