open-pencil

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow includes operating on a user-supplied .fig document (e.g., via open_file/open-pencil ... design.fig and MCP get_page_tree/query_nodes/get_jsx/describe), and that document’s text content (e.g., node names, TEXT nodes) is read and can be injected into the LLM context as free-form design text.