Skip to content
Home › Claude Code Configuration & Workflows

Custom Slash Commands

Claude Code Configuration & Workflows

Definition

User-defined /commands backed by Markdown files in .claude/commands/*.md — now unified with skills: a file at .claude/commands/deploy.md and a skill at .claude/skills/deploy/SKILL.md both create /deploy and work the same way.

Key points

  • Command name = file name without extension (.claude/commands/deploy.md → /deploy). Recognized only at the start of your message; text after the name becomes its arguments.
  • "Custom commands have been merged into skills." Existing .claude/commands/ files keep working and support the same frontmatter as skills; skills add supporting-file directories, invocation-control frontmatter, and automatic model loading. If a skill and a command share a name, the skill takes precedence.
  • Argument substitutions: $ARGUMENTS (all), $ARGUMENTS[N] (0-based index), $N (shorthand), $name (declared in arguments).
  • Dynamic context injection: !`<command>` runs the shell command before the content is sent and replaces the placeholder (preprocessing, not something Claude executes); recognized only when ! is at line start or after whitespace. Supporting files referenced with markdown links / @file-style paths.

Why it matters for the exam

  • D3 (20% of scored content) tests the commands↔skills merge (both .claude/commands/*.md and .claude/skills/<name>/SKILL.md produce /name, skill wins on clash) and the substitution/!bash syntax — likely distractors in "how do you parameterize a command" questions.

Common gotchas

  • A command is only recognized at the start of the message — mid-message /name is literal text.
  • !bash runs as preprocessing before Claude sees the content, once — output is not re-scanned; escape a literal $ before a digit with a backslash (\$1.00).
  • On a name clash, the skill wins, not the .claude/commands/ file.

See also

Sources

Referenced by