Operon
A registry of AI agents for your team. It watches the Claude Code, Codex, and Cursor sessions you already run and turns the moves you keep repeating into agents everyone can reuse.
Overview
This is what I'm building right now. Everyone's running coding agents all day, but the good patterns die in one person's terminal history. Someone figures out the perfect way to review a PR or ship a feature with an agent, and that knowledge never leaves their machine. Operon catches the moves that keep showing up across a team's Claude Code, Codex, and Cursor sessions and turns them into shared agents everyone can run, version, and trust. Your best work stops living in one head. The bet is that the next unit of reuse in engineering isn't a library, it's an agent.