fix(web-fetch): keep spill content truncation UTF-16 safe (#101312)

Alix-007 · web-flow · commit 240487179b69 · 2026-07-07T10:56:14.000+01:00
* Fix UTF-16-safe web fetch spill truncation

* fix(web-fetch): report actual spill character count
diff --git a/src/agents/tools/web-fetch.ts b/src/agents/tools/web-fetch.ts
@@ -9,6 +9,7 @@ import {
   normalizeOptionalLowercaseString,
   normalizeOptionalString,
 } from "@openclaw/normalization-core/string-coerce";
+import { truncateUtf16Safe } from "@openclaw/normalization-core/utf16-slice";
 import { Type } from "typebox";
 import { resolveWebProviderConfig } from "../../../packages/web-content-core/src/provider-runtime-shared.js";
 import type { OpenClawConfig } from "../../config/types.openclaw.js";
@@ -314,7 +315,8 @@ async function spillWebFetchContent(
   }
   // maxChars/maxCharsCap bound the model-visible return text. Recoverable spill
   // uses this fixed file cap so vanished pages can still be read after truncation.
-  const content = value.slice(0, WEB_FETCH_SPILL_MAX_CHARS);
+  const content = truncateUtf16Safe(value, WEB_FETCH_SPILL_MAX_CHARS);
+  const spilledChars = content.length;
   const fullOutputPath = await writePrivateTempFile(
     "openclaw-web-fetch",
     wrapWebContent(content, "web_fetch"),
@@ -324,7 +326,7 @@ async function spillWebFetchContent(
   const spillNote = sourceTruncated
     ? " Spilled available content from truncated response."
     : spillCapped
-      ? ` Spilled first ${WEB_FETCH_SPILL_MAX_CHARS} chars.`
+      ? ` Spilled first ${spilledChars} chars.`
       : "";
   const fullOutputFooter = formatFullOutputFooter(fullOutputPath);
   const footer = `\n\n[Showing truncated web_fetch content. ${fullOutputFooter}.${spillNote}]`;
@@ -344,7 +346,7 @@ async function spillWebFetchContent(
     text,
     wrappedLength: text.length,
     fullOutputPath,
-    spilledChars: content.length,
+    spilledChars,
     spillTruncated,
   };
 }
diff --git a/src/agents/tools/web-tools.fetch.test.ts b/src/agents/tools/web-tools.fetch.test.ts
@@ -503,6 +503,38 @@ describe("web_fetch extraction fallbacks", () => {
     await rm(details.fullOutputPath, { force: true });
   });
 
+  it("does not split an emoji at the web_fetch spill file cap", async () => {
+    const prefix = "x".repeat(WEB_FETCH_SPILL_MAX_CHARS - 1);
+    const fullText = `${prefix}${String.fromCodePoint(0x1f600)}tail`;
+    installPlainTextFetch(fullText);
+
+    const tool = createFetchTool({
+      firecrawl: { enabled: false },
+      maxChars: 500,
+      maxResponseBytes: WEB_FETCH_SPILL_MAX_CHARS + 1_000,
+    });
+
+    const result = await tool?.execute?.("call", { url: "https://example.com/spill-utf16" });
+    const details = result?.details as {
+      text?: string;
+      fullOutputPath?: string;
+      spilledChars?: number;
+      spillTruncated?: boolean;
+    };
+    if (!details.fullOutputPath) {
+      throw new Error("expected fullOutputPath");
+    }
+
+    expect(details.spilledChars).toBe(WEB_FETCH_SPILL_MAX_CHARS - 1);
+    expect(details.text).toContain(`Spilled first ${WEB_FETCH_SPILL_MAX_CHARS - 1} chars.`);
+    expect(details.spillTruncated).toBe(true);
+    const spilledText = await readFile(details.fullOutputPath, "utf8");
+    expect(spilledText).toContain(prefix);
+    expect(spilledText).not.toContain(String.fromCodePoint(0x1f600));
+    expect(spilledText).not.toContain(String.fromCharCode(0xd83d));
+    await rm(details.fullOutputPath, { force: true });
+  });
+
   it("marks byte-capped web_fetch spills as partial", async () => {
     const fullText = "z".repeat(40_000);
     installMockFetch((input: RequestInfo | URL) => {
