@bitkyc08/opencodex 2.14.1 → 2.15.0

package/src/server/responses/collaboration.ts

@@ -102,18 +102,28 @@ import type { TranslatorBudget } from "../../lib/translator-budget";
 
 export function buildToolBridgeMaps(parsed: OcxParsedRequest, budget?: TranslatorBudget): {
   toolNsMap: Map<string, { namespace: string; name: string }>;
+  declaredToolNames: Set<string>;
+  /** Declared parameter schema per request-visible tool name (#1611 integer repair). */
+  toolParameterSchemas: Map<string, Record<string, unknown>>;
   freeformToolNames: Set<string>;
   toolSearchToolNames: Set<string>;
 } {
   const toolNsMap = new Map<string, { namespace: string; name: string }>();
+  const declaredToolNames = new Set<string>();
+  const toolParameterSchemas = new Map<string, Record<string, unknown>>();
   const freeformToolNames = new Set<string>();
   const toolSearchToolNames = new Set<string>();
   const toolAllowed = toolChoiceToolPredicate(parsed.options.toolChoice);
   for (const t of parsed.context.tools ?? []) {
     // Upstream output is untrusted: only restore calls for tools the caller authorized.
     if (!toolAllowed(t)) continue;
+    const wireName = namespacedToolName(t.namespace, t.name);
+    budget?.chargeRetained(new TextEncoder().encode(wireName).byteLength, { kind: "retained_collectors" });
+    declaredToolNames.add(wireName);
+    // Retained by reference (the schema is already resident in parsed.context.tools),
+    // so this adds a map entry rather than a copy of every tool's parameters.
+    if (t.parameters && typeof t.parameters === "object") toolParameterSchemas.set(wireName, t.parameters);
     if (t.namespace) {
-      const wireName = namespacedToolName(t.namespace, t.name);
       budget?.chargeRetained(new TextEncoder().encode(JSON.stringify([wireName, t.namespace, t.name])).byteLength, { kind: "retained_collectors" });
       toolNsMap.set(wireName, { namespace: t.namespace, name: t.name });
     }
@@ -126,7 +136,7 @@ export function buildToolBridgeMaps(parsed: OcxParsedRequest, budget?: Translato
       toolSearchToolNames.add(t.name);
     }
   }
-  return { toolNsMap, freeformToolNames, toolSearchToolNames };
+  return { toolNsMap, declaredToolNames, toolParameterSchemas, freeformToolNames, toolSearchToolNames };
 }
 
 

package/src/server/responses/core.ts

@@ -22,6 +22,7 @@ import {
   markBodyNonPersistable,
   previousResponseProviderState,
   previousResponseReplayFailure,
+  previousResponseScopeMismatch,
   rememberResponseState,
 } from "../../responses/state";
 import {
@@ -1503,8 +1504,12 @@ async function handleResponsesInner(
   let unreadableEncryptedAgentTask = hasUnreadableEncryptedAgentTask(
     (body as { input?: unknown } | undefined)?.input,
   );
+  const inboundClientThreadId = req.headers.get("x-codex-parent-thread-id")?.trim() || undefined;
   const originalBody = body;
-  body = expandPreviousResponseInput(body);
+  body = expandPreviousResponseInput(body, inboundClientThreadId);
+  if (previousResponseScopeMismatch(body)) {
+    console.warn("[opencodex] dropped a previous_response_id with a mismatched client task scope; continuing fresh");
+  }
   if (previousResponseReplayFailure(body)) {
     return formatErrorResponse(
       400,
@@ -1512,7 +1517,8 @@ async function handleResponsesInner(
       "Continuation state is unavailable or corrupt; resend the full conversation without previous_response_id.",
     );
   }
-  const previousResponseInputExpanded = body !== originalBody;
+  const previousResponseInputExpanded = body !== originalBody
+    && typeof (body as { previous_response_id?: unknown }).previous_response_id === "string";
 
   // Spawn-message compatibility (both directions): agent_message task payloads ride in
   // encrypted_content slots as plaintext. Rewrite them to input_text on the RAW body BEFORE
@@ -1529,7 +1535,7 @@ async function handleResponsesInner(
       );
   }
 
