From edd396c62ec349fa686309c1ee55e02ebad6b903 Mon Sep 17 00:00:00 2001
From: makesomethingshit <junsoo1172@gmail.com>
Date: Mon, 7 Sep 2026 15:35:53 +0900
Subject: [PATCH 1/7] fix(opencode-go): carry session affinity through Pi
 native chat [skip ci]

(cherry picked from commit fbb214f92a7dc75f89d0d835fdb95136d61f380a)

Co-authored-by: makesomethingshit <246213378+makesomethingshit@users.noreply.github.com>
(cherry picked from commit 7e45a95f70689dfe1eb313c4af9aecbf2c44cdb6)
---
 src/clients/config-export.ts                  |  2 +
 src/server/chat-completions.ts                |  7 +++
 src/server/responses/core.ts                  |  3 +-
 tests/config/client-config-export.test.ts     |  6 ++-
 .../opencode-go-session-header.test.ts        | 52 ++++++++++++++++++-
 5 files changed, 67 insertions(+), 3 deletions(-)

diff --git a/src/clients/config-export.ts b/src/clients/config-export.ts
index 8fb42f311d6..e1744c2209f 100644
--- a/src/clients/config-export.ts
+++ b/src/clients/config-export.ts
@@ -695,6 +695,7 @@ export interface PiProviderBlock {
   baseUrl: string;
   api: string;
   apiKey: string;
+  compat?: { sendSessionAffinityHeaders: boolean };
   models: PiModelEntry[];
 }
 
@@ -859,6 +860,7 @@ function buildPiClientConfig(ctx: ExportContext): PiGeneratedConfig {
         baseUrl: ctx.baseUrl,
         api: PI_API_DIALECT,
         apiKey: LOOPBACK_API_KEY_PLACEHOLDER,
+        compat: { sendSessionAffinityHeaders: true },
         models,
       },
     },
diff --git a/src/server/chat-completions.ts b/src/server/chat-completions.ts
index e7fd04f42d4..7e69010636d 100644
--- a/src/server/chat-completions.ts
+++ b/src/server/chat-completions.ts
@@ -25,6 +25,8 @@ import { estimateTokens } from "../lib/token-estimate";
 import { NoEligiblePolicyCandidateError, UnknownRoutingPolicyError, routeModel } from "../router";
 import { evidenceFromBody } from "../routing/request-evidence";
 import { resolveWireProtocolOverride } from "./adapter-resolve";
+import { resolveOpenCodeGoTransport } from "../providers/opencode-go-transport";
+import { normalizeLogConversationId, sessionLaneIdFromRequest } from "./request-log-conversation";
 import type { OcxConfig } from "../types";
 import { readJsonRequestBody } from "./request-decompress";
 import {
@@ -136,6 +138,8 @@ async function handleChatCompletionsWithBudget(
   let chatNativeRoute: ReturnType<typeof routeModel> | null = null;
   try {
     const route = routeModel(config, chatBody.model as string, evidenceFromBody(chatBody));
+    route.provider = resolveOpenCodeGoTransport(route.provider,
+      sessionLaneIdFromRequest(req.headers) ?? normalizeLogConversationId(req.headers.get("x-opencode-session")));
     // Settle the wire once so every branch below reads the adapter this model will
     // actually use, not the provider-wide default (#404).
     route.provider = resolveWireProtocolOverride(route.providerName, route.modelId, route.provider, "chat");
@@ -237,6 +241,9 @@ async function handleChatCompletionsWithBudget(
     return chatCompletionsErrorResponse(400, CODEX_RESERVE_HELPER_UNSUPPORTED_MESSAGE, "invalid_request_error");
   }
   const headers = new Headers({ "content-type": "application/json" });
+  // Internal bridge metadata; the Go resolver scopes and hashes it before upstream use.
+  const openCodeSession = req.headers.get("x-opencode-session");
+  if (openCodeSession) headers.set("x-opencode-session", openCodeSession);
   for (const name of FORWARD_HEADERS) {
     if (name === "authorization" && !directRoute) continue;
     const value = req.headers.get(name);
diff --git a/src/server/responses/core.ts b/src/server/responses/core.ts
index 312af7ac43d..594b444c8bf 100644
--- a/src/server/responses/core.ts
+++ b/src/server/responses/core.ts
@@ -2304,7 +2304,8 @@ async function applyFinalRouteRequestNormalization(args: {
 
   // Settle the wire once so logging, fast-mode, auth, and sidecars read the adapter
   // this request will actually use (#404).
-  route.provider = resolveOpenCodeGoTransport(route.provider, sessionLaneIdFromRequest(req.headers));
+  route.provider = resolveOpenCodeGoTransport(route.provider,
+    sessionLaneIdFromRequest(req.headers) ?? normalizeLogConversationId(req.headers.get("x-opencode-session")));
   route.provider = resolveWireProtocolOverride(route.providerName, route.modelId, route.provider, inboundWire);
   if (preserveAnthropicResponseModel) parsed._responseModelId = responseModelId;
   logCtx.model = route.modelId;
diff --git a/tests/config/client-config-export.test.ts b/tests/config/client-config-export.test.ts
index 6a71813e9f0..4d707904033 100644
--- a/tests/config/client-config-export.test.ts
+++ b/tests/config/client-config-export.test.ts
@@ -315,6 +315,7 @@ describe("Pi serializer (accept criterion 2)", () => {
     expect(provider.baseUrl).toBe(BASE_URL);
     expect(provider.api).toBe("openai-completions");
     expect(provider.apiKey).toBe(LOOPBACK_API_KEY_PLACEHOLDER);
+    expect(provider.compat?.sendSessionAffinityHeaders).toBe(true);
   });
 
   test("cost is omitted on every entry — zeros would assert routed models are free", () => {
@@ -899,7 +900,7 @@ describe("EXPORT_CLIENTS registry", () => {
 `);
   });
 
