readwise-chat

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly performs RAG over the user's full highlight library by accessing readwise.io/chat (requiring session cookies) and thus ingests user-generated / third-party highlight content which the agent reads and synthesizes into answers, allowing malicious or web-originated instructions in those highlights to influence its behavior.