vercel-ai-sdk
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill functions as a legitimate developer guide for the Vercel AI SDK.
- [EXTERNAL_DOWNLOADS]: References installation of official packages (@ai-sdk/openai, @ai-sdk/anthropic, ai) from the Vercel organization, a well-known service.
- [COMMAND_EXECUTION]: The provided code examples include tool definitions for database and search operations, which are part of the intended agent functionality.
- [PROMPT_INJECTION]: The skill describes an architecture for processing user-supplied messages. 1. Ingestion points: req.json() in app/api/chat/route.ts. 2. Boundary markers: None present in the code examples. 3. Capability inventory: Database writes via db.tasks.create and data retrieval via vectorSearch in SKILL.md. 4. Sanitization: Employs Zod schemas for validation of tool parameters.
Audit Metadata