cloud
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation directs users to install several Node.js packages including @assistant-ui/react, @assistant-ui/ai-sdk, @assistant-ui/cloud-ai-sdk, @assistant-ui/react-langgraph, and assistant-cloud. These are official packages provided by the skill author (assistant-ui). It also references well-known ecosystem libraries such as ai, @ai-sdk/openai, drizzle-orm, next-auth, @clerk/nextjs, better-auth, @auth0/auth0-react, @supabase/auth-helpers-react, and firebase. These downloads are standard for the skill's functionality.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and persistence of chat messages, which creates a potential surface for indirect prompt injection through untrusted user content stored in history.\n
- Ingestion points: Untrusted data enters the agent context through chat messages managed via useChatRuntime, useCloudChat, and direct API calls to cloud.threads.messages.create.\n
- Boundary markers: The provided documentation does not specify the use of clear delimiters or instructions to the model to ignore potential injection attempts within the persisted message content.\n
- Capability inventory: The skill performs network API requests to the assistant-api.com domain (vendor infrastructure) and manages file uploads using pre-signed URLs.\n
- Sanitization: There are no documented requirements or examples for sanitizing user-provided chat content before it is processed by the AI runtime.
Audit Metadata