ai-sdk

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The required workflow includes fetching ai-sdk.dev documentation over the public web at runtime (outsider-authored free text), which can then be ingested into the agent’s LLM context for answering when local docs aren’t found.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs fetching remote documentation and live model lists at runtime (e.g., https://ai-sdk.dev/api/search-docs?q=your_query and https://ai-gateway.vercel.sh/v1/models), which the agent would use to determine APIs and choose model IDs and thus directly influence its instructions/behavior.