What's Working
The landscape shifts constantly. A new model drops, a framework updates, a practice that worked last month stops holding up. This is a living catalog of what's actually working right now — the tools, frameworks, and practices we're using in production and seeing real results from.
Every entry below ships with a complete, copy-pasteable skill — the full prompt, framework, or playbook you can drop into Claude Code, Cursor, or any agent and run today. No signup, no email, no catch.
- → The full reasoning, ranked and explained
- → A single copyable
.mdfile per entry - → One-click copy. Paste into your agent. Build.
Hydrate the World
AI agents reach for the broadest helper they can find. listEverything() to populate a dropdown. The page works at twelve rows and dies at twelve thousand. A structured audit for the over-fetching, N+1s, and serial waterfalls AI loves to write.
Vibe Coding vs. Agentic Coding
You don't need to be a developer. You need five pillars that keep your project from falling apart. A scoring framework that separates AI code that lasts from AI code that drifts.
Vibecoder Security Review
AI-generated code has predictable security holes. A structured triage that finds them — secrets, auth bypasses, hallucinated packages, and the specific patterns where speed trumped safety.
Scroll-Driven Storytelling
The complete playbook for Snowfall-caliber immersive pages. Every effect ranked S→C, every architecture pattern explained, four built examples — and the full prompt to build your own.