agent-ops

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

• Third-party content exposure detected (high risk: 1.00). The skill's required workflows explicitly call external retrieval functions—e.g., execute_tool uses search_web(...) in references/agent-patterns.md and evaluation flows use retrieve_documents/retrieve_context in references/evaluation.md—and even include sanitize_external_content that mentions "retrieved docs, web pages", so the agent is expected to ingest untrusted public web/third-party content that can influence tool selection and responses.