The Agent Skills Directory

[SAFE]: The core functionality is implemented through modular Python scripts (validate_data.py, transform_data.py, handle_errors.py, and pipeline.py) that perform legitimate data science operations such as normalization, categorical encoding, and data integrity checks.
[COMMAND_EXECUTION]: The orchestration script pipeline.py uses the subprocess module to execute internal utility scripts. It correctly passes arguments as a list and avoids using shell=True, which follows security best practices for preventing shell injection.
[DATA_EXFILTRATION]: The skill does not contain any code for network communication (e.g., requests, urllib, or socket). It operates exclusively on local files provided via command-line arguments, and no hardcoded credentials or sensitive file paths were found.
[PROMPT_INJECTION]: The instructions in SKILL.md are focused on the intended data processing task and do not contain any patterns typical of jailbreaks, role-play injections, or attempts to bypass agent safety guidelines.

preprocessing-data-with-automated-pipelines