web3d-animation
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use
npx @gltf-transform/clifor inspecting and validating 3D assets (GLB/glTF). This is a standard industry-standard tool for 3D model processing and is used here for legitimate technical analysis. - [INDIRECT_PROMPT_INJECTION]: The skill processes metadata from external 3D asset files, which could theoretically contain malicious strings. However, this is a standard requirement for 3D workflows and the risk is minimal given the technical nature of the data extracted.
- Ingestion points: Asset files (.glb, .gltf) inspected via
npxor direct JSON reading. - Boundary markers: None explicitly defined in the prompts.
- Capability inventory: Command execution (
npx), file system access (reading assets), and code generation for Three.js. - Sanitization: Not explicitly implemented; the skill relies on standard tool outputs.
- [SAFE]: External references to 3D service providers (Meshy, Tripo) and libraries (Three.js, glTF-transform) are well-known industry resources. The use of Cyrillic characters in triggers and documentation is legitimate and does not contain homoglyph-based obfuscation.
Audit Metadata