Selected work
Outputs are cheap now. Judgment isn't.
Anyone can generate. What follows is less about what came out and more about the decisions that shaped it — the directions killed, the constraints held, the things checked rather than assumed.
Try it — live, right now
The Prompt Scaffolder
Most AI output disappoints because the prompt was thin. This builds a structured brief for common L&D artifacts — role, context, structure, quality bar, and the anti-patterns worth naming up front. Runs entirely in your browser. Nothing is sent anywhere.
Paste into Claude, ChatGPT, or Gemini. The structure matters more than the model — the same brief will lift output from any of them.