light-research-plan
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill includes local Python scripts (
plan_lint.py,power_check.py) for validating research matrices and calculating statistical power. Analysis shows these scripts perform benign data processing and mathematical calculations on local project files. - [SAFE]: External documentation and repository links point to well-known research and development platforms such as arXiv, OSF, and official documentation for established tools like DVC, MLflow, and Hydra. These are recognized as trusted sources for research metadata and technical guidelines.
- [SAFE]: The skill demonstrates proactive security measures by explicitly instructing users to use offline modes for experiment tracking services (Weights & Biases, MLflow) when processing sensitive data, effectively mitigating potential data exfiltration risks.
- [SAFE]: Instructions regarding network-connected tasks, such as citation verification via DOI, are documented as part of the core research functionality and are directed toward legitimate academic database endpoints.
Audit Metadata