ai-sdk
Warn
Audited by Snyk on Jun 20, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent at runtime to fetch remote documentation and model lists (e.g., curl -s https://ai-gateway.vercel.sh/v1/models and queries against https://ai-sdk.dev/api/search-docs?q=your_query or fetching https://ai-sdk.dev/docs/...), and those fetched resources are used to choose models and drive code/prompt decisions, so they directly control the agent's behavior.
Issues (1)
W012
MEDIUMUnverifiable external dependency detected (runtime URL that controls agent).
Audit Metadata