Repository navigation
Conversation
|
Codex review: needs changes before merge. What this changesThis PR enables Effort Off for qualified Qwen models, migrates managed provider metadata, and preserves session effort during canvas actions. Example: Select Effort Off with Qwen3.8 27B and ask “What is 19 times 21?”
Review scores
ProductKind: Bug fix · Worth it: Yes · Fix scope: Partial Merge readiness⛔ Needs changes before merge - 1 item remains Keep this PR open for a focused compatibility repair; it addresses a confirmed Local AI defect that current main still contains. Priority: P2 Before merge
Findings
Agent review detailsHow this fits togetherCompanion generates llama-server presets and managed Gateway provider configuration; Gateway translates chat and canvas effort choices into inference requests. flowchart LR
A[Chat effort choice] --> B[Gateway]
C[Canvas action] --> B
D[Model catalog] --> E[Provider coordinator]
E --> B
B --> F[llama-server]
F --> G[Model response]
Technical reviewBest possible solution: Retain the focused effort fix and validate replacement compatibility through the existing lifecycle and publication owners before withdrawing any working legacy route. Do we have a high-confidence way to reproduce the issue? Current-main source establishes the effort defects, and contributor runtime evidence demonstrates them; the older-Gateway upgrade failure is source-reproducible against the released strict schema, without a live reproduction in this review. Is this the best way to solve the issue? The catalog-owned effort vocabulary and existing migration owners fit the bug, but replacement compatibility must be established before route withdrawal; recorded requests also resolve the previous default-effort concern by showing Medium before and after. Full review comments:
Overall correctness: patch is incorrect AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning medium; reviewed against aa2089bd64a0. Provenance checked
TestingProof path: shipped entry point. SecurityNone. EvidenceAcceptance criteria:
What I checked:
Likely related people:
Preflight sourceSource unknown. Source comparison at the observed PR snapshot only. Completion is self-reported; body, policy, tests, execution, and report bytes are not verified. Review metrics
LabelsLabel changes:
Label justifications:
Rating scale6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior. WorkflowClawSweeper edits this one comment on every review. Comment HistoryReview history (9 earlier review cycles; latest 8 shown)
Reviewed November 19, 2026, 5:18 PM ET / 22:18 UTC (Revision 10). |
|
🦞👀 Pull request received. I will update this pull request when review starts. |
PR cleanup reviewReviewed head: Thanks for the follow-up work on legacy recovery, exact rollback and localized guidance. This still addresses behavior present on main: the router forces reasoning on, provider entries lack effort metadata, and canvas actions explicitly request Low, so I would keep the PR open. The latest head needs Windows validation and product proof before landing. The currently visible successful checks are ClawSweeper dispatch jobs, and the substantive review targets the older One remaining source path needs attention: direct WSL Settings Start/Restart returns from After reconciling with main, please provide exact-head Windows build/test results and native/WSL migration plus Effort/canvas proof. Maintainer confirmation of the 2026.9.8 floor, Qwen3.8 default change and deferred non-Qwen scope is also still needed. Action in this cleanup pass: leaving this PR open. No branch changes or runtime-proof waiver. The findings above come from source, discussion and recorded-CI review; Windows validation was not rerun in this pass. |
e249e75 to
f4b2dc0
Compare
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
Effort Off still produced reasoning for managed Qwen models because the llama-server preset forced `reasoning = on` and the Gateway provider had no effort vocabulary to map Off onto. - Drop `reasoning = on` from the router preset so llama-server's default auto mode lets each request turn reasoning off. - Tag the qualified Qwen catalog entries with a reasoning profile (Qwen38 / QwenOnOff) and publish effort metadata for those templates only (supportsReasoningEffort, supportedReasoningEfforts, reasoningEffortMap off->none, thinkingLevelMap). Other templates keep the previous provider shape. - Recognize the exact pre-upgrade provider shape in ownership, recovery, rollback and teardown (SetupEngine step, native lifecycle, coordinator) so an unchanged legacy route is not treated as externally edited. - Migrate an exact legacy provider by replacing only its owned `models` array: guarded config.patch on native, dry-run then --expect-current-json write then restart on WSL. The API key is never replayed from a redacted readback. - Native endpoint recovery keeps its early return for a current-shape route and continues through the guarded coordinator only for the exact legacy shape. - Canvas actions no longer send `thinking: "low"`, so they inherit the session's effort. No Gateway version gate: every field written has been in the Gateway config schema since 2026.6.5. Gateways 2026.6.5 to 2026.9.7 accept it and keep today's behavior (Off still reasons); older Gateways reject the write through the existing guarded paths with no config change. