ydc-openai-agent-sdk-integration

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

• Third-party content exposure detected (high risk: 1.00). The skill configures and runs agents that call You.com MCP tools (mcp__ydc__you_search and mcp__ydc__you_contents) via the https://api.you.com/mcp endpoint to fetch raw public web page content which the agent ingests and uses to answer queries, creating a clear indirect prompt injection surface.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill makes runtime requests to https://api.you.com/mcp (configured as the MCP server_url and used by hostedMcpTool / MCPServerStreamableHttp), which is a required dependency and returns raw web/tool results that get injected into the agent's context — providing a direct indirect-prompt-injection surface that can control agent behavior.