deep-interview

Structured requirement gathering through Socratic questioning with mathematical ambiguity scoring. Gates execution behind a 20% ambiguity threshold to prevent expensive work on underspecified tasks.