x-meeting-reports

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

• Third-party content exposure detected (high risk: 0.80). The skill's "Matching existing structure" workflow explicitly tells the agent to fetch the Notion database and example Notion pages via notion-fetch (e.g., the provided https://www.notion.so/... URLs) and to read their properties/structure to determine page properties and content, so it ingests user-generated third-party Notion content that can materially influence how the agent constructs and creates pages.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly calls Notion at runtime (e.g., fetching the database URL https://www.notion.so/d30ba60af8344e99b97a94428f838ccb?v=fbe020906b5c4f98a1b792d87883f76f and example pages) and uses the fetched page/schema content to determine property names and page structure, which directly steers the agent's generated instructions and is required for operation.