The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill performs local statistical analysis (Population Stability Index, KS-test) on user-provided CSV files.
[COMMAND_EXECUTION]: The skill uses uv run to execute a local Python script (scripts/detect_drift.py). This is standard behavior for managing dependencies and executing skill logic within the designated environment.
[EXTERNAL_DOWNLOADS]: The documentation mentions well-known machine learning libraries such as evidently, nannyml, and river. These are industry-standard tools for MLOps and monitoring, and their inclusion in documentation or dependencies is considered safe practice.
[DATA_EXFILTRATION]: No evidence of data exfiltration. The script reads local CSV files and writes a local HTML report. There are no unexpected network calls to non-whitelisted domains or unknown external services.
[PROMPT_INJECTION]: The instructions are focused on guiding the agent to perform drift detection tasks. No attempts to bypass safety filters, override system prompts, or extract internal configuration were identified.

drift-detect