Gate 01
Freeze what is being tested
Record the core build, configuration, app binaries, device profile, and scoring rules before results are opened. A later code or configuration change starts a new candidate.
Evidence standardResults come after
Results come after
reproduction.
We are publishing the standard before any product score. Product numbers will appear only after the relevant core passes a locked, repeated, multi-app review.
No current product score is published here. Historical engineering results are treated as inputs to verification, not as proof that the present core can handle every app.
The gate
What must be true before numbers go public.
The standard covers detection, restraint, repeatability, and the production path. Passing one app or one run is not enough.
Gate 02
Use more than one kind of app
Exercise the supported Android scope across multiple apps and UI frameworks, including workflows that require navigation, saved state, and test data.
Gate 03
Score catches and clean controls
Grade end-to-end outcomes against a locked answer key. Include clean builds and hard negatives so a high catch count cannot hide noisy findings.
Gate 04
Repeat the runs on devices
Run the same candidate repeatedly, disclose incomplete runs, and measure whether it reaches the intended workflows as well as whether it records a finding.
Gate 05
Prove the operating path
Verify review completion, bounded cost and time, evidence retention, monitoring, and rollback using the same configuration intended for production.
Gate 06
Publish the misses with the score
Release the cohort, denominator, exclusions, per-app outcomes, and known gaps beside any summary number. Inconclusive runs remain in the record.
Current status
What we are not claiming yet.
The public site should not make the verification decision before the test record does. These limits remain explicit until a result meets the standard.
- No current-core, multi-app certification score is published on this page.
- We do not publish catch-rate, false-alarm, speed, or reliability scores until the underlying work is independently reproduced.
- Earlier internal campaigns inform engineering work; they are not presented here as certification of the current product.
- Production readiness is assessed separately and includes operating and rollback checks.
Publication record
What each result must carry.
Help make the evidence representative.
We are working with a small number of Android teams whose real release workflows can strengthen the multi-app validation set. Participation is scoped before any review.