ai-sdk-ui
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a utility script
scripts/check-versions.shthat executes thenpmCLI locally to check the versions of installed dependencies. - [EXTERNAL_DOWNLOADS]: The
scripts/check-versions.shscript performs network requests to the npm registry vianpm viewto compare local package versions with the latest available versions. - [INDIRECT_PROMPT_INJECTION]: The React templates provided by the skill are designed to ingest and render data streamed from external AI models, creating a vulnerability surface inherent to chat interfaces.
- Ingestion points: AI-generated message content processed by the
useChat,useCompletion, anduseAssistanthooks across multiple templates likeuse-chat-basic.tsxandnextjs-chat-app-router.tsx. - Boundary markers: No explicit instruction-ignoring delimiters are implemented in the UI logic of the templates.
- Capability inventory: The templates focus on UI state management and display; executable logic (tools, file handling) is defined in server-side API routes as seen in
nextjs-api-route.ts. - Sanitization: The
custom-message-renderer.tsxtemplate provides a secure rendering pattern usingReactMarkdownto sanitize and display markdown and code blocks returned by the AI model.
Audit Metadata