syncfusion-react-ai-assistview
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The AI AssistView component is designed to ingest and display content from external AI models and user inputs, which creates a potential surface for indirect prompt injection.
- Ingestion points: The component accepts untrusted data through the
promptRequestevent handlers (referenced inSKILL.mdandevents-and-interactions.md), theattachedFilesproperty (documented infile-attachments.md), and raw responses from AI providers like OpenAI and Azure OpenAI (detailed inai-integration-setup.md). - Capability inventory: Across all referenced files, the component's capabilities are limited to network operations (via
fetch) for AI service communication and file uploads, alongside UI rendering. It does not possess dangerous capabilities such as local shell command execution or arbitrary file system writes. - Sanitization: The documentation proactively addresses security by recommending input validation and explicitly mentioning XSS prevention when rendering HTML responses in
conversation-management.mdandtemplating-system.md. - Boundary markers: The provided implementation examples do not utilize specific delimiters or prompt engineering techniques to instruct the underlying AI to disregard instructions embedded within user data.
- [EXTERNAL_DOWNLOADS]: The skill documents the installation of the
@syncfusion/ej2-react-interactive-chatandmarkedpackages from the NPM registry. These are standard dependencies from a well-known software vendor and a widely-used markdown parsing library.
Audit Metadata