GHSA-G77H-45RF-HCX4

Vulnerability from github – Published: 2026-07-17 20:19 – Updated: 2026-07-17 20:19
VLAI
Summary
ExifReader HEIC/AVIF ISO-BMFF parser throws uncaught RangeError on truncated boxes
Details

Summary

ExifReader 4.40.0 can throw an uncaught RangeError: Offset is outside the bounds of the DataView while parsing crafted HEIC/AVIF files. The file only needs a valid leading ftyp box with a HEIC/AVIF major brand followed by a malformed ISO-BMFF box, such as an empty 8-byte free box or a truncated extended-size box.

This is reachable through the public ExifReader.load() API for in-memory buffers and through the async file/URL loaders when an application parses attacker-supplied images. In applications that do not wrap every parse in a defensive try/catch, a single uploaded or fetched image can abort the request/worker and cause a denial of service.

Credit requested: Yaohui Wang.

Affected version tested

  • npm package: exifreader
  • Version: 4.40.0
  • Repository commit tested: 8cb0261a26b7d986955fe0a6780f076dcb7902e7

Root cause

The ISO-BMFF parser assumes that every top-level box with at least an 8-byte header also has enough bytes for the fields required by its parsed form. In src/image-header-iso-bmff.js:

  • findMetaBox() calls parseBox(dataView, offset) while only checking that offset + 8 <= dataView.byteLength.
  • parseBox() calls getBoxLength() and then unconditionally reads fields such as the full-box version byte for meta/iloc/iinf/idat boxes.
  • getBoxLength() handles boxLength === 1 by calling hasEmptyHighBits(dataView, offset), which reads dataView.getUint32(offset + 8) without first checking that the 64-bit extended size field is present.

As a result, syntactically small or truncated boxes after a valid HEIC/AVIF ftyp box escape the format-detection catch blocks and throw from the main parsing path.

Reproduction

Run this from the repository root against the committed dist/exif-reader.js bundle:

const ExifReader = require('./dist/exif-reader.js');

function u32be(n) {
  return [(n >>> 24) & 255, (n >>> 16) & 255, (n >>> 8) & 255, n & 255];
}
function ascii(s) {
  return Array.from(Buffer.from(s, 'ascii'));
}
function box(type, content = []) {
  return [...u32be(8 + content.length), ...ascii(type), ...content];
}

for (const brand of ['heic', 'avif']) {
  for (const badBox of ['free', 'abcd']) {
    const bytes = Uint8Array.from([
      ...box('ftyp', ascii(brand)),
      ...box(badBox), // 8-byte box header with no content
    ]);

    try {
      ExifReader.load(bytes.buffer);
      console.log(`${brand}/${badBox}: no throw`);
    } catch (e) {
      console.log(`${brand}/${badBox}: ${e.name}: ${e.message}`);
      console.log(String(e.stack).split('\n').slice(0, 6).join('\n'));
    }
  }
}

Observed output on Node v23.11.0 with ExifReader 4.40.0:

heic/free: RangeError: Offset is outside the bounds of the DataView
RangeError: Offset is outside the bounds of the DataView
    at DataView.prototype.getUint8 (<anonymous>)
    at parseBox (.../dist/exif-reader.js:1:16513)
    at findMetaBox (.../dist/exif-reader.js:1:19032)
    at findOffsets (.../dist/exif-reader.js:1:19101)

heic/abcd: RangeError: Offset is outside the bounds of the DataView
avif/free: RangeError: Offset is outside the bounds of the DataView
avif/abcd: RangeError: Offset is outside the bounds of the DataView

A second variant triggers the extended-size path:

const truncatedExtendedBox = [...u32be(1), ...ascii('free')];
const heic = Uint8Array.from([...box('ftyp', ascii('heic')), ...truncatedExtendedBox]);
ExifReader.load(heic.buffer);

That throws from hasEmptyHighBits() / getBoxLength() because the extended-size high/low fields are not present.

Expected behavior

Malformed/truncated metadata boxes should be handled like other malformed metadata in the project: return only the successfully parsed file type/metadata, return no app markers, or throw a controlled project-specific error. A safe JavaScript bounds error should not escape from the parser for an attacker-controlled image container.

Security impact

This is a denial-of-service issue for services that parse user-provided HEIC/AVIF files with ExifReader. A minimal attacker-controlled image buffer can cause an unhandled exception in the parser and abort the surrounding request/worker if the embedding application does not catch every parse error.

Suggested severity: Medium. Suggested CVSS: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L.

Suggested fix

Add explicit bounds checks before every DataView read in the ISO-BMFF box parser, especially:

  • before reading the 64-bit extended size fields in getBoxLength();
  • before reading the full-box version byte in parseBox();
  • before descending into parseSubBoxes() when a declared box length exceeds available bytes;
  • ensure findMetaBox() breaks on boxes whose declared length is invalid or not fully present.

A regression test should cover ftyp/heic and ftyp/avif followed by an 8-byte empty free/unknown box and by a truncated extended-size box.

