@bitkyc08/opencodex 2.82.0 → 2.84.0

package/src/server/responses/passthrough-dispatch.ts

@@ -1,3 +1,4 @@
+import { spendLedgerStorageErrorResponse } from "./spend-storage-error";
 import { createSteeringSettingsNormalizer } from "./native-steering-policy";
 import { nativeResponseControlEligible } from "./native-response-control";
 import { NativeInjectionReplay } from "./native-injection-replay";
@@ -8,7 +9,7 @@ import type {
   PassthroughAdmissionState,
 } from "./core-options";
 import type { PreparedResponsesRequest } from "./request-prepare";
-import type { ResponsesTransport } from "./request-transport";
+import { CopilotAutoAdapterHandoff, type ResponsesTransport } from "./request-transport";
 import type { ResponsesEffects } from "./response-effects";
 import type { ResponsesSendBudget } from "./request-send-budget";
 import { transientSendCapFor } from "./request-send-budget";
@@ -90,7 +91,7 @@ import {
   safeHostLabel,
 } from "./fetch-helpers";
 import { classifyPoolRecoveryDispatch } from "../../routing/probe-lease";
-import { clientCancelledResponse } from "./core-errors";
+import { anthropicPoolCooldownResponse, clientCancelledResponse } from "./core-errors";
 import {
   upstreamHostCircuitOpenResponse,
   usesCodexForwardPoolAuth,
@@ -100,6 +101,8 @@ import {
   shouldRetryCodexPoolAccountQuota,
   shouldRetryCodexPoolAccountTransient,
   retryCodexPoolOnAlternateAccount,
+  PoolPreOutputRefusals,
+  poolPreOutputRefusalHold,
 } from "./core-codex-account";
 import {
   clearCodexModelDenialEvidence,
@@ -166,6 +169,9 @@ import { isTransientConsoleGoUploadRejection } from "../../providers/opencode-ze
 import { planReasoningEffortDowngrade } from "../../providers/reasoning-metadata";
 import { unboundPoolSpendRefusalResponse } from "../workflow-refusal";
 
+/** Transfer an uncommitted Copilot exchange to the existing translated pipeline without new admission. */
+export interface PassthroughAdapterHandoff { kind: "adapter-handoff" }
+
 /** Prepares and recovers one native Responses exchange before client commitment. */
 export async function preparePassthroughExchange(
   requestContext: Pick<ResponsesRequestContext, "config" | "logCtx" | "options" | "req">,
@@ -192,6 +198,12 @@ export async function preparePassthroughExchange(
     | "passiveQuotaWriterGeneration"
     | "oauthDispatch"
     | "resolveSelectionAdapter"
+    | "resolveCopilotSelection"
+    | "copilotRefusalResponse"
+    | "oauth401ReplayAttempted"
+    | "activeAdapter"
+    | "runTurnAdapter"
+    | "invalidateSameTargetRequest"
     | "isOAuth401ReplayProvider"
     | "sentOAuthSnapshot"
     | "refreshResolvedOAuthSelection"
@@ -248,6 +260,8 @@ export async function preparePassthroughExchange(
     passiveQuotaWriterGeneration,
     oauthDispatch,
     resolveSelectionAdapter,
+    resolveCopilotSelection,
+    copilotRefusalResponse,
     isOAuth401ReplayProvider,
     refreshResolvedOAuthSelection,
     applyFailoverSnapshot,
@@ -732,7 +746,7 @@ export async function preparePassthroughExchange(
     // consumer's cancel to a signalled fetch, so we pass the signal and relay through relayWithAbort,
     // whose cancel() aborts the upstream — preventing leaked connections (RC2, passthrough path).
