codebuddy-cli

What This Skill Does

The CodeBuddy CLI skill integrates Tencent's powerful AI-driven programming assistant directly into your terminal workflow. Designed to accelerate development, CodeBuddy acts as a bridge between natural language prompts and technical terminal actions. It interprets user intent to generate code, troubleshoot errors, and perform file-system operations. By leveraging advanced language models like Gemini, GPT, and DeepSeek, it enables developers to automate mundane coding tasks, refactor functions, or generate boilerplate code without leaving their terminal environment.

Installation

To begin using CodeBuddy, ensure your system has Node.js 18 or higher installed. Install the package globally via npm by running: npm install -g @tencent-ai/codebuddy-code. Once installed, verify the installation by executing codebuddy --version. For authentication, simply run codebuddy in your terminal and select your preferred provider (Google/GitHub for international or WeChat for China-based models).

Use Cases

This skill is perfect for developers needing quick iterative coding support. Use it for generating unit tests for complex API endpoints, debugging stack traces by analyzing logs in real-time, or refactoring existing codebases for performance improvements. It is particularly effective for project initialization, where users can run /init to generate comprehensive documentation via CODEBUDDY.md, or for maintaining consistent coding standards across large teams by using shared custom command files located in global configuration directories.

Example Prompts

1. "@tencent-ai/codebuddy-code, generate a Python unit test for the authentication service located in src/auth.py."
2. "CodeBuddy, identify the memory leak in the current directory's Node.js application and suggest a fix."
3. "AI CLI, help me refactor this function into an asynchronous implementation to improve performance."

Tips & Limitations

To maximize effectiveness, organize custom commands in .codebuddy/commands/ to create reusable AI workflows. Always exercise caution with the -y flag; it bypasses security confirmations which could lead to accidental file deletions or unintended script executions. For security, never execute the dangerously-skip-permissions flag on production machines or sensitive project directories. Regularly run /doctor to diagnose connectivity or configuration issues and use /cost to keep track of your token usage.