Author: Aravind V Version: 1.0.0

A sophisticated command-line coding assistant built with Rust. This agent leverages the power of multiple Large Language Models (LLMs) to understand goals, create plans, and execute them by writing code and using a variety of tools.

• Multi-Provider LLM Support: Seamlessly switch between different AI models using a simple command-line flag.
  • --provider openai
  • --provider gemini
  • --provider claude
  • --provider deepseek
  • --provider ollama (For running local models)
• Intelligent Orchestration: A reasoning agent creates a step-by-step plan for your goal and executes it intelligently.
• Extensible Tool System: The agent can interact with its environment to:
  • Read and write files (ReadFile, WriteFile).
  • Execute arbitrary shell commands (RunCommand).
  • Perform real-time web searches for up-to-date information (Search).
  • List directory contents to understand project structure (ListFiles).
• Context-Aware Operation: Maintains a history of actions and results to make informed decisions and self-correct.
• Asynchronous & Performant: Built on tokio for efficient, non-blocking operations.
• Secure Configuration: Manages API keys and other secrets via a .env file, keeping them out of the source code.

• Rust: Install the Rust toolchain from rustup.rs.
• API Keys: You'll need API keys for the LLM services you intend to use.

1. Clone the Repository:

   git clone <repository-url>
   cd rust-cli-coding-agent

2. Configure Environment Variables:

   • Copy the example .env file:

     cp .env.example .env

   • Open the .env file and add your API keys. Only the keys for the providers you use are required.

     # For OpenAI
     OPENAI_API_KEY="your-openai-api-key"
     
     # For Google Gemini
     GOOGLE_API_KEY="your-google-api-key"
     
     # For Anthropic Claude
     ANTHROPIC_API_KEY="your-anthropic-api-key"
     
     # For DeepSeek
     DEEPSEEK_API_KEY="your-deepseek-api-key"
     
     # For the Search Tool (using Brave Search API)
     BRAVE_SEARCH_API_KEY="your-brave-search-api-key"
     
     # For Ollama (if using a custom base URL)
     OLLAMA_BASE_URL="http://localhost:11434"

3. Build the Project:

   cargo build --release

   The executable will be located at target/release/cli_coding_agent.

After building, the agent runs in an interactive mode, prompting you for goals.

Navigate to the project root and run:

cargo run --

Or, if you've added the executable to your PATH (see below):

cli_coding_agent

Once running, you will be prompted to enter your goal:

Enter your goal (or 'quit' to exit): Create a Rust function that calculates the factorial of a number and write it to a file named `factorial.rs`.

You can specify the LLM provider when starting the agent:

cargo run -- --provider gemini

Or, if using the direct executable:

cli_coding_agent --provider ollama

To run cli_coding_agent from any directory without specifying its full path, you can add its executable to your system's PATH or create a symbolic link.

For Windows:

1. Open Command Prompt as Administrator.
2. Navigate to the project root: cd rust-cli-coding-agent
3. Run the provided script: install_path_windows.bat
4. Important: Restart your command prompt or PowerShell for changes to take effect.

For Linux/macOS:

1. Open your terminal.
2. Navigate to the project root: cd rust-cli-coding-agent
3. Make the script executable: chmod +x install_path_linux.sh
4. Run the script with sudo: sudo ./install_path_linux.sh

This will allow you to simply type cli_coding_agent in any directory to start the interactive agent.

• main.rs: Entry point, CLI parsing.
• orchestrator.rs: The core reasoning engine that manages the plan and state.
• llm/: Module containing all LLM client implementations, unified under the LLMClient trait.
• agents/: Contains specialized agents (PlannerAgent, CoderAgent) responsible for specific tasks.
• tools/: Defines and implements the tools the agent can use.
• state.rs: Manages the application state, including history and context.
• config.rs: Handles loading configuration from the .env file.
• error.rs: Custom error types for robust error handling.