[Zaknin]

Evidence

• Snapshot: C:\Users\FucktoryVR\Documents\PimaxVrcSupervisorDiagnosticsArchive\BaseStationStartup\Phase29C-20260619-085435
• Phase 29B deployment source: d34715177cdde4887ee362d663b08901a8bc8da9
• Analysis base: ac5c37c4509bd94bb164b6ad580aa25d947bab2f
• Supervisor integrity: 10,443/10,443 valid records, 0 malformed
• Configurator integrity: 1,569/1,569 valid records, 0 malformed
• No invalid timestamps, duplicate event IDs, out-of-order timestamps, or uncorrelated records

Selected sessions

• Normal: bs-session-abad6804a98146b18dc7db98d5d9bb3d — 8 attempts, 8 successes, 0 timeouts, no Scan overlap
• Failed/no Scan: bs-session-196da76d16e24dd3ae176f6e37cca03f — 24 attempts, 0 successes, 24 device-resolution timeouts
• Failed/Scan-assisted: bs-session-b391b076b9b947058eae3bb5a0b029d7 — 3 initial device-resolution timeouts, then Supervisor-owned progress during Scan

The earliest reliable divergence is deviceResolution. Adapter lookup succeeded in all three scenarios. In the assisted session, Supervisor started the recovering operation 450 ms before Configurator Scan; device resolution completed 562 ms after Scan began. Configurator only enumerated/watched and matched the same four stations. It emitted no station-command event.

Classification

Primary: G — instrumentation is insufficient, medium confidence.

Active discovery may have warmed OS-visible Bluetooth state, but the evidence cannot distinguish that from elapsed readiness time. Supervisor created and completed the successful operations; Configurator did not resume an operation or issue the successful commands.

Changes

• Add offline-only base-station-startup-analysis-json command with schema base-station-startup-analysis-v1
• Add independent JSONL integrity parsing with line numbers and byte offsets
• Add session segmentation, normalized timelines, comparison, Scan correlation, and classification
• Add minimal sanitized normal, failed, and Scan-assisted fixtures
• Add 16 focused analysis and safety tests
• Add Phase 29C analysis documentation and MkDocs navigation

Validation

• Supervisor Release build: pass, 0 warnings
• .NET tests: 201 passed, 0 failed
• Configurator Release build: pass, 0 warnings
• SteamVR host Release build: pass, 0 warnings
• Rust TUI tests: 12 passed, 0 failed
• MkDocs strict: pass
• Fixture and real-snapshot smoke tests: exactly one valid JSON document
• Static safety scan: no Bluetooth, station-command, scheduled-task, deployment, process, configuration-write, or network path in the analyzer
• Phase 29B deployment hashes and scheduled-task action unchanged
• Protected Phase 28C3D branch remains at 9e466f4e801c396d17c383a878614e9a423af37c

No physical action, Bluetooth command, Configurator Scan, task restart, process restart, service restart, deployment replacement, release, or tag was performed.

Exactly one recommended next phase

Phase 29D-E — improve instrumentation before correction.

Add observation only at existing device-resolution and discovery-watcher transitions. Add no delay, retry, cooldown, discovery watcher, or command source; preserve current scheduler and concurrency guards. Capture OS device-arrival/readiness and resolution-route timing so a later explicitly authorized studio run can distinguish active-discovery warm-up from passive delay.