llm-app-patterns

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md explicitly defines web-retrieval in core workflows—e.g., the ReAct pattern uses actions like search("X") and the FunctionCallingAgent includes a "search_web" tool that searches the open web—so the agent is expected to ingest and act on untrusted public web/search results which could carry indirect prompt injections.