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aa4bc38 to
4ec8e74
Compare
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
…igration A provider in the exact pre-effort shape needs its model metadata upgraded. WSL: the companion never edits it in place. Every runtime start (and every non-ready refresh) quiesces the route first: the provider is unset, the Gateway restarted and the removal read back before any publication, and the llama-server router is in a kill-on-close job object, so it never outlives the app. If a publication still finds the legacy shape, it is withdrawn the same way and published from scratch (API key from the credential store, never from a redacted readback); a failed withdrawal fails the publication and writes nothing new. Native: endpoint recovery deliberately keeps a verified live route and passes the legacy provider to publication, so only the owned model array is patched atomically and read back, as before. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ed6d88f to
34f8085
Compare
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
…hout effort metadata From an independent review of the previous head: - A native legacy migration whose write lands but whose readback does not match (here, a Gateway that drops thinkingLevelMap) now has a test. It reports the localized LocalAi_GatewaySavedRouteVerification failure, makes no second write, and keeps the upgraded route, its credential and the managed primary in place. - The Bonsai receipt test now asserts what removing the forced `reasoning = on` means for a model without a reasoning profile: the preset uses llama-server's auto mode, and the provider entry carries no effort vocabulary, so it is unchanged from the pre-upgrade shape and the Gateway never requests "none" for it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
|
asking for a review @joelagnel |
|
@turutupa could you check whether upgrading a legacy WSL Local AI route against an older Gateway can leave the working provider removed when the new metadata is rejected? At the reviewed head ( A concrete version to check is Gateway My reading is that publication would then fail after the old route has already been removed, and retrying would generate the same rejected configuration. Normal runtime startup also quiesces before publication, so this may extend beyond the direct migration branch. Is there a minimum-version or capability guard elsewhere that rules this out? If not, could we check compatibility before withdrawal or preserve the legacy shape for those Gateways, with a regression test for replacement rejection? I have not reproduced this on a live WSL installation; this concern is based on source review and the schema probe. |
|
@clawsweeper re-review |
|
🦞👀 Re-review progress:
|
…tadata
The companion admits any protocol-4 Gateway, but the effort metadata this
PR publishes needs newer model schemas: supportsReasoningEffort arrived in
2026.1.14, supportedReasoningEfforts and reasoningEffortMap in 2026.4.26,
and thinkingLevelMap only in 2026.6.5. Gateway 2026.6.1-2026.6.4 therefore
rejects the new provider ("Unrecognized key"). PublishAsync withdrew a
pre-effort WSL provider (config unset + Gateway restart) before the
replacement was dry-run, and normal startup also withdraws on its endpoint
cycle before publishing (native: config.patch, same order). On those
Gateways the working route was removed, publication failed, and every retry
generated the same rejected config.
Decide the provider shape before anything is withdrawn or written:
- WSL: dry-run the full batch (new ValidateBatchAsync, dry-run only); if it
is rejected, dry-run the exact pre-effort batch.
- Native: read-only config.schema check (SupportsEffortMetadataAsync); a
Gateway that does not list config.schema is newer than 2026.6.5.
- Full shape when accepted; otherwise the pre-effort shape, as on main, with
one warning that Effort Off cannot disable reasoning on that Gateway. A
pre-effort provider there counts as current: no withdraw/republish churn
and no in-place migration.
