anthropics/cwc-workshops — reverse-engineered prompt
Reverse engineered prompt
Build me a repo of hands on workshop materials for teaching people how to use Claude Code and Claude Managed Agents.
I want each workshop to be self contained, with starter code, clear instructions, and a small finished app or exercise people can run locally. Include workshops for choosing the right Claude model using evals, breaking a big agent prompt into smaller skills and helper agents, building with Claude Code from interview to spec to UI, shipping an incident dashboard with an SRE chat agent, running a game bot competition, adding memory so an agent remembers across sessions, improving a slide deck agent with eval scores, and a production style deal desk chat app with several research agents.
Keep it practical and demo friendly. The goal is for workshop attendees to learn by changing prompts, running evals, watching agent events, and seeing measurable improvements. Use TypeScript where it makes sense, and look up current Claude docs online if needed.
Want more depth? Deep Reverse