homelab

name	homelab
description	Query homelab infrastructure status — check UptimeKuma monitor uptime, get monitor summaries, and read ntfy topic messages (e.g., homelab-watchdog alerts). Use proactively when the user asks about server status, uptime, alerts, or infrastructure health.
allowed-tools	Bash(curl:*)

Homelab Monitoring

Queries the claude-remote-api which proxies UptimeKuma and ntfy. All requests go to $CLAUDE_REMOTE_API_URL with Authorization: Bearer $CLAUDE_REMOTE_API_SECRET.

UptimeKuma

Monitor summary (up/down counts)

curl -s -H "Authorization: Bearer $CLAUDE_REMOTE_API_SECRET" \
  "$CLAUDE_REMOTE_API_URL/homelab/uptime-kuma/status"
# → { "up": 12, "down": 1, "paused": 2, "total": 15 }

All monitors with full details

curl -s -H "Authorization: Bearer $CLAUDE_REMOTE_API_SECRET" \
  "$CLAUDE_REMOTE_API_URL/homelab/uptime-kuma/monitors"
# → UptimeKuma monitorList with heartbeat data

ntfy — topic discovery and message querying

Always discover topics first, then fetch messages for each one:

# Step 1: list all subscribed topics
curl -s -H "Authorization: Bearer $CLAUDE_REMOTE_API_SECRET" \
  "$CLAUDE_REMOTE_API_URL/homelab/ntfy/topics"
# → ["homelab-watchdog", "homelab-watchtower", "vps-watchtower", "uptime-alerts", "claude-remote", ...]

# Step 2: fetch messages for each topic
curl -s -H "Authorization: Bearer $CLAUDE_REMOTE_API_SECRET" \
  "$CLAUDE_REMOTE_API_URL/homelab/ntfy/messages?topic=<topic>"
# → JSON array of recent messages on that topic

Never assume which topics exist — always call /homelab/ntfy/topics first.

Response interpretation

• UptimeKuma status 1 = up, 0 = down
• active: false means the monitor is paused (not a failure)
• ntfy message fields: id, time (Unix timestamp), event, topic, title, message, priority, tags