-  test("pi bytes are unchanged, to the last newline", () => {
+  test("pi bytes include session affinity, to the last newline", () => {
     const built = buildClientConfigText("pi", ctx({ config: cfg() }));
     expect(built.format).toBe("json");
     expect(built.text).toBe(`{
@@ -908,6 +909,9 @@ describe("EXPORT_CLIENTS registry", () => {
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "opencodex-loopback",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
diff --git a/tests/providers/opencode-go-session-header.test.ts b/tests/providers/opencode-go-session-header.test.ts
index 00684f065bd..9b476760cdd 100644
--- a/tests/providers/opencode-go-session-header.test.ts
+++ b/tests/providers/opencode-go-session-header.test.ts
@@ -3,6 +3,7 @@ import { providerConfigSeed } from "../../src/providers/derive";
 import { resolveOpenCodeGoTransport } from "../../src/providers/opencode-go-transport";
 import { getProviderRegistryEntry } from "../../src/providers/registry";
 import { handleResponses } from "../../src/server/responses/core";
+import { handleChatCompletions } from "../../src/server/chat-completions";
 import type { OcxConfig, OcxProviderConfig } from "../../src/types";
 
 const MUSE_MODEL = "muse-spark-1.3-contributor";
@@ -52,6 +53,8 @@ async function captureRequest(input: {
   model?: string;
   child?: string;
   provider?: OcxProviderConfig;
+  nativeChat?: boolean;
+  headers?: Record<string, string>;
 } = {}): Promise<{ url: string; headers: Headers }> {
   const providerName = input.providerName ?? "opencode-go";
   const model = input.model ?? MUSE_MODEL;
@@ -65,7 +68,15 @@ async function captureRequest(input: {
   const config = {
     providers: { [providerName]: input.provider ?? opencodeGo() },
   } as unknown as OcxConfig;
-  const response = await handleResponses(
+  const response = input.nativeChat ? await handleChatCompletions(
+    new Request("http://localhost/v1/chat/completions", {
+      method: "POST",
+      headers: input.headers ?? codexHeaders(input.child),
+      body: JSON.stringify({ model: `${providerName}/${model}`, messages: [{ role: "user", content: "ping" }], stream: false }),
+    }),
+    config,
+    { model: "", provider: "" },
+  ) : await handleResponses(
     new Request("http://localhost/v1/responses", {
       method: "POST",
       headers: codexHeaders(input.child),
@@ -77,6 +88,7 @@ async function captureRequest(input: {
   );
 
   expect(response.status).toBe(200);
+  await response.text();
   expect(requests).toHaveLength(1);
   return requests[0]!;
 }
@@ -85,6 +97,44 @@ describe("OpenCode Go session affinity (#3344)", () => {
   const originalFetch = globalThis.fetch;
   afterEach(() => { globalThis.fetch = originalFetch; });
 
+  test("native Chat ingress preserves stable Go affinity and separates conversations", async () => {
+    const provider = opencodeGo();
+    const input = { nativeChat: true, model: "omen-alpha", provider };
+    const first = await captureRequest(input);
+    const continued = await captureRequest(input);
+    const sibling = await captureRequest({ ...input, child: "child-thread-b" });
+    expect(first.url).toBe("https://opencode.ai/zen/go/v1/chat/completions");
+    expect(first.headers.get(SESSION_HEADER)).toMatch(/^ocx_[0-9a-f]{32}$/);
+    expect(continued.headers.get(SESSION_HEADER)).toBe(first.headers.get(SESSION_HEADER));
+    expect(sibling.headers.get(SESSION_HEADER)).not.toBe(first.headers.get(SESSION_HEADER));
+    expect(provider.headers?.[SESSION_HEADER]).toBeUndefined();
+  });
+
+  test("native Chat honors configured session headers on renamed Go providers", async () => {
+    const captured = await captureRequest({
+      nativeChat: true, model: "omen-alpha", providerName: "renamed-go",
+      provider: opencodeGo({ headers: { "X-OpenCode-Session": "operator-session" } }),
+    });
+    expect(captured.headers.get(SESSION_HEADER)).toBe("operator-session");
+  });
+
+  test("uses a Pi session header without Codex headers on native and bridged Chat", async () => {
+    const headers = { "content-type": "application/json", "x-opencode-session": "pi-conversation-a" };
+    const chat = await captureRequest({ nativeChat: true, model: "omen-alpha", headers });
+    const bridged = await captureRequest({ nativeChat: true, model: MUSE_MODEL, headers });
+    expect(chat.headers.get(SESSION_HEADER)).toMatch(/^ocx_[0-9a-f]{32}$/);
+    expect(chat.headers.get(SESSION_HEADER)).not.toContain("pi-conversation-a");
+    expect(bridged.headers.get(SESSION_HEADER)).toBe(chat.headers.get(SESSION_HEADER));
+  });
+
+  test("native Chat does not send Go affinity to an unrelated destination", async () => {
+    const captured = await captureRequest({
+      nativeChat: true, model: "omen-alpha", providerName: "custom-go",
+      provider: opencodeGo({ baseUrl: "https://opencode.ai.evil.test/zen/go/v1" }),
+    });
+    expect(captured.headers.has(SESSION_HEADER)).toBe(false);
+  });
+
   test("sends one stable opaque session header on Responses and Chat wires", async () => {
     const responses = await captureRequest({ model: MUSE_MODEL });
     const chat = await captureRequest({ model: CHAT_MODEL });

From effba4c646408c8fd4b22936c820de37d370ce9e Mon Sep 17 00:00:00 2001
From: makesomethingshit <junsoo1172@gmail.com>
Date: Mon, 7 Sep 2026 16:11:52 +0900
Subject: [PATCH 2/7] test(opencode-go): define inbound opaque session identity
 contract [skip ci]

(cherry picked from commit 9f15a7c4139de294c5b0d7eddf50d885523f149e)

Co-authored-by: makesomethingshit <246213378+makesomethingshit@users.noreply.github.com>
(cherry picked from commit 65908008e52c5dca2eb1130d1c0cff0e0732873d)
---
 .../opencode-go-session-header.test.ts        | 26 +++++++++++++++++--
 1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/tests/providers/opencode-go-session-header.test.ts b/tests/providers/opencode-go-session-header.test.ts
index 9b476760cdd..b05e526e1c1 100644
--- a/tests/providers/opencode-go-session-header.test.ts
+++ b/tests/providers/opencode-go-session-header.test.ts
@@ -1,9 +1,10 @@
 import { afterEach, describe, expect, test } from "bun:test";
 import { providerConfigSeed } from "../../src/providers/derive";
-import { resolveOpenCodeGoTransport } from "../../src/providers/opencode-go-transport";
+import { deriveOpenCodeGoSessionId, resolveOpenCodeGoTransport } from "../../src/providers/opencode-go-transport";
 import { getProviderRegistryEntry } from "../../src/providers/registry";
 import { handleResponses } from "../../src/server/responses/core";
 import { handleChatCompletions } from "../../src/server/chat-completions";
+import { normalizeLogConversationId } from "../../src/server/request-log-conversation";
 import type { OcxConfig, OcxProviderConfig } from "../../src/types";
 
 const MUSE_MODEL = "muse-spark-1.3-contributor";
@@ -79,7 +80,7 @@ async function captureRequest(input: {
   ) : await handleResponses(
     new Request("http://localhost/v1/responses", {
       method: "POST",
-      headers: codexHeaders(input.child),
+      headers: input.headers ?? codexHeaders(input.child),
       body: JSON.stringify({ model: `${providerName}/${model}`, input: "ping", stream: false }),
     }),
     config,
@@ -127,6 +128,27 @@ describe("OpenCode Go session affinity (#3344)", () => {
     expect(bridged.headers.get(SESSION_HEADER)).toBe(chat.headers.get(SESSION_HEADER));
   });
 
