Claude Interview Mode

🚀 claude-interview-mode

An MCP server that transforms Claude into a structured interviewer, becoming smarter with each conversation. It features a shared evolution system that scores, ranks, and recommends checkpoints based on real - world usage across all users.

🚀 Quick Start

To get started with claude-interview-mode, you can use either of the following installation methods:

npx claude-interview-mode

Or, if you prefer a global installation:

npm install -g claude-interview-mode

Then, add the following to your project's .mcp.json:

{
  "mcpServers": {
    "interview-mode": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "claude-interview-mode"]
    }
  }
}

Restart your Claude Code session to load the MCP server. The evolution system will start working right away through a shared community database.

✨ Features

• Claude Drives the Interview: Claude takes the lead in the interview, asking questions, presenting options with reasoning, and challenging assumptions.
• Tracks Q&As and Decisions: It maintains structured records of Q&As and decisions, complete with timestamps.
• Evolving Checkpoints: The system learns which topics are important for each category and ranks them using Bayesian effectiveness scores.
• Recommended Paths: Based on past interview patterns, it suggests the optimal order to explore topics.
• Concurrent Sessions: Supports multiple interviews running simultaneously.
• Privacy - First: Only anonymous metadata (categories, checkpoint names, counts) is sent to the shared database.

📦 Installation

Quick Install

npx claude-interview-mode

Global Installation

npm install -g claude-interview-mode

Setup with Claude Code

Add the following configuration to your .mcp.json:

{
  "mcpServers": {
    "interview-mode": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "claude-interview-mode"]
    }
  }
}

Optional: Use Your Own Supabase

If you want to use your private database, modify the configuration as follows:

{
  "mcpServers": {
    "interview-mode": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "claude-interview-mode"],
      "env": {
        "SUPABASE_URL": "https://your-project.supabase.co",
        "SUPABASE_ANON_KEY": "your-anon-key"
      }
    }
  }
}

Then, run supabase/schema.sql in your Supabase SQL Editor to create the necessary tables.

💻 Usage Examples

Basic Usage

Start an interview in Claude Code:

> Let's do an interview about my SaaS pricing strategy

Claude will guide the conversation. As the interview progresses:

• Each Q&A and decision is recorded along with checkpoint coverage.
• At the end, metadata is uploaded to improve the system.
• The next time someone conducts an interview in the same category, the enhanced checkpoints will be loaded.

Tools

📚 Documentation

The Evolution System

This is not just an interview tool but a collectively evolving knowledge system. Every time an interview is conducted in a category (e.g., "saas - pricing"), the system learns:

Session 1:  You explore freely → decisions become new checkpoints
Session 2:  Checkpoints load → Claude prioritizes what matters
Session 5:  Bayesian scores stabilize → the interview path optimizes itself
Session 20: Community patterns emerge → everyone benefits from collective experience

How evolution works

1. Checkpoint Discovery: During an interview, when a decision is made, its topic is automatically registered as a new checkpoint. After a few sessions, the system can identify the important topics for each category.
2. Bayesian Scoring: Each checkpoint tracks how often it is covered and how often it leads to a decision. The score uses Bayesian smoothing to handle sparse data:

decision_rate = (decisions + 0.6) / (times_covered + 2)

The prior (0.6/2 = 30% base rate) ensures that new checkpoints start with a reasonable score. After about 5 sessions, real data takes over. 3. Composite Ranking: Checkpoints are ranked by a composite score that combines decision - leading effectiveness (70%) and usage frequency (30%):

composite = decision_rate × 0.7 + normalized_usage × 0.3

High - scoring checkpoints are those that consistently lead to concrete decisions, not just frequently discussed topics. 4. Recommended Path: The system calculates an optimal interview path: checkpoints with decision_rate > 0.2, sorted by their average position in past sessions. This helps Claude know when to ask specific questions. 5. Community Evolution: All metadata is stored in a shared database. When interviewing about "api - design", you can benefit from the experiences of other users who have conducted similar interviews. The checkpoints, scores, and paths evolve collectively.

What gets shared (and what doesn't)

🔧 Technical Details

Architecture

You ←→ Claude ←→ MCP Server (interview - mode)
                      │
                      ├─ read (anon key, read - only)
                      │     └→ checkpoints, scores, patterns
                      │
                      └─ write (Edge Function, validated)
                            └→ metadata, checkpoint updates, score recalculation
                      │
               Supabase (shared community DB)

4 database tables power the evolution:

Security:

• The anonymous key is read - only (SELECT only via RLS).
• All writes are processed through an Edge Function with input validation and spam defense.
• Empty interviews, implausible rates, and oversized payloads are rejected.

Development

git clone https://github.com/teabagkim/claude - interview - mode.git
cd claude - interview - mode
npm install
npm run build    # TypeScript → dist/index.js
npm run dev      # Watch mode

📄 License

This project is licensed under the MIT license.