creating-visual-notes

Pass

Audited by Gen Agent Trust Hub on Apr 4, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted text input to generate visual diagrams, creating a surface for indirect prompt injection.
  • Ingestion points: The script scripts/generate-mindmap.js reads text data from input files provided as command-line arguments.
  • Boundary markers: No delimiters or 'ignore instructions' warnings are present in the script or usage instructions.
  • Capability inventory: The script scripts/generate-mindmap.js performs file read and write operations (fs.readFileSync, fs.writeFileSync).
  • Sanitization: The script performs minimal character escaping (removing quotes and brackets) for Mermaid syntax compatibility but does not sanitize the input for potential agent instructions.
  • [COMMAND_EXECUTION]: The skill includes instructions to execute Node.js scripts that perform file system operations (read and write) based on user-supplied file paths, which the agent could be directed to use on sensitive locations.
  • [NO_CODE]: The SKILL.md documentation references several scripts (scripts/create-timeline.js, scripts/build-comparison-table.js) and resource files (e.g., resources/visual-note-patterns.md) that are not included in the provided skill package.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 4, 2026, 01:44 PM
Security Audit — agent-trust-hub — creating-visual-notes