shanraisshan/claude-code-best-practice

This repository is a comprehensive reference guide for getting the most out of Claude Code, Anthropic's AI-powered coding tool. The project bridges the gap between casual "vibe coding" (using AI informally and intuitively) and disciplined "agentic engineering", a more structured approach where AI agents carry out complex, multi-step coding tasks with minimal manual intervention. The repository documents best practices, working examples, and implementations for every major Claude Code feature. These features include subagents (specialized AI workers that handle specific tasks), custom commands (shortcuts that tell Claude what to do), skills (reusable workflows), hooks (automated actions triggered at certain points), MCP servers (external tools Claude can connect to), memory management (how Claude retains context across sessions), and settings configuration. Each feature has its own explanation page, best practice guide, and often a working implementation file you can copy into your own project. Someone would use this repository when they want to work with Claude Code more effectively, for instance, when building software with AI assistance and wanting to set up repeatable workflows, automate repetitive steps, or give Claude clear rules to follow. It is especially useful for developers who want Claude to behave consistently across large projects or teams. The tech stack is primarily HTML (for documentation pages and badges) with Markdown files for the guides themselves. The repository does not contain a backend or application code, it is a living documentation and template project, continuously updated with Claude Code itself. The README functions as a structured index linking to deeper guides, implementations, and official Anthropic documentation.