🐒 Wukong Stack

WukongStack Wukong is a revolutionary developer toolkit that empowers engineers with intelligent coding assistance, project scaffolding, and automated code review capabilities. Inspired by the legendary Monkey King's wisdom and magic, Sun Wukong, from Chinese mythology, Wukong helps developers create, refactor, and review code with unparalleled intelligence.Wukong streamlines development by delivering a secure, testable, and modern project structure for rapid prototyping and production-ready applications.

🔥 "I've got all the magic you need for your coding journey!"

🌟 Features

🔧 Project Generation & Initialization

• Create project structures for multiple frameworks (Flask, FastAPI, React, Vue.js, etc.)
• Initialize existing projects with pre-configured templates
• Support for Python packages, Node.js modules, Go applications, and more

💻 AI-Assisted Coding

• Interactive shell with persistent conversation history
• Code generation from natural language prompts
• Intelligent code extraction and file saving
• Multi-file project scaffolding powered by LLMs

🕵️ Code Review & Refactoring

• Automated code review for logic flaws, vulnerabilities, and best practices
• Smart refactoring suggestions to improve code quality
• Unit test generation for existing codebases
• Detailed explanations of complex code sections

🔍 Intelligence Engine

• Conversation history management with save/load capabilities
• Configurable LLM settings (local Ollama support)
• Security-conscious file operations with validation
• Multi-language source code parsing and handling

📦 Installation

Install via pip:

pip install wukong-stack

Note: Requires Python 3.10+

🚀 Quick Start

Initialize a new project

wukong init

Create a full-stack web app skeleton

wukong create

Review existing code for vulnerabilities or logic errors

wukong review --path ./src/app.py

Generate unit tests automatically from current codebase

wukong unittest --path ./src/

Refactor problematic sections in your code

wukong refactor --path ./src/old_module.py

Start the interactive shell for continuous assistance

wukong shell

Use AI Code Assistant with custom prompts or files

wukong code --prompt "Write a Python function that sorts a list of dictionaries by their 'name' key"

Or use a file:

wukong code --prompt-file ./prompts/api_setup.txt

Save LLM output and extract code blocks:

wukong code --prompt "Generate REST API routes for user authentication" --save-llm-output --extract-code

AI Assistant Configuration Guide

Welcome to the AI Assistant setup instructions. You have two options to configure your AI assistant:

Option 1: Use Remote OpenAI Compatible LLM (Recommended for beginners)

If you want to use a remote service like OpenAI, Anthropic, or other providers:

Step 1: Configure Basic Settings

# Set the API key (replace YOUR_API_KEY with your actual key)
wukong config set llm.api_key YOUR_API_KEY --global

# Set the timeout in seconds (optional)
wukong config set llm.timeout 30 --global

# If using a custom base URL, set it here
# wukong config set llm.base_url https://api.openai.com/v1 --global

Step 2: Verify Configuration

# Check current configuration
wukong config get llm.api_key
wukong config get llm.timeout
wukong config get llm.base_url

Step 3: Test Your Setup

Run a simple test to ensure everything works:

wukong chat "Hello, how are you?"

---

Option 2: Install Local LLM (For Users with GPU)

Installing Ollama

1. Visit the official Ollama website: https://ollama.com/download

2. Download and install Ollama for your operating system

3. Follow the installation instructions specific to your OS:

   • Windows: Run the installer and follow prompts
   • macOS: Use Homebrew or download from website: brew install ollama
   • Linux: Install via package manager or curl script

4. Once installed, run Ollama server:

ollama serve

5. Pull a model (example with llama3):

ollama pull llama3

6. Configure your AI assistant to use local LLM:

wukong config set llm.base_url http://localhost:11434/v1 --global
wukong config set llm.api_key placeholder --global  # Not required for Ollama
wukong config set llm.timeout 60 --global

---

Installing LM Studio

1. Visit the official LM Studio website: https://lmstudio.ai

2. Download and install LM Studio for your operating system:

   • Windows: Download .exe installer
   • macOS: Download .dmg file
   • Linux: Use package manager or download from website

3. Launch LM Studio and follow the setup wizard to:

   • Install models (download preferred models like Llama, Mistral, etc.)
   • Configure local server settings
   • Set up API access if needed

4. Once installed, configure your AI assistant:

wukong config set llm.base_url http://localhost:1234/v1 --global
wukong config set llm.api_key placeholder --global  # Not required for LM Studio
wukong config set llm.timeout 60 --global

---

Configuration Parameters Explained

• llm.base_url: The endpoint URL of your LLM service (optional for OpenAI, mandatory for local)
• llm.api_key: Your API authentication key
• llm.timeout: Maximum time in seconds to wait for responses from the AI

Quick Setup Commands Summary

For Remote Services:

wukong config set llm.api_key YOUR_API_KEY --global
wukong config set llm.timeout 30 --global

For Local Ollama:

wukong config set llm.base_url http://localhost:11434/v1 --global
wukong config set llm.timeout 60 --global

For Local LM Studio:

wukong config set llm.base_url http://localhost:1234/v1 --global
wukong config set llm.timeout 60 --global

---

Troubleshooting Tips

1. Connection Issues: Ensure your local server is running before configuring
2. Authentication Problems: Double-check that your API keys are correct
3. Timeout Errors: Increase timeout values if you're experiencing slow responses

Choose the option that best fits your setup and requirements!

📄 License

MIT © Sunny Liu