How does AgentHub handle terminal confirmation prompts and tool approval when I am away?

[xiaoran7]

Claude Code frequently asks for permission before running bash commands or editing critical files. If I am not at my terminal, the agent hangs indefinitely. Can AgentHub surface these approval prompts remotely?

[xiaoran7]

Yes. AgentHub natively hooks into the official Claude Code and Codex protocols:

• Whenever the agent emits a permission request or confirmation prompt, agenthub-runner forwards an actionable, opaque approval descriptor to your Gateway.
• The Web client (desktop or phone) displays a dedicated, touch-friendly approval banner showing the exact command or action.
• Clicking "Approve" or "Reject" sends an idempotent signed receipt back to the workstation runner, which signals the native CLI to proceed.
• If an approval times out or is rejected, the session gracefully pauses without crashing the agent.