mmk-notion-page

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the vendor-provided mmk CLI to interact with the Notion API. It facilitates administrative tasks such as inviting users with specific roles, revoking access, and modifying page configurations (e.g., full-width, small-text).
  • [COMMAND_EXECUTION]: Provides access to destructive operations via the mmk notion page update command. This includes the --erase-content flag, which permanently deletes all content on a page, and flags for trashing, archiving, or locking pages.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it retrieves and processes external data from Notion pages and meeting transcripts which is then provided to the agent.
  • Ingestion points: Data enters the agent context through the markdown and transcript commands in SKILL.md.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard potential instructions embedded within the retrieved Notion content.
  • Capability inventory: The agent has the ability to perform high-impact actions through the update (erase content, archive, trash) and invite commands in SKILL.md.
  • Sanitization: No sanitization or filtering of the retrieved Notion content is specified before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 02:24 PM