Monitoring
Dormant bell with trailing sensors; wakes on signal.
A sentinel agent sits dormant and watches: it consumes a stream of signals and stays idle until a trigger condition fires, then wakes, acts, and returns to watching. It trades continuous reasoning for cheap, event-gated attention, which suits monitoring, alerting and security work where most of the time nothing should happen. In anatomy terms it is a transport-heavy, low-duty-cycle core wired to sensors and a wake condition.
Sensors and transports feed a dormant reasoning core; a wake condition (a rule or a model) gates activation; effectors fire only on trigger; the perimeter governs the response actions.
- ·Security monitoring: watch logs and traffic, wake on anomaly (Charlotte, Dropzone, Sardine).
- ·Operations alerting and incident triage.
- ·Fraud and compliance surveillance over high-volume streams.
- ·Cheap: it reasons only when triggered.
- ·Scales to high-volume signal streams.
- ·Fast, focused response to the event that matters.
- ·Clean separation of watching from acting.
- ·Only as good as its trigger definition: it misses what it was not told to watch.
- ·False positives and negatives tune against each other.
- ·Cold-start latency on wake.
- ·Limited context at the moment of firing.
Mapping classified by the Field Book from each agent's documented control flow.
Description, use cases and trade-offs: Field Book analysis, synthesised from established agent-architecture literature.