readwise-mcp

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly ingests and displays arbitrary public web content and feeds (e.g., reader_create_document(url="...") — "Reader scrapes it automatically", reader_list_documents(location="feed"), and response_fields like "html_content"/"content") and the SKILL.md workflows call reader_get_document_details and then act on that content (triage, move/archive), so untrusted third-party pages/RSS items can influence agent decisions.