+  for (const session of ["client-session-a", "ocx_0123456789abcdef0123456789abcdef"]) {
+    test(`treats inbound ${session.startsWith("ocx_") ? "ocx-prefixed" : "raw"} identity as client input on every ingress`, async () => {
+      const headers = { "content-type": "application/json", [SESSION_HEADER]: session };
+      const expected = deriveOpenCodeGoSessionId(normalizeLogConversationId(session)!);
+      const native = await captureRequest({ nativeChat: true, model: "omen-alpha", headers });
+      const bridged = await captureRequest({ nativeChat: true, model: MUSE_MODEL, headers });
+      const responses = await captureRequest({ model: MUSE_MODEL, headers });
+      expect(native.url).toEndWith("/chat/completions");
+      expect(bridged.url).toEndWith("/responses");
+      for (const request of [native, bridged, responses]) {
+        expect(request.headers.get(SESSION_HEADER)).toBe(expected);
+        expect(request.headers.get(SESSION_HEADER)).not.toBe(session);
+      }
+      const override = await captureRequest({
+        nativeChat: true, model: "omen-alpha", headers,
+        provider: opencodeGo({ headers: { "X-OpenCode-Session": session } }),
+      });
+      expect(override.headers.get(SESSION_HEADER)).toBe(session);
+    });
+  }
+
   test("native Chat does not send Go affinity to an unrelated destination", async () => {
     const captured = await captureRequest({
       nativeChat: true, model: "omen-alpha", providerName: "custom-go",

From eb8cc9ab735a3c2396cb5393bbf62dc473d5716e Mon Sep 17 00:00:00 2001
From: makesomethingshit <junsoo1172@gmail.com>
Date: Mon, 7 Sep 2026 16:24:45 +0900
Subject: [PATCH 3/7] fix(pi): scope generated session affinity to the Pi
 client [skip ci]

(cherry picked from commit 23d869350e9a90359f65624e41ae8e93797ae398)

Co-authored-by: makesomethingshit <246213378+makesomethingshit@users.noreply.github.com>
(cherry picked from commit 1e65ffd4bc9ba8befdf6e3546643b428106f4650)
---
 src/clients/config-export.ts              |  8 ++++----
 tests/clients/prime-client.test.ts        | 20 ++++++++------------
 tests/config/client-config-export.test.ts |  2 ++
 3 files changed, 14 insertions(+), 16 deletions(-)

diff --git a/src/clients/config-export.ts b/src/clients/config-export.ts
index e1744c2209f..6a94b74d709 100644
--- a/src/clients/config-export.ts
+++ b/src/clients/config-export.ts
@@ -817,7 +817,7 @@ export interface GajaeGeneratedConfig {
  * model. The rest of this contract (omitting `cost`) is still ours rather than
  * a claim about Pi's acceptance.
  */
-function buildPiClientConfig(ctx: ExportContext): PiGeneratedConfig {
+function buildPiClientConfig(ctx: ExportContext, sendSessionAffinityHeaders = false): PiGeneratedConfig {
   const models: PiModelEntry[] = [];
   for (const model of normalizeExportModels(ctx.models)) {
     // Text is the one modality every routed model supports; anything richer must come
@@ -860,7 +860,7 @@ function buildPiClientConfig(ctx: ExportContext): PiGeneratedConfig {
         baseUrl: ctx.baseUrl,
         api: PI_API_DIALECT,
         apiKey: LOOPBACK_API_KEY_PLACEHOLDER,
-        compat: { sendSessionAffinityHeaders: true },
+        ...(sendSessionAffinityHeaders ? { compat: { sendSessionAffinityHeaders: true } } : {}),
         models,
       },
     },
@@ -1033,7 +1033,7 @@ function buildOpencodeContribution(ctx: ExportContext): ManagedContribution {
 }
 
 function buildPiContribution(ctx: ExportContext): ManagedContribution {
-  const doc = buildPiClientConfig(ctx);
+  const doc = buildPiClientConfig(ctx, true);
   return singleFragment("pi", ["providers", OPENCODE_PROVIDER_ID], doc.providers[OPENCODE_PROVIDER_ID]);
 }
 
@@ -1129,7 +1129,7 @@ export const EXPORT_CLIENTS: Record<ExportClientId, ExportClientSpec> = {
     destination: env => piConfigPath(env),
     apiKeyEnv: "",
     exportHint: "Pi reads a non-secret placeholder from models.json; loopback needs no key.",
-    build: buildPiClientConfig,
+    build: ctx => buildPiClientConfig(ctx, true),
     format: "json",
     summarize: summarizePi,
     buildContribution: buildPiContribution,
diff --git a/tests/clients/prime-client.test.ts b/tests/clients/prime-client.test.ts
index c88c77508af..6c7c0a2f76e 100644
--- a/tests/clients/prime-client.test.ts
+++ b/tests/clients/prime-client.test.ts
@@ -37,18 +37,14 @@ function context(): ExportContext {
 }
 
 describe("Prime Agent client config", () => {
-  /**
-   * The load-bearing claim of this client: Prime Agent is the pi coding agent
-   * under a different brand, so it reads the SAME models.json contract rather
-   * than a lookalike. Locking the two documents together is what keeps that
-   * claim true — if a future Pi-only change diverges, this fails here instead
-   * of silently shipping Prime users a config their agent rejects.
-   */
-  test("generates byte-for-byte the document Pi generates", () => {
-    const prime = buildClientConfigText("prime", context());
-    const pi = buildClientConfigText("pi", context());
-    expect(prime.format).toBe("json");
-    expect(prime.text).toBe(pi.text);
+  test("shares Pi's model contract without opting Prime into session headers", () => {
+    const prime = buildClientConfig("prime", context()) as PiGeneratedConfig;
+    const pi = buildClientConfig("pi", context()) as PiGeneratedConfig;
+    expect(pi.providers[OPENCODE_PROVIDER_ID]!.compat).toEqual({ sendSessionAffinityHeaders: true });
+    delete pi.providers[OPENCODE_PROVIDER_ID]!.compat;
+    expect(prime).toEqual(pi);
+    expect(buildClientContribution("prime", context()).fragments[0]!.value)
+      .toEqual(prime.providers[OPENCODE_PROVIDER_ID]);
   });
 
