CVE-2023-48219

TinyMCE is an open source rich text editor. A mutation cross-site scripting (mXSS) vulnerability was discovered in TinyMCE’s core undo/redo functionality and other APIs and plugins. Text nodes within specific parents are not escaped upon serialization according to the HTML standard. If such text nodes contain a special character reserved as an internal marker, they can be combined with other HTML patterns to form malicious snippets. These snippets pass the initial sanitisation layer when the content is parsed into the editor body, but can trigger XSS when the special internal marker is removed from the content and re-parsed. his vulnerability has been patched in TinyMCE versions 6.7.3 and 5.10.9. Users are advised to upgrade. There are no known workarounds for this vulnerability.

TinyMCE is an open source rich text editor. A mutation cross-site scripting (mXSS) vulnerability was discovered in TinyMCE’s core undo/redo functionality and other APIs and plugins. Text nodes within specific parents are not escaped upon serialization according to the HTML standard. If such text nodes contain a special character reserved as an internal marker, they can be combined with other HTML patterns to form malicious snippets. These snippets pass the initial sanitisation layer when the content is parsed into the editor body, but can trigger XSS when the special internal marker is removed from the content and re-parsed. his vulnerability has been patched in TinyMCE versions 6.7.3 and 5.10.9. Users are advised to upgrade. There are no known workarounds for this vulnerability.

Impact

A mutation cross-site scripting (mXSS) vulnerability was discovered in TinyMCE’s core undo/redo functionality and other APIs and plugins. Text nodes within specific parents are not escaped upon serialization according to the HTML standard. If such text nodes contain a special character reserved as an internal marker, they can be combined with other HTML patterns to form malicious snippets. These snippets pass the initial sanitisation layer when the content is parsed into the editor body, but can trigger XSS when the special internal marker is removed from the content and re-parsed. Such mutations occur when serialised HTML content is processed before being stored in the undo stack, or when the following APIs and plugins are used:

• tinymce.Editor.getContent({ format: 'raw' })
• tinymce.Editor.resetContent()
• Autosave Plugin

Patches

This vulnerability has been patched in TinyMCE 6.7.3 by:

• ensuring that any unescaped text nodes which contain the special internal marker are emptied before removing the marker from the rest of the HTML, and
• removing the special internal marker from content strings passed to Editor.setContent, Editor.insertContent, and Editor.resetContent APIs to prevent them from being loaded into the editor as user-provided content.

Fix

To avoid this vulnerability:

• Upgrade to TinyMCE 6.7.3 or higher for TinyMCE 6.x.
• Upgrade to TinyMCE 5.10.9 or higher for TinyMCE 5.x.

Acknowledgements

Tiny Technologies would like to thank Masato Kinugawa of Cure53 for discovering this vulnerability.

References

• TinyMCE 5.10.9 Release Notes
• TinyMCE 6.7.3 Release Notes

For more information

Email us at [email protected] Open an issue in the TinyMCE repo

Impact

A cross-site scripting (XSS) vulnerability was discovered in the URL sanitization logic of the core parser for form elements. The vulnerability allowed arbitrary JavaScript execution when inserting a specially crafted piece of content into the editor using the clipboard or APIs, and then submitting the form. However, as TinyMCE does not allow forms to be submitted while editing, the vulnerability could only be triggered when the content was previewed or rendered outside of the editor. This impacts all users who are using TinyMCE 5.7.0 or lower.

Patches

This vulnerability has been patched in TinyMCE 5.7.1 by improved URL sanitization logic.

Workarounds

To work around this vulnerability, either:

• Upgrade to TinyMCE 5.7.1 or higher
• Manually sanitize form URL attributes using a TinyMCE node filter.
• Disable form elements in your content using the invalid_elements setting.

Example: Sanitizing using a node filter

editor.parser.addNodeFilter('form', function(nodes) {
  nodes.forEach(function(node) {
    if (node.attributes) {
      node.attributes.forEach(function(attr) {
        var name = attr.name;
        var value = attr.value;
        // Sanitize the attribute value here or remove it entirely
        var sanitizedValue = ...;
        node.attr(name, santizedValue);
      });
    }
  });
});

Example: Using invalid_elements

invalid_elements: 'form'

Acknowledgements

Tiny Technologies would like to thank Mikhail Khramenkov at Solar Security Research Team for discovering this vulnerability.

References

https://www.tiny.cloud/docs/release-notes/release-notes571/#securityfixes

For more information

If you have any questions or comments about this advisory:

• Open an issue in the TinyMCE repo
• Email us at [email protected]

TinyMCE is an open source rich text editor. A mutation cross-site scripting (mXSS) vulnerability was discovered in TinyMCE’s core undo/redo functionality and other APIs and plugins. Text nodes within specific parents are not escaped upon serialization according to the HTML standard. If such text nodes contain a special character reserved as an internal marker, they can be combined with other HTML patterns to form malicious snippets. These snippets pass the initial sanitisation layer when the content is parsed into the editor body, but can trigger XSS when the special internal marker is removed from the content and re-parsed. his vulnerability has been patched in TinyMCE versions 6.7.3 and 5.10.9. Users are advised to upgrade. There are no known workarounds for this vulnerability.

TinyMCE is an open source rich text editor. A mutation cross-site scripting (mXSS) vulnerability was discovered in TinyMCE’s core undo/redo functionality and other APIs and plugins. Text nodes within specific parents are not escaped upon serialization according to the HTML standard. If such text nodes contain a special character reserved as an internal marker, they can be combined with other HTML patterns to form malicious snippets. These snippets pass the initial sanitisation layer when the content is parsed into the editor body, but can trigger XSS when the special internal marker is removed from the content and re-parsed. his vulnerability has been patched in TinyMCE versions 6.7.3 and 5.10.9. Users are advised to upgrade. There are no known workarounds for this vulnerability.