Why must commands be written for 'agent consumption'?

Because the content of a command is delivered directly to Claude as an instruction. It should tell the AI exactly what to do rather than explaining the command's purpose to the human user.

Can Claude commands run terminal operations?

Yes, by using the bash execution syntax (!`command`), commands can dynamically gather system context, git status, or environment variables before Claude processes the prompt.

Where should I store my custom commands?

Project-specific commands should be stored in .claude/commands/, while personal global commands that work across all projects belong in ~/.claude/commands/.

How do I pass arguments to a custom command?

You can use $1, $2, etc., for positional arguments, or $ARGUMENTS to capture the entire string passed after the command name for use within your prompt.

What is a Claude Code slash command?

A slash command is a Markdown file containing a specific prompt that Claude executes when invoked, allowing for reusable, consistent, and sharable developer workflows.

Claude Code Command Development

Name: Claude Code Command Development
Author: zach-source

byzach-source

0•

개발자 도구

Guides developers in creating, configuring, and optimizing custom slash commands for Claude Code environments.

The Command Development skill provides a comprehensive framework for building powerful, reusable slash commands within Claude Code. It enables developers to define custom workflows using Markdown and YAML frontmatter, handle dynamic arguments, and integrate live system context through bash execution. Whether you are automating complex code reviews, running test suites, or generating documentation, this skill ensures your custom commands are structured correctly for agent-centric execution and project-wide efficiency.

주요 기능

01Standardized YAML frontmatter configuration for command metadata and tool permissions

02Dynamic argument handling using positional markers and full string capture

03Context injection via intelligent file references and inline bash execution

04Hierarchical command organization and namespacing for large-scale projects

050 GitHub stars

06Best practices for writing agent-focused instructions that optimize Claude's performance

사용 사례

01Creating documentation generators that reference specific project files to ensure accuracy

02Automating multi-step code reviews that combine git diffs with specific security checklists

03Building interactive test-and-fix loops that execute terminal commands and process failures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zach-source/claude-plugins command-development

For use in Claude.ai and ChatGPT

Download Skill