Building shared coding guidelines for AI (and people too)

As AI coding agents become central to software development, engineering teams need to adapt their coding guidelines to work for both humans and AI. Unlike onboarding a junior developer, agents require explicit, unambiguous instructions covering naming conventions, formatting, error handling, and comments. Good guidelines should include concrete examples of correct and incorrect code, ideally a 'gold standard' file showing all conventions applied together. Treat agent mistakes as feedback to iteratively improve the guidelines, and keep the standards file as a living document maintained collaboratively by the engineering team. Traditional tools like linters and static analysis still play a role alongside agent-specific rules files like agents.md.