   test("adds only providers.opencodex, wired to the loopback proxy", () => {
diff --git a/tests/config/client-config-export.test.ts b/tests/config/client-config-export.test.ts
index 4d707904033..c5a5840b824 100644
--- a/tests/config/client-config-export.test.ts
+++ b/tests/config/client-config-export.test.ts
@@ -11,6 +11,7 @@ import {
   LOOPBACK_API_KEY_PLACEHOLDER,
   SCHEMA_REQUIRED_OUTPUT_BUDGET,
   buildClientConfig,
+  buildClientContribution,
   buildClientConfigText,
   isExportClientId,
   normalizeExportModels,
@@ -316,6 +317,7 @@ describe("Pi serializer (accept criterion 2)", () => {
     expect(provider.api).toBe("openai-completions");
     expect(provider.apiKey).toBe(LOOPBACK_API_KEY_PLACEHOLDER);
     expect(provider.compat?.sendSessionAffinityHeaders).toBe(true);
+    expect(buildClientContribution("pi", ctx()).fragments[0]!.value).toEqual(provider);
   });
 
   test("cost is omitted on every entry — zeros would assert routed models are free", () => {

From 697a500339f323157ff20370e6be6bb65803e318 Mon Sep 17 00:00:00 2001
From: lidge-jun <243035832+lidge-jun@users.noreply.github.com>
Date: Mon, 7 Sep 2026 18:29:47 +0900
Subject: [PATCH 4/7] test: pin Go affinity identity and header precedence
 [skip ci]

Co-authored-by: makesomethingshit <246213378+makesomethingshit@users.noreply.github.com>
---
 .../opencode-go-session-header.test.ts        | 27 ++++++++++++++++---
 1 file changed, 23 insertions(+), 4 deletions(-)

diff --git a/tests/providers/opencode-go-session-header.test.ts b/tests/providers/opencode-go-session-header.test.ts
index b05e526e1c1..ab28c8475f5 100644
--- a/tests/providers/opencode-go-session-header.test.ts
+++ b/tests/providers/opencode-go-session-header.test.ts
@@ -1,10 +1,9 @@
 import { afterEach, describe, expect, test } from "bun:test";
 import { providerConfigSeed } from "../../src/providers/derive";
-import { deriveOpenCodeGoSessionId, resolveOpenCodeGoTransport } from "../../src/providers/opencode-go-transport";
+import { resolveOpenCodeGoTransport } from "../../src/providers/opencode-go-transport";
 import { getProviderRegistryEntry } from "../../src/providers/registry";
 import { handleResponses } from "../../src/server/responses/core";
 import { handleChatCompletions } from "../../src/server/chat-completions";
-import { normalizeLogConversationId } from "../../src/server/request-log-conversation";
 import type { OcxConfig, OcxProviderConfig } from "../../src/types";
 
 const MUSE_MODEL = "muse-spark-1.3-contributor";
@@ -128,10 +127,13 @@ describe("OpenCode Go session affinity (#3344)", () => {
     expect(bridged.headers.get(SESSION_HEADER)).toBe(chat.headers.get(SESSION_HEADER));
   });
 
-  for (const session of ["client-session-a", "ocx_0123456789abcdef0123456789abcdef"]) {
+  // Fixed vectors independently calculated with SHA-256, including the domain separator.
+  for (const [session, expected] of [
+    ["client-session-a", "ocx_516d593899f34b7baca2db37c7b0c8c5"],
+    ["ocx_0123456789abcdef0123456789abcdef", "ocx_60bcbfb9a85d3dc23b9b2b1cef3b0882"],
+  ] as const) {
     test(`treats inbound ${session.startsWith("ocx_") ? "ocx-prefixed" : "raw"} identity as client input on every ingress`, async () => {
       const headers = { "content-type": "application/json", [SESSION_HEADER]: session };
-      const expected = deriveOpenCodeGoSessionId(normalizeLogConversationId(session)!);
       const native = await captureRequest({ nativeChat: true, model: "omen-alpha", headers });
       const bridged = await captureRequest({ nativeChat: true, model: MUSE_MODEL, headers });
       const responses = await captureRequest({ model: MUSE_MODEL, headers });
@@ -149,6 +151,23 @@ describe("OpenCode Go session affinity (#3344)", () => {
     });
   }
 
+  test("operator override precedes the Codex lane, which precedes client fallback on every ingress", async () => {
+    const headers = { ...codexHeaders(), [SESSION_HEADER]: "different-client-fallback" };
+    for (const ingress of [
+      { nativeChat: true, model: "omen-alpha" },
+      { nativeChat: true, model: MUSE_MODEL },
+      { model: MUSE_MODEL },
+    ]) {
+      const codex = await captureRequest({ ...ingress, headers });
+      expect(codex.headers.get(SESSION_HEADER)).toBe("ocx_67b70584fb755130286eff5488a3be9d");
+      const operator = await captureRequest({
+        ...ingress, headers,
+        provider: opencodeGo({ headers: { "X-OpenCode-Session": "different-operator-override" } }),
+      });
+      expect(operator.headers.get(SESSION_HEADER)).toBe("different-operator-override");
+    }
+  });
+
   test("native Chat does not send Go affinity to an unrelated destination", async () => {
     const captured = await captureRequest({
       nativeChat: true, model: "omen-alpha", providerName: "custom-go",

From feb1855df91e5db351699d959d88d7e6c21c0f93 Mon Sep 17 00:00:00 2001
From: lidge-jun <243035832+lidge-jun@users.noreply.github.com>
Date: Mon, 7 Sep 2026 18:59:23 +0900
Subject: [PATCH 5/7] docs: include Pi session affinity in every guide example
 [skip ci]

