How does a Skie-managed device prove it is in a good state?
After every policy apply, the device reads each protection back from Android and reports the result, rather than just saying it applied the rules. Those readings become a compliance score on the dashboard. The device also reports on a regular rhythm, so a missing check-in is itself flagged.
Telling a server that you obeyed an instruction is not the same as having obeyed it. Skie closes that gap by reading back. After applying a policy, the device asks Android for the actual state of each protection, for example whether factory reset is blocked or whether the accounts lock is on, and reports what Android says.
The result is a compliance score per device, which rolls up to group, branch, school and organisation. Insights then shows readiness, fully compliant devices, average compliance and the share on the latest policy and latest app. Devices that have not synced recently are visible through the Due, Overdue and silent measures.
There are other layers around this. Policy bundles are signed and the device rejects anything that does not verify. A per-device key and identity tie reports to the right device. A trusted clock gives sync timing that a changed device date cannot distort. And the undo of Reset all settings is done within seconds by re-applying protections, which is the self-heal.
What it does not do is promise that nothing can ever go wrong. Skie is built to be tamper-proof, and if anything ever did get through, the admin is alerted immediately, down to the exact device. A compliance run can also be started on demand from the dashboard.
Practical tip: treat a drop in average compliance for one model or one group as a prompt to look at the device page and the audit log, not as a verdict on any person.