mcpField
D
deep-interview

deep-interview

github.com/zereightAgent workflows2.1k57 installsUpdated today
Repository

About deep-interview

deep-interview is a requirements-gathering skill for vague or complex ideas, using Socratic questions and ambiguity scoring to develop a specification. It can bridge the resulting spec to refinement and execution workflows.

Key features

  1. 01Asks one targeted question per round
  2. 02Scores ambiguity across clarity dimensions
  3. 03Challenges assumptions in later rounds
  4. 04Writes a structured specification
  5. 05Offers paths to refinement and execution

Use cases

  1. 01Clarify a loosely described product idea
  2. 02Gather requirements before a complex coding task
  3. 03Expose assumptions before selecting an execution path

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/zereight/gitlab-mcp --skill deep-interview

After installing, type/deep-interviewin Claude Code to confirm it loaded.

Details

From the project README

Deep Interview

Ouroboros-inspired Socratic questioning with mathematical ambiguity scoring. Replaces vague ideas with crystal-clear specifications by asking targeted questions that expose hidden assumptions.

Pipeline

deep-interview → ralplan (consensus refinement) → omg-autopilot (execution)

When to Use

  • User has a vague idea and wants thorough requirements gathering
  • Task is complex enough that jumping to code would waste cycles
  • User wants mathematically-validated clarity before execution

When NOT to Use

  • Detailed specific request with file paths → execute directly
  • Quick fix → delegate to @executor or /ralph
  • User says "just do it" → respect their intent

Phases

Phase 1: Initialize
  1. Parse the user's idea
  2. Detect brownfield vs greenfield (use @explore to check codebase)
  3. For brownfield: map relevant codebase areas
  4. Initialize ambiguity score at 100%
Phase 2: Interview Loop

Repeat until ambiguity <= 20% or user exits early:

  1. Generate question targeting the WEAKEST clarity dimension
  2. Ask ONE question at a time with current ambiguity context
  3. Score ambiguity across dimensions:
    • Goal Clarity (40% weight for greenfield, 35% brownfield)
    • Constraint Clarity (30% / 25%)
    • Success Criteria (30% / 25%)
    • Context Clarity (N/A / 15% for brownfield)
  4. Report progress with dimension scores and gaps
  5. Track ontology (key entities, stability ratio)
Phase 3: Challenge Agents
  • Round 4+: Contrarian - challenge core assumptions
  • Round 6+: Simplifier - probe for complexity removal
  • Round 8+: Ontologist - find the essence (if ambiguity > 30%)
Phase 4: Crystallize Spec

When ambiguity <= threshold, generate spec to .omc/specs/deep-interview-{slug}.md:

  • Goal, Constraints, Non-Goals, Acceptance Criteria
  • Assumptions Exposed & Resolved
  • Ontology (Key Entities) with convergence tracking
  • Interview Transcript
Phase 5: Execution Bridge

Present options:

  1. Ralplan → OMG Autopilot (Recommended): consensus-refine then execute
  2. Execute with omg-autopilot (skip ralplan)
  3. Execute with ralph: persistence loop
  4. Execute with team: parallel agents
  5. Refine further: continue interviewing

Rules

  • Ask ONE question at a time
  • Target the WEAKEST clarity dimension each round
  • Gather codebase facts via @explore BEFORE asking user
  • Score ambiguity after every answer
  • Do not proceed until ambiguity <= threshold (default 20%)
  • Hard cap at 20 rounds, soft warning at 10

Comments

Markdown supported