data-cleaning

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides best practices for data normalization and validation. It includes deterministic patterns for deduplication and handling missing values.
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing standard, well-known data science libraries including pandas, pandera, pyarrow, polars, and ftfy. These are common dependencies for the stated purpose of data cleaning.
  • [COMMAND_EXECUTION]: A verification script (scripts/verify.sh) is provided to perform static analysis and a small local runtime test using a temporary directory. The script is used to ensure the data validation gate works as intended.
  • [DATA_EXFILTRATION]: No network operations or sensitive file access patterns were detected. Data handling is limited to reading raw tables and writing to local quarantine or output artifacts.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 09:37 PM
Security Audit — agent-trust-hub — data-cleaning