Show HN: My LLM CLI tool can run tools now, from Python code or plugins

Simon Willison's LLM CLI tool now supports running tools through Python code or plugins, enabling more complex interactions and automation. This enhancement allows users to extend the tool's capabilities by integrating custom Python functions or pre-built plugins.

Key Points

• •LLM CLI tool now supports tool execution capabilities, enabling LLMs to call external functions and plugins
• •Tools can be defined directly in Python code, allowing developers to extend functionality without modifying core tool
• •Plugin system enables modular tool creation and distribution, making it easy to share and reuse custom tools
• •LLMs can dynamically invoke tools during conversations, expanding their capabilities beyond text generation
• •Tool integration follows a structured interface, ensuring compatibility and consistent behavior across different tool implementations
• •Developers can create tools for API calls, data processing, file operations, and other external integrations
• •Tool execution happens within the LLM CLI context, maintaining conversation state and enabling multi-step workflows
• •Plugin architecture supports both built-in tools and community-contributed extensions

Workflow Diagram

Concepts