Co-authored-by: makesomethingshit <246213378+makesomethingshit@users.noreply.github.com>
---
 docs-site/src/content/docs/fr/guides/pi.md    | 5 +++++
 docs-site/src/content/docs/guides/pi.md       | 5 +++++
 docs-site/src/content/docs/ja/guides/pi.md    | 5 +++++
 docs-site/src/content/docs/ko/guides/pi.md    | 5 +++++
 docs-site/src/content/docs/ru/guides/pi.md    | 5 +++++
 docs-site/src/content/docs/tr/guides/pi.md    | 5 +++++
 docs-site/src/content/docs/zh-cn/guides/pi.md | 5 +++++
 docs-site/src/content/docs/zh-tw/guides/pi.md | 5 +++++
 8 files changed, 40 insertions(+)

diff --git a/docs-site/src/content/docs/fr/guides/pi.md b/docs-site/src/content/docs/fr/guides/pi.md
index eab89600636..030d91679c3 100644
--- a/docs-site/src/content/docs/fr/guides/pi.md
+++ b/docs-site/src/content/docs/fr/guides/pi.md
@@ -27,6 +27,9 @@ d’exportation de la variable d’environnement et le nombre de modèles dotés
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -41,6 +44,8 @@ d’exportation de la variable d’environnement et le nombre de modèles dotés
 }
 ```
 
+Les fournisseurs Pi générés activent `compat.sendSessionAffinityHeaders`. Conservez ce réglage lors de la fusion ou de la modification manuelle du fournisseur : Pi transmet un identifiant de session stable, dont OpenCodex dérive l’affinité pour la destination canonique OpenCode Go. Pi peut omettre cet identifiant lorsque `cacheRetention` vaut `none`.
+
 Les identifiants de modèle sont les sélecteurs canoniques du proxy : les modèles routés apparaissent donc sous la forme `provider/model`
 (`anthropic/claude-opus-5`) et les slugs natifs OpenAI restent sans préfixe (`gpt-5.6-sol`). Le `name`
 suffixe — `(anthropic)`, `(native)`, `(routed)` — permet de distinguer, dans le sélecteur de Pi, deux modèles de même nom
diff --git a/docs-site/src/content/docs/guides/pi.md b/docs-site/src/content/docs/guides/pi.md
index c44b97f12a6..f44e4be381d 100644
--- a/docs-site/src/content/docs/guides/pi.md
+++ b/docs-site/src/content/docs/guides/pi.md
@@ -27,6 +27,9 @@ export line, and how many models carry authoritative context limits.
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -41,6 +44,8 @@ export line, and how many models carry authoritative context limits.
 }
 ```
 
+Generated Pi providers enable `compat.sendSessionAffinityHeaders`. Keep this flag when merging or manually editing the provider: Pi supplies a stable session identity and OpenCodex derives canonical OpenCode Go affinity from it. Pi may omit the identity when `cacheRetention` is `none`.
+
 Model ids are the proxy's canonical selectors, so routed models appear as `provider/model`
 (`anthropic/claude-opus-5`) and native OpenAI slugs stay unprefixed (`gpt-5.6-sol`). The `name`
 suffix — `(anthropic)`, `(native)`, `(routed)` — is what makes two same-named models from
diff --git a/docs-site/src/content/docs/ja/guides/pi.md b/docs-site/src/content/docs/ja/guides/pi.md
index 788fe48c603..9b637e84e4b 100644
--- a/docs-site/src/content/docs/ja/guides/pi.md
+++ b/docs-site/src/content/docs/ja/guides/pi.md
@@ -23,6 +23,9 @@ ocx export --client pi
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -37,6 +40,8 @@ ocx export --client pi
 }
 ```
 
+生成される Pi プロバイダーでは `compat.sendSessionAffinityHeaders` が有効です。設定をマージしたり手動で編集したりする際も、このフラグを保持してください。Pi が送る安定したセッション識別子から、OpenCodex が正規の OpenCode Go 接続先用の affinity を生成します。`cacheRetention` が `none` の場合、Pi は識別子を送信しないことがあります。
+
 モデル ID はプロキシの正規セレクターであるため、ルーティングされたモデルは `provider/model` (`anthropic/claude-opus-5`) として表示され、ネイティブ OpenAI スラグはプレフィックスなし (`gpt-5.6-sol`) のままになります。 `name` サフィックス (`(anthropic)`、`(native)`、`(routed)`) により、異なるアップストリームの 2 つの同じ名前のモデルが Pi のピッカーで区別できるようになります。
 
 ## どこへ行くのか
diff --git a/docs-site/src/content/docs/ko/guides/pi.md b/docs-site/src/content/docs/ko/guides/pi.md
index 648d71060e0..6bda9c2b363 100644
--- a/docs-site/src/content/docs/ko/guides/pi.md
+++ b/docs-site/src/content/docs/ko/guides/pi.md
@@ -27,6 +27,9 @@ ocx export --client pi
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -41,6 +44,8 @@ ocx export --client pi
 }
 ```
 
+생성된 Pi provider에는 `compat.sendSessionAffinityHeaders`가 활성화됩니다. provider를 병합하거나 직접 수정할 때 이 설정을 유지하세요. Pi가 안정적인 세션 식별자를 보내면 OpenCodex가 이를 바탕으로 정규 OpenCode Go 대상의 affinity를 계산합니다. `cacheRetention`이 `none`이면 Pi가 식별자를 보내지 않을 수 있습니다.
+
 모델 id는 프록시의 정규 선택자이므로, 라우팅된 모델은 `provider/model`
 (`anthropic/claude-opus-5`) 형태로 나타나고, 네이티브 OpenAI slug는 접두사 없이
 (`gpt-5.6-sol`) 유지됩니다. `name` 접미사인 `(anthropic)`, `(native)`, `(routed)`는
diff --git a/docs-site/src/content/docs/ru/guides/pi.md b/docs-site/src/content/docs/ru/guides/pi.md
index 0960ecf49ae..e36a73da7e7 100644
--- a/docs-site/src/content/docs/ru/guides/pi.md
+++ b/docs-site/src/content/docs/ru/guides/pi.md
@@ -27,6 +27,9 @@ ocx export --client pi
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -41,6 +44,8 @@ ocx export --client pi
 }
 ```
 
+В создаваемой конфигурации Pi включён `compat.sendSessionAffinityHeaders`. Сохраняйте этот флаг при объединении или ручном редактировании провайдера: Pi передаёт стабильный идентификатор сессии, из которого OpenCodex формирует affinity для канонического OpenCode Go. При `cacheRetention: none` Pi может не передавать идентификатор.
+
 Id моделей — это канонические селекторы прокси, поэтому маршрутизируемые модели появляются как
 `provider/model` (`anthropic/claude-opus-5`), а нативные slug OpenAI остаются без префикса
 (`gpt-5.6-sol`). Суффикс в `name` — `(anthropic)`, `(native)`, `(routed)` — как раз и позволяет
diff --git a/docs-site/src/content/docs/tr/guides/pi.md b/docs-site/src/content/docs/tr/guides/pi.md
index 0741f7be518..fe6044de281 100644
--- a/docs-site/src/content/docs/tr/guides/pi.md
+++ b/docs-site/src/content/docs/tr/guides/pi.md
@@ -31,6 +31,9 @@ export line, and how many models carry authoritative context limits.
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -45,6 +48,8 @@ export line, and how many models carry authoritative context limits.
 }
 ```
 