     const upstream = new AbortController();
-    linkAbortSignal(upstream, options.abortSignal);
+    const cleanupUpstreamAbort = linkAbortSignal(upstream, options.abortSignal ?? req.signal);
     const connectMs = config.connectTimeoutMs ?? 200_000;
     let upstreamResponse: Response;
     /**
@@ -822,7 +836,7 @@ export async function preparePassthroughExchange(
         httpOnly: true,
         providerName: route.providerName,
         modelId: route.modelId,
-        dispatchOverride: oauthDispatch(request),
+        dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
         beforeDispatch: headers => {
           if (signal.aborted) throw signal.reason;
           if (!transportState.selectionIsCurrent(transportState.requestBindings.get(request))) {
@@ -896,9 +910,10 @@ export async function preparePassthroughExchange(
         describeOutboundBodyRefusal(result),
       );
     };
-    const transportFailureResponse = (err: unknown): Response => {
+    const transportFailureResponse = (err: unknown): Response | PassthroughAdapterHandoff => {
+      cleanupUpstreamAbort();
       upstream.abort();
-      if (options.abortSignal?.aborted) {
+      if (options.abortSignal?.aborted || req.signal.aborted) {
         releaseUpstreamHostAdmission(nativeHostState.lease);
         nativeHostState.lease = null;
         releaseCodexAuthContextProbeLease(admissionState.authCtx);
@@ -915,6 +930,17 @@ export async function preparePassthroughExchange(
           ?? formatErrorResponse(429, "request_send_budget_exhausted", err.message);
       }
       const refusal = unwrapUpstreamRetryEvidenceError(err);
+      if (refusal instanceof CopilotAutoAdapterHandoff) {
+        bindRefreshedAdapter(refusal.adapter);
+        return handoffToChat(refusal.adapter);
+      }
+      const copilotRefusal = copilotRefusalResponse(refusal);
+      if (copilotRefusal) {
+        releaseUpstreamHostAdmission(nativeHostState.lease);
+        nativeHostState.lease = null;
+        releaseCodexAuthContextProbeLease(admissionState.authCtx);
+        return copilotRefusal;
+      }
       // Pacing may outlive the selected account's admission. No fetch occurred, so do
       // not turn an operator pause into a 502 or charge it to host/account health.
       if (refusal instanceof OAuthAccountPausedError || refusal instanceof OAuthLoginRequiredError || refusal instanceof AnthropicAccountCooldownError) {
@@ -925,12 +951,11 @@ export async function preparePassthroughExchange(
           return formatErrorResponse(401, "authentication_error", publicOAuthAuthenticationErrorMessage(refusal));
         }
         if (refusal instanceof AnthropicAccountCooldownError) {
-          return formatErrorResponse(429, "rate_limit_error", refusal.message,
-            refusal.retryAfterSeconds === null ? undefined : { retryAfter: String(refusal.retryAfterSeconds) });
+          return anthropicPoolCooldownResponse(refusal, { comboAttempt: options.comboAttempt, logCtx });
         }
         return formatErrorResponse(403, "permission_error", publicOAuthAuthenticationErrorMessage(refusal));
       }
-      const localRefusal = mapCodexAuthContextErrorToResponse(refusal, {
+      const localRefusal = spendLedgerStorageErrorResponse(err, logCtx) ?? mapCodexAuthContextErrorToResponse(refusal, {
         now: Date.now(), accountSelector: route.codexAccountNamespace,
       });
       if (localRefusal) {
@@ -981,6 +1006,9 @@ export async function preparePassthroughExchange(
     // Stable per request: the Combo booking (`reserveDispatch`) already started and froze the spend
     // policy, so `spendEnforced` cannot change before dispatch even though no reporter starts it here.
     const receiptMode = !sendBudgetState.adapterSendBudget?.spendEnforced && Boolean(options.comboInitialSend);
+    // #6969: Pool sends hold empty lifecycle frames so a pre-output refusal reaches the ladder below.
+    const preOutputRefusals = new PoolPreOutputRefusals();
+    const preOutputRefusalHold = () => poolPreOutputRefusalHold(admissionState.authCtx, route.provider, connectMs, options.comboAttempt);
     try {
       // Transient-5xx pre-stream retry (devlog/_plan/260716_claudecode_hardening/010):
       // the ChatGPT backend emits transient 502/520s that an immediate retry absorbs.
