skills/jeremylongshore/tons-of-skills-marketplace/building-classification-models/Gen Agent Trust Hub
building-classification-models
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external datasets, creating a surface for potential indirect prompt injection. 1. Ingestion points: The skill reads dataset paths and model configurations provided by the user (SKILL.md). 2. Boundary markers: The instructions lack explicit delimiters or safety markers to differentiate between data and instructions during processing. 3. Capability inventory: The skill is configured to use the Bash tool for command execution and has file system operations enabled (SKILL.md). 4. Sanitization: There is no evidence of sanitization or filtering of dataset content to prevent instruction injection.
- [DYNAMIC_EXECUTION]: The skill is designed to generate Python code for training classification models and then execute that code to evaluate performance (SKILL.md). This runtime script generation and execution represents a dynamic execution pattern.
Audit Metadata