+Oluşturulan Pi sağlayıcılarında `compat.sendSessionAffinityHeaders` etkinleştirilir. Sağlayıcıyı birleştirirken veya elle düzenlerken bu ayarı koruyun: Pi sabit bir oturum kimliği gönderir ve OpenCodex bu kimlikten kanonik OpenCode Go hedefi için oturum yakınlığı üretir. `cacheRetention` değeri `none` olduğunda Pi kimliği göndermeyebilir.
+
 Model ids are the proxy's canonical selectors, so routed models appear as
 `provider/model`
 (`anthropic/claude-opus-5`) and native OpenAI slugs stay unprefixed
diff --git a/docs-site/src/content/docs/zh-cn/guides/pi.md b/docs-site/src/content/docs/zh-cn/guides/pi.md
index ad868e31946..c9ebf7b4a62 100644
--- a/docs-site/src/content/docs/zh-cn/guides/pi.md
+++ b/docs-site/src/content/docs/zh-cn/guides/pi.md
@@ -23,6 +23,9 @@ ocx export --client pi
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -37,6 +40,8 @@ ocx export --client pi
 }
 ```
 
+生成的 Pi 提供方配置启用了 `compat.sendSessionAffinityHeaders`。合并或手动编辑提供方时请保留该设置：Pi 提供稳定的会话标识，OpenCodex 据此为规范的 OpenCode Go 目标生成会话亲和标识。`cacheRetention` 为 `none` 时，Pi 可能不发送会话标识。
+
 模型 id 是代理的规范选择器，因此已路由模型会显示为 `provider/model`（`anthropic/claude-opus-5`），而原生 OpenAI slug 会保持不带前缀（`gpt-5.6-sol`）。`name` 后缀 - `(anthropic)`、`(native)`、`(routed)` - 负责让两个同名但来自不同上游的模型在 Pi 的选择器中可区分。
 
 ## 放置位置
diff --git a/docs-site/src/content/docs/zh-tw/guides/pi.md b/docs-site/src/content/docs/zh-tw/guides/pi.md
index 0353338574c..d8e9b625107 100644
--- a/docs-site/src/content/docs/zh-tw/guides/pi.md
+++ b/docs-site/src/content/docs/zh-tw/guides/pi.md
@@ -23,6 +23,9 @@ ocx export --client pi
       "baseUrl": "http://127.0.0.1:10100/v1",
       "api": "openai-completions",
       "apiKey": "$OPENCODEX_API_KEY",
+      "compat": {
+        "sendSessionAffinityHeaders": true
+      },
       "models": [
         {
           "id": "anthropic/claude-opus-5",
@@ -37,6 +40,8 @@ ocx export --client pi
 }
 ```
 
+產生的 Pi 供應商設定會啟用 `compat.sendSessionAffinityHeaders`。合併或手動編輯供應商時請保留此設定：Pi 提供穩定的工作階段識別碼，OpenCodex 據此為標準 OpenCode Go 目標產生工作階段親和識別碼。當 `cacheRetention` 為 `none` 時，Pi 可能不傳送識別碼。
+
 模型 id 是代理的規範選擇器，因此路由模型顯示為 `provider/model`（`anthropic/claude-opus-5`），而原生 OpenAI slug 保持無前綴（`gpt-5.6-sol`）。`name` 後綴 — `(anthropic)`、`(native)`、`(routed)` — 正是讓來自不同上游的兩個同名模型在 Pi 的 picker 中可區分的關鍵。
 
 ## 放置位置

From e0e8d26bb86927fac378d88807f191843a9cf339 Mon Sep 17 00:00:00 2001
From: lidge-jun <243035832+lidge-jun@users.noreply.github.com>
Date: Mon, 7 Sep 2026 18:31:24 +0900
Subject: [PATCH 6/7] fix: fall back from native compact 404 to routed
 compaction [skip ci]

Reimplements only the residual compaction portion of #3769; quota attribution remains the implementation already landed through #3791.

Co-authored-by: ideabib <221318067+ideabib@users.noreply.github.com>
---
 .../content/docs/reference/proxy-formats.md   |   7 +
 src/server/responses/compact.ts               |   5 +-
 .../responses-compaction-routing.test.ts      | 141 ++++++++++++++++++
 3 files changed, 151 insertions(+), 2 deletions(-)

diff --git a/docs-site/src/content/docs/reference/proxy-formats.md b/docs-site/src/content/docs/reference/proxy-formats.md
index 0c7b8257bc5..593c8ae7a32 100644
--- a/docs-site/src/content/docs/reference/proxy-formats.md
+++ b/docs-site/src/content/docs/reference/proxy-formats.md
@@ -405,6 +405,13 @@ conversation.
 | Canonical ChatGPT or official OpenAI route | Forwards the request to the native `/responses/compact` endpoint with the resolved account and model authentication |
 | Other routed model | Runs an internal, non-streaming, no-tools compaction turn with a `compaction_trigger`; requires exactly one synthetic `compaction` item whose `encrypted_content` is an `ocx1:` envelope; decodes that summary into v1 replacement history |
 
