Release notes confirm `/clear` wiping sessions, MCP user configs now respected, multi-extension hook merges, and fresh macOS/Windows/Linux assets — proof of active iteration after GA.
GitHub Copilot CLI
activeGitHub's official terminal coding agent. GA Feb 25, 2026; v1.0.11 (Mar 23) shipped `/clear`, multi-extension hook merges, MCP config respect, and cross-platform builds. Multi-model (Opus 4.6, Sonnet 4.6, GPT-5.3-Codex, Gemini 3 Pro). Enterprise Agent Control Plane. 15M Copilot subscriber distribution. 9.5K stars, 1,296 forks.

Shipped last month
Copilot CLI jumps to #3 in coding-clis
The v1.0.11 release, Reddit GA thread, and demo tweet proved Copilot CLI's adoption signal, so it now sits ahead of Codex CLI despite the ongoing PromptArmor security warning. · 2026-03-25
Copilot CLI demoted to #5 in coding-clis (was #4)
200-250K actual CLI installs from a 15M subscriber base is a 1.5% conversion rate — developers aren't switching. Malware incident and request limits dampen enthusiasm. · 2026-03-25
Where it wins
v1.0.11 (Mar 23) shipped `/clear`, multi-extension hook merges, MCP config fixes, and refreshed macOS/Windows/Linux builds — fast post-GA iteration
15M paid Copilot subscribers get it automatically — largest distribution in category
Multi-model: Claude Opus 4.6, Sonnet 4.6, GPT-5.3-Codex, Gemini 3 Pro — users pick per task
Enterprise Agent Control Plane — most mature enterprise admin controls in category
Feature set: Plan mode, Autopilot, /research deep research, Agent Skills
GA since Feb 25, 2026 — shipping as a full product, not beta
Where to be skeptical
CVE-2026-29783 hit 2 days after GA — arbitrary code execution via shell expansion, 'no confirmation dialog, no approval' (PromptArmor)
No published SWE-bench Pro or Terminal-Bench scores
Requires Copilot subscription — not free standalone
Low community signal: 24 HN points on launch, 9.4K stars (lowest among serious contenders)
Quality of the agent itself is unproven — distribution ≠ capability
Editorial verdict
#3 coding CLI — distribution moat (15M Copilot subscribers) and the Enterprise Agent Control Plane are unique advantages, and the v1.0.11 release shows GitHub is iterating quickly. But PromptArmor’s remote-code advisory plus the lack of public benchmarks keep it below Claude/Gemini until security policies catch up.
How to get started
Install via npm install -g @github/copilot (or brew install copilot-cli/winget install GitHub.Copilot), run github-copilot-cli and use the /login slash command the first time to authorize your GitHub account or export COPILOT_GITHUB_TOKEN. Once authenticated, try /task or /tests inside a repo to let Copilot plan, then approve or edit the generated command queue before running it.
Videos
Reviews, tutorials, and comparisons from the community.
The ultimate guide to the GitHub Copilot CLI | Full demo | GitHub Checkout
Power agentic workflows in your terminal with GitHub Copilot CLI
Demo: Using GitHub Copilot CLI and yolo mode
Related

Claude Code
98Anthropic's official agentic coding CLI. v2.1.81 (Mar 20) shipped `--bare`, smarter worktree resume, and improved MCP OAuth while the repo crossed 82,204 stars and logged ~14 commits/week across 10+ maintainers. Terminal-native, tool-use-driven, with deep file system + shell access, #1 SWE-bench Pro standardized (45.89%), ~4% of GitHub public commits (SemiAnalysis), $2.5B annualized revenue. 8M+ npm weekly downloads. Opus 4.6 with 1M context.

OpenHands
88Category leader in multi-agent orchestration — 69,352 stars (verified), $18.8M Series A, AMD hardware partnership, 455 contributors, 1M downloads/month PyPI (3.4M all-time). SWE-Bench Verified 72% with Claude 4.5 Extended Thinking (updated 2026-03-19), Multi-SWE-Bench #1 across 8 languages. Gap to #2 is enormous on every axis.
OpenCode
88Open-source AI coding agent from SST. v1.2.27 active (2026-03-16) — development resumed after a gap. OpenAI official partnership following the Anthropic OAuth block controversy. 126K+ GitHub stars (star surge driven by Anthropic controversy). Known unauthenticated RCE fixed in v1.1.10+ (CVE, 432 HN pts). CVE-2026-22812 (CVSS 8.8-10.0) is a second serious security incident.

