Historical release · v0.4.0 · public summary revision 2
The record behind the assessment.
Local, synthetic findings. Results and limitations remain separate from production qualification.
Separate counts.
Retained negative findings.
| Population | Recorded finding |
|---|---|
| Unit / regression methods | 201 passed per complete run. |
| Reference executions | 36 scenarios × 4 modes = 144. 71 PASS / 33 FAIL / 40 INCONCLUSIVE. |
| Legacy executions | 120. Another 31 targeted executions occur inside tests; these populations are separate. |
| Actual buyer paths / native provider calls | 0 / 0. |
A passing test means the expected finding occurred. FAIL and INCONCLUSIVE remain in the record. Repeat runs and replays are not independent observations or a reliability percentage.
Two assessment defects were identified after this release. The recorded counts do not establish their detection; see v0.4.1.
A local record has a local boundary.
No customer integration, native-provider finality, production acceptance or external independent reproduction is established. Real-money production remains NO_GO. These assessment results do not validate the execution-control engine.
Local cases do not establish exhaustive distributed-race coverage, production persistence or disaster recovery. Historical local storage reliability remains unresolved. Partial payments, FX and refunds require separate evaluation.
Current evidence and limitsFollow the summary
to the source.
Supporting records, source provenance and reproduction prerequisites are available for scoped technical review. Sharing terms and access to the matching kit are agreed before delivery.
Source archive identity
HALTSEAL Qualification Kit v0.4.0 · SHA-256
c4ad73759e85670685ca7eccd60f6a6f35e5a09bc6661e62315edc7cc8570c0cHashes identify bytes against a reference. They do not authenticate evidence or certify its conclusions.
Original records are preserved separately. This revision changes the public summary, not the historical findings.