/

CVE-2026-55085

Severity:

CRITICAL RISK

(9.6)

Stored XSS (CWE-79) in Etherpad List Rendering Bug

CVE-2026-55085

Deferred

Vulnerability type

Category:

Cross-Site Scripting

CWE:

CWE-79

Impact:

Stored Cross-Site Scripting

Severity:

CRITICAL RISK

(9.6)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A stored cross-site scripting vulnerability in Etherpad allows attacker-controlled list metadata from a crafted import to be inserted into generated HTML without adequate output encoding. A user with write access to a pad may cause script execution when another user, including an administrator, opens the pad or its timeslider.

Affected context

The vulnerability occurs in Etherpad deployments that process imported pad content and render numbered lists through the affected client-side code. It becomes risky when a user who can write to a pad can influence imported attribute-pool values and another user later views the affected pad or timeslider.

Exploit preconditions

  • Attack surface: Etherpad pad and timeslider rendering
  • Required access: Write access to a pad and the ability to introduce crafted imported content
  • Trigger point: A victim opens the pad or views its timeslider
  • Why exploitation is possible: A list start value is interpolated into an unquoted HTML attribute and assigned through innerHTML without sufficient contextual encoding

Why this happens

The underlying bug is improper neutralization of user-controlled data in dynamically generated HTML, mapping to CWE-79: Improper Neutralization of Input During Web Page Generation. Etherpad constructs an ordered-list element by directly placing an attacker-influenced value into an HTML attribute and then parses the resulting string with innerHTML. Because the value is not safely constrained or encoded for the specific HTML attribute context, browser markup interpretation can turn stored content into executable script.

Potential impact

  • Arbitrary script execution in another user's browser session
  • Unauthorized actions performed using a victim's Etherpad privileges
  • Exposure of pad content or other data available to the victim
  • Account or administrative workflow abuse
  • Modification of collaborative documents or disruption of editing activity

General mitigation guidance

Mitigation generally focuses on treating imported metadata and pad content as untrusted throughout the rendering pipeline.

Common defensive patterns include:

  • Constructing DOM elements with safe property or attribute APIs instead of concatenating HTML strings
  • Applying context-aware output encoding before inserting values into markup
  • Constraining list attributes to an allowlisted numeric representation
  • Validating imported attribute-pool data against a strict schema before storage and rendering
  • Minimizing use of innerHTML for values that can be influenced by users
  • Applying least privilege to pad authors and administrative sessions
  • Using browser security controls such as a restrictive Content Security Policy as defense in depth

How teams detect issues like this

Detection can include static analysis for string interpolation into HTML attributes, unsafe innerHTML assignments, and DOM construction that combines trusted markup with imported values. Code review should examine import handlers, attribute-pool processing, list rendering, and every boundary where stored content reaches browser parsing APIs. Policy enforcement can require contextual output encoding, schema validation for imported metadata, and security tests covering stored content in pads and timeslider views. Automated application-security review tools may help identify related DOM-based injection patterns.

Sources

SourceLink
github.comgithub.com/ether/etherpad/commit/d828185efcac9abc2e46e6da1c171393500fa319
github.comgithub.com/ether/etherpad/pull/7937
github.comgithub.com/ether/etherpad/releases/tag/v3.3.1
github.comgithub.com/ether/etherpad/security/advisories/GHSA-f7h5-v9hm-548j

Frequently asked

questions

An attacker generally needs write access to a pad or another permitted method of introducing crafted imported content. The victim only needs to open the affected pad or its timeslider.

No. Administrative access is not required to plant the content, although an administrator who views it may be exposed to script execution with elevated session privileges.

Avoid assembling HTML with untrusted values. Build elements through safe DOM APIs, validate list metadata against a narrow schema, and apply context-appropriate encoding before rendering.

Treat imported content as untrusted, test all stored-content rendering paths, restrict dangerous DOM sinks, enforce secure coding policies, and use browser security controls as defense in depth.

Find out what's already
exploitable in your codebase.