Deep Interview

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

Interactive Hook Protocol

MANDATORY: Use vscode_askQuestions for ALL user-facing questions in this skill (when available). If vscode_askQuestions is NOT available (e.g., Copilot CLI), present numbered options in markdown and ask the user to respond with a number or freeform text. This ensures structured input collection with selectable options, consistent UX, and clear decision tracking.

When to Fire Hooks

| Trigger Point | Question Type | Options Required | |---------------|---------------|------------------| | Phase 2 each round | Ambiguity-targeted question | 3-5 options + freeform | | Phase 3 challenges | Assumption validation | Yes/No + "It depends..." | | Phase 4 spec review | Confirm crystallized spec | Approve / Revise / Add constraints | | Phase 5 execution bridge | Choose next workflow | 5 predefined options |

Hook Format Rules

• header: Short unique ID, e.g. "interview-round-3", "spec-approval"
• question: The Socratic question targeting the weakest clarity dimension
• options: Provide 3-5 selectable answers that represent likely user intents
• First option: most common/expected answer (mark as recommended)
• Last option: always include an "Other / Let me explain..." escape hatch
• allowFreeformInput: Always true — user can override any option with their own words
• After receiving the answer, score ambiguity immediately and report progress

Example Hook Call

vscode_askQuestions([{
  header: "goal-scope",
  question: "What's the primary output you expect? (Ambiguity: 72% → targeting Goal Clarity)",
  options: [
    { label: "A working CLI tool", recommended: true },
    { label: "A library/SDK for other developers" },
    { label: "A web application with UI" },
    { label: "A background service/daemon" },
    { label: "Other / Let me explain..." }
  ],
  allowFreeformInput: true
}])

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 30%)

Phase 4: Crystallize Spec

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

• Goal, Constraints, Non-Goals, Acceptance Criteria
• Assumptions Exposed & Resolved
• Ontology (Key Entities) with convergence tracking
• Interview Transcript
• HOOK: Spec approval — present spec summary via vscode_askQuestions:

`` header: "spec-approval" question: "Review the crystallized spec. Ready to proceed?" options: [Approve & continue, Revise specific sections, Add more constraints, Restart interview] ``

Phase 5: Execution Bridge

HOOK: Present execution options via vscode_askQuestions:

header: "execution-bridge"
question: "Spec is complete (ambiguity: X%). Choose execution path:"
options: [
  { label: "Ralplan → OMG Autopilot", description: "Consensus-refine then execute", recommended: true },
  { label: "Execute with omg-autopilot", description: "Skip ralplan, direct execution" },
  { label: "Execute with ralph", description: "Persistence loop" },
  { label: "Execute with team", description: "Parallel agents" },
  { label: "Refine further", description: "Continue interviewing" }
]

Rules

• Ask ONE question at a time
• Use vscode_askQuestions for user-facing questions when available; in CLI, present numbered markdown options
• 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
• Include ambiguity % in every question's text so user sees progress

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: jmstar85
• Source: jmstar85/oh-my-githubcopilot
• License: MIT
• Homepage: https://www.linkedin.com/in/minsungjung/

Install and usage instructions live in the source repository linked above.