svelte
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- Indirect Prompt Injection Surface: The skill is designed to process JSON specifications from external sources, such as AI-generated streams. This creates a potential surface for untrusted data to enter the application context.
- Ingestion points: The Renderer component processes specifications provided via props, which can be sourced from createUIStream or createChatUI functions (SKILL.md).
- Boundary markers: The skill relies on the structure of the JSON specification; explicit delimiters or instructions to ignore embedded commands within the data itself are not shown in the documentation.
- Capability inventory: The system can modify application state (setState), render diverse UI components, and execute actions through an event-based system (emit).
- Sanitization: The skill demonstrates the use of zod schemas for validating component properties within the registry, providing a layer of structured validation for the expected data formats (SKILL.md).
- Dynamic UI Rendering: The skill facilitates the runtime assembly of Svelte component trees based on JSON input. While this involves dynamic execution patterns, it is the primary intended functionality of the library and is mediated through a developer-defined component registry that restricts the set of available components to those explicitly provided.
Audit Metadata