-  let parsed;
+  let parsed: OcxParsedRequest;
   let toolBridgeMaps: ReturnType<typeof buildToolBridgeMaps>;
   try {
     parsed = parseRequest(body);
@@ -1537,10 +1543,9 @@ async function handleResponsesInner(
     if (previousResponseInputExpanded) parsed._previousResponseInputExpanded = true;
     parsed._providerContinuation = previousResponseProviderState(parsed.previousResponseId);
     parsed._cursorConversationId = parsed._providerContinuation?.cursor?.conversationId;
-    const clientThreadId = req.headers.get("x-codex-parent-thread-id")?.trim();
-    if (clientThreadId) {
-      parsed._clientThreadId = clientThreadId;
-      parsed._reasoningReplayScope = { clientThreadId };
+    if (inboundClientThreadId) {
+      parsed._clientThreadId = inboundClientThreadId;
+      parsed._reasoningReplayScope = { clientThreadId: inboundClientThreadId };
     }
   } catch (err) {
     if (isTranslatorBudgetExceededError(err)) {
@@ -1550,6 +1555,10 @@ async function handleResponsesInner(
     }
     return formatErrorResponse(400, "invalid_request_error", err instanceof Error ? err.message : String(err));
   }
+  const responseStateOptions = (force = false): { force?: boolean; clientThreadId?: string } => ({
+    ...(force ? { force: true } : {}),
+    ...(parsed._clientThreadId ? { clientThreadId: parsed._clientThreadId } : {}),
+  });
   // Prefer a pre-populated id (routed Claude) over Responses headers that may be
   // absent or synthetically injected (session_id from prompt_cache_key).
   if (!logCtx.conversationId) {
@@ -2137,7 +2146,7 @@ async function handleResponsesInner(
       && (!parsed.previousResponseId || parsed._previousResponseInputExpanded === true);
     const rememberPassthroughResponse = passthroughRecordEligible
       ? (response: { id?: unknown; output?: unknown; status?: unknown }) =>
-        rememberResponseState(parsed._rawBody, response, undefined, { force: true })
+        rememberResponseState(parsed._rawBody, response, undefined, responseStateOptions(true))
       : undefined;
     if (parsed.previousResponseId && !parsed._previousResponseInputExpanded) {
       console.warn(
@@ -2962,7 +2971,7 @@ async function handleResponsesInner(
           parsed._rawBody,
           response,
           continuationStateForResponse(providerState),
-          adapterNeedsForcedContinuation(adapter.name) ? { force: true } : undefined,
+          responseStateOptions(adapterNeedsForcedContinuation(adapter.name)),
         ),
     });
     if (imgResponse.body) {
@@ -3075,7 +3084,7 @@ async function handleResponsesInner(
       }
     };
 
