Proof and Method
This page is the public summary of reliability and diagnostic quality evidence used in our current release narrative.
Public Gate Snapshot
- Captured events: 135
- Reviewed events: 51
- Correct-enough rate: 100.0%
- Actionable rate: 100.0%
- Wrong-cause rate: 0.0%
- Snapshot generated at: 2026-02-20T01:10:52Z UTC
Method Summary
- Capture events from local app lifecycle and termination pathways.
- Correlate local evidence near the event window.
- Classify cause and confidence through deterministic rules.
- Review event outputs against structured scoring criteria.
Scoring Criteria
| Metric | Definition | Target |
|---|---|---|
| Correct enough | Diagnosis aligns with observed cause | >= 80% |
| Actionable | Output provides a usable next action | >= 80% |
| Wrong cause | Diagnosis points to incorrect primary cause | <= 10% |
Why This Is Public
We publish this proof summary so teams can evaluate diagnostic quality before rollout decisions, instead of relying on generic product claims.