@@ -1004,18 +1032,19 @@ export async function preparePassthroughExchange(
               nativeControl: nativeResponseControlEligible(route.provider, options.nativeControl) && options.inboundTransport === "websocket" && !options.comboAttempt
                 && responseEffects.plaintextV2AgentMessageToolNames.size === 0
                 ? options.nativeControl : undefined,
-              dispatchOverride: oauthDispatch(request),
+              dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
               providerName: route.providerName,
               modelId: route.modelId,
               onPhysicalDispatch: receiptMode ? sendBudgetState.noteInitialDispatch : undefined,
               onCodexWsQuota: codexWsQuotaObserver(admissionState.authCtx, route.provider, route.modelId),
               beforeDispatch: isCanonicalOpenAiForwardProvider(route.provider)
                 ? createCodexAuthDispatchGuard(admissionState.authCtx, options.codexAuthPolicy ?? config, route.modelId, options.admission, options.visionDescribeTerminal === true) : undefined,
+              preOutputRefusalHold: preOutputRefusalHold(),
             }),
             route.provider.authMode === "forward")
             // Every real attempt response — including an intermediate 5xx the
             // retry wrapper replaces — proves the host was reached (#914 review).
-            .then(adoptObservedResponse);
+            .then(adoptObservedResponse).then(preOutputRefusals.observe);
         },
         { abortSignal: upstream.signal, label: safeHostLabel(request.url),
           attempts: sendBudgetState.initialSendAllowance(transientSendAttempts()),
@@ -1026,6 +1055,8 @@ export async function preparePassthroughExchange(
         },
       );
     } catch (err) {
+      preOutputRefusals.recordOnLocalExit(err, options.abortSignal?.aborted === true || req.signal.aborted,
+        config, admissionState.authCtx, route.provider, route.modelId);
       return transportFailureResponse(err);
     } finally {
       request.releaseBodyObservation?.();
@@ -1034,7 +1065,33 @@ export async function preparePassthroughExchange(
     const opaqueBlobRecoveryGuard: OpaqueBlobRecoveryGuard = { attempted: false };
     // At most one reasoning-effort downgrade per request.
     const reasoningEffortDowngradeGuard: { attempted: boolean } = { attempted: false };
-    let oauth401ReplayAttempted = false;
+    /** Rebind the refreshed wire before transferring this request to the Chat pipeline. */
+    function bindRefreshedAdapter(refreshedAdapter: typeof transportState.adapter) {
+      bindRouteReasoningReplayScope({
+        parsed,
+        providerName: route.providerName,
+        provider: route.provider,
+        adapterName: refreshedAdapter.name,
+        oauthCredentialSnapshot: transportState.replayOAuthCredentialSnapshot,
+      });
+      logCtx.providerAdapter = refreshedAdapter.name;
+      sealRequestAttemptIdentity(
+        logCtx.activeAttempt,
+        logCtx.provider,
+        refreshedAdapter.name,
+        logCtx.accountLogLabel,
+      );
+      recordAttemptCredentialSource(logCtx.activeAttempt, route.providerName, route.provider, refreshedAdapter.name);
+    }
+    /** Transfer adapter ownership without retaining the native abort link or request cache. */
+    function handoffToChat(refreshedAdapter: typeof transportState.adapter) {
+      cleanupUpstreamAbort();
+      upstream.abort();
+      transportState.adapter = transportState.activeAdapter = transportState.runTurnAdapter = refreshedAdapter;
+      transportState.invalidateSameTargetRequest();
+      return { kind: "adapter-handoff" as const };
+    }
+    let oauth401ReplayAttempted = transportState.oauth401ReplayAttempted;
     let codex401ReplayKind: "main" | "stored" | null = null;
     // Console Go answers a transient 400 "Invalid upload request." for bodies it accepts
     // moments later; at most one byte-identical replay is allowed per request.
