Custom Slash Command Creator
Overview
Claude Code slash commands are markdown files with optional YAML frontmatter that create reusable /command workflows. Commands and skills are unified: .claude/commands/ files and .claude/skills/ directories both create slash commands. Skills are the recommended approach since they support additional features like supporting files, but single-file commands still work.
When to use: Repeatable prompts, team workflow standardization, guardrailed operations (deploy, commit), multi-phase tasks, dynamic context injection with bash output.
When NOT to use: Complex capabilities needing multiple files and scripts (use a full skill directory instead), one-off prompts, built-in commands that already exist (/compact, /help, /init).
Quick Reference
| Feature |
Syntax / Location |
Key Points |
| Project command |
.claude/commands/name.md |
Shows "(project)" in /help |
| Project skill |
.claude/skills/name/SKILL.md |
Recommended over commands, supports extra files |
| Personal command |
~/.claude/commands/name.md |
Available across all projects |
| Personal skill |
~/.claude/skills/name/SKILL.md |
Available across all projects |
| Plugin command |
<plugin>/skills/name/SKILL.md |
Namespaced as plugin-name:skill-name |
| Subdirectory commands |
.claude/commands/git/commit.md |
Shows "(project:git)" in /help |
| All arguments |
$ARGUMENTS |
Entire argument string |
| Positional arguments |
$0, $1, $2 or $ARGUMENTS[0] |
Zero-based index |
| Bash injection |
!`git status` |
Runs before prompt is sent, output replaces placeholder |
| File reference |
@src/file.ts |
Inlines file contents into prompt |
| Extended thinking |
Include "ultrathink" in content |
Triggers deeper reasoning mode |
| Session ID |
${CLAUDE_SESSION_ID} |
Current session identifier for logging |
| Frontmatter description |
description: What it does |
Required for /help listing and Skill tool |
| Tool restrictions |
allowed-tools: Read, Grep, Glob |
Limits tools without per-use approval |
| Manual-only invocation |
disable-model-invocation: true |
Prevents Claude from auto-triggering |
| Hide from menu |
user-invocable: false |
Background knowledge only Claude loads |
| Argument hint |
argument-hint: [issue-number] |
Shown in autocomplete |
| Model override |
model: claude-sonnet-4-20250514 |
Specific model for this command |
| Subagent execution |
context: fork |
Runs in isolated context |
| Subagent type |
agent: Explore |
Built-in or custom agent when context: fork |
Priority Order
When commands share the same name across levels, higher-priority locations win:
- Enterprise (managed settings)
- Personal (
~/.claude/)
- Project (
.claude/)
Plugin commands use namespacing (plugin:name) so they never conflict. If a skill and a command share the same name, the skill takes precedence.
Common Mistakes
| Mistake |
Impact |
Correct Pattern |
Adding name field in command files |
Ignored for .md commands |
Name is inferred from filename |
Missing description frontmatter |
Invisible in /help and Skill tool |
Always include description |
Using category or tags fields |
Silently ignored |
Use subdirectories for organization |
$ARGUMENTS without handling empty case |
Unexpected behavior |
Check if empty and provide default |
!`bash` without allowed-tools |
Commands fail to execute |
Add allowed-tools: Bash(...) to frontmatter |
| Expecting project to override personal |
Personal wins over project |
Personal commands take priority; rename to avoid conflicts |
Using context: fork without a task |
Subagent has no actionable work |
Only fork skills with explicit step-by-step instructions |
| Overriding built-in command names |
Built-ins cannot be overridden |
Use different names (/my-help not /help) |
Delegation
- Command pattern discovery: Use
Explore agent to find existing commands in .claude/commands/ or .claude/skills/
- Complex workflow skills: For multi-file capabilities, create a full skill directory with
SKILL.md and supporting files
References
- Command anatomy: markdown structure, YAML frontmatter fields, file locations, and priority order
- Arguments and references: $ARGUMENTS, positional args, bash injection, and file references
- Templates: ready-to-use command templates for common patterns
1---2name: meta-command-creator3description: Create custom slash commands for Claude Code with markdown and YAML frontmatter. Use when building workflow automations, creating reusable prompts, or defining custom /commands. Use for slash commands, argument handling, file references, bash execution.4license: MIT5---6
7# Custom Slash Command Creator
8
9## Overview
10
11Claude Code slash commands are markdown files with optional YAML frontmatter that create reusable `/command` workflows. Commands and skills are unified: `.claude/commands/` files and `.claude/skills/` directories both create slash commands. Skills are the recommended approach since they support additional features like supporting files, but single-file commands still work.
