What problem does it solve?
Deep Interview protects autonomous execution from misalignment by turning vague ideas into rigorous specifications through structured Socratic questioning, ambiguity scoring, and gating before any build phase.
Core Features & Use Cases
- Socratic questioning loop targets the weakest clarity dimension each round, demands one question at a time, and highlights assumptions that must be settled before moving forward.
- Mathematical ambiguity scoring combines goal, constraint, and success criteria clarity (plus context for brownfield) to calculate a threshold (default 20%) that must be met, with ontology tracking to show convergence.
- Pipeline-ready delivery persists interview state, writes crystalized specs, and hands off to omc-plan → autopilot pipelines or autoresearch when prompted, so a CTO-level workflow yields reliable requirements; for example, use it when a stakeholder only offers a high-level vision and you need a precise spec before coding.
Quick Start
Ask Claude to start a deep interview on your idea so it can question assumptions, score clarity, and produce a spec before execution.