Skip to main content

Create a Personal GitHub Assistant Project

This tutorial demonstrates how to create a project that interacts with GitHub through obot's Chat Interface and MCP Gateway. You'll build an AI assistant that can help you track GitHub issues, pull requests, and other development tasks.

This tutorial showcases:

  • Chat Interface: Creating and configuring a project for conversational AI
  • MCP Gateway: Connecting to GitHub through the Model Context Protocol
  • Integration: How the chat interface and MCP gateway work together
note

As you configure the project, changes will be saved and applied automatically.

What You'll Learn​

By the end of this tutorial, you'll understand how to:

  1. Create and configure a project in the chat interface
  2. Connect external services through MCP servers
  3. Authenticate with GitHub using OAuth
  4. Test and iterate on your AI assistant

Prerequisites​

  • Access to an obot deployment (self-hosted or demo)
  • A GitHub account with a personal access token
  • Basic familiarity with GitHub concepts (issues, pull requests, repositories)

1. Creating Your Project​

The Chat Interface is where you'll create and interact with your GitHub assistant.

Access the Chat Interface​

  1. Go to your obot homepage
  2. Click on your profile picture in the top right
  3. Choose Chat from the dropdown menu

If you don't have an existing project, one will be created automatically. If you already have projects, click the + button in the left sidebar to create a new one.

Configure Project Basics​

  1. Name: Set a descriptive name like "GitHub Assistant" or "My GitHub Helper"
  2. Description: Write a brief description such as "Personal assistant for GitHub tasks and status updates"

Define the System Prompt​

Click the gear icon next to your project name in the sidebar to configure the project instructions. This system prompt defines your AI assistant's behavior and capabilities.

Try this example prompt:

You are a smart assistant with access to the GitHub API through MCP tools.

Your main responsibilities:
- Help me track GitHub issues assigned to me
- Monitor pull requests where my review is requested
- Provide status updates on my GitHub activity
- Answer questions about repositories, commits, and GitHub workflows

When I ask for a "status update", provide:
1. All issues currently assigned to me
2. Pull requests where my review is requested  
3. Recent activity on repositories I care about

Be concise but informative, and always provide relevant links to GitHub.

2. Connecting GitHub via MCP Gateway​

The MCP Gateway enables your project to securely connect with external services like GitHub.

Add the GitHub MCP Server​

  1. In the left sidebar, click the + button next to MCP Servers
  2. Search for "GitHub" in the server catalog
  3. Select the GitHub MCP server from the results
  4. Click Connect To Server

Configure Authentication​

  1. You'll need a GitHub Personal Access Token:

    • Go to GitHub → Settings → Developer settings → Personal access tokens
    • Generate a token with appropriate permissions (repo, read:user, read:org)
    • Copy the token

  2. In obot, paste your token in the GitHub Personal Access Token field

  3. Click Update to save the configuration

The MCP Gateway will now handle secure communication between your project and GitHub's API.

3. Testing Your GitHub Assistant​

Now that your project has access to GitHub through the MCP Gateway, test the integration:

Basic Functionality Test​

Start with a simple request to verify the connection:

Get the star count for the repository "torvalds/linux"

Personal GitHub Data​

Test with your own GitHub data:

Show me all repositories I own

Status Update Feature​

If you included the status update instructions, try:

Give me a status update on my GitHub activity

Advanced Queries​

Explore more complex interactions:

What are the most recent issues I've been assigned across all repositories?

4. Understanding the Integration​

How the Parts Work Together​

  1. Chat Interface: You interact with your GitHub assistant through natural conversation
  2. MCP Gateway: Securely proxies requests to GitHub's API using the MCP protocol
  3. GitHub MCP Server: Translates between MCP calls and GitHub API requests
  4. Your Project: Uses the available tools to fulfill your requests

Data Flow​

You → Chat Interface → Project → MCP Gateway → GitHub MCP Server → GitHub API

The response flows back through the same path, with the MCP Gateway ensuring secure, authenticated access to your GitHub data.

5. Customization Ideas​

Enhance your GitHub assistant by:

Extended System Prompt​

Add more specific instructions for your workflow:

  • Repository priorities
  • Notification preferences
  • Specific GitHub workflows you use

Additional Context​

Upload files with:

  • Team member information
  • Project priorities
  • Coding standards or guidelines

Tool Integration​

Connect additional MCP servers for:

  • Slack notifications about GitHub activity
  • Email summaries of GitHub updates
  • Integration with project management tools

Best Practices​

Security​

  • Use personal access tokens with minimal required permissions
  • Regularly rotate your GitHub tokens
  • Don't share projects that contain sensitive repository access

Effective Prompts​

  • Be specific about what information you want
  • Use consistent terminology for commands like "status update"
  • Provide context about which repositories matter most to you

Project Organization​

  • Use clear, descriptive project names
  • Document your project's purpose in the description
  • Keep system prompts focused but comprehensive

Troubleshooting​

Authentication Issues​

  • Verify your GitHub token has the correct permissions
  • Check that the token hasn't expired
  • Ensure the token is correctly entered in the MCP server configuration

API Limitations​

  • GitHub API has rate limits; the MCP Gateway handles this automatically
  • Some operations require specific permissions on repositories
  • Private repositories require appropriate token permissions

This tutorial demonstrates how obot's three main components work together to create powerful, connected AI assistants that can interact with your development workflow through familiar conversational interfaces.