Useful GitHub Copilot CLI Features

GitHub Copilot CLI includes many helpful features that can make coding faster and easier.
Here are some of the most useful ones.

1. /tasks
This command shows the background jobs or agents that are currently running.

For example, if Copilot is using multiple models to review or refactor code, /tasks lets you see their progress and open a running session.

2. Autopilot mode
Autopilot mode means Copilot keeps working until the assigned tasks are finished.

3. Plugins
Plugins make Copilot CLI easier to extend. A plugin can package several features together, such as:

• MCP servers
• skills
• custom instructions
• custom agents

This makes things simpler for users. Instead of deciding whether they need an MCP server, a skill, or an agent, they can install one plugin and get everything they need.

4. plugin marketplace browse and plugin install
These commands work like an app store for Copilot CLI extensions.

You can browse available plugins in the marketplace, then install one with a command. After that, the plugin becomes available in your Copilot sessions.

5. Chronicle
Chronicle is one of the most interesting Copilot CLI features. It keeps a local record of your Copilot sessions in an SQLite database on your machine.

It can remember:

• prompts you used
• repeated patterns
• past interactions

Two important Chronicle commands are:

chronicle improve
This reviews your past sessions and suggests ways to improve your Copilot instructions or repository instructions. For example, if you often remind Copilot how to format pull requests, it may suggest adding that as a permanent instruction.

chronicle tips
This gives personal tips on how you use Copilot. For example, if you often paste links and ask Copilot to analyze them, it may suggest using /research instead.

A key security point is that this history stays on your machine, inside your .copilot directory. It is not shared with your team or manager.

6. /research
This command is used for deeper analysis.

Instead of only responding to one pasted link, /research can search more widely across the web, GitHub, and similar repositories to provide better results.

7. Multiple models working in parallel
Copilot CLI can run multiple background agents using different models, such as:

• Claude Opus
• GPT-5.4
• Gemini

Each model reviews the same codebase and gives its own recommendations. Then the main agent combines those results into one final report.

This is useful because different models may notice different problems. If several models find the same issue, that is a strong sign that the issue should be fixed.

8. Refactoring recommendations in tiers
After the background agents finish, Copilot can group the findings into tiers:

• issues found by all models
• issues found by two models
• issues found by only one model

This helps you decide what to fix first.

9. Fleet mode
Fleet mode allows Copilot to split work across multiple sub-agents and run tasks in parallel.

Instead of handling one refactoring task at a time, it can work on several independent tasks at once.

10. /experimental show
This command shows preview features that are still being tested and are not yet generally available.

If you want early access to new features, you can turn on experimental mode and use this command to see what is available.

11. Future feature: remote session monitoring
It would allow users to check a running Copilot session remotely while away from their machine.

Copilot CLI is growing from a simple terminal assistant into a powerful multi-agent coding tool.