01
Every score points to a sentence
Script Analysis scores your text on 16 fixed sub-criteria, and the meaning of each score from 0 to 4 is written down in advance. The AI has to back every score with a verbatim quote; the system checks each quote against your original text and lowers any high score whose evidence cannot be found. The overall score is calculated by code, not guessed by the model. Policy Audit and Scenario Analysis follow the same rule: no quote, no finding.

