Summary
Amundsen frontend versions through 4.3.0 render table, dashboard, and feature descriptions as HTML without sanitizing the content. An attacker who can place malicious markup in metadata may execute JavaScript in the browser of every user who views affected search results.
Affected context
The vulnerability occurs in Amundsen frontend ResourceListItem components when descriptions retrieved from the metadata service or Elasticsearch are rendered with dangerouslySetInnerHTML. Risk depends on whether an attacker can create or modify indexed descriptions and whether users subsequently view the affected tables, dashboards, or features in search results.
Exploit preconditions
- Attack surface: Amundsen web frontend search-result rendering
- Required access: Ability to submit or alter metadata that is returned as a description
- Trigger point: A user views a search result containing the attacker-controlled description
- Why exploitation is possible: The frontend inserts description content into the document without applying HTML sanitization or equivalent output encoding
Why this happens
The underlying issue is the treatment of untrusted description content as trusted HTML through dangerouslySetInnerHTML. When user-controlled metadata reaches a browser-rendering sink without context-appropriate sanitization, markup and event-handler content may be interpreted as executable script. This behavior maps to CWE-79: Improper Neutralization of Input During Web Page Generation, commonly known as cross-site scripting.
Potential impact
- An attacker may execute JavaScript in the security context of an affected user
- Session-related data or accessible application information could be exposed
- Search results and other frontend content may be altered or impersonated
- Administrative users could be targeted for broader application actions permitted by their sessions
- The vulnerability could support phishing, unauthorized requests, or further compromise of connected systems
General mitigation guidance
Mitigation generally focuses on ensuring that metadata is treated as text unless trusted HTML is explicitly required.
Common defensive patterns include:
- Replacing raw HTML insertion with safe text rendering where formatting is unnecessary
- Applying a well-maintained, allowlist-based HTML sanitizer before rendering any permitted markup
- Encoding output according to the browser context and avoiding inline event handlers
- Validating and constraining description content at ingestion and service boundaries
- Reducing metadata write permissions and limiting administrative privileges
- Applying restrictive browser security policies, such as an appropriately configured Content Security Policy, as defense in depth
How teams detect issues like this
Organizations can detect this class of issue through:
- Static analysis rules identifying
dangerouslySetInnerHTMLand comparable HTML injection sinks - Code review checks tracing metadata fields from ingestion or search services to browser rendering
- Tests containing harmless markup and event-handler probes to verify sanitization and encoding behavior
- Dependency and policy checks requiring approved sanitization libraries for rich-text fields
- Runtime monitoring for unexpected script execution, content-security-policy violations, and suspicious metadata changes

