From 15258921ee1c55323f962becb40c9ae3f0b28cc5 Mon Sep 17 00:00:00 2001
From: zhulijin1991 <167979819+zhulijin1991@users.noreply.github.com>
Date: Sun, 12 Apr 2026 08:16:03 +0800
Subject: [PATCH] fix(codex): avoid re-exposing image tool on vision turns

---
 .../codex/src/app-server/run-attempt.ts       | 30 +++++++++++++++----
 .../run-attempt.vision-tools.test.ts          | 20 +++++++++++++
 2 files changed, 45 insertions(+), 5 deletions(-)
 create mode 100644 extensions/codex/src/app-server/run-attempt.vision-tools.test.ts

diff --git a/extensions/codex/src/app-server/run-attempt.ts b/extensions/codex/src/app-server/run-attempt.ts
index f8d7238523d..67eb23f294b 100644
--- a/extensions/codex/src/app-server/run-attempt.ts
+++ b/extensions/codex/src/app-server/run-attempt.ts
@@ -365,10 +365,14 @@ async function buildDynamicTools(input: DynamicToolBuildParams) {
       input.runAbortController.abort("sessions_yield");
     },
   });
+  const visionFilteredTools = filterToolsForVisionInputs(allTools, {
+    modelHasVision,
+    hasInboundImages: (params.images?.length ?? 0) > 0,
+  });
   const filteredTools =
     params.toolsAllow && params.toolsAllow.length > 0
-      ? allTools.filter((tool) => params.toolsAllow?.includes(tool.name))
-      : allTools;
+      ? visionFilteredTools.filter((tool) => params.toolsAllow?.includes(tool.name))
+      : visionFilteredTools;
   return normalizeProviderToolSchemas({
     tools: filteredTools,
     provider: params.provider,
@@ -381,6 +385,19 @@ async function buildDynamicTools(input: DynamicToolBuildParams) {
   });
 }
 
+function filterToolsForVisionInputs<T extends { name?: string }>(
+  tools: T[],
+  params: {
+    modelHasVision: boolean;
+    hasInboundImages: boolean;
+  },
+): T[] {
+  if (!params.modelHasVision || !params.hasInboundImages) {
+    return tools;
+  }
+  return tools.filter((tool) => tool.name !== "image");
+}
+
 async function withCodexStartupTimeout<T>(params: {
   timeoutMs: number;
   timeoutFloorMs?: number;
@@ -495,6 +512,9 @@ function handleApprovalRequest(params: {
   });
 }
 
-export const __testing = createCodexAppServerClientFactoryTestHooks((factory) => {
-  clientFactory = factory;
-});
+export const __testing = {
+  filterToolsForVisionInputs,
+  ...createCodexAppServerClientFactoryTestHooks((factory) => {
+    clientFactory = factory;
+  }),
+} as const;
diff --git a/extensions/codex/src/app-server/run-attempt.vision-tools.test.ts b/extensions/codex/src/app-server/run-attempt.vision-tools.test.ts
new file mode 100644
index 00000000000..8c3e9fc5eb9
--- /dev/null
+++ b/extensions/codex/src/app-server/run-attempt.vision-tools.test.ts
@@ -0,0 +1,20 @@
+import { describe, expect, it } from "vitest";
+import { __testing } from "./run-attempt.js";
+
+describe("Codex dynamic tool filtering", () => {
+  it("drops the image tool when the model already has inbound vision input", () => {
+    const toolNames = __testing
+      .filterToolsForVisionInputs(
+        [{ name: "image" }, { name: "read" }, { name: "write" }],
+        {
+          modelHasVision: true,
+          hasInboundImages: true,
+        },
+      )
+      .map((tool) => tool.name);
+
+    expect(toolNames).toContain("read");
+    expect(toolNames).toContain("write");
+    expect(toolNames).not.toContain("image");
+  });
+});