ai-sdk
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches documentation content and model metadata from official Vercel and AI SDK domains (ai-sdk.dev and vercel.sh). These sources are recognized as well-known services and are used for their intended purpose of documentation retrieval and API interaction.
- [COMMAND_EXECUTION]: The instructions include the use of 'pnpm' for package installation, 'grep' for searching local source code, and 'curl' to interact with the Vercel AI Gateway. These operations are standard for development workflows and target trusted infrastructure.
- [INDIRECT_PROMPT_INJECTION]: The skill features a surface for indirect prompt injection as it ingests data from external websites and local project files.
- Ingestion points: Documentation from ai-sdk.dev, model lists from ai-gateway.vercel.sh, and local project files in node_modules/ai/.
- Boundary markers: No specific delimiters or warnings are used to isolate untrusted external content from the agent's core instructions.
- Capability inventory: The agent can execute shell commands like 'curl', 'pnpm', and 'grep' based on the retrieved information.
- Sanitization: There are no specific mechanisms for sanitizing or validating the content fetched from external documentation sources.
Audit Metadata