Quick thing before we start. If your use case is "I want the agent to know my setup commands, how to run tests, and basic project structure" — you don't need Rules for that. A single AGENTS.md file in your repo root handles it, no configuration required.
Rules are for the stuff AGENTS.md can't do.
Things like: automatically injecting your security standards only when the agent touches an auth file. Attaching your database migration rules only when it opens a schema — not when it's writing a React component. Enforcing a hard constraint that the agent can never install a package without asking, in every project, globally, across your entire machine. Giving your whole engineering team the same AI behaviour from a single config in the dashboard — without asking anyone to set anything up locally.
That's the difference. AGENTS.md is a readme for the agent. Rules are a targeting system. And in this video I'm going to show you exactly when and how to use them — four rule types, three real production examples, and the one team setup pattern that scales to every developer without any per-machine configuration.
