This gives Claude a persistent tmux session to run commands in, with the ability to watch output in real time and dispatch commands across SSH host groups. You'd reach for this when you need an AI agent to interact with long-running processes, tail logs, or coordinate operations across multiple remote machines without spawning ephemeral shells that disappear after each command. The real-time spectating means you can see exactly what the agent is doing as it happens, which is handy for debugging or auditing automated tasks. Think infrastructure automation, deployment workflows, or any scenario where you need stateful terminal access that survives across multiple AI interactions.
claude mcp add --transport stdio sanchirico-camacho-mcp -- npx -y camacho-mcp