exploratory-data-analysis

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is well-structured and focuses on providing descriptive analysis of scientific datasets without performing high-risk operations.
  • [DATA_INGESTION]: The skill processes a wide range of external scientific data formats (FASTA, CSV, HDF5, etc.). This represents an inherent attack surface for indirect prompt injection; however, the scripts/eda_analyzer.py script treats these files as data for statistical analysis rather than executing their content, which mitigates the risk.
  • [CODE_SAFETY]: The Python analysis script uses standard, reputable libraries like NumPy, Pandas, and Biopython. It performs safe file handling by mapping extensions to a hardcoded whitelist before attempting to load reference information or perform specialized analysis.
  • [DOCUMENTATION]: The reference files include comprehensive guidance on handling various formats, including explicit warnings about security validation when dealing with serialized formats like Python Pickles.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 02:23 AM