machine-learning
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external datasets for model training, hyperparameter optimization, and interpretability analysis, creating a potential surface for indirect prompt injection if the data contains malicious instructions or is used to influence the agent's decision-making flow.
- Ingestion points: Dataset processing within the
hyperparameter_searchandexplain_modelfunctions inSKILL.md. - Boundary markers: Absent; the implementation uses standard function arguments for numpy arrays without specific delimiters or warnings against embedded instructions.
- Capability inventory: The skill possesses capabilities for heavy computation (training loops, hyperparameter trials) but lacks capabilities for network operations, file system writes, or subprocess execution.
- Sanitization: No data validation or sanitization logic is implemented for the processed arrays.
Audit Metadata