Our review

Creates or updates a skill (slash command) for Claude Code following best practices.

When to use it

When you need to create a new slash command skill or update an existing one in a Claude project.

When not to use it

For general configuration tasks or modifications unrelated to Claude skills.

name: create-skill description: > Create a new skill (slash command) following Anthropic's best practices. Use when the user wants to create a new skill, command, or update an existing one. Reads the best practices guide first to ensure quality. argument-hint: <skill-name> <description of what it should do>

Create a Skill

Before creating or modifying any skill, read the best practices guide:

Read .claude/references/skills-best-practices.md first.

Understand what the skill should do — ask for concrete usage examples if unclear
Plan what resources it needs (scripts, references, assets)
Check if similar functionality already exists in .claude/commands/
Create the skill following these principles:
- YAML frontmatter with name and comprehensive description (this is the trigger)
- Body under 500 lines — split to references/ if approaching limit
- Move large inline content (scripts, API docs, schemas) to .claude/references/
- Use imperative form in instructions
- Only include what Claude doesn't already know
Verify the skill works by mentally walking through a usage scenario

[ ] Frontmatter description includes "when to use" triggers
[ ] Body is concise — no redundant explanations
[ ] Large content moved to reference files with clear pointers
[ ] disable-model-invocation: true if the skill should only be user-triggered (deploy, commit, destructive actions)
[ ] No extraneous files (README, CHANGELOG, etc.)

Project-specific: .claude/commands/<name>.md (legacy) or .claude/skills/<name>/SKILL.md
Personal: ~/.claude/skills/<name>/SKILL.md

For this project, use .claude/commands/<name>.md for consistency with existing commands.