- If neither shape validates or the check fails, nothing is withdrawn and
the operation fails with a localized message, keeping the current route.
An endpoint-cycle quiesce decides the shape and the following publication
reuses it; teardown needs no check. Publication, verification and native
migration use the decided shape. Setup's WSL step already dry-runs before
applying; it now falls back to the pre-effort batch when a dry-run confirms
the Gateway rejects the full one, and its rollback recognizes that shape.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An independent review of 7bdbfe9 found that endpoint-cycle quiesce ran the provider-shape check before withdrawing. The runtime also withdraws through an endpoint cycle to take the route off an untrusted or conflicting listener (LlamaServerRuntimeService ownership-incomplete and conflict paths), so a failed check (config.schema timeout, dry-run indeterminate) left the Gateway routing prompts to that listener. - Quiesce withdraws unconditionally again, as on main. The shape is decided only in PublishAsync, once per publication, before anything is written. This also removes the cached decision, so native quiesce and publish on separate coordinator instances no longer re-check after a withdrawal, and a failed quiesce cannot leave a stale shape behind. - WSL treats the effort shape as rejected only when its dry-run fails twice and the pre-effort shape validates, so one failed CLI run cannot downgrade a working route. - The rejected and check-failed messages now say nothing was written instead of claiming the current route was kept (all six languages). Tests: endpoint-cycle withdrawal never waits on the check (WSL and native, native publish through a separate coordinator), publication rejecting both shapes writes nothing, and a one-off dry-run failure keeps the current shape. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
@joelagnel thanks, you were right, and it reproduces on a real install. Repro (Gateway 2026.6.1, managed WSL, fresh v2026.11.3 install, then this branch): at Root cause: the companion admits any protocol-4 Gateway, but Fix (
Live, same machine:
I also measured every stable Gateway release from 2026.6.1 to 2026.11.6 (50), with both entries. Off sends |
Related: #1976
What Problem This Solves
Effort Off does not stop managed Qwen Local AI models from thinking. The llama-server preset forces
reasoning = on, and the Gateway provider entry has no effort vocabulary, so Off is never sent asnone. Canvas actions also sendthinking: "low", which switches the session's Effort.User Impact
With a qualified Qwen model, Effort Off now sends
reasoning_effort: "none"and the model answers without reasoning. Qwen3.8 offers Off, Low, Medium, High and Ultra. Qwen3.6 and Qwen3.5 offer Off, Low, Medium and Ultra: Low and Ultra are sent asmedium, the only other effort these templates accept. Canvas actions keep the user's Effort. Existing installs are upgraded automatically without changing their credentials. Bonsai, Nemotron Bolt and custom models keep their current provider entry (tracked in #1976).Why This Change Was Made
none, and auto mode keeps a thinking template reasoning as before. A test now pins this down.supportedReasoningEfforts,reasoningEffortMap(offtonone) andthinkingLevelMapfor those templates only.LlamaServerRuntimeService, endpoint-cycle quiesce). The new shape is then published from scratch, with the API key from the companion's credential store, never from a redacted readback.modelsarray, atomically (config.patchagainst the captured hash), then reads it back. In practice this happens on explicit endpoint recovery ("Recover and use Local AI"), which deliberately keeps the verified route; startup has already withdrawn it otherwise. If the readback doesn't match, the failure is reported and the upgraded route is left in place.thinkingLevelMaponly exists from Gateway 2026.6.5. Gateway 2026.6.1 rejects this PR's entry, and the earlier code published it anyway, so Local AI ended with no route.openclaw config set --batch-file … --dry-run), then the pre-upgrade entry if needed; a rejection counts only when the dry-run fails twice. Native readsconfig.schema(config.patchhas no dry-run).main.main(commit 5).Five commits on current
main(17794662):LocalAiManagedProcessHost). A pre-upgrade WSL provider is withdrawn and republished instead.7bdbfe9d).4c902b0b): withdrawal no longer waits on the shape check, the decision is made once per publication, one failed dry-run can't downgrade a route, and the two failure messages say nothing was written.Default effort (the open decision in #1976)
Measured on every stable OpenClaw Gateway release from 2026.6.1 to 2026.11.6 (50 releases, none skipped), with Qwen3.8 27B and the companion's GGUF and preset. The method and every release's numbers are under Gateway versions.
