Claude Code Custom Command Arguments and Skills

Learn how Claude Code custom command arguments work in slash commands and when to use skills with argument-hint, named arguments, and supporting files.

Claude Code Custom Command Arguments

The simplest argument is just the text you type after the slash command.

Last checked on July 7, 2026. Current Claude Code docs position skills as the richer path for custom workflows, while existing .claude/commands/*.md prompts still work for simple slash commands.

Quick Answer

/explain src/app/page.tsx

In that example, src/app/page.tsx is the argument text the custom command should use.

For structured arguments, use a skill:

---
name: explain-file
description: Explain a file and identify risky changes
argument-hint: '<file path>'
arguments:
  - name: file
    description: File to inspect
    required: true
---

Slash Command vs Skill Arguments

Need	Use
One prompt with free text after the command name	`.claude/commands/*.md`
Named or documented arguments	Skill
Autocomplete hint for expected input	Skill `argument-hint`
Supporting files, scripts, or examples	Skill
Team workflow that should be portable	Skill

Practical Pattern

Use a short command for simple transformations:

/rewrite-title Make this clearer for a docs homepage

Use a skill when you need repeatable behavior:

/explain-file src/server/auth.ts

The skill can document what the file argument means, what checks to run, and what output format to return.

Common Mistakes

Expecting .claude/commands/*.md to validate argument types automatically.
Hiding required inputs inside prose instead of documenting them.
Building a complex workflow as a single prompt when it needs supporting scripts.
Using a global command for a repository-specific workflow.

Claude Code Custom Command Arguments and Skills