Safety & security

GrowRig controls physical equipment, so safety and security are architectural requirements — not features added later. Responsibility is split across layers, each with an explicit job.

Safety layers

Electrical & mechanical — future hardware should include fusing, reverse-polarity protection, current limits, transient protection, thermal design, certified external mains switching, and moisture-aware enclosures.

Controller firmware — enforces safe boot outputs, watchdog behavior, command timeout, fallback fan speed, maximum pump runtime, minimum ventilation, sensor-loss behavior, and emergency-temperature behavior.

Grow Core — enforces target limits, stale-sensor detection, conflicting-actuator prevention, override expiration, policy validation, alerting, and an audit history.

User interface — always shows why an output changed: which policy is active, whether control is local or remote, whether a fallback is active, which sensor caused an action, and when a manual override expires.

Network security

Grow devices sit behind a Grow Gateway. Default assumptions: cameras and IoT devices are untrusted, devices don’t need home-LAN access, internet access is denied unless explicitly required, and the Hub is the main user-facing endpoint.

Device identity

The direction is unique identity per controller, per-device credentials, restricted MQTT permissions, secure pairing, signed firmware, and rollback-safe OTA. The prototype may start with simpler LAN credentials, but protocols should not assume one global password forever.

Data privacy

By default, measurements and photos stay on the Hub, no cloud account is required, remote access is optional, and hosted notifications receive minimal data.

Responsible use

GrowRig does not present environmental target values as guaranteed agricultural advice. Recipes must declare their origin, version, verification status, and assumptions.