Gemini CLI
88Google's open-source terminal agent with Gemini 3 models, 1M token context, built-in Google Search grounding, and the best free tier in the category (60 req/min, 1K req/day). v0.35.0 (Mar 24) shipped keybinding, policy, and telemetry fixes while the repo hit 98,957 stars and 12,593 forks. Terminal-Bench 2.0: 78.4% (#1). SWE-bench Pro standardized 43.30% (#3). Plan Mode added March 2026. First-pass correctness ~50-60% (Educative.io).
Public evidence
Community-run repo recreates GitHub’s GA demo so developers can run the CLI end-to-end. Confirms real teams are experimenting despite Copilot CLI’s low star count.
Tweeted walkthrough of the GA repo showing `/task`, `/splits`, and `/tests` usage, giving practitioners a reproducible path to evaluate the CLI.
A crafted env curl command slips past the validator, downloads a payload from an attacker URL, and pipes it to sh. No confirmation dialog, no approval. Hit 2 days after GA (Feb 27, 2026).
GA launch with multi-model support (Claude, GPT, Gemini), Plan mode, Autopilot, Agent Skills. Enterprise Agent Control Plane for org admin controls. 15M subscriber distribution.
Morph's independent tier classification puts Copilot CLI in Tier 2. Distribution advantage (15M subscribers) is acknowledged but capability is unproven without benchmark scores.
Raw GitHub source
GitHub README peek
Constrained peek so you can sanity-check the source material without leaving the site.
GitHub Copilot in the CLI has been deprecated on October 25, 2025 in favor of GitHub Copilot CLI, an agentic assistant that brings AI-powered coding assistance directly to your command line, enabling you to build, debug, and understand code through natural language conversations. Powered by the same agentic harness as GitHub's Copilot coding agent, it provides intelligent assistance while staying deeply integrated with your GitHub workflow. We suggest migrating to the GitHub Copilot CLI, which can explain and suggest shell commands and much more!
GitHub Copilot in the CLI
GitHub Copilot in the CLI is an extension for GitHub CLI which provides a chat-like interface in the terminal that allows you to ask questions about the command line. You can ask Copilot in the CLI to suggest a command for your use case with gh copilot suggest, or to explain a command you're curious about with gh copilot explain.
For enabling and using, see "Using GitHub Copilot in the command line".
For use cases and limitations, see "Responsible use of GitHub Copilot in the CLI".
For what data is collected, used, and shared, see "Privacy Policies".
For help troubleshooting connectivity, see "Troubleshooting network errors for GitHub Copilot".
Quickstart
[!NOTE] To use and install GitHub Copilot in the CLI, you must have an active GitHub Copilot subscription, have GitHub CLI installed, and authenticate using the GitHub CLI OAuth app.
Classic and fine-grained PATs are currently unsupported and might require clearing the
GITHUB_TOKENandGH_TOKENenvironment variables.
[!IMPORTANT] GitHub Copilot in the CLI does not currently have plans to support 32-bit Android distributions.
For more information, see https://github.com/github/gh-copilot/issues/122.
- Authenticate with GitHub CLI OAuth app
gh auth login --web - Install / upgrade extension
gh extension install github/gh-copilot --force - Suggest a command
gh copilot suggest "Install and configure git lfs" - Explain a command
gh copilot explain 'git lfs migrate import --everything --include="*.gz,*.png,*.jar"'
Usage
$ gh copilot --help
Your AI command line copilot.
Usage:
copilot [command]
Examples:
$ gh copilot suggest "Install git"
$ gh copilot explain "traceroute github.com"
Available Commands:
alias Generate shell-specific aliases for convenience
config Configure options
explain Explain a command
suggest Suggest a command
Flags:
-h, --help help for copilot
--hostname string The GitHub host to use for authentication
-v, --version version for copilot
Use "copilot [command] --help" for more information about a command.
Multi-account notes
Multi-account users can use --hostname flag or GH_HOST environment variable to specify the GitHub hostname for commands that would otherwise assume the github.com host.
Set up optional helpers
Is gh copilot suggest ... too many keystrokes? Do you want support for executing suggestions and keeping them in history?
v1.0.0 introduces gh copilot alias, which generates shell configuration for ghcs and ghce aliases that wrap gh copilot suggest and gh copilot explain. These aliases provide faster invocation and support executing suggested commands if applicable. Executed suggestions are added to your shell history for reuse later.
$ ghcs print "Hello world"
Welcome to GitHub Copilot in the CLI!
version 1.0.0 (2024-03-21)
I'm powered by AI, so surprises and mistakes are possible. Make sure to verify any generated code or suggestions, and share feedback so that we can learn and improve. For more information, see https://gh.io/gh-copilot-transparency
Suggestion:
echo "Hello world"
? Select an option
> Execute command
? Are you sure you want to execute the suggested command?
> Yes
Hello world
To setup these optional helpers, see below for setup instructions for supported shells:
- bash
- powershell
- zsh
For more information, run gh copilot alias --help.
For changing the execute command confirmation behavior, run gh copilot config to change Default value for confirming command execution.
Bash
echo 'eval "$(gh copilot alias -- bash)"' >> ~/.bashrc
PowerShell
[!NOTE] PowerShell users might need to update the file containing
gh-copilothelpers after updating the extension.
$GH_COPILOT_PROFILE = Join-Path -Path $(Split-Path -Path $PROFILE -Parent) -ChildPath "gh-copilot.ps1"
gh copilot alias -- pwsh | Out-File ( New-Item -Path $GH_COPILOT_PROFILE -Force )
echo ". `"$GH_COPILOT_PROFILE`"" >> $PROFILE
Zsh
echo 'eval "$(gh copilot alias -- zsh)"' >> ~/.zshrc
Metrics sent to GitHub
GitHub Copilot in the CLI sends certain metrics to our analytics system. We want you to understand what is being sent and why it's important to our ability to continue to improve the product and provide you with a better experience