ColdCode - AI Coding Assistant

Overview

ColdCode is a powerful AI assistant that lives in your terminal. It can understand your codebase, edit files, run commands, and handle entire workflows for you.

⚠️ Security Notice: By default, ColdCode executes all operations without confirmation for maximum productivity. For important projects, use coldcode --safe to enable manual approval for file modifications and command execution.

Installation

npm install -g @coolbeeai/coldcode

After installation, you can use these commands:

• coldcode - Primary command
• cold - Short alias

Windows Users

• Install Git for Windows: https://git-scm.com/download/win
  • ColdCode automatically prefers Git Bash/MSYS or WSL Bash when available
  • If neither is available, it will fall back to your default shell, but many features work best with Bash
• Use VS Code's integrated terminal rather than Command Prompt (cmd)
• Optional: Avoid spaces in the global npm prefix path to prevent installation issues
  • Example: npm config set prefix "C:\\npm" and reinstall global packages

Quick Start

Interactive Mode

coldcode
# or
cold

Non-Interactive Mode

coldcode -p "explain this function" path/to/file.js

Configuration

On first run, ColdCode will guide you through the setup:

• Configure AI model settings
• Set up API keys
• Choose default models for different purposes

You can also manually configure using:

• /model - Change AI model settings
• /config - Open configuration panel

Configuration is stored in:

• ~/.cold.json - Global configuration file
• ~/.cold/ - Additional data directory

Core Features

AI-Powered Development

• 🤖 Multi-Model Support - Works with 20+ AI models (Claude, GPT, Gemini, Qwen, etc.)
• 📝 Code Editing - Directly edit files with intelligent suggestions
• 🔍 Codebase Understanding - Analyzes project structure and code relationships
• 🚀 Command Execution - Run shell commands and see results in real-time

Advanced Capabilities

• 🦜 Expert Model Consultation - Use @ask-model-name to consult specific AI models
• 👤 Intelligent Agent System - Use @run-agent-name to delegate tasks to specialized subagents
• 📋 AGENTS.md Integration - Use # prefix to auto-generate and maintain project documentation
• 🔄 Multi-Model Collaboration - Switch between models dynamically for optimal results

Intelligent Completion System

• Smart Fuzzy Matching - Type dao to match run-agent-dao-qi-harmony-designer
• Auto-Prefix Addition - Tab/Enter automatically adds @ for agents and models
• Unix Command Optimization - 500+ common commands with smart prioritization

Usage Examples

Expert Model Consultation

@ask-claude-sonnet-4 How should I optimize this React component for performance?
@ask-gpt-5 What are the security implications of this authentication method?
@ask-o1-preview Analyze the complexity of this algorithm

Specialized Agent Delegation

@run-agent-simplicity-auditor Review this code for over-engineering
@run-agent-architect Design a microservices architecture for this system
@run-agent-test-writer Create comprehensive tests for these modules

Documentation Generation

# How do I set up the development environment?
# What are the testing procedures for this project?
# Explain the deployment pipeline and requirements

Commands

• /help - Show available commands
• /model - Change AI model settings
• /config - Open configuration panel
• /cost - Show token usage and costs
• /clear - Clear conversation history
• /init - Initialize project context

Multi-Model Collaboration

ColdCode implements true multi-model collaboration, allowing you to leverage the unique strengths of different AI models.

Key Features

• Model Profiles: Each model has independent configuration (API endpoints, authentication, context window, costs)
• Model Pointers: Configure default models for different purposes (main, task, reasoning, quick)
• Dynamic Switching: Switch models at runtime without restarting sessions
• Parallel Processing: Multiple subagents can work on different tasks simultaneously

Example Workflow

# Architecture Design
"Use o3 model to design a high-concurrency message queue system"

# Multi-Model Collaboration
"First use GPT-5 to analyze this performance issue, then use Claude Sonnet 4 to write optimization code"

# Parallel Task Processing
"Use Qwen Coder model as subagent to refactor these three modules simultaneously"

License

© 2025 CoolBeeAI LLC. All rights reserved.

This software is proprietary and confidential. Unauthorized copying, distribution, or use is strictly prohibited.

Support

For technical support, please contact CoolBeeAI support team.