Can I use Deep Interview on existing projects?

Yes. It includes a 'brownfield' mode that uses an explore agent to map your current codebase, ensuring that the interview questions are grounded in your existing technical context.

How does this differ from the omc-plan skill?

Deep Interview is for the discovery phase when requirements are still fuzzy. omc-plan is used once the requirements are clear and you need a specific technical roadmap for implementation.

What is the Deep Interview skill for Claude Code?

It is a specialized capability that uses Socratic questioning to help users refine vague ideas into detailed, executable technical specifications before any code is written.

How does the ambiguity scoring work?

The skill evaluates your project across four dimensions: Goal Clarity, Constraint Clarity, Success Criteria, and Context. It calculates a weighted score and usually requires 80% clarity before proceeding to execution.

Deep Interview

Name: Deep Interview
Author: Yeachan-Heo

byYeachan-Heo

•

28,000

•

Productivity & Workflow

Refines vague project ideas into precise specifications using Socratic questioning and mathematical ambiguity scoring.

Deep Interview is a requirements-gathering powerhouse for Claude Code that eliminates project drift before execution begins. Inspired by the Ouroboros project, it replaces broad assumptions with a Socratic questioning loop that probes the weakest areas of a project's definition. By measuring clarity across four weighted dimensions—Goal, Constraints, Success Criteria, and Context—the skill ensures that the AI possesses a mathematically validated understanding of your intent. Whether starting a new project or modifying a complex codebase, Deep Interview bridges the gap between a vague concept and a production-ready implementation plan, ensuring autonomous agents build exactly what you need.

Key Features

01Socratic Questioning Loop: Iteratively probes hidden assumptions with targeted, one-at-a-time questions.

02Mathematical Ambiguity Gating: Real-time scoring of project clarity with a configurable threshold for execution readiness.

03Brownfield Codebase Analysis: Automatically explores existing repositories to align new requirements with current architecture.

0428,000 GitHub stars

05Multi-Agent Handoff: Generates detailed spec files that feed directly into planning and autopilot execution pipelines.

06Autoresearch Integration: Specialized mode to define missions and evaluators for zero-learning-curve research setups.

Use Cases

01Converting a high-level 'vague idea' into a concrete, executable project specification.

02Defining rigorous success criteria and mission briefs for autonomous AI research tasks.

03Aligning new features with existing complex codebases to avoid architectural inconsistencies.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yeachan-heo/oh-my-claudecode deep-interview

For use in Claude.ai and ChatGPT

Key Features

01Socratic Questioning Loop: Iteratively probes hidden assumptions with targeted, one-at-a-time questions.

02Mathematical Ambiguity Gating: Real-time scoring of project clarity with a configurable threshold for execution readiness.

03Brownfield Codebase Analysis: Automatically explores existing repositories to align new requirements with current architecture.

0428,000 GitHub stars

05Multi-Agent Handoff: Generates detailed spec files that feed directly into planning and autopilot execution pipelines.

06Autoresearch Integration: Specialized mode to define missions and evaluators for zero-learning-curve research setups.

Use Cases

01Converting a high-level 'vague idea' into a concrete, executable project specification.

02Defining rigorous success criteria and mission briefs for autonomous AI research tasks.

03Aligning new features with existing complex codebases to avoid architectural inconsistencies.