dataset-curator

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely informational and consists of documentation, architectural guidance, and illustrative code snippets for data science practitioners. No automated execution or dangerous command patterns were detected.
  • [DATA_EXFILTRATION]: The instructions demonstrate a strong security posture by recommending that users strip Personally Identifiable Information (PII) including names, emails, and account numbers from datasets before they are used for training or annotation.
  • [EXTERNAL_DOWNLOADS]: The text references well-known, reputable Python libraries such as datasketch, cleanlab, and Hugging Face datasets as recommended tools for deduplication and quality filtering. These are industry-standard utilities in the machine learning ecosystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 04:46 PM
Security Audit — agent-trust-hub — dataset-curator