model-drift-detector

Quick Review

This skill is essentially a template placeholder with no substantive content. The description is circular and vague ('model drift detector' repeated without explanation of what it does). It lacks any actual task knowledge - no guidance on statistical tests for drift detection (KS test, PSI, etc.), no code for comparing distributions, no monitoring setup, and no thresholds or alerting logic. While the structure is clear, it's only because there's minimal content. The novelty score is low because detecting model drift (comparing training vs. production data distributions) is a well-defined ML task that a capable LLM could guide through without a dedicated skill. A useful version would include: drift detection algorithms, feature/prediction distribution comparison code, visualization of drift metrics, integration with monitoring systems, and actionable remediation steps.