From 8813b79990f44bf038655fddd1e5dcc59d5ae53c Mon Sep 17 00:00:00 2001
From: Vincent Koc <vincentkoc@ieee.org>
Date: Thu, 14 May 2026 19:43:34 +0800
Subject: [PATCH] fix(openai): wrap malformed embedding batch jsonl

---
 CHANGELOG.md                              |  1 +
 extensions/openai/embedding-batch.test.ts | 10 ++++++++++
 extensions/openai/embedding-batch.ts      | 10 ++++++++--
 3 files changed, 19 insertions(+), 2 deletions(-)
 create mode 100644 extensions/openai/embedding-batch.test.ts

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 41081c7ae18..d11b69bcdcf 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -62,6 +62,7 @@ Docs: https://docs.openclaw.ai
 - Node host: report malformed built-in invoke `paramsJSON` with stable invalid-request errors instead of leaking raw parser failures.
 - Amazon Bedrock embeddings: report malformed provider response JSON with provider-owned errors instead of leaking raw parser failures.
 - QQBot: report malformed access-token JSON with provider-owned errors instead of leaking raw parser failures.
+- OpenAI embeddings: report malformed batch output JSONL with provider-owned errors instead of leaking raw parser failures.
 - Models config/auth: stop inferring provider env-var markers from broad `^[A-Z_][A-Z0-9_]*$` strings, and resolve config-backed provider `apiKey` values only through structured env SecretRefs (`secrets.providers[id]` / `secrets.defaults`), so unrelated env vars cannot accidentally become provider credentials. Thanks @sallyom.
 - Media fetch: skip allocating and buffering the response body for bodyless media responses (HEAD probes and 204-style empty bodies), avoiding wasted heap on streams that carry no payload. Thanks @shakkernerd.
 - CLI/onboarding: forward provider-specific auth flags (e.g. `--openai-api-key`) through the onboarding wizard so they reach provider auth methods via `ctx.opts`, letting `--openai-api-key "$OPENAI_API_KEY"` skip the redundant "use existing env var?" prompt in non-interactive harnesses. (#81669) Thanks @sjf.
diff --git a/extensions/openai/embedding-batch.test.ts b/extensions/openai/embedding-batch.test.ts
new file mode 100644
index 00000000000..7dacf1310ae
--- /dev/null
+++ b/extensions/openai/embedding-batch.test.ts
@@ -0,0 +1,10 @@
+import { describe, expect, it } from "vitest";
+import { parseOpenAiBatchOutput } from "./embedding-batch.js";
+
+describe("OpenAI embedding batch output", () => {
+  it("wraps malformed JSONL output", () => {
+    expect(() => parseOpenAiBatchOutput('{"custom_id":"ok"}\n{not json')).toThrow(
+      "OpenAI embedding batch output contained malformed JSONL",
+    );
+  });
+});
diff --git a/extensions/openai/embedding-batch.ts b/extensions/openai/embedding-batch.ts
index 0b1366c1019..ab119aaa0bd 100644
--- a/extensions/openai/embedding-batch.ts
+++ b/extensions/openai/embedding-batch.ts
@@ -122,7 +122,7 @@ async function fetchOpenAiBatchResource<T>(params: {
   });
 }
 
-function parseOpenAiBatchOutput(text: string): OpenAiBatchOutputLine[] {
+export function parseOpenAiBatchOutput(text: string): OpenAiBatchOutputLine[] {
   if (!text.trim()) {
     return [];
   }
@@ -130,7 +130,13 @@ function parseOpenAiBatchOutput(text: string): OpenAiBatchOutputLine[] {
     .split("\n")
     .map((line) => line.trim())
     .filter(Boolean)
-    .map((line) => JSON.parse(line) as OpenAiBatchOutputLine);
+    .map((line) => {
+      try {
+        return JSON.parse(line) as OpenAiBatchOutputLine;
+      } catch {
+        throw new Error("OpenAI embedding batch output contained malformed JSONL");
+      }
+    });
 }
 
 async function readOpenAiBatchError(params: {