TSFM Forecast Skill

Generates Python + DuckDB forecasting pipeline code using foundation models. Produces runnable code for local execution; does NOT run models.

When to Use This Skill

Activate when: Generating zero-shot time-series forecasting pipelines, selecting between TimesFM / Chronos / MOIRAI / Lag-Llama, preparing time-series data with DuckDB, building backtesting harnesses, comparing model accuracy, or producing client forecast deliverables.

Don't use for:

• ML model training or fine-tuning → use python-data-engineering
• Real-time or streaming forecasts → use event-streaming
• Scheduling or orchestrating forecast jobs → use data-pipelines
• Loading raw files into DuckDB without forecasting intent → use duckdb

Scope Constraints