multi-agent-patterns

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly instructs and examples researcher agents that perform "web search, document retrieval" and uses tools like [search, browse, read] (see SKILL.md "Researcher (web search, document retrieval)" and references/frameworks.md create_agent with tools [search, browse, read]), so agents are expected to ingest and act on open/public third‑party web content that can influence downstream tool use and decisions.