Pydantic AI: GitHub Repository Analysis Agent

An intelligent GitHub repository analysis agent built using Pydantic AI, capable of analyzing GitHub repositories to answer user questions. The agent can fetch repository information, explore directory structures, and analyze file contents using the GitHub API.

Features

• Repository information retrieval (size, description, etc.)
• Directory structure analysis
• File content examination
• Support for both OpenAI and OpenRouter models
• Available as both API endpoint and command-line interface

Prerequisites

• Python 3.11+
• GitHub Personal Access Token (for private repositories)
• OpenRouter API key

Installation

1. Clone the repository:

git clone https://github.com/coleam00/ottomator-agents.git
cd ottomator-agents/pydantic-github-agent

2. Install dependencies (recommended to use a Python virtual environment):

pip install -r requirements.txt

3. Set up environment variables:
   • Rename .env.example to .env
   • Edit .env with your API keys and preferences:

   GITHUB_TOKEN=your_github_token  # Required for private repos
   OPEN_ROUTER_API_KEY=your_openrouter_api_key
   LLM_MODEL=your_chosen_model  # e.g., deepseek/deepseek-chat
   SUPABASE_URL=your_supabase_url  # Only needed for endpoint
   SUPABASE_SERVICE_KEY=your_supabase_key  # Only needed for endpoint

Usage

FastAPI Endpoint

To run the agent as an API endpoint (also compatible with the oTTomator Live Agent Studio), run the following command:

python github_agent_endpoint.py

The endpoint will be available at http://localhost:8001

Command Line Interface

For a simpler interactive experience, you can use the command-line interface:

python cli.py

Example queries you can ask:

• "What's the structure of repository https://github.com/username/repo?"
• "Show me the contents of the main Python file in https://github.com/username/repo"
• "What are the key features of repository https://github.com/username/repo?"

Configuration

LLM Models

You can configure different LLM models by setting the LLM_MODEL environment variable. The agent uses OpenRouter as the API endpoint, supporting various models:

LLM_MODEL=deepseek/deepseek-chat  # Default model

API Keys

• GitHub Token: Generate a Personal Access Token from GitHub Settings
• OpenRouter API Key: Get your API key from OpenRouter

Project Structure

• github_agent_ai.py: Core agent implementation with GitHub API integration
• cli.py: Command-line interface for interacting with the agent
• requirements.txt: Project dependencies

Live Agent Studio Version

If you're interested in seeing how this agent is implemented in the Live Agent Studio, check out the studio-integration-version directory. This contains the production version of the agent that runs on the platform.

Error Handling

The agent includes built-in retries for API calls and proper error handling for:

• Invalid GitHub URLs
• Rate limiting
• Authentication issues
• File not found errors