[PrimeBuild-pc]

Summary

Implements the scope explicitly approved by the owner in this session; supersedes the original recommend-only scope in #32.

• Independent cleanup of unused older driver packages across the store: retain newest/equal versions and all device-bound versions; revalidate in the elevated worker; no /uninstall or /force; explicit user confirmation and no automatic undo claim. Optional official RAPR link.
• Initial ICMP baseline followed by five overlapping ETW/ICMP windows. Require repeated unique dominance and jitter above baseline/instrument resolution before affinity/disable suggestions; missing replies, single spikes and ambiguous devices remain inconclusive.
• Ancillary-only device disable checked both in UI and worker; protect display/GPU, storage, system/chipset, NIC, input and unknown classes.
• Fix WPF dispatcher deadlock reproduced with the real worker process, plus TraceEvent native dependency extraction in single-file builds and .sys module mapping.
• Revised specification: docs/HOST-DIAGNOSTICS.md; release notes: docs/release-notes/0.9.0-alpha.8.md.
• Seven freshly captured alpha.8 screenshots, cache-safe paths, explicit source/version provenance, and preserved historical Research: ETW DPC/ISR capture technique and per-core CPU usage reader #49 research. Original research branch already removed after preservation.

Verification

• Full Release tests: 956 passed, 14 live hardware tests skipped, 0 failed.
• Worker regression demonstrated failing before the dispatcher fix and passing afterward.
• Self-contained alpha.8 candidate tested on clean NeuroTune-W11: 12 baseline replies, 20 concurrent replies across five windows, rejected invalid cleanup, rejected System-class disable, driver store unchanged. VM restored to Clean-NeuroTune-Alpha2.
• tools/Test-HostDiagnostics.ps1 retains the packaged-worker smoke test.
• Alpha.8 desktop launch/preferences version and all seven screenshots inspected; all 14 image references validated.
• graphify update . and git diff --check completed.

No real driver removal/device-disable or affinity benefit has been validated on physical hardware. Repeated co-occurrence is not proof of causation. The release remains an alpha preview.

Closes #32.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	⚠️ Failed 2026-10-30T15:44:26.470862Z We couldn’t complete this review. Comment @codex review to try again.	b44dd07	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[PrimeBuild-pc]

Branch cleanup: git cherry verified the original research commit is preserved in this PR (- 44bd300). The redundant research/dpc-isr-capture-technique branch has now been deleted locally and remotely; the research document remains in this pushed PR branch.

[PrimeBuild-pc]

Final local checks pass: dotnet format --verify-no-changes; 2 Python archive tests and regeneration check; 956 Release tests passed, 14 opt-in live tests skipped. Updated packaged VM smoke additionally requires resolved module names: vmbus.sys, ntoskrnl.exe, tcpip.sys, storvsp.sys, fastfat.sys, storport.sys, classpnp.sys, symcryptk.dll, ntfs.sys observed. No eligible removal/disable was requested; driver store unchanged and VM reset.