pydantic-ai

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly documents fetching and ingesting open web and user-generated content — e.g., references/agents.md shows ImageUrl/VideoUrl/DocumentUrl usage and references/tools.md plus references/tools.md and references/models.md list built-in WebSearchTool/WebFetchTool and Exa/duckduckgo/YouTube examples that the agent consumes for summarization/search — meaning untrusted third-party pages can be read and fed into the agent's workflow.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes an MCP server configured with a remote URL (MCPServerTool url='https://mcp.deepwiki.com/mcp') which is used at runtime to invoke external tools on that server — allowing remote code/tool execution that can directly affect the agent's behavior and responses.