Initial objective and assumption
The objective was a local native shell for an analytical Observatory view while retaining authentication and bounded navigation. Initial rendering succeeded, but rendering alone was not the acceptance condition.
Original assumption, test and root cause
The original assumption was that successful rendering would be sufficient. The observation was retained helper-process memory after closure while Workshop remained open; the root cause was incomplete lifecycle cleanup.
Failure or contradiction and design response
Lifecycle testing observed retained WebEngine helper memory after the embedded view closed while Workshop remained open. The evidence treated cleanup as part of correctness. Native integration was therefore deferred rather than described as active.
- Claim
- A working render was insufficient to pass Observatory integration.
- Method
- Local shell load and post-close lifecycle observation.
- Result
- The cleanup gate failed; native integration was deferred.
- Limitation
- This does not mean the local analytical stack failed. The browser and CLI paths remain available; the embedded path is not current.
- Source period
- 2026 integration result.
Validation after correction, current state and lesson
The correction was deliberate deferral of the embedded path. Validation after correction retained the browser and CLI routes with authentication enabled. Remaining limitation: the embedded shell is not active.
Current canonical state and transferable engineering lesson
Native System Pulse remains the primary in-application view. The local browser/CLI Observatory remains a bounded analytical option. Transferable lesson: lifecycle cleanup is a functional correctness requirement, not post-launch polish.
Related evidence: KEJ-EV-011.
Source and redaction note: Public summary derived from the canonical integration result. Process detail, addresses and configuration are withheld.