oh-my-openagent

Why we included this project

Once a codebase grows past what a single model context can hold, a lone coding agent tends to stall. Oh My OpenAgent works around that by acting as a harness: it splits one agent session into a team of specialized sub-agents, each taking a slice of the work while sharing state and what they learn across tasks. It sits on top of tools you may already use, like OpenCode and Codex, and brings LSP and AST tooling so the agents work from the real project structure instead of guessing. A large set of lifecycle hooks also makes it easy to script or customize agent behavior. If you already work with a CLI coding agent and want more parallelism without switching ecosystems, this is a strong fit.