Purpose

Translate validated mathematical modeling plans into executable code thinking.

This skill does not write runnable code itself. It turns the selected modeling route (from the candidate method pool) into a language-neutral coding plan: what each subquestion needs, which data enters each stage, which intermediate artifacts should be saved, how code should be split by subquestion, how the experiments/roundN/ output structure must be organized, and which language-specific generator should implement the plan next.

A critical function of this skill is defining the output structure so that experiment reproducibility, multi-method comparison, and downstream report generation are possible.

When to use

Use this skill:

• After data-auditor-cleaner has confirmed that required data is ready or ready with documented warnings.
• After method-selector has produced the candidate method pool per subquestion.
• Before python-model-code-generator or matlab-model-code-generator.
• When the team needs a code-thinking document under code/ before writing executable scripts.
• When a new round of experiments is being planned and the code structure needs to be defined.

Preconditions