visualize-in-3d
Pass
Audited by Gen Agent Trust Hub on Nov 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads 3D asset packs and terrain data from Hugging Face (
codepawlaccount) and OpenStreetMap. It also installs Node.js packages such as@codepawlhq/cruciblefrom the official npm registry. - [COMMAND_EXECUTION]: The instructions involve executing system commands for Blender, Node.js, Python, and FFmpeg to process assets, perform audits (e.g.,
scripts/arm-audit.mjs), and render media. - [DYNAMIC_EXECUTION]: The skill dynamically generates and executes scripts for Blender and web applications (Vite/React) to create customized 3D scenes based on user input.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, including user-provided subjects, 3D model metadata from various sources, and map data. This represents an ingestion surface where embedded instructions could theoretically influence the agent's script generation or tool invocation, though the risk is mitigated by the specific technical nature of the task.
Audit Metadata