AI Lab
Prompts, Claude skills and AGENTS.md files I use every day — copy them, remix them, ship with them.
User Rubber Duck
A Claude Code skill that drives your live app through Playwright as a real persona — first-time visitor, new signup, daily-ops user — screenshots every state, watches the console, and writes the friction log your unit tests never will.
Rubber-Duck Review
A paste-anywhere review prompt that turns the model into a sceptical senior colleague: evidence over agreement, severity-ranked findings, and an explicit instruction that 'looks fine' is a forbidden answer.
AGENTS.md for a static SPA
The real AGENTS.md running this very website. Build-pipeline order, routing gotchas, deploy traps and ten hard rules — everything an agent needs to ship a Blazor WASM static site without breaking prod.
Open toolbox
Everything here is in real daily use on my own projects. Take what works, adapt what doesn't.