AI Agent Design Skill

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly shows agents performing web research and tool calls (e.g., "step_1_research" with "tool_call: search_web" in the observability trace and a "WebSearch" agent in the Hierarchical Teams diagram), which indicates the agent fetches and interprets open/public web content as part of its workflow and that content could materially influence planning and tool use.