@@ -1043,7 +1100,7 @@ export async function preparePassthroughExchange(
     let rateLimitRetries = 0;
     const rebuildAndRefetch = async (
       recovery: AttemptRecoveryKind,
-    ): Promise<Response | { failed: Response }> => {
+    ): Promise<Response | { failed: Response | PassthroughAdapterHandoff }> => {
       const retryAdapter = resolveSelectionAdapter(
         resolveWireProtocolOverride(route.providerName, route.modelId, route.provider, inboundWire, route.staticPolicy),
         config.cacheRetention,
@@ -1113,7 +1170,7 @@ export async function preparePassthroughExchange(
                 nativeControl: nativeResponseControlEligible(route.provider, options.nativeControl) && options.inboundTransport === "websocket" && !options.comboAttempt
                   && responseEffects.plaintextV2AgentMessageToolNames.size === 0
                   ? options.nativeControl : undefined,
-                dispatchOverride: oauthDispatch(request),
+                dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
                 providerName: route.providerName,
                 modelId: route.modelId,
                 onCodexWsQuota: codexWsQuotaObserver(admissionState.authCtx, route.provider, route.modelId),
@@ -1220,7 +1277,7 @@ export async function preparePassthroughExchange(
             nativeControl: nativeResponseControlEligible(route.provider, options.nativeControl) && options.inboundTransport === "websocket" && !options.comboAttempt
               && responseEffects.plaintextV2AgentMessageToolNames.size === 0
               ? options.nativeControl : undefined,
-            dispatchOverride: oauthDispatch(request),
+            dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
             providerName: route.providerName,
             modelId: route.modelId,
             onCodexWsQuota: codexWsQuotaObserver(admissionState.authCtx, route.provider, route.modelId),
@@ -1282,7 +1339,7 @@ export async function preparePassthroughExchange(
       // fault and throw away the credential evidence the client needs.
       && !sendBudgetExhausted(transientSendAttempts())
     ) {
-      oauth401ReplayAttempted = true;
+      oauth401ReplayAttempted = transportState.oauth401ReplayAttempted = true;
       try { void upstreamResponse.body?.cancel().catch(() => {}); } catch { /* already consumed/closed */ }
       let refreshed: OAuthAccessSnapshot;
       try {
@@ -1305,7 +1362,7 @@ export async function preparePassthroughExchange(
       if (route.providerName === "kiro") {
         parsed._kiroAuthContext = { ...(refreshed.kiro ?? {}) };
       }
-      const refreshedProvider = resolveProviderTransport(
+      let refreshedProvider = resolveProviderTransport(
         route.providerName,
         {
           ...route.provider,
@@ -1318,29 +1375,21 @@ export async function preparePassthroughExchange(
           : undefined,
       );
       route.provider = refreshedProvider;
+      try { await resolveCopilotSelection(parsed); }
+      catch (err) { return transportFailureResponse(err); }
+      refreshedProvider = route.provider;
       const refreshedAdapter = resolveSelectionAdapter(
         resolveWireProtocolOverride(route.providerName, route.modelId, refreshedProvider, inboundWire, route.staticPolicy),
         config.cacheRetention,
       );
+      bindRefreshedAdapter(refreshedAdapter);
       if (!("passthrough" in refreshedAdapter) || !refreshedAdapter.passthrough) {
+        if (route.providerName === "github-copilot" && refreshedAdapter.name === "openai-chat") {
+          return handoffToChat(refreshedAdapter);
+        }
         upstream.abort();
         return formatErrorResponse(502, "upstream_error", "OAuth refresh changed the provider wire unexpectedly");
       }
-      bindRouteReasoningReplayScope({
-        parsed,
-        providerName: route.providerName,
-        provider: refreshedProvider,
-        adapterName: refreshedAdapter.name,
-        oauthCredentialSnapshot: transportState.replayOAuthCredentialSnapshot,
-      });
-      logCtx.providerAdapter = refreshedAdapter.name;
-      sealRequestAttemptIdentity(
-        logCtx.activeAttempt,
-        logCtx.provider,
-        refreshedAdapter.name,
-        logCtx.accountLogLabel,
-      );
