llamaparse

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

• Third-party content exposure detected (high risk: 0.85). The skill parses user-provided documents by uploading their file bytes to LlamaParse and then ingesting the returned extracted text/markdown (e.g., result.text_full / result.markdown_full) into the agent’s context; since the document content is not authored by the operating user, it is outsider-authored free text at runtime.