From e220d444a304630db4eb5acb240a6afc72d220ab Mon Sep 17 00:00:00 2001 From: cliffhall Date: Tue, 27 Oct 2026 00:25:49 -0400 Subject: [PATCH] fix(everything): drop the corrupt iCCP chunk from the tiny image PNG The PNG embedded as MCP_TINY_IMAGE carried an iCCP chunk whose stored CRC-32 did not match its bytes, and whose compressed profile also failed its own zlib Adler-32 check, so correcting the CRC alone would have left a broken profile. Remove the chunk; every other chunk, the pixels included, is byte-for-byte unchanged. A protocol-level test walks every chunk of the image the client receives and verifies each CRC. Closes #5208 Co-Authored-By: Claude Opus 5.5 Signed-off-by: cliffhall --- .changeset/everything-tiny-image-crc.md | 5 +++++ src/everything/__tests__/tools.test.ts | 29 +++++++++++++++++++++++++ src/everything/tools/get-tiny-image.ts | 2 +- 3 files changed, 35 insertions(+), 1 deletion(-) create mode 100644 .changeset/everything-tiny-image-crc.md diff --git a/.changeset/everything-tiny-image-crc.md b/.changeset/everything-tiny-image-crc.md new file mode 100644 index 0000000000..434ac872e5 --- /dev/null +++ b/.changeset/everything-tiny-image-crc.md @@ -0,0 +1,5 @@ +--- +"@modelcontextprotocol/server-everything": patch +--- + +`get-tiny-image` (and `get-annotated-message`'s image) now returns a fully valid PNG: the embedded logo's corrupt `iCCP` color-profile chunk, whose CRC-32 did not match and which strict decoders rejected, is removed. The pixels are unchanged. diff --git a/src/everything/__tests__/tools.test.ts b/src/everything/__tests__/tools.test.ts index 13ff88e368..6ad300c26b 100644 --- a/src/everything/__tests__/tools.test.ts +++ b/src/everything/__tests__/tools.test.ts @@ -5,6 +5,7 @@ * to the client (sampling, elicitation, roots), the task tools, logging, * subscriptions and gzip each have their own file. */ +import { crc32 } from "node:zlib"; import { afterEach, beforeEach, describe, expect, it, vi } from "vitest"; import { INCLUDE_SECRETS_VAR, REDACTED } from "../tools/get-env.js"; import { MCP_TINY_IMAGE } from "../tools/get-tiny-image.js"; @@ -171,6 +172,34 @@ describe("get-tiny-image", () => { Buffer.from(MCP_TINY_IMAGE, "base64").subarray(1, 4).toString(), ).toBe("PNG"); }); + + // #5208: the embedded PNG once carried an iCCP chunk whose CRC did not + // match, which strict decoders reject. Walk every chunk of the image the + // client receives and check its stored CRC-32 against its type and data. + it("returns a PNG whose every chunk has a valid CRC", async () => { + const result = await call("get-tiny-image"); + const image = contentOfType(contentOf(result)[1], "image"); + const png = Buffer.from(image.data, "base64"); + expect(png.subarray(0, 8)).toEqual( + Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]), + ); + const chunks: string[] = []; + let offset = 8; + while (offset < png.length) { + const length = png.readUInt32BE(offset); + const type = png.subarray(offset + 4, offset + 8).toString("latin1"); + const stored = png.readUInt32BE(offset + 8 + length); + expect(crc32(png.subarray(offset + 4, offset + 8 + length)), type).toBe( + stored, + ); + chunks.push(type); + offset += 12 + length; + } + expect(offset).toBe(png.length); + expect(chunks[0]).toBe("IHDR"); + expect(chunks.at(-1)).toBe("IEND"); + expect(chunks).toContain("IDAT"); + }); }); describe("get-structured-content", () => { diff --git a/src/everything/tools/get-tiny-image.ts b/src/everything/tools/get-tiny-image.ts index cb931cf941..972906b1b9 100644 --- a/src/everything/tools/get-tiny-image.ts +++ b/src/everything/tools/get-tiny-image.ts @@ -2,7 +2,7 @@ import { McpServer, CallToolResult } from "@modelcontextprotocol/server"; // A tiny encoded MCP logo image export const MCP_TINY_IMAGE = - "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"; + "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"; // Tool configuration const name = "get-tiny-image";