exa-search

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

• Third-party content exposure detected (high risk: 1.00). The skill clearly fetches and ingests untrusted public web content (e.g., scripts/exa_contents.py using Exa's /contents with arbitrary URLs and livecrawl, scripts/exa_search.py using /search with get_context/outputSchema, and exa_research.py using /answer to synthesize sources), and that content is incorporated into RAG/context, summaries, and structured outputs that the agent will read and act on, so third-party pages could carry indirect prompt injections.