notion-meeting-intelligence

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

• Third-party content exposure detected (medium risk: 0.30). In the “Meeting Intelligence” workflow, the skill performs Notion:notion-search and then Notion:notion-fetch to read prior meeting context and specs from a user’s Notion workspace, which could contain outsider-authored free text if the user/workspace ingests it.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes a runtime Notion MCP endpoint (https://mcp.notion.com/mcp) that the agent uses to fetch Notion pages which are injected into the model context to drive prompt generation and outputs, and the skill relies on that MCP as a required dependency.