Commands¶
Commands are quick, single-purpose actions that gather information or perform a focused task. Unlike skills, which are multi-step interactive workflows with decision points, commands run a well-defined sequence of steps and report back.
| Aspect | Commands | Skills |
|---|---|---|
| Purpose | Quick checks, reports, single actions | Multi-step workflows with decisions |
| Duration | Seconds to a minute | Minutes, often interactive |
| Example | /status shows project health |
/pr walks through lint, test, branch, commit, push, PR |
Available Commands¶
| Command | Description |
|---|---|
/status |
Git status, recent commits, open PRs, lint and test health |
/test |
Run the test suite, diagnose and fix failures |
/coverage |
Run tests with coverage, highlight untested code |
/deps |
Audit dependencies for outdated, vulnerable, or missing packages |
/release |
Semver version bump, git tag, and GitHub release |
/db-schema |
Inspect, compare, or document database schemas |
/cron |
List, add, validate, or remove cron jobs |
/env-setup |
Set up or verify a Python development environment |
/log-triage |
Parse Loguru logs, group errors, and summarize an unattended run |
/healthcheck |
Verify a scheduled job landed: freshness, volume, continuity |
Usage¶
Type any command name as a slash command in Claude Code:
Commands require no arguments. They auto-detect project context (language, framework, test runner, database type) and adapt accordingly.
Combine commands for a full health check
Run /status, then /test, then /deps in sequence to get a comprehensive view of project health before starting work.