scenario-3d
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
curl -Lto download 3D assets (e.g., GLB, FBX files) from URLs generated by theasset_downloadtool. These downloads are directed toward the service's own infrastructure for asset retrieval. - [COMMAND_EXECUTION]: The skill utilizes shell commands for its core functionality, specifically
curlfor asset retrieval andnpxfor installing supplementary skills from the vendor's own repository (scenario-labs/skills). - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input (prompts for 3D generation and mesh editing) and interpolates them into model execution calls.
- Ingestion points: User-supplied prompts are used as parameters for the
model_runandrecommendtools inSKILL.md. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands within user input are provided.
- Capability inventory: The skill possesses capabilities to perform network downloads (
curl), install additional executable skills (npx), and interact with 3D generation models. - Sanitization: No evidence of sanitization, escaping, or validation of user-provided prompt strings is present in the instructions.
Audit Metadata