Purpose

Generate MATLAB / 北太天元 compatible modeling code from a validated candidate method pool and cleaned data.

This skill implements the approved candidate methods (baseline, main candidates, optional improved) as .m scripts. It should produce code that reads cleaned data, runs the planned models, saves results in the experiments/roundN/ output structure, generates a run_summary.json, and produces portable artifacts for downstream experiment reports, robustness checks, and paper writing.

This skill does not choose the model, clean raw data, review code, run final QA, or write paper sections.

When to use

Use this skill:

• After model-code-analyzer hands off to matlab-model-code-generator.
• When implementation.target = matlab.
• When the contest requires MATLAB / 北太天元 compatible scripts.
• When the candidate method pool and cleaned data are ready.
• When executable .m scripts are needed for one or more candidate methods in a specific experiment round.

Preconditions