-      recordAttemptCredentialSource(logCtx.activeAttempt, route.providerName, route.provider, refreshedAdapter.name);
       try {
         request = await refreshedAdapter.buildRequest(parsed, {
           headers: requestState.selectedForwardHeaders,
@@ -1373,7 +1422,7 @@ export async function preparePassthroughExchange(
                 nativeControl: nativeResponseControlEligible(route.provider, options.nativeControl) && options.inboundTransport === "websocket" && !options.comboAttempt
                   && responseEffects.plaintextV2AgentMessageToolNames.size === 0
                   ? options.nativeControl : undefined,
-                dispatchOverride: oauthDispatch(request),
+                dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
                 providerName: route.providerName,
                 modelId: route.modelId,
                 onCodexWsQuota: codexWsQuotaObserver(admissionState.authCtx, route.provider, route.modelId),
@@ -1414,38 +1463,50 @@ export async function preparePassthroughExchange(
         true,
       );
       if (hop.allowed) {
-        const nextAccountId = rotateGenericOAuthAccountOn429(
-          config, route.providerName, transportState.genericFailoverAccountId,
-          upstreamResponse.headers.get("retry-after"),
-          Date.now(),
-          route.modelId,
-        );
-        let snapshot: OAuthAccessSnapshot | undefined;
-        if (nextAccountId) {
-          try { snapshot = await failoverAccountSnapshot(route.providerName, nextAccountId); }
-          catch { /* Keep the original 429 body readable when the next credential is unavailable. */ }
-        }
-        if (snapshot && await applyFailoverSnapshot(snapshot)) {
-          transportState.genericFailovers += 1;
-          route.provider = resolveProviderTransport(
-            route.providerName, route.provider, parsed.options.promptCacheKey, transportState.sentOAuthSnapshot?.apiBaseUrl,
+        let transferred = false;
+        try {
+          const nextAccountId = rotateGenericOAuthAccountOn429(
+            config, route.providerName, transportState.genericFailoverAccountId,
+            upstreamResponse.headers.get("retry-after"),
+            Date.now(),
+            route.modelId,
           );
-          bindRouteReasoningReplayScope({
-            parsed, providerName: route.providerName, provider: route.provider,
-            adapterName: "openai-responses", oauthCredentialSnapshot: transportState.replayOAuthCredentialSnapshot,
-          });
+          let snapshot: OAuthAccessSnapshot | undefined;
+          if (nextAccountId) {
+            try { snapshot = await failoverAccountSnapshot(route.providerName, nextAccountId); }
+            catch { /* Keep the original 429 body readable when the next credential is unavailable. */ }
+          }
+          if (snapshot && await applyFailoverSnapshot(snapshot)) {
+            transportState.genericFailovers += 1;
+            route.provider = resolveProviderTransport(
+              route.providerName, route.provider, parsed.options.promptCacheKey, transportState.sentOAuthSnapshot?.apiBaseUrl,
+            );
+            const rotatedAdapter = resolveSelectionAdapter(
+              resolveWireProtocolOverride(route.providerName, route.modelId, route.provider, inboundWire, route.staticPolicy),
+              config.cacheRetention,
+            );
+            bindRefreshedAdapter(rotatedAdapter);
+            try { void upstreamResponse.body?.cancel().catch(() => {}); } catch { /* already closed */ }
+            sendBudgetState.pendingHopPermit = hop.permit;
+            if (route.providerName === "github-copilot" && rotatedAdapter.name === "openai-chat") {
+              transferred = true;
+              return handoffToChat(rotatedAdapter);
+            }
+            const result = await rebuildAndRefetch("oauth-account-429");
+            sendBudgetState.pendingHopPermit = undefined;
+            if ("failed" in result) return result.failed;
+            upstreamResponse = result;
+            continue passthroughRecovery;
+          }
+        } catch (err) {
           try { void upstreamResponse.body?.cancel().catch(() => {}); } catch { /* already closed */ }
-          // The replay IS this hop's send, so the rebuild spends the reservation instead of
-          // asking for one of its own.
