fix(scripts): preserve UTF-16 surrogate pairs in firecrawl truncation (#109561)

* fix(scripts): preserve UTF-16 surrogate pairs in firecrawl truncation

* test(scripts): cover UTF-16-safe firecrawl truncation

---------

Co-authored-by: Peter Steinberger <steipete@gmail.com>

scripts/firecrawl-compare.ts

@@ -1,5 +1,6 @@
 // Firecrawl Compare script supports OpenClaw repository automation.
 import { pathToFileURL } from "node:url";
+import { truncateUtf16Safe } from "@openclaw/normalization-core/utf16-slice";
 import { fetchFirecrawlContent } from "../extensions/firecrawl/api.ts";
 import { formatErrorMessage } from "../src/infra/errors.ts";
 import { extractReadableContent } from "../src/web-fetch/content-extractors.runtime.js";
@@ -32,7 +33,7 @@ function truncate(value: string, max = 180): string {
   if (!value) {
     return "";
   }
-  return value.length > max ? `${value.slice(0, max)}…` : value;
+  return value.length > max ? `${truncateUtf16Safe(value, max)}…` : value;
 }
 
 function readBoundedResponseText(
@@ -176,4 +177,5 @@ export const testing = {
   FETCH_HTML_MAX_BYTES,
   fetchHtml,
   readBoundedResponseText,
+  truncate,
 };

test/scripts/firecrawl-compare.test.ts

@@ -3,6 +3,10 @@ import { describe, expect, it } from "vitest";
 import { testing as firecrawlCompareTesting } from "../../scripts/firecrawl-compare.ts";
 
 describe("firecrawl-compare", () => {
+  it("does not split surrogate pairs when truncating output", () => {
+    expect(firecrawlCompareTesting.truncate(`ab🤖cd`, 3)).toBe("ab…");
+  });
+
   it("fetches local HTML under the byte cap", async () => {
     const result = await firecrawlCompareTesting.fetchHtml("https://example.test/page", {
       fetchImpl: (() =>