quark-onnx-model-intake
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to intake and analyze external ONNX model files, creating a surface for indirect prompt injection through malicious model metadata or configurations.
- Ingestion points: External .onnx files specified by the user.
- Capability inventory: Inspection of model properties like opset, IR version, and input/output shapes.
- Boundary markers: None defined in this metadata stub.
- Sanitization: Not specified for model metadata analysis.
- [NO_CODE]: The provided file contains only YAML frontmatter and a redirection instruction, with no internal scripts or executable logic to evaluate for direct security threats.
Audit Metadata