[Bug]: Native Gateway Local AI setup fails its response test despite retries #1713

Activity

1. github-project-automation

   added this to Windows Companion Appon Oct 12, 2026
2. github-project-automation

   moved this to Backlog in Windows Companion Appon Oct 12, 2026
3. clawsweeper

   added

   P0Emergency: data loss, security bypass, crash loop, or unusable core runtime.Emergency: data loss, security bypass, crash loop, or unusable core runtime.

   clawsweeper:needs-live-reproClawSweeper needs live local, crabbox, or manual validation to confirm this issue.ClawSweeper needs live local, crabbox, or manual validation to confirm this issue.

   clawsweeper:needs-maintainer-reviewClawSweeper marked this issue as needing maintainer review before automation.ClawSweeper marked this issue as needing maintainer review before automation.

   clawsweeper:no-new-fix-prClawSweeper does not recommend queueing a new automated fix PR for this issue.ClawSweeper does not recommend queueing a new automated fix PR for this issue.

   impact:ux-release-blockerA non-technical user is blocked without terminal, logs, config, or support.A non-technical user is blocked without terminal, logs, config, or support.

   issue-rating: 🐚 platinum hermitGood issue quality with a plausible reproduction path needing some confirmation.Good issue quality with a plausible reproduction path needing some confirmation.

   on Oct 12, 2026
4. clawsweeper commented on Oct 12, 2026

   

   clawsweeperBot

   on Oct 12, 2026 – with ClawSweeper

   Last edited by clawsweeper

   More actions

   Codex review: this still needs some work.

   Summary
   Keep open: screenshots establish blocked onboarding, current main still obscures verification failures, and the same-author diagnostics PR explicitly leaves the underlying inference failure unresolved.

   Reproducibility: Screenshots establish the reported onboarding failure and source proves lost diagnostic feedback, but missing native-run versions and logs prevent a high-confidence reproduction of persistent inference failure.

   Next step
   Review the existing diagnostics PR, collect versions and correlated verification logs from the affected native run, and establish the inference failure before selecting a repair with guided in-product recovery.

   Review details

   Best possible solution:

   Restore exact-model verification and provide guided in-product recovery, such as a Doctor or Fix action, with sanitized failure details and an accessible diagnostic log location.

   Do we have a high-confidence way to reproduce the issue?

   Screenshots establish the reported onboarding failure and source proves lost diagnostic feedback, but missing native-run versions and logs prevent a high-confidence reproduction of persistent inference failure.

   Is this the best way to solve the issue?

   Preserving sanitized errors is a useful partial repair already proposed in the linked PR; the inference repair should follow correlated verification and runtime evidence rather than an assumed Gateway or CUDA cause.

   AGENTS.md: found and applied where relevant.

   Remaining risk / open question:

   • The affected native run's Companion and Gateway versions, model ID, Windows build, verification response, and correlated runtime logs remain missing.
   • Persistent inference failure has not been independently reproduced; outdated Gateway, CUDA failure, and earlier startup failure remain unconfirmed explanations.

   Codex review notes: model internal, reasoning medium; reviewed against 817b0b586c1e.

   Label changes

   Label changes:

   • add issue-rating: 🦪 silver shellfish: Current issue advisory state selects this label.
   • remove clawsweeper:no-new-fix-pr: Current issue advisory state no longer selects this label.
   • remove clawsweeper:needs-maintainer-review: Current issue advisory state no longer selects this label.
   • remove issue-rating: 🐚 platinum hermit: Current issue advisory state no longer selects this label.
   • remove clawsweeper:needs-live-repro: Current issue advisory state no longer selects this label.

   Label justifications:

   • P0: The recommended first-run Local AI route reportedly remains blocked after retries, with no demonstrated successful in-product recovery.
   • impact:ux-release-blocker: The screenshots leave a community user with generic rejection, terminal access, and retry actions that reportedly fail.

   Evidence reviewed

   What I checked:

   • Inspected reporter screenshots: Both downloaded failure screenshots show generic provider rejection and Gateway recovery actions; the first marks the real-response test Needs attention after the preceding checks complete. They establish the visible failure, without identifying its runtime cause.
   • Current-main verification boundary: VerifyConfiguredAsync enters Rejected on unsuccessful verification; VerifyRouteAsync requires failure Status and Error but retains neither as client failure state. (src/OpenClaw.SetupEngine/GatewayAiSetupClient.cs:301, 817b0b586c1e)
   • Failure feedback overwritten: Initialization and refresh display VerificationFailed without returned error details; RunCoreAsync subsequently displays Rejected for the rejected phase. (src/OpenClaw.SetupEngine.UI/Pages/AiSetupPage.xaml.cs:943, 817b0b586c1e)
   • Existing retry coverage: ConfiguredModelMode_RetriesExactVerificationWithoutOfferingOtherProviders supplies an unavailable response followed by success for demo/model; it does not explain persistent native inference failure or verify visible diagnostic retention. (tests/OpenClaw.SetupEngine.Tests/GatewayAiSetupClientTests.cs:529, 817b0b586c1e)
   • Open same-author diagnostic work: fix(setup): preserve sanitized Local AI verification errors #1716 (fix(setup): preserve sanitized Local AI verification errors) remains open and explicitly improves diagnostics only; joelagnel reports controlled WinUI-host proof and says native inference reproduction is blocked by the test machine's isolated-session capability check.
   • Related report has different evidence: Store 2026.9.803.0: AI setup reports ready despite verification error, then Gateway restarts #1717 reports Store 2026.9.803.0 with a WSL Gateway, unavailable blender-mcp executable, and repeated restarts; its versions and trigger cannot be assigned to this native Gateway report.

   How this review workflow works

   • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
   • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
   • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
   • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
   • Maintainers can also comment @clawsweeper review to request a fresh review only.
   • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge; those execution lanes are retired.
   • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

   Reviewed October 13, 2026, 2:32 PM ET / 18:32 UTC.
5. clawsweeper

   added

   issue-rating: 🦪 silver shellfishThin issue quality; more reproduction proof or environment detail is needed.Thin issue quality; more reproduction proof or environment detail is needed.

   and removed

   clawsweeper:no-new-fix-prClawSweeper does not recommend queueing a new automated fix PR for this issue.ClawSweeper does not recommend queueing a new automated fix PR for this issue.

   clawsweeper:needs-maintainer-reviewClawSweeper marked this issue as needing maintainer review before automation.ClawSweeper marked this issue as needing maintainer review before automation.

   issue-rating: 🐚 platinum hermitGood issue quality with a plausible reproduction path needing some confirmation.Good issue quality with a plausible reproduction path needing some confirmation.

   clawsweeper:needs-live-reproClawSweeper needs live local, crabbox, or manual validation to confirm this issue.ClawSweeper needs live local, crabbox, or manual validation to confirm this issue.

   on Oct 13, 2026
6. github-actions commented on Oct 28, 2026

   

   github-actionsBot

   on Oct 28, 2026 – with GitHub Actions

   Contributor

   More actions

   This issue has been automatically marked as stale due to inactivity.
   Please add updated OpenClaw Windows details or it will be closed.
7. github-actions

   added

   stale

   on Oct 28, 2026