MCP Dev Agent

🚀 MCP Development Agent

The MCP Development Agent is a powerful tool based on the Model Context Protocol (MCP). It offers various functional modules, including memory management, document search, Git/GitHub integration, and code analysis. This tool supports two operation modes: Command Line Interface (CLI) and Server-Sent Events (SSE) mode.

✨ Features

1. Memory Management

• Storage and Retrieval based on ChromaDB: Utilize a vector database for efficient information management and retrieval.
• Context-Aware Search: Support precise content search based on context.
• Temporal History: Maintain operation history for easy backtracking and analysis.

2. Document Search

• Document Query: Use natural language processing technology to quickly locate the required document content.
• Context Integration: Provide relevant document suggestions based on the current working environment.

3. Git/GitHub Integration

• Git Tool Support: Seamlessly integrate Git commands to improve version control efficiency.
• GitHub API Calls: Interact directly with GitHub services to manage code repositories, Issues, and Pull Requests.

4. Code Analysis

• Syntax Parsing: Use Tree-sitter technology to perform structured analysis on multiple programming languages (e.g., Python, JavaScript, TypeScript).
• Code Comprehension: Generate a high-level overview of the code, identify key components and their responsibilities, and provide improvement suggestions and security considerations.

📦 Installation

1. Installation Methods

• Global Installation: It is recommended to use this method for unified management and automatic environment configuration.

  python setup.py install
  

• Manual Installation: Suitable for users who need custom paths or dependency management.

  pip install .
  

2. Environment Variable Setup

Create and edit the config.ini file and add the following configuration:

[Paths]
PYTHON_EXEC = /usr/bin/python3

[GIT]
DEFAULT_GITHUB_USER = your_github_username

Ensure that Git is correctly configured with a global username and email:

git config --global user.name "Your Name"
git config --global user.email "your.email@example.com"

3. Module Dependencies

Install the necessary Python packages:

pip install chromadb gitpython fastapi python-multipart

💻 Usage Examples

1. Basic Commands

• Start CLI Mode:

  mcp-agent --mode cli
  

• Code Analysis Example:

  > /code analyze src/main.py
  

2. Advanced Usage

a. Code Structure Analysis

> /code analyze src/main.py

The output includes a module overview, key component identification, and improvement suggestions.

b. Document Search

> /docs search "How to use the requests library"

c. Git Operations

> /git clone https://github.com/username/repository.git

3. Server Mode (SSE)

Start the SSE service:

mcp-agent --mode server

Send a POST request via an HTTP client to trigger an analysis task, and the response will be transmitted in the event stream format.

🔧 Technical Details

Project Structure

.
├── main.py              # Application entry point
├── setup_parsers.py     # Code parser configuration
├── tools/
│   ├── memory_tool.py   # Memory management module
│   ├── doc_tool.py      # Document search tool
│   ├── git_tool.py      # Git integration tool
│   └── github_tool.py   # GitHub integration and code analysis
└── docs/
    └── api_reference.md # API reference documentation

Model Configuration

1. Configuration File .agent.json

{
  "default_model": "codellama",
  "context_window": 32768,
  "temperature": 0.1,
  "top_p": 0.9
}

2. Model Selection and Task Allocation

• Code Documentation Generation: CodeLlama model
• Problem Summarization: CodeLlama model
• General Queries: Llama 2 model

Summary

The MCP Development Agent integrates various tools and services to provide developers with efficient and intelligent auxiliary functions. Whether it's code analysis, document search, or version control, this tool can significantly improve development efficiency and experience.

Start using the MCP Development Agent to optimize your development process!