ai-sdk
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- COMMAND_EXECUTION (SAFE): The skill directs the agent to use grep and glob within specific node_modules/ directories to retrieve documentation and source code for the AI SDK. These actions are scoped to the project dependencies.
- EXTERNAL_DOWNLOADS (LOW): The skill performs HTTP requests to https://ai-sdk.dev to search and fetch documentation. Since Vercel is a trusted organization, this finding is downgraded to LOW per the [TRUST-SCOPE-RULE].
- CREDENTIALS_UNSAFE (SAFE): The file references/ai-gateway.md contains a placeholder 'your_api_key_here' for an API key environment variable, which is a standard documentation practice and not a security vulnerability.
- PROMPT_INJECTION (LOW): This finding identifies a surface for Indirect Prompt Injection (Category 8) as the skill processes external documentation. 1. Ingestion points: Markdown content fetched from https://ai-sdk.dev/docs/. 2. Boundary markers: No delimiters or ignore instructions are used when processing fetched content. 3. Capability inventory: File system searching (grep/glob) and network fetches. 4. Sanitization: No sanitization is performed on the ingested documentation data.
Audit Metadata