Lack of automatic polling/persistent status for Codex agents after task completion.

This issue highlights a limitation in Codex's operational model: it terminates after task completion instead of maintaining a persistent, polling status for subsequent commands or results. This behavior contradicts the promise of "Full Automation" and "Agent Swarm Intelligence," where continuous operation and dynamic interaction are expected. Market implication: for agent systems aiming for autonomy and swarm intelligence, single-shot execution models are insufficient. The inability to maintain a persistent state or automatically poll for new instructions creates significant operational overhead and limits the complexity of tasks agents can handle. This friction point hinders the development of truly autonomous and collaborative agent systems.