Skip to content

Arbitrary DOM attributes in element.attributes and leaf.attributes

High
12joan published GHSA-73rg-f94j-xvhx Sep 20, 2024

Package

npm @udecode/plate-core (npm)

Affected versions

>= 37.0.0, < 38.0.6
>= 22.0.0, < 36.5.9
< 21.5.1

Patched versions

38.0.6
36.5.9
21.5.1

Description

Impact

One longstanding feature of Plate is the ability to add custom DOM attributes to any element or leaf using the attributes property. These attributes are passed to the node component using the nodeProps prop.

Note: The attributes prop that is typically rendered alongside nodeProps is unrelated.

[{
  type: 'p',
  attributes: { 'data-my-attribute': 'This will be rendered on the paragraph element' },
  children: [{
    bold: true,
    attributes: { 'data-my-attribute': 'This will be rendered on the bold leaf element' },
    text: 'Bold text',
  }],
}]
const ParagraphElement = ({ attributes, nodeProps, children }) => (
  <p
    {...attributes}
    {...nodeProps} // Arbitrary DOM attributes are injected here
  >
    {children}
  </p>
);

const BoldLeaf = ({ attributes, nodeProps, children }) => (
  <strong
    {...attributes}
    {...nodeProps} // Arbitrary DOM attributes are injected here
  >
    {children}
  </strong>
);

It has come to our attention that this feature can be used for malicious purposes, including cross-site scripting (XSS) and information exposure (specifically, users' IP addresses and whether or not they have opened a malicious document).

Note that the risk of information exposure via attributes is only relevant to applications in which web requests to arbitrary URLs are not ordinarily allowed. Plate editors that allow users to embed images from arbitrary URLs, for example, already carry the risk of leaking users' IP addresses to third parties.

All Plate editors using an affected version of @udecode/plate-core are vulnerable to these information exposure attacks via the style attribute and other attributes that can cause web requests to be sent.

In addition, whether or not a Plate editor is vulnerable to cross-site scripting attacks using attributes depends on a number of factors. The most likely DOM attributes to be vulnerable are href and src on links and iframes respectively. Any component that spreads {...nodeProps} onto an <a> or <iframe> element and does not later override href or src will be vulnerable to XSS.

<a
  href={sanitizedHref}
  {...attributes}
  {...nodeProps} // Definitely vulnerable to XSS since `href` can be overridden
>
<a
  {...attributes}
  {...nodeProps} // Probably not vulnerable to XSS via `href`
  href={sanitizedHref}
>
<a
  {...attributes}
  {...nodeProps} // May be vulnerable to XSS via `href` if `href` is sometimes omitted from `sanitizedLinkProps`
  {...sanitizedLinkProps}
>

React does not allow passing a string to event handler props like onClick, so these are unlikely (but not impossible) to be vulnerable.

The attack surface is larger for users running older browsers, which may be vulnerable to XSS in DOM attributes that are less dangerous (although still vulnerable to information exposure) in modern browsers such as style or background.

Potential attack vectors for delivering malicious Slate content to users include:

  • Opening a malicious document stored on the server
  • Pasting a malicious Slate fragment into a document
  • Receiving malicious Slate operations on a collaborative document

Patches

In patched versions of Plate, we have disabled element.attributes and leaf.attributes for most attribute names by default, with some exceptions including target, alt, width, height, colspan and rowspan on the link, image, video, table cell and table header cell plugins.

If this is a breaking change for you, you can selectively re-enable attributes for certain plugins as follows. Please carefully research and assess the security implications of any attribute you allow, as even seemingly innocuous attributes such as style can be used maliciously.

Plate >= 37

For custom plugins, specify the list of allowed attribute names in the node.dangerouslyAllowAttributes plugin configuration option.

const ImagePlugin = createPlatePlugin({
  key: 'image',
  node: {
    isElement: true,
    isVoid: true,
    dangerouslyAllowAttributes: ['alt'],
  },
});

To modify an existing plugin, use the extend method.

const MyImagePlugin = ImagePlugin.extend({
  node: {
    dangerouslyAllowAttributes: ['alt'],
  },
});

Plate < 37

Note that the patch has been backported to versions @udecode/plate-core@21.5.1 and @udecode/plate-core@36.5.9 only.

For custom plugins, specify the list of allowed attribute names in the dangerouslyAllowAttributes plugin configuration option.

const createImagePlugin = createPluginFactory({
  key: 'image',
  isElement: true,
  isVoid: true,
  dangerouslyAllowAttributes: ['alt'],
});

To modify an existing plugin, pass dangerouslyAllowAttributes to the plugin factory.

createImagePlugin({
  dangerouslyAllowAttributes: ['alt'],
});

Workarounds

If you are unable to upgrade to any of the patched versions, you should use a tool like patch-package or yarn patch to remove the logic from @udecode/plate-core that adds attributes to nodeProps.

This logic can be found in the getRenderNodeProps function and looks something like this. The entire if statment can safely be removed.

  if (!newProps.nodeProps && attributes) {
    newProps.nodeProps = attributes;
  }

After applying the patch, be sure to test its effectiveness by rendering a Slate value containing an attributes property on some element.

[{
  type: 'p',
  attributes: { 'data-vulnerable': true },
  children: [{ text: 'My paragraph' }],
}]

If the patch was successful, the data-vulnerable="true" attribute should not be present on any DOM element when the Plate editor is rendered in the browser.

Severity

High

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Network
Attack complexity
Low
Privileges required
Low
User interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
Low

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:L

CVE ID

CVE-2024-47061

Weaknesses