shap
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_EXECUTION]: The code examples in references/workflows.md and references/explainers.md demonstrate loading machine learning models using joblib.load(), torch.load(), and tf.keras.models.load_model(). These methods are susceptible to arbitrary code execution if the input files are from an untrusted source.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze user-provided models and datasets, creating a surface for indirect prompt injection.
- Ingestion points: User-provided model objects and datasets in SKILL.md and references/workflows.md.
- Boundary markers: No markers are used to isolate untrusted data from the agent's instructions.
- Capability inventory: File system interaction (model loading/saving) and library installations via shell commands.
- Sanitization: The skill does not implement validation or filtering of the ingested models or data.
- [EXTERNAL_DOWNLOADS]: The skill documentation includes commands to install the shap library and related dependencies from official package repositories.
Audit Metadata