My global instructions for AI agents like OpenCode, Claude Code, Codex, and friends.
See AGENTS.md for the actual rules.
This is an anonymized version of the AGENTS.md file I keep on my own machine. It loads into every coding session as the system-level guide for how I want agents to behave: what to verify, what to skip, how to write commits and PR bodies, when to push back, and so on.
Feel free to copy, fork, or steal any of it. Conventions like AGENTS.md are documented at https://agents.md.