The Agent Skills Directory

[SAFE]: The skill utilizes well-known and trusted Python packages, specifically statsforecast (maintained by Nixtla) and pandas, for its core forecasting logic. These are standard industry tools for time-series analysis.
[SAFE]: File operations are limited to reading input data from history.csv and writing forecast results to forecast.csv, which is consistent with the skill's stated purpose of time-series projection.
[SAFE]: The included utility script scripts/verify.sh performs read-only validation of CSV headers and row counts. It does not perform any network operations, administrative actions, or destructive file system modifications.
[SAFE]: The skill implements strong security and data science best practices, including requirements for backtesting (rolling-origin cross-validation) and mandatory accuracy reporting (WAPE, Bias, MASE) to ensure model validity before deployment.

forecasting