ml-engineer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions designed to override AI safety constraints or extract system prompts were detected.
- [Data Exposure & Exfiltration] (SAFE): The skill does not contain hardcoded credentials, access sensitive file paths, or perform unauthorized data transmissions.
- [External Downloads] (SAFE): References to libraries such as scikit-learn, PyTorch, and TensorFlow are legitimate and fall under trusted ecosystem usage.
- [Indirect Prompt Injection] (LOW): A data ingestion surface exists as the skill is designed to process user datasets for ML tasks. Evidence: 1. Ingestion points: Feature engineering and data processing stages (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: No dangerous execution capabilities (subprocess, eval, network) are present in the provided snippets. 4. Sanitization: Absent. Result: The lack of exploitable capabilities makes this surface low risk.
Audit Metadata