scientific-eda-correlation

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements standard data science workflows using well-known libraries like pandas, seaborn, and matplotlib.
  • [DATA_EXPOSURE]: The skill writes output to local directories (results/ and figures/). This is standard behavior for data analysis tasks and does not involve sensitive system paths or external exfiltration.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided dataframes, the operations are limited to statistical calculations and visualization. The ingestion of untrusted data is confined to data analysis tools with no path to command execution or network exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 03:24 AM
Security Audit — agent-trust-hub — scientific-eda-correlation