mmk-notion

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and processes Notion user-generated content (e.g., page get/update, list-published, "AI summary"/"AI summary auto" on databases, and "List AI meeting notes") via the mmk CLI as described in SKILL.md, so the agent would read untrusted third‑party (published/guest) Notion content that could influence actions.