tools-ui

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

• Third-party content exposure detected (high risk: 0.80). The Agent example and ToolResult/ToolCall components explicitly include a "search_web" tool and show rendering of tool results (e.g., ToolResult with results and the Agent config listing name: 'search_web' — i.e., web search results from the open web would be consumed and displayed), so the skill would surface untrusted, third‑party web content to the agent.