-    const { toolNsMap, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
+    const { toolNsMap, declaredToolNames, toolParameterSchemas, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
     if (parsed.stream) {
       void runTurn();
       let eventSource: AsyncIterable<AdapterEvent> = queue.stream();
@@ -3101,6 +3110,8 @@ async function handleResponsesInner(
           ...(options.forceEmptyResponseId ? { responseId: "" } : {}),
           stallTimeoutSec: config.stallTimeoutSec,
           hideThinkingSummary: parsed.options.hideThinkingSummary,
+          declaredToolNames,
+          toolParameterSchemas,
           ...(options.onFirstOutput ? { onFirstOutput: options.onFirstOutput } : {}),
           ...(routedCompaction ? { compaction: true } : {}),
           onUsage: usage => {
@@ -3118,7 +3129,7 @@ async function handleResponsesInner(
                 parsed._rawBody,
                 response,
                 continuationStateForResponse(providerState),
-                adapterNeedsForcedContinuation(adapter.name) ? { force: true } : undefined,
+                responseStateOptions(adapterNeedsForcedContinuation(adapter.name)),
               ),
           }),
         },
@@ -3147,6 +3158,8 @@ async function handleResponsesInner(
       replayCacheScope: parsed._reasoningReplayScope,
       hideThinkingSummary: parsed.options.hideThinkingSummary,
       toolNsMap,
+      declaredToolNames,
+      toolParameterSchemas,
       freeformToolNames,
       toolSearchToolNames,
       ...(routedCompaction ? { compaction: true } : {}),
@@ -3164,7 +3177,7 @@ async function handleResponsesInner(
         parsed._rawBody,
         json,
         continuationStateForResponse(providerState),
-        adapterNeedsForcedContinuation(adapter.name) ? { force: true } : undefined,
+        responseStateOptions(adapterNeedsForcedContinuation(adapter.name)),
       );
     }
     return new Response(JSON.stringify(json), { headers: { "Content-Type": "application/json" } });
@@ -3835,7 +3848,7 @@ async function handleResponsesInner(
           continuation: fetchTerminalGuardContinuation,
         })
       : initialEventStream;
-    const { toolNsMap, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
+    const { toolNsMap, declaredToolNames, toolParameterSchemas, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
     const sseStream = bridgeToResponsesSSE(
       eventStream, parsed._responseModelId ?? parsed.modelId, toolNsMap, freeformToolNames, toolSearchToolNames,
       () => upstream.abort(), 2_000,
@@ -3845,6 +3858,8 @@ async function handleResponsesInner(
         ...(options.forceEmptyResponseId ? { responseId: "" } : {}),
         stallTimeoutSec: config.stallTimeoutSec,
         hideThinkingSummary: parsed.options.hideThinkingSummary,
+        declaredToolNames,
+      toolParameterSchemas,
         ...(options.onFirstOutput ? { onFirstOutput: options.onFirstOutput } : {}),
         ...(routedCompaction ? { compaction: true } : {}),
         onUsage: usage => {
@@ -3864,7 +3879,7 @@ async function handleResponsesInner(
               parsed._rawBody,
               response,
               continuationStateForResponse(providerState),
-              activeAdapter.name === "kiro" ? { force: true } : undefined,
+              responseStateOptions(activeAdapter.name === "kiro"),
             ),
         }),
       },
@@ -3895,13 +3910,15 @@ async function handleResponsesInner(
     } finally {
       cleanupUpstreamAbort();
     }
-    const { toolNsMap, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
+    const { toolNsMap, declaredToolNames, toolParameterSchemas, freeformToolNames, toolSearchToolNames } = toolBridgeMaps;
     let providerState: OcxProviderContinuationState | undefined;
     const json = buildResponseJSON(events, parsed._responseModelId ?? parsed.modelId, {
       translatorBudget,
       replayCacheScope: parsed._reasoningReplayScope,
       hideThinkingSummary: parsed.options.hideThinkingSummary,
       toolNsMap,
+      declaredToolNames,
+      toolParameterSchemas,
       freeformToolNames,
       toolSearchToolNames,
       ...(routedCompaction ? { compaction: true } : {}),
@@ -3920,7 +3937,7 @@ async function handleResponsesInner(
         parsed._rawBody,
         json,
         continuationStateForResponse(providerState),
-        activeAdapter.name === "kiro" ? { force: true } : undefined,
+        responseStateOptions(activeAdapter.name === "kiro"),
       );
     }
     return new Response(JSON.stringify(json), { headers: { "Content-Type": "application/json" } });

package/src/server/startup-health-cache.ts

@@ -10,6 +10,7 @@ import { truncateRetainedUtf8 } from "../lib/admission";
 
 const CACHE_TTL_MS = 30_000;
 const PROBE_TIMEOUT_MS = 5_000;
+const INITIAL_PROBE_WAIT_MS = 5_500;
 const MAX_DIAGNOSTIC_VALUE_BYTES = 8 * 1024;
 let cached: { timestamp: number; value: StartupHealth } | null = null;
 let inflight: Promise<StartupHealth> | null = null;
@@ -109,6 +110,17 @@ function refreshInBackground(config: Pick<OcxConfig, "codexAutoStart">): void {
 export async function getCachedStartupHealth(config: Pick<OcxConfig, "codexAutoStart">): Promise<StartupHealth> {
   if (cached && Date.now() - cached.timestamp < CACHE_TTL_MS) return cached.value;
   refreshInBackground(config);
+  // An expired or empty read is an explicit protection check. Wait for the
+  // isolated probe instead of presenting a synthetic failure while that probe
+  // is still running. The probe remains child-process isolated and hard-capped
+  // at 5s; stale state is returned only if that bounded probe cannot settle.
+  if (inflight) {
+    const settled = await Promise.race([
+      inflight,
+      new Promise<null>(resolve => setTimeout(() => resolve(null), INITIAL_PROBE_WAIT_MS)),
+    ]);
+    if (settled) return settled;
+  }
   return cached ? markStartupHealthDiagnosticStale(cached.value) : conservativeFallback(config);
 }
 

