Rawdog is a CLI assistant that generates and auto-executes Python scripts to answer user prompts, offering a novel alternative to RAG for tasks requiring dynamic context gathering. It's designed for developers and power users who need to interact with their system and data programmatically via natural language.

How It Works

Rawdog operates by recursively augmenting its context. When given a prompt, it generates a Python script, executes it, captures the output, adds it to the conversation history, and then calls itself again with the updated context. This self-correction loop allows it to dynamically gather information, such as file system details or active ports, to fulfill complex requests.

Quick Start & Requirements

• Install with: pip install rawdog-ai
• Requires an OpenAI API key (or other providers via configuration).
• Supports conversational or direct execution modes.
• See official documentation for model selection and configuration.

Highlighted Details

• Generates and executes Python scripts directly in the CLI.
• Supports recursive context gathering via script output.
• Integrates with various LLM providers through litellm (default: gpt-4-turbo-preview).
• Optional --leash flag for manual script approval before execution.

Maintenance & Community

• Project maintained by AbanteAI.
• Community channels and roadmap details are available on the GitHub repository.

Licensing & Compatibility

• License details are not explicitly stated in the provided README.
• Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project explicitly warns of potential harm due to its auto-execution capabilities, advising users to proceed with caution. The default mode executes scripts without user approval, posing a security risk if not managed carefully.