The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill utilizes the browser_run_code tool to execute custom JavaScript within the browser context to interact with Notion. This provides powerful control over the session but introduces a significant attack surface if used improperly. Evidence is found throughout SKILL.md (Step 4) and REFERENCE.md.
[COMMAND_EXECUTION]: The documentation and code templates in REFERENCE.md and EXAMPLES.md demonstrate unsafe string interpolation where user-controlled variables (e.g., ${content}) are directly embedded into JavaScript code strings. This creates a vulnerability where malicious input could break out of string literals to execute arbitrary code within the browser context (XSS-like attack).
[INDIRECT_PROMPT_INJECTION]: The skill is designed to navigate to and process data from external Notion pages provided by users without implementing boundary markers or sanitization. This makes the agent susceptible to instructions embedded in those external pages. 1. Ingestion points: External Notion pages via URL navigation and DOM snapshots. 2. Boundary markers: Absent. 3. Capability inventory: Includes browser navigation, element interaction, and arbitrary JavaScript execution. 4. Sanitization: No validation or escaping logic is prescribed for content retrieved from external sources.

notion-docs-enhancer