GitHub Copilot CLI 入門：よく使うスラッシュコマンド一覧と使い方 GitHub Copilot CLI for Beginners: Overview of common slash commands

GitHub Copilot CLI has quietly become one of the more practical ways to bring AI assistance into the terminal, and slash commands are the feature that makes it genuinely useful rather than merely novel. This beginner-oriented guide walks developers through the most common commands available in Copilot CLI and explains how each one slots into a real development workflow.

Slash commands are short, intent-declaring inputs — /explain, /fix, /tests — that tell Copilot CLI exactly what kind of help you need without requiring a full sentence. Instead of typing "can you explain what this shell script does," you invoke /explain and pass the relevant context. The brevity is deliberate: the command line is an environment where terseness is culturally valued, and this interaction model respects that. Among the most commonly used commands, /explain returns a plain-language breakdown of code or command output; /fix takes an error message or broken snippet and suggests a correction; and /tests scaffolds unit tests based on existing code, which proves especially useful when navigating an unfamiliar codebase.

GitHub first released Copilot CLI as a technical preview in 2023, later folding it into GitHub CLI as an extension. For developers already using the gh tool in their daily work, the path to adoption is short. The integration reflects GitHub's broader ambition to embed AI not only inside code editors but across the full development toolchain — terminals, CI pipelines, pull request workflows, and beyond. Adoption appears to be strongest among DevOps engineers, SREs, and data engineers who live primarily in the terminal.

The slash command pattern itself is worth contextualizing. It has emerged as something close to a de facto standard across AI-assisted developer tooling. JetBrains AI Assistant, VS Code's Copilot Chat, and various Slack-based developer bots have all converged on similar designs. The common thread is the idea of giving the model a clear task frame — reducing ambiguity so the AI can return a focused, useful response rather than a generalized one. This framing matters more than it might initially appear, because it also makes AI behavior more predictable and auditable.

Looking ahead, there is speculation within the developer community that future iterations of Copilot CLI could allow teams to define custom slash commands tied to their own codebases, toolchains, or internal conventions. If that extensibility materializes, Copilot CLI would shift from a static assistant to a programmable automation layer — a meaningful evolution. For now, that remains speculative, but the architectural direction seems to point in that general trajectory.

For beginners, the practical guidance is straightforward: start with /explain, /fix, and /tests, learn when each one is the right tool, and build from there. The more durable skill is understanding slash commands as a category — a way of communicating clear intent to an AI agent — rather than memorizing a static list. As the tooling evolves, that underlying principle is likely to remain stable even as specific commands are added, changed, or retired. Getting fluent in that grammar now is an investment that will pay off as AI-assisted development continues to mature.