+If the native compact endpoint returns HTTP 404, OpenCodex retries compaction through a regular
+Responses turn with the same model selector and session headers. Canonical ChatGPT fallback
+turns use upstream SSE; the compact caller still receives JSON. A completed native opaque
+compaction item is preserved, while an `ocx1:` summary is decoded into replacement user history.
+Failed or incomplete fallback turns return an error instead of replacement history. Other
+native compact statuses retain their existing handling.
+
 Codex names a bare OpenAI-family model (for example `gpt-5.6-sol`) for its compaction turns
 regardless of which provider the operator routes ordinary turns to. Ordinary requests reserve
 such ids for the canonical `openai` provider. On the compaction surface only — `POST
diff --git a/src/server/responses/compact.ts b/src/server/responses/compact.ts
index c914dcb4b2b..4e7481bb2e2 100644
--- a/src/server/responses/compact.ts
+++ b/src/server/responses/compact.ts
@@ -1094,7 +1094,8 @@ export async function handleResponsesCompact(
         }
       }
     }
-    return buffered;
+    // A native compact 404 falls back to a regular Responses compaction turn.
+    if (buffered.status !== 404) return buffered;
     } finally {
       releaseUpstreamHostAdmission(compactHostAdmissionLease);
       releaseCodexAuthContextProbeLease(authCtx);
@@ -1111,7 +1112,7 @@ export async function handleResponsesCompact(
     // the completed event back into the v1 compact JSON contract below. Combo-dispatched
     // turns also go out as SSE: failover can land on a canonical child that rejects a
     // non-streaming turn, and every combo-capable provider already serves streaming traffic.
-    stream: accountGatedCompactWireModel || route.combo ? true : false,
+    stream: isCanonicalOpenAiForwardProvider(route.provider) || accountGatedCompactWireModel || route.combo ? true : false,
     input: [...inputItems, { type: "compaction_trigger" }],
   };
   const internalHeaders = new Headers({ "content-type": "application/json" });
diff --git a/tests/responses/responses-compaction-routing.test.ts b/tests/responses/responses-compaction-routing.test.ts
index a787024d95f..2a1f69be1ed 100644
--- a/tests/responses/responses-compaction-routing.test.ts
+++ b/tests/responses/responses-compaction-routing.test.ts
@@ -9,6 +9,7 @@ import { mkdtempSync, writeFileSync } from "node:fs";
 import { tmpdir } from "node:os";
 import { join } from "node:path";
 import { handleResponses, handleResponsesCompact } from "../../src/server/responses";
+import { OPAQUE_COMPACTION_NOTE, SUMMARY_PREFIX } from "../../src/responses/compaction";
 import { looksLikeBackendCiphertext } from "../../src/server/responses/encrypted-payload";
 import * as adapterResolveModule from "../../src/server/adapter-resolve";
 import * as visionModule from "../../src/vision";
@@ -948,6 +949,146 @@ describe("compact alternate-account attempt (#913)", () => {
     });
   }
 