mainmainmainmediumhighsent, reasonsmediumnone, no reasoningmediummediumnone, no reasoning"Nothing sent" means the chat template's default applies. Its system text reads "Reasoning effort is set to xhigh".
On Gateway 2026.10.5 and later (setup installs the latest Gateway; 2026.11.6 in the runs below), this PR does not change the default effort:
mainalready sendsmedium. It fixes Off.On Gateway 2026.8.1 – 2026.10.4, the default changes from the template's
xhightomedium. This is the trade-off Local AI: Effort Off doesn't stop llama-server thinking, and newer gateways return HTTP 500 for unsupported levels #1976 leaves open.On 2026.6.5 – 2026.7.35, those Gateways send Off as
high, notnone, so Off still reasons.mainsends nothing there, so Off reasons onmaintoo.With
main's preset (reasoning = on), llama-server also ignoresnone: an Off request spent 98 tokens (median) reasoning. This branch removes that, andnonegives 4.12-prompt benchmark (details below):
medium(this branch's default)none(this branch's Off)Off trades accuracy for speed, as chosen by the user. It is not the default.
Change Type
Scope
winnodeRequired proof pools
windows-wsl-dgx-blackwell: Local AI on a Blackwell GPU. Inventory recipe run, plus fixed-prompt inference, below.windows-wsl-gateway-e2e: a real managed WSL Gateway upgraded from v2026.11.3 (refused restarts and recovery included), plus the inventory recipe, below.windows-clean-installer-upgrade: v2026.11.3's own setup (headlessSetupEnginefor WSL; the release app's onboarding UI for native), then this branch on the same data, plus the installer source assertions recipe. Signed installers: see Not verified.windows-11-arm64: every run below is native ARM64 (RTX Spark N1X).windows-winui-interactive: Effort menu, chat, recovery UI and canvas in this branch's app, plus the inventory recipe, below.Validation
Windows 11 ARM64 (NVIDIA RTX Spark N1X, Blackwell), head
4c902b0b:OpenClaw.Tray.WinUI, Debug win-arm64): 0 warnings, 0 errors. ProductVersion…Sha.4c902b0b3310….Tray (
GatewayActionTransportTests,LocalAiGatewayProviderCoordinatorTests,LocalAiGatewayRpcConfigurationTransportTests,LocalAiPageViewModelTests): 147 passed. Includes the shape tests:config.schema.It also includes
RpcPublish_LegacyMigrationThatFailsReadbackKeepsTheUpgradedRouteAndReportsLocalizedFailure.Connection (
LocalAiPortLifecycleTests): 131 passed, 1 skipped. Includes the Bonsai preset assertions.SetupEngine (
NativeLocalAiLifecycleTests,LocalAiGatewayUninstallTests,LocalAiOnboardingTests,SetupPipelineTests): 296 passed. Includes the setup fallback to the legacy batch and its rollback.scripts/Test-Localization.ps1: 44 passed.git diff --check: clean.scripts/run-proof-tests.ps1, checked withscripts/validate-proof-evidence.ps1), at4c902b0b:windows-wsl-dgx-blackwell/run-local-ai-gpu-tests, exactly as inventoried: 15 discovered, 15 passed, 0 skipped.validExecutionEvidence: true. The summary also binds the app binary and the eight screenshots below, including the two older-Gateway ones.windows-wsl-dgx-blackwell/record-gpu-model:NVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), 621.24.The
