threejs-3d-generator
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined prompts and images to generate 3D assets through an external API. This creates an attack surface where maliciously crafted inputs could potentially lead to unintended behavior if reflected in agent outputs or used in multi-step workflows.\n
- Ingestion points: The text and image commands in scripts/threejs_3d_asset.py ingest untrusted text and image files via CLI arguments.\n
- Boundary markers: The skill does not use specific delimiters to wrap user input when sending it to the external API.\n
- Capability inventory: The agent can perform network requests to api.tripo3d.ai and write files to the local filesystem.\n
- Sanitization: Input is used directly in API request payloads with minimal content validation.\n- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute a local Python script (scripts/threejs_3d_asset.py) to manage the asset generation lifecycle.\n- [DATA_EXFILTRATION]: The skill performs network operations to api.tripo3d.ai to interact with the Tripo 3D generation service. This is a functional requirement but involves sending data to a non-whitelisted external domain.\n- [EXTERNAL_DOWNLOADS]: The skill communicates with api.tripo3d.ai to download generated 3D assets (GLB, FBX) and preview images. It also references the use of fflate from the JSDelivr CDN in the integration guidelines.
Audit Metadata