+  for (const [model, account] of [["gpt-5.5", "pool-a"], ["side/gpt-5.5", "pool-b"]] as const) {
+    test(`native 404 falls back to canonical SSE with ${model} account and session identity`, async () => {
+      await withPoolEnv("ocx-compact-404-canonical-", async config => {
+        config.codexAccountNamespaces = { side: "pool-b" };
+        const item = { type: "compaction", id: "cmp_native_3769", encrypted_content: "native-opaque-3769" };
+        const calls: Array<{ url: string; headers: Headers; body: Record<string, unknown> }> = [];
+        globalThis.fetch = (async (input: string | URL | Request, init?: RequestInit) => {
+          const request = new Request(input, init);
+          calls.push({ url: request.url, headers: request.headers, body: await request.json() as Record<string, unknown> });
+          if (request.url.endsWith("/responses/compact")) return Response.json({ detail: "Not Found" }, { status: 404 });
+          return sseResponse([{ type: "response.completed", response: {
+            id: "resp_compact_3769", status: "completed", output: [item],
+          } }]);
+        }) as typeof fetch;
+        const headers = { "session-id": "compact-3769-session", "thread-id": `compact-3769-${account}`, "x-codex-parent-thread-id": "compact-3769-parent" };
+        const response = await handleResponsesCompact(compactionRequest({
+          model, input: [{ role: "user", content: "retain this history" }],
+        }, undefined, headers), config, { model: "", provider: "" });
+        expect(response.status).toBe(200);
+        expect(response.headers.get("content-type")).toContain("application/json");
+        expect(await response.json()).toEqual({ output: [item] });
+        expect(calls.map(call => call.url)).toEqual([
+          "https://chatgpt.com/backend-api/codex/responses/compact",
+          "https://chatgpt.com/backend-api/codex/responses",
+        ]);
+        expect(calls[1]!.body.stream).toBe(true);
+        expect(calls[1]!.body.model).toBe("gpt-5.5");
+        expect((calls[1]!.body.input as Array<{ type?: string }>).filter(value => value.type === "compaction_trigger")).toHaveLength(1);
+        for (const call of calls) {
+          expect(call.headers.get("authorization")).toBe(`Bearer ${account}-access-token`);
+          expect(call.headers.get("chatgpt-account-id")).toBe(account === "pool-a" ? "pool_acc_a" : "pool_acc_b");
+          for (const [name, value] of Object.entries(headers)) expect(call.headers.get(name)).toBe(value);
+        }
+      });
+    });
+  }
+
+  test("official key-auth native 404 decodes synthetic fallback into replacement user history", async () => {
+    const config = { providers: { "openai-apikey": {
+      adapter: "openai-responses", baseUrl: "https://api.openai.com/v1", authMode: "key", apiKey: "test-key",
+    } } } as OcxConfig;
+    const calls: Array<{ url: string; body: Record<string, unknown> }> = [];
+    globalThis.fetch = (async (input: string | URL | Request, init?: RequestInit) => {
+      const request = new Request(input, init);
+      calls.push({ url: request.url, body: await request.json() as Record<string, unknown> });
+      return request.url.endsWith("/responses/compact")
+        ? Response.json({ detail: "Not Found" }, { status: 404 })
+        : jsonResponse(completedPayload("handoff-3769"));
+    }) as typeof fetch;
+    const response = await handleResponsesCompact(compactionRequest({
+      model: "openai-apikey/gpt-5.5", input: [{ role: "user", content: "retain-3769" }],
+      tools: [{ type: "function", name: "shell", parameters: { type: "object" } }],
+    }), config, { model: "", provider: "" });
+    expect(response.status).toBe(200);
+    expect(await response.json()).toEqual({ output: [
+      { type: "message", role: "user", content: [{ type: "input_text", text: "retain-3769" }] },
+      { type: "message", role: "user", content: [{ type: "input_text", text: `${SUMMARY_PREFIX}\nhandoff-3769` }] },
+    ] });
+    expect(calls.map(call => call.url)).toEqual(["https://api.openai.com/v1/responses/compact", "https://api.openai.com/v1/responses"]);
+    expect(calls[1]!.body.tools).toBeUndefined();
+    expect(JSON.stringify(calls[1]!.body.input)).not.toContain("compaction_trigger");
+    expect(JSON.stringify(calls[1]!.body.input)).toContain("CONTEXT CHECKPOINT COMPACTION");
+  });
+
+  for (const status of [200, 400]) {
+    test(`native compact ${status} retains its body without the 404 fallback`, async () => {
+      await withPoolEnv("ocx-compact-404-control-", async config => {
+        const payload = status === 200 ? { output: [{ type: "compaction", encrypted_content: "native-control" }] } : { error: { message: "invalid compact" } };
+        const urls: string[] = [];
+        globalThis.fetch = (async (input: string | URL | Request) => {
+          urls.push(typeof input === "string" ? input : input instanceof URL ? input.toString() : input.url);
+          return Response.json(payload, { status });
+        }) as typeof fetch;
+        const response = await handleResponsesCompact(compactionRequest({ model: "gpt-5.5", input: [] }), config, { model: "", provider: "" });
+        expect(response.status).toBe(status);
+        expect(await response.json()).toEqual(payload);
+        expect(urls).toEqual(["https://chatgpt.com/backend-api/codex/responses/compact"]);
+      });
+    });
+  }
+
+  for (const status of ["failed", "incomplete"] as const) {
+    test(`native 404 followed by ${status} SSE does not install replacement history`, async () => {
+      await withPoolEnv("ocx-compact-404-terminal-", async config => {
+        let calls = 0;
+        globalThis.fetch = (async () => {
+          calls++;
+          if (calls === 1) return Response.json({ detail: "Not Found" }, { status: 404 });
+          return sseResponse([{ type: `response.${status}`, response: {
+            id: "resp_compact_rejected_3769", status, output: [],
+          } }]);
+        }) as typeof fetch;
+        const response = await handleResponsesCompact(compactionRequest({ model: "gpt-5.5", input: [] }), config, { model: "", provider: "" });
+        expect(response.status).toBe(502);
+        const payload = await response.json() as { output?: unknown; error?: unknown };
+        expect(payload.output).toBeUndefined();
+        expect(payload.error).toBeDefined();
+        expect(calls).toBe(2);
+      });
+    });
+  }
+
+  test("404 fallback records the compaction serving account for subsequent opaque replay", async () => {
+    await withPoolEnv("ocx-compact-404-replay-", async config => {
+      config.codexAccountNamespaces = { side: "pool-b", first: "pool-a" };
+      const headers = { "thread-id": `compact-replay-${crypto.randomUUID()}` };
+      const item = { type: "compaction", encrypted_content: "native-account-b-3769" };
+      const calls: Array<{ body: Record<string, unknown>; headers: Headers }> = [];
+      let compacting = false;
+      globalThis.fetch = (async (input: string | URL | Request, init?: RequestInit) => {
+        const request = new Request(input, init);
+        if (request.url.endsWith("/responses/compact")) return Response.json({ detail: "Not Found" }, { status: 404 });
+        calls.push({ body: await request.json() as Record<string, unknown>, headers: request.headers });
+        return sseResponse([{ type: "response.completed", response: compacting
+          ? { id: "resp_identity_compact_3769", status: "completed", output: [item] }
+          : completedPayload("ordinary turn") }]);
+      }) as typeof fetch;
+      const turn = async (model: string, input: unknown[]) => {
+        const response = await handleResponses(compactionRequest({ model, input, stream: true, store: false }, undefined, headers), config, { model: "", provider: "" });
+        expect(response.status).toBe(200);
+        await response.text();
+      };
+      await turn("first/gpt-5.5", [{ role: "user", content: "seed account A" }]);
+      compacting = true;
+      const compact = await handleResponsesCompact(compactionRequest({ model: "side/gpt-5.5", input: [{ role: "user", content: "compact on B" }] }, undefined, headers), config, { model: "", provider: "" });
+      expect(compact.status).toBe(200);
+      const output = (await compact.json() as { output: unknown[] }).output;
+      expect(output).toEqual([item]);
+      compacting = false;
+      await turn("side/gpt-5.5", [...output, { role: "user", content: "continue on B" }]);
+      expect(calls.at(-1)!.headers.get("authorization")).toBe("Bearer pool-b-access-token");
+      expect(JSON.stringify(calls.at(-1)!.body.input)).toContain("native-account-b-3769");
+      await turn("first/gpt-5.5", [...output, { role: "user", content: "switch back to A" }]);
+      expect(calls.at(-1)!.headers.get("authorization")).toBe("Bearer pool-a-access-token");
+      expect(JSON.stringify(calls.at(-1)!.body.input)).not.toContain("native-account-b-3769");
+      expect(JSON.stringify(calls.at(-1)!.body.input)).toContain(OPAQUE_COMPACTION_NOTE);
+      expect(calls).toHaveLength(4);
+    });
+  });
+
   test("native compact headers followed by a stalled body return 504 without retry and release account cleanup", async () => {
     await withPoolEnv("ocx-compact-body-deadline-", async config => {
       config.stallTimeoutSec = 2;

From c84cc3b3e076b6c49cf93a6ca267f6303a698f5b Mon Sep 17 00:00:00 2001
From: lidge-jun <243035832+lidge-jun@users.noreply.github.com>
Date: Mon, 7 Sep 2026 18:33:18 +0900
Subject: [PATCH 7/7] docs: describe native compact 404 fallback in route table
 [skip ci]

Co-authored-by: ideabib <221318067+ideabib@users.noreply.github.com>
---
 docs-site/src/content/docs/reference/proxy-formats.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs-site/src/content/docs/reference/proxy-formats.md b/docs-site/src/content/docs/reference/proxy-formats.md
index 593c8ae7a32..1f2e589252a 100644
--- a/docs-site/src/content/docs/reference/proxy-formats.md
+++ b/docs-site/src/content/docs/reference/proxy-formats.md
@@ -402,7 +402,7 @@ conversation.
 
 | Route type | Behavior |
 | --- | --- |
-| Canonical ChatGPT or official OpenAI route | Forwards the request to the native `/responses/compact` endpoint with the resolved account and model authentication |
+| Canonical ChatGPT or official OpenAI route | Tries the native `/responses/compact` endpoint with the resolved account and model authentication; HTTP 404 falls back to a regular Responses compaction turn |
 | Other routed model | Runs an internal, non-streaming, no-tools compaction turn with a `compaction_trigger`; requires exactly one synthetic `compaction` item whose `encrypted_content` is an `ocx1:` envelope; decodes that summary into v1 replacement history |
 
 If the native compact endpoint returns HTTP 404, OpenCodex retries compaction through a regular