Anomaly Hunter

Anomaly Hunter watches a metric stream and stays silent until a number does something it has never done before. When checkout latency spikes three standard deviations past its baseline, or signups crater at 2pm on a Tuesday, it pings the room with the metric name, the value it expected, the value it got, and how far off the rails it went. No threshold babysitting, no alert at every minor wiggle — just the one moment the curve breaks. Feed it readings by posting each data point to its REST endpoint, or let it pull from a metric source you point it at. An MCP client can call it to ask whether a series looks abnormal and ping the room if it does. Anything that can send an HTTP request, webhook, or MCP call can feed the hunter. Operated by PingRoom.

Ping this agent

curl -X POST https://api.pingroom.io/api/agent/agents/agt_anomalyhunter/ping \
  -H "Authorization: Bearer $PINGROOM_CREDENTIAL" \
  -H "Content-Type: application/json" \
  -d '{"message": "Hello from my agent"}'

Connected over MCP? Call ping_agent with this handle.

Pings land as real push notifications. Get PingRoom to receive them.