Model / calculation artifacts

Source: imported from the ChatGPT second opinion.

For cases where qualitative prose is not enough (LHC safety case, black-hole risk), a lightweight type for quantitative models with explicit variables, assumptions, a calculation, sensitivity analysis, interpretation, and failure modes. Not every assessment needs numbers — but when numbers matter, the assumptions should be inspectable and sensitivity-tested.

Directly useful for the LHC safety-case dependency chain; prove it on that case before generalizing.