-          sendBudgetState.pendingHopPermit = hop.permit;
-          const result = await rebuildAndRefetch("oauth-account-429");
-          sendBudgetState.pendingHopPermit = undefined;
-          if ("failed" in result) return result.failed;
-          upstreamResponse = result;
-          continue passthroughRecovery;
+          return transportFailureResponse(err);
+        } finally {
+          if (!transferred) {
+            if (sendBudgetState.pendingHopPermit === hop.permit) sendBudgetState.pendingHopPermit = undefined;
+            hop.permit?.release();
+          }
         }
-        // No credential moved, so the reservation costs nothing.
-        hop.permit?.release();
       } else {
         // The activation quorum ignores cooldowns; prove that the selector has a live alternate
         // before describing this as a recovery that only the shared request budget withheld.
@@ -1508,7 +1569,7 @@ export async function preparePassthroughExchange(
                 nativeControl: nativeResponseControlEligible(route.provider, options.nativeControl) && options.inboundTransport === "websocket" && !options.comboAttempt
                   && responseEffects.plaintextV2AgentMessageToolNames.size === 0
                   ? options.nativeControl : undefined,
-                dispatchOverride: oauthDispatch(request),
+                dispatchOverride: oauthDispatch(request, parsed, "passthrough"),
                 providerName: route.providerName,
                 modelId: route.modelId,
                 onCodexWsQuota: codexWsQuotaObserver(admissionState.authCtx, route.provider, route.modelId),
@@ -1622,6 +1683,7 @@ export async function preparePassthroughExchange(
           passthroughEstimate,
           stream: upstreamRequestsStream,
           httpOnly: canonicalBufferedJson,
+          preOutputRefusalHold: preOutputRefusalHold(),
           onResponse: (response, retryAuthCtx, retryRequest) => {
             adoptCodexWsStage(response);
             captureAffinityResponse(
@@ -1648,6 +1710,8 @@ export async function preparePassthroughExchange(
         }
       }
     }
+    // #6969: a projection nothing above recovered goes on as the original stream, before any reader.
+    upstreamResponse = preOutputRefusals.restore(upstreamResponse);
     // The deterministic route record cannot classify history it never observed (restart, expiry,
     // eviction, or an older transcript). Inspect only a bounded clone of a 4xx whose exact outbound
     // Responses body still carries opaque state, then rebuild once through the ordinary adapter
@@ -1875,6 +1939,7 @@ export async function preparePassthroughExchange(
   const localUpstream = isLocalUpstream(route.provider.baseUrl);
 
   return {
+    kind: undefined,
     codexSafetyBufferingOptions,
     imageGenCallAliases,
     routedCustomToolNames,
@@ -1914,7 +1979,8 @@ export async function preparePassthroughExchange(
     connectMs,
     localUpstream,
     upstreamResponse,
+    terminalOutcomeRecorded: preOutputRefusals.outcomeRecorded,
   };
 }
 
-export type PassthroughExchange = Exclude<Awaited<ReturnType<typeof preparePassthroughExchange>>, Response>;
+export type PassthroughExchange = Exclude<Awaited<ReturnType<typeof preparePassthroughExchange>>, Response | PassthroughAdapterHandoff>;

package/src/server/responses/passthrough-execution.ts

@@ -8,7 +8,7 @@ import type { ResponsesTransport } from "./request-transport";
 import type { ResponsesSidecarAuth } from "./request-sidecar-auth";
 import type { ResponsesEffects } from "./response-effects";
 import type { ResponsesSendBudget } from "./request-send-budget";
-import { preparePassthroughExchange } from "./passthrough-dispatch";
