Oh My OpenAgent

This is a note from my public notes. View the canonical version: Oh My OpenAgent.

Oh My OpenAgent (omo) is an agent-orchestration harness built on top of OpenCode; it coordinates multiple specialized agents to handle complex development work end-to-end with limited supervision. It was previously called oh-my-opencode.

The headline mode is "Ultra Work" (ulw); it triggers autonomous execution with auto-planning, deep research, parallel agents, and self-correction loops. A primary coordinator (Sisyphus) tracks session continuity via a boulder.json file; specialized agents (Prometheus for planning, Atlas for execution, Hephaestus for deep work, plus dynamic Oracle / Librarian / Explore / Metis / Momus roles) take on the right slice of the job.

Team mode runs up to 8 parallel agents with real-time communication and tmux visualization. A category router picks the best model per intent (ultrabrain, visual-engineering, quick, deep, writing, git); under the hood it mixes Claude Opus / Sonnet for orchestration and planning, GPT-5 family models for specialized tasks, and Gemini 3.1 Pro for visual work. 54+ lifecycle hooks expose customization points.

It is TypeScript-first, installs via bunx, and ships under SUL-1.0. It sits in the same space as other agent harnesses but pushes harder on planning-execution separation, verified multi-agent coordination, accumulated task learnings, and session resumption.

References


About Sébastien

Ready to get to the next level?

Found this valuable? Share it with someone who needs it.

Join 6,000+ readers. Get practical systems for knowledge & AI. Free.

Subscribe ✨

Free: Knowledge System Checklist

A clear roadmap to building your own knowledge system. Subscribe and get it straight to your inbox.

6,000+ readers. No spam. Unsubscribe anytime.

Subscribe