diff --git a/test/scripts/e2e-mock-config-limits.test.ts b/test/scripts/e2e-mock-config-limits.test.ts
index a994cc19e044..8fd2bcce62ea 100644
--- a/test/scripts/e2e-mock-config-limits.test.ts
+++ b/test/scripts/e2e-mock-config-limits.test.ts
@@ -173,7 +173,7 @@ async function withMockServer(
     },
   ) => Promise<void>,
 ) {
-  const port = env.MOCK_PORT === undefined ? await getFreePort() : Number(env.MOCK_PORT);
+  const port = Number(env.MOCK_PORT ?? "0");
   let stderr = "";
   let stdout = "";
   const child = spawn(process.execPath, [scriptPath], {
@@ -1148,8 +1148,9 @@ describe("mock OpenAI response markers", () => {
 });
 
 describe("e2e mock and config helper numeric limits", () => {
-  it("reports the actual OS-assigned port for MOCK_PORT=0", async () => {
-    await withMockServer(mockOpenAiPath, { MOCK_PORT: "0" }, async (baseUrl, output) => {
+  it.each([undefined, "0"])("reports the bound port for MOCK_PORT=%s", async (port) => {
+    const env: Record<string, string> = port === undefined ? {} : { MOCK_PORT: port };
+    await withMockServer(mockOpenAiPath, env, async (baseUrl, output) => {
       expect(Number(new URL(baseUrl).port)).toBeGreaterThan(0);
       expect(output.stdout()).not.toContain("mock-openai listening on 0\n");
       const response = await fetch(`${baseUrl}/v1/responses`, {
@@ -1274,6 +1275,7 @@ describe("e2e mock and config helper numeric limits", () => {
       await withMockServer(
         webSearchMockPath,
         {
+          MOCK_PORT: String(await getFreePort()),
           MOCK_REQUEST_LOG: requestLogDirectory,
           RAW_SCHEMA_ERROR: "400 schema rejected",
           SUCCESS_MARKER: "OPENCLAW_SCHEMA_E2E_OK",