The Agent Skills Directory

[SAFE]: The skill implements structured methodologies for HR analytics, incorporating explicit legal and ethical constraints to prevent the use of protected attributes.
[SAFE]: Code snippets provided for feature engineering and disparate impact auditing utilize standard, well-known data science libraries (pandas, scikit-learn, numpy, xgboost) without any remote execution or network exfiltration logic.
[SAFE]: The documentation includes 'Iron Law' constraints and validation gates to ensure that model outputs are treated as probabilistic risks rather than definitive employment decisions, effectively mitigating the risk of automated decision-making abuse and ensuring compliance with Taiwan PDPA and international standards.

algo-hr-turnover