Openclaw Health Mcp

Gives you a single tool call to check whether your AI agent deployment is actually healthy right now. Surfaces gateway status, CPU/RAM/swap usage, recent errors from journalctl, skill registry integrity, upgrade outcomes, cron state, and disk usage with a HEALTHY/DEGRADED/CRITICAL classification per component plus ranked critical findings. Built for the OpenClaw ecosystem but the linux-proc backend works on any host via psutil. Pairs with silentwatch-mcp for deeper cron monitoring. If you're running production agents on a VPS and want infrastructure observability inside the same Claude conversation that's running your agent, this slots in where Datadog feels like overkill.