ai-sdk

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • External Network Operations: The skill fetches up-to-date model definitions from the official Vercel AI Gateway at ai-gateway.vercel.sh. This network access is used to ensure the agent uses the most current model versions and follows vendor best practices.
  • Development Command Execution: The instructions utilize standard development tools and package managers (such as pnpm, npm, and npx) to manage dependencies and run utility scripts like @ai-sdk/devtools. This is expected behavior for a developer-oriented tool.
  • Local File System Interaction: The skill searches through the project's node_modules directory using standard utilities like grep to find current documentation and API definitions. This allows the agent to provide accurate guidance based on the locally installed version of the SDK.
  • Data Processing Surface: The skill analyzes project metadata, such as package.json, and processes external model lists to provide tailored implementation advice. While this represents a data ingestion surface, it is necessary for the skill's primary function of providing context-aware guidance.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 06:04 PM