A Slash Command in 📝Claude Code is a user-typed /foo that invokes a built-in command, a registered skill, or a plugin-provided workflow without requiring natural-language prompting.
Slash commands give Claude Code users a direct invocation surface for repeatable workflows. Built-ins like /help, /clear, /init, /model, and /remember ship with the CLI itself and handle session-level operations — viewing help, resetting context, generating a starter 📝CLAUDE.md as Infrastructure, switching models, or saving an item to memory.
Beyond built-ins, slash commands extend into the skills and plugin ecosystem. Any markdown file in ~/.claude/skills/ with the appropriate YAML frontmatter becomes a slash command — typing /skillname injects the skill's content as a system prompt and triggers Claude to execute the predefined workflow. Plugins bundle multiple skills, hooks, and subagents under namespaced commands.
The pattern distinguishes Claude Code from chat-only AI interfaces: rather than describing what you want every session, users package recurring workflows once and invoke them by name. Common practitioner patterns include /commit for conventional commit messages, /review-pr for pull request reviews, /loop for recurring tasks, /schedule for cron-scheduled remote runs, and /ultrareview for multi-agent cloud review of a branch. The slash namespace is shared across built-ins, skills, and plugins — collisions resolve by precedence, with built-ins taking priority.