The Agent Skills Directory

[SAFE]: The skill provides guidelines and scripts for auditing sports modeling data for leakage, which is a common and legitimate data science activity.
[REMOTE_CODE_EXECUTION]: The script audit_pregame_features.py utilizes the pandas library for data analysis. This is an industry-standard, well-known, and trusted package.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided data files (CSV, Parquet, JSON) through its audit script. While this represents an ingestion point for untrusted data, the script only performs statistical heuristics and does not possess capabilities that would allow for malicious command execution or data exfiltration.

leakage-audit