Summary
A high-severity cross-site scripting vulnerability in Paella Player allows malicious closed-caption cue text to be interpreted as active content when displayed by the video player. The issue becomes risky when an attacker can influence caption data that is later rendered in a viewer's browser.
Affected context
The vulnerability occurs in Paella Player deployments that render closed-caption cue text without adequately treating it as untrusted content. It may affect applications embedding vulnerable Paella Player libraries, including Opencast environments that use affected Paella Player releases. Exploitability depends on whether an attacker can create, modify, or supply caption data that reaches the player.
Exploit preconditions
- Attack surface: Web-based video playback and caption rendering
- Required access: Ability to influence closed-caption or subtitle content, directly or through an upstream content-management workflow
- Trigger point: Rendering of attacker-controlled cue text in a user's browser
- Why exploitation is possible: Caption content is not sufficiently constrained or output-encoded before being inserted into the document
Why this happens
The underlying bug is improper neutralization of user-controlled input in a web page context. When closed-caption cue text is treated as trusted markup rather than plain text, browser interpretation may allow injected script-capable content to execute. This maps to CWE-79: Improper Neutralization of Input During Web Page Generation, commonly known as cross-site scripting.
Potential impact
- Arbitrary script execution in a viewer's browser session
- Theft of accessible session data or page content
- Unauthorized actions performed through the affected application
- Defacement or manipulation of the video-player interface
- Phishing or redirection to attacker-controlled content
General mitigation guidance
Mitigation generally focuses on ensuring that caption data remains inert throughout its processing and rendering lifecycle.
Common defensive patterns include:
- Treating cue text as text rather than trusted HTML
- Applying context-appropriate output encoding at the rendering boundary
- Sanitizing markup with a restrictive, allowlist-based policy when formatted captions are required
- Validating caption files and metadata before publication
- Restricting who can create or modify caption content
- Applying a restrictive Content Security Policy as defense in depth
- Limiting application privileges and exposed session capabilities
How teams detect issues like this
Organizations can identify related issues through:
- Static analysis rules for unsafe DOM insertion and template rendering
- Code review checks for caption or subtitle text inserted as HTML
- Tests using representative markup and script-like input in cue text
- Content-management policies that require validation and review of uploaded captions
- Browser security telemetry and application logs for unexpected caption content or client-side errors
Automated security review tools may help flag unsafe rendering patterns, but results should be confirmed against the application's actual output context.

