That’s exactly what we’re doing. Connect the agent to your GitHub issues, tell it implement each one and spin on a loop until it’s finished. There’s more nuance to it than that but at a high level yeah, that’s how some people are using it.