package/src/usage/expected-prices.ts

@@ -52,6 +52,10 @@ const DAYBREAK_RED: Cost4 = { input: 12.5, output: 75, cacheRead: 1.25, cacheWri
 const GPT56_TERRA: Cost4 = { input: 2, output: 12, cacheRead: 0.2, cacheWrite: 2.5 };
 const GPT56_LUNA: Cost4 = { input: 0.2, output: 1.2, cacheRead: 0.02, cacheWrite: 0.25 };
 const GEMINI_36_FLASH: Cost4 = { input: 1.5, output: 7.5, cacheRead: 0.15, cacheWrite: 0 };
+// Gemini 3.7 Flash launch promotion: Google publishes $0.75 in / $3.75 out per 1M
+// through 2026-12-31, stepping up to $1.50 / $7.50 on 2027-01-01. Revisit this row
+// then — the promotional rate is dated on the pricing page, not open-ended.
+const GEMINI_37_FLASH: Cost4 = { input: 0.75, output: 3.75, cacheRead: 0.075, cacheWrite: 0 };
 const MINIMAX_M21_HIGHSPEED: Cost4 = { input: 0.6, output: 2.4, cacheRead: 0.03, cacheWrite: 0.375 };
 const KIMI_K3: Cost4 = { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3 };
 const KIMI_K27_CODE: Cost4 = { input: 0.95, output: 4, cacheRead: 0.19, cacheWrite: 0.95 };
@@ -70,6 +74,7 @@ const CLAUDE_OPUS_5_DERIVED_SOURCE =
 const ANTHROPIC_PRICING = "https://platform.claude.com/docs/en/about-claude/pricing (official; 5m cache-write tier)";
 
 const GEMINI_PRICING = "https://ai.google.dev/gemini-api/docs/pricing (2026-07-22); cacheWrite=0: storage is billed per-hour, not per-token";
+const GEMINI_37_PRICING = "https://ai.google.dev/gemini-api/docs/pricing (2026-08-14); promotional rate through 2026-12-31, rises to 1.50/7.50 on 2027-01-01; cacheWrite=0: storage is billed per-hour, not per-token";
 const MINIMAX_PRICING = "https://platform.minimax.io/docs/guides/pricing-paygo";
 const OPENAI_GPT56_PRICING = "https://developers.openai.com/api/docs/pricing";
 const DEEPSEEK_PRICING = "https://api-docs.deepseek.com/quick_start/pricing-details-usd; V4 Flash alias transition scheduled 2026-07-24 — re-verify after";
@@ -102,6 +107,12 @@ export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   // Google Antigravity effort-suffix variants — derived from the verified base-model
   // price (Google does not publish per-suffix prices; Agent inference bills at the
   // base model's standard rate per the official Billing FAQ).
+  // 3.7 Flash rides CCA, whose billing equivalence to the Developer API list price is
+  // not published, so this is `verified-derived` rather than `verified`: the number is
+  // proven, the claim that Antigravity charges it is inferred.
+  { provider: "google-antigravity", modelId: "gemini-3.7-flash", cost4: GEMINI_37_FLASH, source: `derived: Gemini 3.7 Flash promotional rate through 2026-12-31 ${GEMINI_37_PRICING}`, verifiedAt: "2026-08-14", status: "verified-derived" },
+  // Retained after the 3.6 retirement: historical usage.jsonl rows still carry these
+  // ids, and dropping the row would silently zero the cost of requests already made.
   { provider: "google-antigravity", modelId: "gemini-3.6-flash", cost4: GEMINI_36_FLASH, source: `collapsed base ID ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified" },
   { provider: "google-antigravity", modelId: "gemini-3.1-pro", cost4: GEMINI_31_PRO, source: `collapsed base ID ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified" },
   // OpenAI GPT-5.6 `-pro` virtual selections. The virtual resolver keeps the SELECTED id in
@@ -130,6 +141,8 @@ export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   { provider: "google-antigravity", modelId: "gemini-3-flash-agent", cost4: GEMINI_36_FLASH, source: `compat alias -> gemini-3.6-flash-high ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified-derived" },
   // Direct Google Gemini API current model (verified — published table).
   { provider: "google", modelId: "gemini-3.6-flash", cost4: GEMINI_36_FLASH, source: GEMINI_PRICING, verifiedAt: "2026-07-22", status: "verified" },
+  // Developer API row: the price IS published for this surface, so `verified`.
+  { provider: "google", modelId: "gemini-3.7-flash", cost4: GEMINI_37_FLASH, source: GEMINI_37_PRICING, verifiedAt: "2026-08-14", status: "verified" },
   { provider: "google-antigravity", modelId: "gemini-3.1-pro-preview", cost4: GEMINI_31_PRO, source: GEMINI_PRICING, verifiedAt: "2026-07-20", status: "verified" },
   // Antigravity-bundled third-party models — derived from the underlying vendor's
   // official API price (Antigravity itself bills via subscription quota).