agent-patterns

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

• Third-party content exposure detected (high risk: 0.80). The skill's examples and code (notably resources/augmented-llm.md and dynamic-orchestration.md) explicitly retrieve and embed external/search results and third-party tool outputs (e.g., vector store/search, traditional search engines, external APIs/MCP) into the model's prompt context, meaning the agent will read and act on untrusted public/web-sourced content at runtime.