Skip to content

IDE Software Prior Testing Summary

Status: Reviewer summary Updated: 2026-09-08

Testing Body

The frozen software is supported by unit, integration, UI, simulation, pod-simulation, algorithm characterization, differential, static-analysis, cybersecurity, formative real-device, and operator evidence. Development evidence was used to identify and correct field issues before the freeze; formal claims use the controlled exact-freeze evidence lane.

Exact-Freeze Engineering Results

Area Result
Algorithm protocol Behavioral suites passed; Algo2015 coverage 96.12% line / 87.77% branch; bridge coverage 90.27% line / 65.85% branch; one host regression passed
App unit/integration target 997 total / 996 passed / 1 intentional IFU reference-table generation helper skip gated by an export directory / 0 failed
UI target 41 passed / 3 reproducible test-harness deviations
Simulation Core, pod, alert, and real-engine lanes passed
Scoped local cybersecurity 68/68 passed
Exact-freeze dependency companion 68/68 scoped security, 8/8 OmniBLE, 9/9 app unlock, and 19/19 core unlock tests passed

Representative High-Risk Coverage

  • runtime cadence, pump-state gates, and degraded operation
  • CGM acquisition, outage behavior, fingerstick BG, and G7 replacement
  • meal announcement, cancellation, relaunch, and interrupted-delivery recovery
  • programmed backup basal, masked fallback, and bounded replay
  • request-step delivery attribution and exactly-once consumption
  • suspension/resumption and Temporary Target lifecycle
  • clinical unlock, participant reset, and persisted state
  • alert lifecycle, priority, notification, and recovery
  • telemetry durability, sequencing, export, and session isolation
  • primary/safety algorithm isolation and total-insulin host conformance

Formative and Physical Evidence

Formative real-device and operator testing informed defect discovery and design refinement. Build 843 working evidence now records subject JRC-FORMAL-VV, installed build 1.0 (843), the 2026-08-24/25 execution window, normal real-Pod operation, outage/reconciliation drills, manual BG, Temporary Target, suspend/resume, and an exact-session recovery ZIP. The ZIP is the primary device-local technical record through its 09:33:48 EDT cutoff; later exercises use operator evidence with Scout telemetry as corroborating backup. Appendix A10B records hashes, checker results, and limitations.

These observations remain supporting evidence. Build 843 is distinct from the original exact-freeze identity; its controlled record identifies the accepted exact-freeze evidence relationship and the separate Build 843 evidence.

Evidence Classification

  • exact-freeze bundles under Evidence/Formal/ are formal execution evidence
  • development and working runs are supporting evidence only
  • signed physical/operator evidence is formal only after protocol completion and sponsor-designated acceptance

Engineering execution is complete, and evidence acceptance is recorded under D08. No new freeze rerun is implied unless a deviation decision or baseline change requires one. The controlled multi-G7 exercise remains a predeployment condition under D06.