mindmap
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The mind map template references and downloads font files from Google Fonts (fonts.googleapis.com and fonts.gstatic.com). These resources are only used for styling the mind map and represent a safe use of a well-known service.
- [COMMAND_EXECUTION]: The skill includes a local validation utility (
scripts/validate.py) and instructs the agent to execute it to verify generated mind map artifacts. This script performs safety and quality checks, such as ensuring no unauthorized external resources are loaded and that labels do not overlap. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied topics and descriptions, representing an attack surface for untrusted data ingestion.
- Ingestion points: Topic, audience, and depth parameters defined in
SKILL.md. - Boundary markers: The agent is instructed to encapsulate the processed data within a specific JSON
<script>block in the HTML output. - Capability inventory: The skill can write files and execute a local validation script using the Playwright browser automation library.
- Sanitization: The instructions mandate that the agent summarize and reformat user input into a rigid JSON schema with strict length constraints, preventing raw input from controlling the execution environment.
Audit Metadata