Canonical runtime-backed view of the artifact, independent from metadata governance.
{
"frontmatter": {
"description": "Proactive Service Monitoring and Alerting",
"name": "service_watchdog"
},
"kind": "skill",
"playbook": "# Service Watchdog Skill\n\nThis skill implements the **Service Watchdog** for the ImagineOS ecosystem.\n\n## Goal\nTo proactively monitor the health of all running services (Vetra, Latinos, Cartera, Paulovila, Garcero, etc.), and detect any crashes, port conflicts, or VRAM exhaustion automatically.\n\n## How It Works\nInstead of passively waiting for a user to report an error (e.g. \"Vetra no funciona\"), the Watchdog runs an automated diagnostic pass (`diagnose_services`) every 5 minutes and checks for anomalies.\n\n## Setup & Execution\nThe watchdog runs as a background script deployed via PM2. \n\n1. `cd /home/paulo/Programs/apps/OS/Skills/service_watchdog/scripts`\n2. `pm2 start watchdog.sh --name \"service-watchdog\" --interpreter bash`\n\n## Remediation Policy\n- Currently, the watchdog acts as an **observer**. It logs the diagnostic report and any detected anomalies.\n- In future iterations, it can trigger Telegram alerts or auto-remediate via `pm2 restart`.\n",
"summary": {
"description": "Proactive Service Monitoring and Alerting",
"id": "service_watchdog",
"kind": "skill",
"path": "Skills/service_watchdog/SKILL.md",
"title": "service_watchdog"
},
"validation_issues": []
}