Open questions and evidence gaps
- Requirement completeness: the bounded HDRFP month explains current movement but is not the complete historical/customer requirement corpus. Which epics/specifications are authoritative for each customer family?
- Firmware compatibility evidence: firmware implementation is correctly external, but joint command/update compatibility still needs released firmware identities and end-to-end traces/results. Which team artifact is authoritative for command/status/version semantics beyond host headers?
- Runtime validation: configured source and Actions prove bounded checks, not that all 13 packages currently build, sign, pass HLK/accuracy tests and operate correctly on hardware.
- Skipped release jobs: the 2026-08-11 all-customer orchestrator succeeded while customer build/release jobs were skipped because no source change was detected. A fresh all-customer build remains separate evidence.
- Rust traceability: many Rust adapter/ETW/replay commits lack HDRFP keys; their relation to HDRFP-12243/12256 is plausible but not direct.
- Blocked backlog: 25 of 69 current unresolved HDRFP issues are Blocked. The Wiki needs blocker/dependency reasons before prioritization can be considered complete.
- External tuning dependencies: QMAutoTrain references legacy external repositories; they remain outside the Windows driver runtime baseline.
- Branch intent: Jira improves current intent coverage, but not every long-lived customer branch difference has a linked decision/issue.
- Secrets/assets in source: source trees include signing/licensing-related assets. This Wiki records only process boundaries and never copies their contents.