← back to v134 · llm-powered-on-device-detection-of-abusive-notifications-on-android
conformance · v134
v134 · llm-powered-on-device-detection-of-abusive-notifications-on-android — conformance probe
6 assertions drawn from the spec. Each is a single contract the spec text makes. Pass/fail reflects what this browser executed; blocked means the contract was not run because it needs user mediation, hardware, or another unavailable precondition. Open the page in Chrome stable / canary / Firefox / Safari to compare.
0
pass
0
fail
0
blocked
6
total
| id | contract | kind | verdict | detail |
|---|---|---|---|---|
notification-api-exists |
Notification API is exposed — the surface the on-device classifier observes. | exists | … | |
notification-permission |
Notification.permission is a string ('default' | 'granted' | 'denied') unaffected by the on-device classifier. | script | … | |
service-worker-registration |
navigator.serviceWorker is exposed — push notifications targeted by the classifier flow through SW registrations. | exists | … | |
push-manager-supported |
ServiceWorkerRegistration.prototype.pushManager exists (entry point for the push channel evaluated for abuse). | exists | … | |
no-classifier-js-api |
The abuse classifier intentionally has no JS surface — there is no navigator.abusiveNotificationStatus or similar to query. | script | … | |
notification-requestpermission-fn |
Notification.requestPermission remains a function (the classifier may revoke under the hood, but the prompt API is unchanged). | typeof | … |