From 856c88f25fd5a51b7b13bddb63729b745efce72c Mon Sep 17 00:00:00 2001
From: Frank Yang <frank.ekn@gmail.com>
Date: Wed, 15 Apr 2026 17:07:39 +0800
Subject: [PATCH] fix: validate full host-read text payload

---
 src/media/web-media.test.ts | 24 ++++++++++++++++++++++++
 src/media/web-media.ts      | 28 +++++++++++++---------------
 2 files changed, 37 insertions(+), 15 deletions(-)

diff --git a/src/media/web-media.test.ts b/src/media/web-media.test.ts
index fbfd6d062f9..a35a6db94d7 100644
--- a/src/media/web-media.test.ts
+++ b/src/media/web-media.test.ts
@@ -250,6 +250,30 @@ describe("loadWebMedia", () => {
     });
   });
 
+  it.each([
+    { label: "CSV", fileName: "prefix-tail.csv" },
+    { label: "Markdown", fileName: "prefix-tail.md" },
+  ])(
+    "rejects %s files with a text prefix and binary tail after the old sample window",
+    async ({ fileName }) => {
+      const fakeTextFile = path.join(fixtureRoot, fileName);
+      const textPrefix = Buffer.from(`name,value\n${"row,1\n".repeat(1400)}`, "utf8");
+      expect(textPrefix.length).toBeGreaterThan(8192);
+      const binaryTail = Buffer.from([0x00, 0xff, 0x10, 0x80]);
+      await fs.writeFile(fakeTextFile, Buffer.concat([textPrefix, binaryTail]));
+      await expect(
+        loadWebMedia(fakeTextFile, {
+          maxBytes: 1024 * 1024,
+          localRoots: "any",
+          readFile: async (filePath) => await fs.readFile(filePath),
+          hostReadCapability: true,
+        }),
+      ).rejects.toMatchObject({
+        code: "path-not-allowed",
+      });
+    },
+  );
+
   it("rejects traversal-style canvas media paths before filesystem access", async () => {
     await expect(
       loadWebMedia(`${CANVAS_HOST_PATH}/documents/../collection.media/tiny.png`),
diff --git a/src/media/web-media.ts b/src/media/web-media.ts
index 9fa01e942e1..b2689164a6b 100644
--- a/src/media/web-media.ts
+++ b/src/media/web-media.ts
@@ -89,7 +89,6 @@ const HOST_READ_ALLOWED_DOCUMENT_MIMES = new Set([
 // file-type returns undefined (no magic bytes) for plain-text formats like CSV and
 // Markdown, so host-read needs an explicit "this really decodes as text" fallback.
 const HOST_READ_TEXT_PLAIN_ALIASES = new Set(["text/csv", "text/markdown"]);
-const HOST_READ_TEXT_SAMPLE_BYTES = 8192;
 const MB = 1024 * 1024;
 const WORDISH_CHAR = /[\p{L}\p{N}]/u;
 
@@ -155,30 +154,29 @@ function getTextStats(text: string): { printableRatio: number; wordishRatio: num
   return { printableRatio: printable / total, wordishRatio: wordish / total };
 }
 
-function decodeHostReadTextSample(buffer: Buffer): string | undefined {
-  const sample = buffer.subarray(0, Math.min(buffer.length, HOST_READ_TEXT_SAMPLE_BYTES));
-  if (sample.length === 0) {
+function decodeHostReadText(buffer: Buffer): string | undefined {
+  if (buffer.length === 0) {
     return "";
   }
-  const utf16Charset = resolveUtf16Charset(sample);
+  const utf16Charset = resolveUtf16Charset(buffer);
   try {
     if (utf16Charset === "utf-16be") {
-      const evenSample = sample.length % 2 === 0 ? sample : sample.subarray(0, sample.length - 1);
-      if (evenSample.length === 0) {
+      const evenBuffer = buffer.length % 2 === 0 ? buffer : buffer.subarray(0, buffer.length - 1);
+      if (evenBuffer.length === 0) {
         return "";
       }
-      const swapped = Buffer.alloc(evenSample.length);
-      for (let i = 0; i + 1 < evenSample.length; i += 2) {
-        swapped[i] = evenSample[i + 1];
-        swapped[i + 1] = evenSample[i];
+      const swapped = Buffer.alloc(evenBuffer.length);
+      for (let i = 0; i + 1 < evenBuffer.length; i += 2) {
+        swapped[i] = evenBuffer[i + 1];
+        swapped[i + 1] = evenBuffer[i];
       }
       return new TextDecoder("utf-16le").decode(swapped);
     }
     if (utf16Charset === "utf-16le") {
-      const evenSample = sample.length % 2 === 0 ? sample : sample.subarray(0, sample.length - 1);
-      return new TextDecoder("utf-16le").decode(evenSample);
+      const evenBuffer = buffer.length % 2 === 0 ? buffer : buffer.subarray(0, buffer.length - 1);
+      return new TextDecoder("utf-16le").decode(evenBuffer);
     }
-    return new TextDecoder("utf-8", { fatal: true }).decode(sample);
+    return new TextDecoder("utf-8", { fatal: true }).decode(buffer);
   } catch {
     return undefined;
   }
@@ -191,7 +189,7 @@ function isValidatedHostReadText(buffer?: Buffer): boolean {
   if (buffer.length === 0) {
     return true;
   }
-  const text = decodeHostReadTextSample(buffer);
+  const text = decodeHostReadText(buffer);
   if (text === undefined) {
     return false;
   }