swiftui-uikit-interop
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill documentation includes recipes for components like
RemotePDFViewandMapViewthat process data from external sources (URLs and annotations). This represents a potential surface for indirect prompt injection where malicious instructions embedded in the fetched content could attempt to influence the agent's behavior. 1. Ingestion points: TheRemotePDFViewcomponent inreferences/representable-recipes.mdfetches data from an external URL usingURLSession. 2. Boundary markers: The provided code templates do not include explicit markers or instructions to delimit or ignore content within the fetched data. 3. Capability inventory: The skill demonstrates network data retrieval (URLSession) and UI rendering of rich text and PDF content. 4. Sanitization: No content validation or sanitization logic is demonstrated for the retrieved data. - [EXTERNAL_DOWNLOADS]: The skill includes code snippets that demonstrate fetching content from remote URLs, such as loading a PDF from a placeholder URL (
example.com). This is a standard implementation of networking functionality for the described use cases.
Audit Metadata