Show details on source website

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 4.40.0"
      },
      "package": {
        "ecosystem": "npm",
        "name": "exifreader"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "4.40.1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-53496"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-248",
      "CWE-755"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-07-17T20:19:39Z",
    "nvd_published_at": null,
    "severity": "MODERATE"
  },
  "details": "## Summary\n\nExifReader 4.40.0 can throw an uncaught `RangeError: Offset is outside the bounds of the DataView` while parsing crafted HEIC/AVIF files. The file only needs a valid leading `ftyp` box with a HEIC/AVIF major brand followed by a malformed ISO-BMFF box, such as an empty 8-byte `free` box or a truncated extended-size box.\n\nThis is reachable through the public `ExifReader.load()` API for in-memory buffers and through the async file/URL loaders when an application parses attacker-supplied images. In applications that do not wrap every parse in a defensive try/catch, a single uploaded or fetched image can abort the request/worker and cause a denial of service.\n\nCredit requested: Yaohui Wang.\n\n## Affected version tested\n\n- npm package: `exifreader`\n- Version: `4.40.0`\n- Repository commit tested: `8cb0261a26b7d986955fe0a6780f076dcb7902e7`\n\n## Root cause\n\nThe ISO-BMFF parser assumes that every top-level box with at least an 8-byte header also has enough bytes for the fields required by its parsed form. In `src/image-header-iso-bmff.js`:\n\n- `findMetaBox()` calls `parseBox(dataView, offset)` while only checking that `offset + 8 \u003c= dataView.byteLength`.\n- `parseBox()` calls `getBoxLength()` and then unconditionally reads fields such as the full-box version byte for `meta`/`iloc`/`iinf`/`idat` boxes.\n- `getBoxLength()` handles `boxLength === 1` by calling `hasEmptyHighBits(dataView, offset)`, which reads `dataView.getUint32(offset + 8)` without first checking that the 64-bit extended size field is present.\n\nAs a result, syntactically small or truncated boxes after a valid HEIC/AVIF `ftyp` box escape the format-detection catch blocks and throw from the main parsing path.\n\n## Reproduction\n\nRun this from the repository root against the committed `dist/exif-reader.js` bundle:\n\n```js\nconst ExifReader = require(\u0027./dist/exif-reader.js\u0027);\n\nfunction u32be(n) {\n  return [(n \u003e\u003e\u003e 24) \u0026 255, (n \u003e\u003e\u003e 16) \u0026 255, (n \u003e\u003e\u003e 8) \u0026 255, n \u0026 255];\n}\nfunction ascii(s) {\n  return Array.from(Buffer.from(s, \u0027ascii\u0027));\n}\nfunction box(type, content = []) {\n  return [...u32be(8 + content.length), ...ascii(type), ...content];\n}\n\nfor (const brand of [\u0027heic\u0027, \u0027avif\u0027]) {\n  for (const badBox of [\u0027free\u0027, \u0027abcd\u0027]) {\n    const bytes = Uint8Array.from([\n      ...box(\u0027ftyp\u0027, ascii(brand)),\n      ...box(badBox), // 8-byte box header with no content\n    ]);\n\n    try {\n      ExifReader.load(bytes.buffer);\n      console.log(`${brand}/${badBox}: no throw`);\n    } catch (e) {\n      console.log(`${brand}/${badBox}: ${e.name}: ${e.message}`);\n      console.log(String(e.stack).split(\u0027\\n\u0027).slice(0, 6).join(\u0027\\n\u0027));\n    }\n  }\n}\n```\n\nObserved output on Node v23.11.0 with ExifReader 4.40.0:\n\n```text\nheic/free: RangeError: Offset is outside the bounds of the DataView\nRangeError: Offset is outside the bounds of the DataView\n    at DataView.prototype.getUint8 (\u003canonymous\u003e)\n    at parseBox (.../dist/exif-reader.js:1:16513)\n    at findMetaBox (.../dist/exif-reader.js:1:19032)\n    at findOffsets (.../dist/exif-reader.js:1:19101)\n\nheic/abcd: RangeError: Offset is outside the bounds of the DataView\navif/free: RangeError: Offset is outside the bounds of the DataView\navif/abcd: RangeError: Offset is outside the bounds of the DataView\n```\n\nA second variant triggers the extended-size path:\n\n```js\nconst truncatedExtendedBox = [...u32be(1), ...ascii(\u0027free\u0027)];\nconst heic = Uint8Array.from([...box(\u0027ftyp\u0027, ascii(\u0027heic\u0027)), ...truncatedExtendedBox]);\nExifReader.load(heic.buffer);\n```\n\nThat throws from `hasEmptyHighBits()` / `getBoxLength()` because the extended-size high/low fields are not present.\n\n## Expected behavior\n\nMalformed/truncated metadata boxes should be handled like other malformed metadata in the project: return only the successfully parsed file type/metadata, return no app markers, or throw a controlled project-specific error. A safe JavaScript bounds error should not escape from the parser for an attacker-controlled image container.\n\n## Security impact\n\nThis is a denial-of-service issue for services that parse user-provided HEIC/AVIF files with ExifReader. A minimal attacker-controlled image buffer can cause an unhandled exception in the parser and abort the surrounding request/worker if the embedding application does not catch every parse error.\n\nSuggested severity: Medium. Suggested CVSS: `CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L`.\n\n## Suggested fix\n\nAdd explicit bounds checks before every `DataView` read in the ISO-BMFF box parser, especially:\n\n- before reading the 64-bit extended size fields in `getBoxLength()`;\n- before reading the full-box version byte in `parseBox()`;\n- before descending into `parseSubBoxes()` when a declared box length exceeds available bytes;\n- ensure `findMetaBox()` breaks on boxes whose declared length is invalid or not fully present.\n\nA regression test should cover `ftyp/heic` and `ftyp/avif` followed by an 8-byte empty `free`/unknown box and by a truncated extended-size box.",
  "id": "GHSA-g77h-45rf-hcx4",
  "modified": "2026-07-17T20:19:39Z",
  "published": "2026-07-17T20:19:39Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/mattiasw/ExifReader/security/advisories/GHSA-g77h-45rf-hcx4"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/mattiasw/ExifReader"
    },
    {
      "type": "WEB",
      "url": "https://github.com/mattiasw/ExifReader/releases/tag/v4.40.1"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "ExifReader HEIC/AVIF ISO-BMFF parser throws uncaught RangeError on truncated boxes"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…

Loading…