Check that Home Security is ready
Review Home Security readiness before you rely on it. Use the authenticated API for the selected tenant and home, read current readiness and confirmed mode separately from any request acknowledgement, and treat pending, unavailable, or unconfirmed states as not yet protected. Home Security can keep local functions working during a cloud outage, but cloud-dependent connectors still require the provider and network.
What readiness means
Section titled “What readiness means”Home Security uses selected devices by capability, not by brand-specific UI.
Recognized devices remain normal room devices. Your integration lets the user choose which devices participate and which rooms count as sleeping quarters.
Mode behavior is:
| Mode | Behavior |
|---|---|
| Away | Uses selected detectors. |
| Home | Uses perimeter contacts and vibration. |
| Sleep | Excludes sleeping quarters and arms selected sensors everywhere else. |
| OFF | Disarms. |
Camera video alone is not a detection signal. If a camera appears without a detector signal, do not treat it as an intrusion source.
An empty zone, an unavailable detector, a queued command, or a responding app does not prove protection is active.
Read state independently from requests
Section titled “Read state independently from requests”Do not infer success from a successful write acknowledgement, cached state, brand, model, or a prior assumption about device support. Read readiness and confirmed mode independently of request acknowledgement, then follow correlated operation results and the current reported state.
Keep these outcomes distinct in your UI and logic:
- mode change requested
- mode change confirmed
- alert dismissed
- disarmed
- arming readiness
- unavailable detectors
- restoration complete or pending
Preserve settings fields your client does not edit when sending updates.
Respect server-returned:
- permissions
- availability
- versions
- pending versus confirmed results
Public clients must not call internal transport endpoints or bypass home authorization.
Configure participation and sleeping quarters
Section titled “Configure participation and sleeping quarters”Use the authenticated API plus capability and settings descriptors for the selected tenant and home.
Security inclusion uses normalized capabilities, not brand-specific UI. Device-type settings are shared across brands exposing the required capabilities.
In the current product behavior:
- Compatible devices are included automatically.
- Users can include or exclude devices while they remain in their usual rooms.
- Rooms can be marked as sleeping quarters for Sleep mode.
If a device has no detection signal, present that clearly. A camera-only device must be shown as not contributing a detection signal.
Entry, exit, and disarm behavior
Section titled “Entry, exit, and disarm behavior”Current supported timing limits for editable delays are:
| Setting | Range |
|---|---|
| Away exit delay | 0 to 180 seconds |
| Door / window entry delay | 0 to 180 seconds |
Away exit delay applies when arming to Away from OFF. Switching from an armed mode to Away applies immediately.
Entry delay gives time to disarm after a door or window opens.
Provider lock events must distinguish supported correct-code keypad unlock and mechanical inside unlock from ambiguous state changes; a generic unlocked report must not invent an origin.
Current disarm expectations are:
- A verified keypad unlock disarms every zone.
- A verified inside unlock also disarms the home whenever perimeter protection is on.
Only show keypad or inside-unlock disarm support when the reported device support confirms it.
Test safely and verify confirmation
Section titled “Test safely and verify confirmation”Test detectors, supported keypad behavior, and inside-unlock behavior only while an authorized person is present and can disarm.
For any readiness or acceptance flow, verify in the app that:
- the expected mode is confirmed
- the expected event is confirmed
Connector tests and source builds are not attended physical acceptance.
Alarm outputs and restoration
Section titled “Alarm outputs and restoration”Home Security can use connected sirens, supported speakers, and lights, depending on the home configuration.
Current visible behavior includes:
- Sonos alarm output uses all available Sonos rooms at 50% volume when enabled.
- Spoken deterrent and siren repeat until the alarm is turned off.
- Flashing lights run at full intensity after dark, after sunset and before sunrise at the home.
- Original volumes, playback, light colors, brightness, and power are restored after the alarm stops.
Arrange any Sonos alarm or flashing-light test with everyone in the home, then turn the alarm off and verify original volume, playback, light color, brightness and power have restored.
Keep restoration outcomes distinct from alarm stop state. An alarm can be stopped while restoration is still pending.
Notifications and disarming
Section titled “Notifications and disarming”Dismissing one notification does not disarm the alarm.
If your client supports actionable notifications, keep actionable notification navigation scoped to the identified device and home.
Do not conflate notification dismissal with OFF mode, disarm, or restoration completion.
Permissions and access boundaries
Section titled “Permissions and access boundaries”Home Security availability in the household app is for owners and household members, but configuration permissions can be narrower. Respect the permissions returned for the current user and home.
Owner approval governs TV security permissions and automatic routine changes. TV linking does not grant security access: use explicit owner viewing/control permission.
Where credentials matter:
- DEV credentials act in the developer sandbox.
- LIVE credentials only for their own installed connector.
Do not assume staff-level access from household membership.
What not to infer
Section titled “What not to infer”Do not infer any of the following:
- protection from camera video without a detector
- support for a disarm path from an ambiguous unlock event
- a supported action from brand, model, cached state, or request acknowledgement
- actual law-enforcement integration from a spoken deterrent
Follow confirmed state and correlated results instead of assumptions.