win-x64recipes ran exactly as inventoried, on an x64 .NET runtime installed side by side (DOTNET_ROOT_X64):windows-wsl-gateway-e2e/run-gateway-product-path(OPENCLAW_RUN_E2E=1,OPENCLAW_RUN_OLLAMA_E2E=1,SetupAndConnectTests,win-x64): 78 discovered, 78 passed, 0 skipped,validExecutionEvidence: true. It used a fresh app-owned WSL distro, removed afterwards. Two earlier runs at this head failed 1 and 4 cases (an MCP connection refused, then node pairing not completing). Both ran while the test distro from the live runs below still had its own Gateway running. With that distro stopped, the next run passed 78/78. I did not isolate the cause further; at7bdbfe9dthis recipe passed 78/78 on its first run.windows-clean-installer-upgrade/run-installer-source-assertions(InstallerIssAssertionTests): 21 passed, 0 skipped,validExecutionEvidence: true.windows-winui-interactive/run-accessibility-scan(Category=Accessibility,win-x64, private desktop), run atf096746c(commits 4 and 5 change no XAML), repeated on both builds:f096746cmain17794662The one failure is the same intermittent
DashboardsPageButtonShouldHavePatternsfinding on both builds, and this PR doesn't touch that page. Every run also has 14 Gateway-fixture cases skipped (opt-in throughscripts/test-gateway-fixture.ps1). The runner treats skips as incomplete proof, so it writes no valid summary for this recipe on either build.main's production code,17794662):RouterPreset_BoundsOmittedGenerationAtGatewayMaximumandQwen38Provider_HidesUnsupportedMinimalButKeepsFourSupportedEfforts.GatewayActionTransportTests): 2 failures,BuildAgentRequestPayload_InheritsSessionEffortand…DeliverIsFalse_WithoutChangingEffort.main: legacy-shape recognition doesn't exist there.f096746c):LocalAiGatewayUninstallTests): 1 failure,Configure_GatewayRejectingEffortMetadataPublishesAndRollsBackLegacyShape(setup fails instead of falling back); 29 passed.SupportsEffortMetadataAsyncand the shape check don't exist).7bdbfe9d): 6 of 68 coordinator tests fail. They are the two endpoint-cycle withdrawal tests, the one-failed-dry-run test, the rejected-both-shapes test, and the two legacy-shape tests whose dry-run count changed.Independent review
An independent reviewer (a separate Claude Code agent given only the diff, with no access to this description) reviewed
17794662..34f80858. It found no blocking bugs and checked:models, WSL republishes from the credential store, and rollback rebuilds from the known key;.reswfiles.Its three P3 notes are addressed:
It also flagged one thing it could not check without a Gateway: whether a live Gateway keeps
thinkingLevelMap'snullvalues. The readbacks below show it does ("thinkingLevelMap":{"minimal":null}).A second independent review, again with only the diff, covered commit 4 (
f096746c..7bdbfe9d). It found one P1, three P2s and two P3s; commit 5 addresses them:LlamaServerRuntimeServiceownership paths), so a failed check left the route on that listener. Withdrawal is unconditional again, as onmain.config.schematext. Measured instead of assumed: on all 50 releases, a liveconfig.schemaresponse gives the same verdict as the dry-run. 2026.6.1 doesn't namethinkingLevelMapanywhere; every later release names it once, as a model property (Gateway versions). The set of Gateways older than the effort keys is closed, so the match can't misfire on them.config.schemais treated as newer. Checked:config.schema, measured on all 50;server-plugins.ts);It confirmed the rest held: no early withdrawal on publish, setup fallback and rollback, the ownership matching, no key in dry-runs or RPC checks, and the six translations.
