Browse all Claude Code slash commands and skills. Click any card to see description, trigger conditions, and example usage.
What are Claude Skills? Skills are reusable slash commands you invoke inside Claude Code. Type /skill-name in any Claude Code session to trigger a specialized workflow — code review, security audit, scheduled agents, and more. Each skill has focused context and trigger logic built in.
Claude Code skills are slash commands that load specialized instructions on demand. Here is how to get started:
Run claude in your terminal inside any project directory. Claude Code reads your project files and CLAUDE.md configuration automatically.
At the Claude Code prompt, type a slash command like /review, /debug, or /test. Claude will load the relevant skill and begin the task.
Most skills work with zero arguments — they read your current branch, open files, or recent changes automatically. For targeted use, pass a file path or PR number: /review 42, /test src/auth.ts.
Some skills (like /deploy or /migrate) need shell access. Use /update-config to add permissions once — Claude remembers them in your project settings.json.
/init to generate CLAUDE.md, then /explain to understand the architecture./review before every merge, /security-review before shipping auth or payments./debug for systematic root-cause analysis, /test to add coverage for the fixed case./loop for polling tasks, /schedule for recurring agents, /update-config to reduce permission prompts.