How I use every Claude Code feature (blog.sshh.io)

Article summary

The article discusses the author's experience using Claude Code, an AI-powered coding tool, and shares tips and best practices for effective usage. The author emphasizes the importance of maintaining a well-structured CLAUDE.md file, which serves as a guide for the AI agent. The article also covers various features of Claude Code, including custom slash commands, subagents, and hooks. The author shares their preferred approaches to using these features and highlights the benefits of using Claude Code for coding tasks.

Main themes

• Claude Code usage
• CLAUDE.md file management
• AI-powered coding tools
• Customization and optimization
• Coding workflows and productivity

What commenters say

• Using a symlink for AGENTS.md is a viable alternative to Anthropic's recommended approach of referencing it in CLAUDE.md.
• The CLAUDE.md file content is automatically included in the system prompt without requiring explicit file read operations.
• Some users find that Claude Code often ignores their AGENTS.md guidance, despite the file being properly referenced.
• The use of MCPs is not always practical, particularly when it comes to scoping and limiting access to specific resources.
• Agents are capable of solving problems on their own, and MCPs should focus on authentication rather than providing a wide range of tools.
• The field of AI-powered coding tools is rapidly evolving, and guides and tutorials can become outdated quickly.
• Customization and optimization of Claude Code features are crucial for effective usage and productivity.
• Different approaches to using Claude Code features, such as subagents and hooks, can have varying degrees of success and may depend on individual workflows and preferences.