visualize
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's build system and templates fetch required visualization engines and geographic data from well-known content delivery networks.
- References to D3, Vega-Embed, TopoJSON, and Mermaid are sourced from
cdn.jsdelivr.net. - Map data (US Atlas) is loaded at runtime from official CDN sources.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing user-supplied datasets and labels, which are transformed into HTML and Markdown artifacts.
- Ingestion Points: User data entering via Phase 2 (Research) and Phase 3 (Build) for chart generation.
- Boundary Markers: The skill uses templated sections and scoped scripts to separate chart logic, though raw data values are interpolated into JavaScript strings.
- Capability Inventory: Python scripts (
visualizer.py,build_viz.py) perform file write and delete operations within a dedicated user directory (~/.visualizations). - Sanitization: Filenames are sanitized via a slugify function, and static analysis is provided by
check_render.pyto identify potential rendering defects. - [COMMAND_EXECUTION]: The skill includes management scripts and provides clear instructions for shell-based workflows.
scripts/build_viz.py: Assembles visualization fragments.scripts/visualizer.py: Manages stored visualization files.scripts/check_render.py: Performs static verification of generated artifacts.
Audit Metadata