+import { preparePassthroughExchange, type PassthroughAdapterHandoff } from "./passthrough-dispatch";
 import { deliverPassthroughResponse } from "./passthrough-delivery";
 import { guardDirectPassthroughBodyInactivity } from "../../lib/response-body-inactivity";
 import { resolveStallTimeoutMs } from "../../stall-timeout";
@@ -24,7 +24,7 @@ export async function executePassthroughResponse(
   sidecarState: ResponsesSidecarAuth,
   responseEffects: ResponsesEffects,
   sendBudgetState: ResponsesSendBudget,
-): Promise<Response> {
+): Promise<Response | PassthroughAdapterHandoff> {
   const nativeHostState: PassthroughAdmissionState = { lease: admissionState.pendingHostAdmissionLease };
   admissionState.pendingHostAdmissionLease = null;
   try {
@@ -37,7 +37,7 @@ export async function executePassthroughResponse(
       responseEffects,
       sendBudgetState,
     );
-    if (nativeExchange instanceof Response) return nativeExchange;
+    if (nativeExchange instanceof Response || nativeExchange.kind === "adapter-handoff") return nativeExchange;
     const response = await deliverPassthroughResponse(
       requestContext,
       admissionState,

package/src/server/responses/request-prepare.ts

@@ -33,6 +33,7 @@ import { resolveContextPrincipal } from "../auth-cors";
 import { resolveSkillsSnapshotScopeKey, snapshotSkillsCatalogInBody } from "./skills-snapshot";
 import {
   isShadowSourceModel,
+  shadowCallPreservesSelectedModel,
   shadowSourceModelPrefix,
   shouldInterceptShadowCall,
 } from "../../lib/shadow-call";
@@ -275,10 +276,11 @@ export async function prepareResponsesRequest(
     }
   }
   let shadowCallIntercepted = false;
-  // A spawned sub-agent turn names its model on purpose; gpt-6-luna is both the helper
-  // slug and a default sub-agent model, so neither intercept site may rewrite that turn.
+  // Luna is also selectable by users/children. Both intercept sites use the same
+  // per-turn provenance; WS frames must not inherit the handshake's classification.
   const threadSpawn = isThreadSpawnRequest(req.headers);
-  if (!concreteSelection && !options.compactionRoutingOverride && !options.internalDecisionCall && !threadSpawn && !memoryModelApplies && body && typeof body === "object" && !Array.isArray(body)) {
+  const preserveShadowModel = shadowCallPreservesSelectedModel(body, req.headers, options.inboundTransport);
+  if (!concreteSelection && !options.compactionRoutingOverride && !options.internalDecisionCall && !preserveShadowModel && !memoryModelApplies && body && typeof body === "object" && !Array.isArray(body)) {
     const shadowIntercept = config.shadowCallIntercept;
     const rawShadowModel = (body as { model?: unknown }).model;
     if (shadowIntercept?.enabled && shadowIntercept.model && typeof rawShadowModel === "string"
@@ -613,7 +615,7 @@ export async function prepareResponsesRequest(
     }
     const _sci = config.shadowCallIntercept;
     let shadowRoute: RouteResult | undefined;
-    if (!concreteSelection && !memoryRoute && !options.memoryModelPhase && !options.compactionRoutingOverride && !options.internalDecisionCall && !threadSpawn && _sci?.enabled && _sci.model && isShadowSourceModel(parsed.modelId, _sci.sourceModels)) {
+    if (!concreteSelection && !memoryRoute && !options.memoryModelPhase && !options.compactionRoutingOverride && !options.internalDecisionCall && !preserveShadowModel && _sci?.enabled && _sci.model && isShadowSourceModel(parsed.modelId, _sci.sourceModels)) {
       const sourcePrefix = shadowSourceModelPrefix(parsed.modelId, _sci.sourceModels)!;
       let sourceIdentity = { providerName: OPENAI_CODEX_PROVIDER_ID, modelId: sourcePrefix };
       try {