Real Behavior Proof
All app runs are on a private desktop (
scripts/Invoke-PrivateDesktop.ps1), driven through UI Automation. API keys are shown only as SHA-256 prefixes.Before and after on the same machine, Gateway and runtime (heads
17794662andf096746c)mainsets up Local AI and writes its provider.What is 19 times 21? Reply with only the number.main17794662f096746creasoning = oncompat: {supportsTools, supportsUsageInStreaming}supportedReasoningEfforts: ["none","low","medium","high"],reasoningEffortMap: {"off":"none"},thinkingLevelMap: {"minimal":null}; same key98ea79e6020fand primary399, 18 tokens generated (reasoning)399, 4 tokens (answer only)main, Effort menu:main, Effort Off:This branch, Effort menu:
This branch, Effort Off:
This branch, Local AI running after the upgrade:
Older Gateway: 2026.6.1 before and after the fix (heads
f096746cand4c902b0b)The case from @joelagnel's review, on a real install.
SetupEngine(headless, isolated data roots) withGateway.Versionset to2026.6.1: a new app-owned WSL distro, OpenClaw Gateway 2026.6.1 as a systemd service, Local AI with Qwen3.8 27B. The v2026.11.3 app wrote the pre-upgrade provider (key8f93e66f6416). Then each branch build was started on that data.openclaw config set --batch-file … --dry-run, throwaway profiles):f096746c(before the fix)4c902b0bThe verified Local AI route could not be published to the app-owned gateway. The just-written route was removed.Nollamacppprovider left; routerFailedmain), decides the shape, then republishes the pre-upgrade entry:The Gateway does not accept Local AI effort metadata (it needs OpenClaw Gateway 2026.6.5 or later). Publishing the earlier provider shape, so Effort Off cannot turn reasoning off on this Gateway.Same key8f93e66f6416, same primary; routerHealthy. The same on a second startf096746cthe upgrade was shown on Gateway 2026.11.4 (WSL upgrade below)8f93e66f6416and primary; routerHealthyf096746con Gateway 2026.6.1, Local AI page after startup (red toast: an unrelated private-desktop notification):4c902b0bon Gateway 2026.6.1, Local AI page after startup (same unrelated toast):On Gateway 2026.6.1, v2026.11.3's own setup fails at its
pair-nodestep (Node approval failed), in both runs. That step is in the released setup, not in this PR; the operator connection and Local AI were set up.Native endpoint recovery: models-only in-place patch (head
f096746c)main1779466298ea79e6020f, primaryllamacpp/qwen3.8-27b-mtp-ud-q4-k-mopenclaw config set agents.defaults.timeoutSeconds 601config.patchwithchangedPaths=models.providers.llamacpp.modelsagents.defaults.timeoutSecondskeptRecovery guidance (test Gateway id blacked out):
The recovery card (captured at
34f80858; commit 3 changes only tests):Gateway log, config writes (redacted)
Provider readbacks (hashes only)
Upgrade from v2026.11.3, managed WSL Gateway (head
34f80858, production code unchanged since)SetupEngine, run headless (astests/OpenClaw.E2ETestsdoes) with isolated data roots: a new app-owned WSL distro, OpenClaw Gateway 2026.11.4 as a systemd service, and Local AI with Qwen3.8 27B (UD-Q4_K_M).8f93e66f6416, routerHealthyThe managed Local AI route was removed from Gateway configuration, but the Gateway restart was not confirmed…; startupFailed; nothing new writtenHealthy; effort fields present; same key and primary; preset withoutreasoning = onStored data: every value in the seven files v2026.11.3 wrote was hashed by key path (values never printed). 147 of 148 values were unchanged. The only change is
gateways.json :: gateways[0].lastConnected. No new stored fields or files.Effort Off chat on the upgraded WSL install (head
f096746c):8f93e66f6416and the same primary.399, 16 tokens generated, no reasoning. The Gateway session reportsthinkingLevel=off.The default-effort turns in these WSL runs got "Agent error" because llama-server could not load the model on this test laptop. It has 31 GB of GPU memory, and the model at its preset's 262k context needs about 37 GB: 15 GB of weights, 16 GB of KV cache and the compute buffers. It loads only when Windows can spill into shared memory. The log line is
failed to allocate CUDA0 buffer of size 2718040192 … exiting due to model loading error. The preset's size settings are the same onmainand v2026.11.3; this PR only removesreasoning = on. Default turns are covered natively above and on Linux below.Upgrade from v2026.11.3, native Gateway (head
34f80858)v2026.11.3's own onboarding: native Gateway, "Set up a local model" with Qwen3.8 27B. Then this branch: route withdrawn at startup and republished with the effort fields, same key
98ea79e6020f. The same happened with the native Gateway stopped beforehand (the app starts it).Rollback: v2026.11.3 started after this branch sees the new shape as "changed outside the companion". It keeps the provider untouched and refuses to cycle it (Error on the Local AI page); nothing is overwritten or deleted. "Release Gateway ownership", or removing the provider, lets it set up again.
