Multi-method assessment comparison

Source: imported from the ChatGPT second opinion.

Run two or more assessment methodologies (qualitative-Bayes, expert-deference, causal-model, adversarial red-team) on the same subquestions, then a compare artifact naming where they agree, where they diverge, which evidence each overweights, and which cruxes are method-sensitive. This tracks FLF's "stepping-stone entry" framing.

Pure verdict-layer work, so it lives entirely in epistack-casebooks, downstream of the neutral map (see adjudication as a separate layer).