panel-custom-components

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the component runtime to load and execute untrusted third-party content from public CDNs and URLs (e.g., numerous _importmap entries like "https://esm.sh/d3@7", "https://esm.sh/@google/model-viewer@3.4.0", inline ESM imports, and examples that fetch model/data from external URLs such as fetch(model.data_url) or example.com), meaning the agent will fetch and run arbitrary external code/data that can influence behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill includes runtime ESM CDN imports (e.g., https://esm.sh/d3@7 and other https://esm.sh/... URLs) that are fetched and executed in the browser as required component libraries, so remote code is executed at runtime and the skill depends on those external resources.