Effort menu, canvas and request path (head
4ec8e743; the production code for these is unchanged since)mainprovider entryreasoning_effort, 14 reasoning charsreasoning_effort: "none", 0 reasoning charshighsenthighsentmainapplowsent, 638 reasoning charsnonesent, 0 reasoning charsthinkingLevelafterwardslowoffThis branch's app after the canvas action (Effort still "off"; host name and device instance id blacked out):
main's app after the canvas action:Gateway request path for every level (
openclaw agent --thinking <level>, logging proxy):reasoning_effort, and the reply carried reasoning.reasoning_effort: "none", 0 reasoning chars.low/medium/highas chosen; Qwen3.6 sentmediumfor Low and Medium.minimalandxhighwere rejected by the Gateway.Gateway-side check on Linux (DGX Spark)
The same model file and Gateway version on Linux, with no memory limits. This shows the Gateway request and the model behavior for both provider entries; it is not the Windows product path, which is covered above.
unsloth/Qwen3.8-27B-GGUFUD-Q4_K_M, revision313447f, SHA-256322e194f…, as the Hub's LFS hash.reasoning-format deepseek, sampling as shipped).ctx-size65536 instead of 262144, on llama.cpp b12367.main's entry is the same without the effort fields.reasoning_effortand the reply's reasoning and answer sizes (no message text).openclaw agent --thinking <level>, the prompt "What is 19 times 21? Reply with only the number.":mainentry,mainpreset (reasoning = on)mediumsent; 18 tokensmediumsent; 285 tokensnonesent; 4 tokens, 0 reasoning charslow; 63 tokenslow; 40 tokensmedium; 22 tokensmedium; 19 tokenshigh; 76 tokenshigh; 279 tokensSingle turns at temperature 1, so the token counts vary from turn to turn; what matters is what is sent and what Off does. Every reply was
399except this branch's Off turn, which answered429without reasoning.Benchmark. 12 arithmetic and logic prompts with known answers, one request per prompt and effort, sent straight to llama-server:
reasoning_effortsentmainpreset: correct, median tokensmain's Off)none(this branch's Off)lowmedium(default)highGateway versions: what Effort Off and the default send
Every stable OpenClaw Gateway release from 2026.6.1 to 2026.11.6: 50 releases from npm, none skipped. Each ran in an isolated loopback profile on the Linux setup above (same GGUF, preset and logging proxy), once with
main's provider entry and once with this branch's. Each got a default turn and an Off turn (openclaw agent --thinking off), "What is 19 times 21? Reply with only the number.". The proxy records thereasoning_effortthe Gateway sent and the reply's reasoning size.config.schemaRPC on a live Gateway, the native check's input):Unrecognized key: "thinkingLevelMap"; the schema doesn't name it. A config that already contains it stops 2026.6.1 from starting (Invalid config). The fix keepsmain's entry there.thinkingLevelMapexactly once, as a property ofmodels.providers.*.models[].config.schemaresponse) gives the same answer as the dry-run on all 50 releases. Every release servesconfig.schema.nonehad 0 reasoning chars.npm install openclaw@2026.7.33leaves out its declared dependency@openclaw/ai@2026.7.33, so its CLI doesn't start. It was measured with that exact declared version installed next to it.chmod 700;noneand then hit the 500 s turn timeout.Every release
