ai-sdk-v6
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- EXTERNAL_DOWNLOADS (SAFE): The skill references installation of official packages from the
@ai-sdkandaiscopes. These are managed by Vercel, which is a recognized trusted organization. - COMMAND_EXECUTION (SAFE): The code snippets demonstrate tool execution and explicitly include safety features like
needsApproval: truefor sensitive operations such as shell command execution. - CREDENTIALS_UNSAFE (SAFE): API key usage in examples is limited to descriptive placeholders like
my-api-keyrather than actual secrets. - PROMPT_INJECTION (SAFE): The skill body consists of educational content and code samples without any instructions intended to override agent safety protocols or bypass constraints.
Audit Metadata