Copy a production-ready example below, or generate one tuned to your exact Ruby on Rails setup in about thirty seconds — free, in your browser.
# your-rails-app > Ruby on Rails project. Instructions for AI coding agents working in this repository. ## Setup - Install dependencies: `bundle install` - Start dev server: `bin/dev` - Run tests: `bin/rails test` ## Code style - Follow Rails conventions over configuration. - Skinny controllers; push logic into models or service objects. - Run RuboCop before committing. ## Guardrails Things agents get wrong here. Follow these strictly: - Never edit `db/schema.rb` by hand. - Avoid N+1 queries; use `includes` for associations.
Dropped into a Ruby on Rails repo with no context, an AI agent guesses: it reaches for the wrong package manager, ignores your structure, and edits files it shouldn’t. AGENTS.md is the briefing it reads first — so the very next prompt behaves like a teammate who already knows your conventions.
The example above covers the three sections that matter most for Ruby on Rails: the real setup commands (bundle install, bin/dev, bin/rails test), your code-style conventions, and the guardrails agents reliably get wrong. The generator fills these in from your actual manifest so the commands match your project, not a generic template.
It is a markdown file at your repo root that tells AI coding agents how your Ruby on Rails project works — setup commands, code style and guardrails — so they follow your conventions instead of guessing.
AGENTS.md is a shared convention read by Cursor, Claude Code, OpenAI Codex, GitHub Copilot, Gemini, Windsurf, Zed and others. For tools with their own filename you can export that format or symlink it to AGENTS.md.
At minimum: install/dev/test commands (bundle install, bin/dev, bin/rails test), your code-style conventions, and the guardrails agents get wrong — like Never edit db/schema.rb by hand.
Detect your stack, tune the guardrails, export to every agent format. Free.
Open the generator