mainentry: defaultmainentry: OffInvalid config(the fix keepsmain's entry here)medium; 36 tokens, 129 reasoning charshigh; 126 tokens, 547 reasoning charsmedium; 262 tokens, 1197 reasoning charshigh; 280 tokens, 1265 reasoning charsmedium; 1046 tokens, 4294 reasoning charshigh; 841 tokens, 3537 reasoning charsmedium; 368 tokens, 1555 reasoning charshigh; 230 tokens, 990 reasoning charsmedium; 141 tokens, 515 reasoning charshigh; 322 tokens, 1390 reasoning charsmedium; 28 tokens, 59 reasoning charshigh; 783 tokens, 3203 reasoning charsmedium; 252 tokens, 1150 reasoning charshigh; 533 tokens, 2350 reasoning charsmedium; 128 tokens, 548 reasoning charshigh; 201 tokens, 885 reasoning charsmedium; 1474 tokens, 6514 reasoning charshigh; 40 tokens, 40 reasoning charsmedium; 449 tokens, 1817 reasoning charshigh; 386 tokens, 1673 reasoning charsmedium; 458 tokens, 1800 reasoning charshigh; 427 tokens, 1853 reasoning charsmedium; 194 tokens, 860 reasoning charshigh; 345 tokens, 1468 reasoning charsmedium; 179 tokens, 656 reasoning charshigh; 59 tokens, 167 reasoning charsmedium; 92 tokens, 344 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 22 tokens, 15 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 11 tokens, 4 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 24 tokens, 36 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 112 tokens, 410 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 30 tokens, 84 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 68 tokens, 247 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 40 tokens, 82 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 23 tokens, 53 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 27 tokens, 54 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 278 tokens, 994 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 61 tokens, 115 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 136 tokens, 525 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 271 tokens, 1130 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 385 tokens, 1490 reasoning charsnone; 45 tokens, 0 reasoning charsmedium; 153 tokens, 455 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 105 tokens, 396 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 93 tokens, 379 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 30 tokens, 57 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 169 tokens, 650 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 23 tokens, 40 reasoning charsnone; 87 tokens, 0 reasoning charsmedium; 34 tokens, 78 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 99 tokens, 354 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 112 tokens, 412 reasoning charsmedium; 32 tokens, 66 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 30 tokens, 69 reasoning charsmedium; 65 tokens, 108 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 144 tokens, 530 reasoning charsmedium; 134 tokens, 536 reasoning charsnone; 3 tokens, 0 reasoning charsmedium; 144 tokens, 457 reasoning charsmedium; 27 tokens, 49 reasoning charsnone; 91 tokens, 0 reasoning charsmedium; 112 tokens, 451 reasoning charsmedium; 99 tokens, 388 reasoning charsnone; 28 tokens, 0 reasoning charsmedium; 64 tokens, 246 reasoning charsmedium; 105 tokens, 259 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 25 tokens, 43 reasoning charsmedium; 58 tokens, 186 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 19 tokens, 14 reasoning charsmedium; 87 tokens, 307 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 102 tokens, 367 reasoning charsmedium; 38 tokens, 72 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 77 tokens, 255 reasoning charsmedium; 78 tokens, 307 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 25 tokens, 42 reasoning charsmedium; 366 tokens, 1593 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 88 tokens, 266 reasoning charsmedium; 28 tokens, 59 reasoning charsnone; 4 tokens, 0 reasoning charsmedium; 19 tokens, 14 reasoning charsmedium; 470 tokens, 1732 reasoning charsnone; 4 tokens, 0 reasoning charsNot verified / blocked
windows-clean-installer-upgrade:verify-installer-signatures,verify-installed-payload,exercise-upgrade-matrix):Security Impact
main. The native in-place patch changes only themodelsarray, so the key stays as stored.config.patch.Compatibility and Migration
nonefrom 2026.8.1 (2026.6.5 – 2026.7.35 send it ashigh).Review Conversations