12
13**When to use:** Repeatable prompts, team workflow standardization, guardrailed operations (deploy, commit), multi-phase tasks, dynamic context injection with bash output.
14
15**When NOT to use:** Complex capabilities needing multiple files and scripts (use a full skill directory instead), one-off prompts, built-in commands that already exist (`/compact`, `/help`, `/init`).
16
17## Quick Reference
18
19| Feature | Syntax / Location | Key Points |
20| ----------------------- | ----------------------------------- | ------------------------------------------------------- |
21| Project command | `.claude/commands/name.md` | Shows "(project)" in `/help` |
22| Project skill | `.claude/skills/name/SKILL.md` | Recommended over commands, supports extra files |
23| Personal command | `~/.claude/commands/name.md` | Available across all projects |
24| Personal skill | `~/.claude/skills/name/SKILL.md` | Available across all projects |
25| Plugin command | `<plugin>/skills/name/SKILL.md` | Namespaced as `plugin-name:skill-name` |
26| Subdirectory commands | `.claude/commands/git/commit.md` | Shows "(project:git)" in `/help` |
27| All arguments | `$ARGUMENTS` | Entire argument string |
28| Positional arguments | `$0`, `$1`, `$2` or `$ARGUMENTS[0]` | Zero-based index |
29| Bash injection | `` !`git status` `` | Runs before prompt is sent, output replaces placeholder |
30| File reference | `@src/file.ts` | Inlines file contents into prompt |
31| Extended thinking | Include "ultrathink" in content | Triggers deeper reasoning mode |
32| Session ID | `${CLAUDE_SESSION_ID}` | Current session identifier for logging |
33| Frontmatter description | `description: What it does` | Required for `/help` listing and Skill tool |
34| Tool restrictions | `allowed-tools: Read, Grep, Glob` | Limits tools without per-use approval |
35| Manual-only invocation | `disable-model-invocation: true` | Prevents Claude from auto-triggering |
36| Hide from menu | `user-invocable: false` | Background knowledge only Claude loads |
37| Argument hint | `argument-hint: [issue-number]` | Shown in autocomplete |
38| Model override | `model: claude-sonnet-4-20250514` | Specific model for this command |
39| Subagent execution | `context: fork` | Runs in isolated context |
40| Subagent type | `agent: Explore` | Built-in or custom agent when `context: fork` |
41
42## Priority Order
43
44When commands share the same name across levels, higher-priority locations win:
45
461. **Enterprise** (managed settings)
472. **Personal** (`~/.claude/`)
483. **Project** (`.claude/`)
49
50Plugin commands use namespacing (`plugin:name`) so they never conflict. If a skill and a command share the same name, the skill takes precedence.
51
52## Common Mistakes
53
54| Mistake | Impact | Correct Pattern |
55| ---------------------------------------- | ----------------------------------- | ---------------------------------------------------------- |
56| Adding `name` field in command files | Ignored for `.md` commands | Name is inferred from filename |
57| Missing `description` frontmatter | Invisible in `/help` and Skill tool | Always include `description` |
58| Using `category` or `tags` fields | Silently ignored | Use subdirectories for organization |
59| `$ARGUMENTS` without handling empty case | Unexpected behavior | Check if empty and provide default |
60| `` !`bash` `` without `allowed-tools` | Commands fail to execute | Add `allowed-tools: Bash(...)` to frontmatter |
61| Expecting project to override personal | Personal wins over project | Personal commands take priority; rename to avoid conflicts |
62| Using `context: fork` without a task | Subagent has no actionable work | Only fork skills with explicit step-by-step instructions |
63| Overriding built-in command names | Built-ins cannot be overridden | Use different names (`/my-help` not `/help`) |
64
65## Delegation
66
67- **Command pattern discovery**: Use `Explore` agent to find existing commands in `.claude/commands/` or `.claude/skills/`
68- **Complex workflow skills**: For multi-file capabilities, create a full skill directory with `SKILL.md` and supporting files
69
70## References
71
72- [Command anatomy: markdown structure, YAML frontmatter fields, file locations, and priority order](references/command-anatomy.md)
73- [Arguments and references: $ARGUMENTS, positional args, bash injection, and file references](references/arguments-and-refs.md)
74- [Templates: ready-to-use command templates for common patterns](references/templates.md)
oakoss