Anyone experimenting with AI-powered development knows the problem: the generated code rarely meets your own quality standards. But LLM agents can do more — when configured properly, they follow team rules and execute complex workflows.

In this talk, you'll learn how to develop rule files for LLM agents that reflect your development process: from TDD cycles to code style to architectural guardrails. You'll discover how to refine rule files iteratively — letting the LLM help you in the process — including common pitfalls and concrete demos.

You'll go home with templates, do's and don'ts, and a blueprint for operationalizing your own workflows as rule files and integrating AI-powered development into your production workflow.