activity-monitor

To automatically restart Claude AI, you need a guardian service that monitors its operational status and triggers automatic restarts upon detecting process failure or unexpected stoppage. This Skill uses PM2 for process management to ensure continuous uptime and minimize task disruption.

A Claude agent health check continuously monitors the agent's liveness by tracking heartbeat signals and system status. If the heartbeat stops, the guardian mode detects the failure and initiates an automatic restart to restore availability and update the status file.

Yes, you need PM2 and Node.js to run automated Claude monitoring scripts. PM2 handles the underlying process management required for automatic restarts, while Node.js executes the health check, heartbeat detection, and state monitoring scripts.

Yes, you can monitor Claude context window usage to trigger a session handoff. The context management feature tracks context capacity and enqueues handoffs to prevent session overflow during long-running tasks and maintain operational continuity.

To schedule daily upgrades for Claude Code, you use a monitoring service that enqueues maintenance tasks and manages upgrade awareness. This ensures the agent stays updated without requiring manual intervention during active operation or downtime windows.

Your Claude AI agent may stop running due to process crashes or task completion. You can prevent downtime by deploying a guardian service that continuously checks operational status, detects stoppage, and automatically restarts the agent to ensure continuous availability.