Remote Control Daemons — Development skill for Claude Code

Remote Control Daemons & Watchdog

[](https://opensource.org/licenses/MIT) [](https://www.kernel.org/) [](https://systemd.io/) [](https://www.python.org/)

Automated, persistent, and self-healing systemd services paired with an active health-monitoring watchdog for **Claude Code Remote Control** (`claude rc`) and **OpenAI Codex Remote Control** (`codex remote-control`).

---

💡 Why This Exists

Remote control daemons for AI coding assistants are essential for managing local workspace sessions from mobile apps or web UI clients. However, default daemon setups often fail silently when:

1. System Reboots: Services try to launch before network interface initialization or DNS resolution finish, triggering startup timeout failures.
2. Network Drops: Temporary network disconnections cause daemons to crash without automatic reconnection.
3. Static Oneshot Units: oneshot systemd units do not monitor underlying background processes once spawned, remaining marked as "active" even after process crashes.

This project solves these pain points with a **dual-layer self-healing architecture**: network gating + systemd process supervision + periodic cron health checks.

---

🏗️ Architecture

flowchart TD
    A[System Boot / Event] --> B{Network Online Check}
    B -- Network Waiting --> B
    B -- Reachable --> C[Launch Remote Control Daemons]
    
    subgraph Systemd Supervision Layer
        C --> D1[claude-remote-control.service]
        C --> D2[codex-remote-control.service]
        D1 -- Process Crash / Exit --> D1
        D2 -- Process Crash / Exit --> D2
    end
    
    subgraph Active Watchdog Layer
        E[Cron Cronjob: every 5 min] --> F[watchdog.py Health Audit]
        F -- Service Ina