Copilot

Copilot plugin for Claude Code

Use GitHub Copilot (GPT-5.4) from inside Claude Code for code reviews, adversarial reviews, or to delegate tasks.

This plugin is for Claude Code users who have a GitHub Copilot subscription and want GPT-5.4 as a second brain for code review and task delegation.

Each slash command costs at most 1 Copilot premium request. Status, result, cancel, and setup cost 0.

What You Get

• /copilot:review for a native Copilot code review
• /copilot:adversarial-review for a steerable challenge review with structured findings
• /copilot:rescue, /copilot:status, /copilot:result, and /copilot:cancel to delegate work and manage background jobs
• /copilot:ask for quick one-shot questions
• /copilot:task for delegating agentic tasks (foreground or background)
• /copilot:plan for generating structured implementation plans
• /copilot:review-plan and /copilot:adversarial-plan-review for reviewing plans

Requirements

• GitHub Copilot subscription with access to the Copilot CLI
• Node.js 18.18 or later

Install

Add the marketplace in Claude Code:

/plugin marketplace add mvpasarel/gh-copilot-plugin-cc

Install the plugin:

/plugin install copilot@mvpasarel-copilot

Reload plugins:

/reload-plugins

Then run:

/copilot:setup

/copilot:setup will tell you whether the Copilot CLI is installed and authenticated.

If Copilot CLI is not installed, follow the GitHub Copilot CLI docs to install it.

If Copilot is installed but not logged in yet, run:

!copilot login

After install, you should see:

• the slash commands listed below
• the copilot:copilot-rescue subagent in /agents

One simple first run is:

/copilot:review --background
/copilot:status
/copilot:result

Update

claude plugin update copilot@mvpasarel-copilot

Then restart Claude Code to apply.

Usage

/copilot:setup

Checks whether the Copilot CLI is installed and authenticated.

You can also use /copilot:setup to manage the optional review gate.

Enabling review gate

/copilot:setup --enable-review-gate
/copilot:setup --disable-review-gate

When the review gate is enabled, the plugin uses a Stop hook to run a targeted Copilot review based on Claude's response. If that review finds issues, the stop is blocked so Claude can address them first.

[!WARNING] The review gate costs 1 premium request per stop and can create a long-running Claude/Copilot loop. Only enable it when you plan to actively monitor the session.

/copilot:review

Runs a native Copilot review on your current work. Copilot's built-in review agent automatically detects uncommitted changes, staged changes, and recent commits.

[!NOTE] Code review especially for multi-file changes might take a while. It's generally recommended to run it in the background.

Use it when you want:

• a review of your current uncommitted changes
• a review of your branch compared to a base branch like main

Use --base <ref> for branch review. It also supports --wait and --background. It is not steerable and does not take custom focus text. Use /copilot:adversarial-review when you want to challenge a specific decision or risk area.

Examples:

/copilot:review
/copilot:review --base main
/copilot:review --background

This command is read-only and will not perform any changes. When run in the background you can use /copilot:status to check on the progress and /copilot:cancel to cancel the ongoing task.

/copilot:adversarial-review

Runs a steerable review that questions the chosen implementation and design.

It can be used to pressure-test assumptions, tradeoffs, failure modes, and whether a different approach would have been safer or simpler.

It uses the same review target selection as /copilot:review, including --base <ref> for branch review. It also supports --wait and --background. Unlike /copilot:review, it can take extra focus text after the flags.

Use it when you want:

• a review before shipping that challenges the direction, not just the code details
• review focused on design choices, tradeoffs, hidden assumptions, and alternative approaches
• pressure-testing around specific risk areas like auth, data loss, rollback, race conditions, or reliability

Examples:

/copilot:adversarial-review
/copilot:adversarial-review --base main challenge whether this was the right caching and retry design
/copilot:adversarial-review --background look for race conditions and question the chosen approach

This command is read-only. It does not fix code.

/copilot:rescue

Hands a task to GPT-5.4 through the copilot:copilot-rescue subagent.

Use it when you want Copilot to:

• investigate a bug
• try a fix
• continue a previous Copilot session
• take a different perspective on a problem

[!NOTE] Depending on the task and the model you choose these tasks might take a long time and it's generally recommended to force the task to be in the background or move the agent to the background.

It supports --background, --wait, --resume, and --fresh. If you omit --resume and --fresh, the plugin can offer to continue the latest rescue session for this repo.

Examples:

/copilot:rescue investigate why the tests started failing
/copilot:rescue fix the failing test with the smallest safe patch
/copilot:rescue --resume apply the top fix from the last run
/copilot:rescue --effort medium investigate the flaky integration test
/copilot:rescue --background investigate the regression

You can also just ask for a task to be delegated to Copilot:

Ask Copilot to redesign the database connection to be more resilient.

Notes:

• if you do not pass --effort, Copilot uses its default
• follow-up rescue requests can continue the latest Copilot session in the repo

/copilot:ask

One-shot foreground question to Copilot. Read-only, always foreground.

Supports --model and --effort.

Examples:

/copilot:ask what does the auth middleware do
/copilot:ask explain the retry logic in lib/http.mjs

Costs 1 premium request.

/copilot:task

Delegate an agentic task to Copilot. Without --write, tasks run in a read-only sandbox.

Supports --background, --write, --resume, --fresh, --model, and --effort.

Examples:

/copilot:task refactor the error handling in lib/api.mjs
/copilot:task --background --write implement input validation for the signup form

Costs 1 premium request.

/copilot:plan

Generate a structured implementation plan with steps, risks, and open questions. The output includes a verdict (ready or needs-clarification), ordered steps with file lists, risks, and open questions.

Supports --background, --prompt-file, --model, and --effort.

Examples:

/copilot:plan add WebSocket support to the notification system
/copilot:plan --prompt-file feature-spec.md

Costs 1 premium request.

/copilot:review-plan

Review a plan for correctness, completeness, ordering, and feasibility. Plan content can be provided via positional args, --plan-file <path>, or stdin.

Supports --background, --model, and --effort.

Examples:

/copilot:review-plan --plan-file plan.md
/copilot:review-plan the plan is to add a caching layer in front of the database

/copilot:adversarial-plan-review

Challenge a plan's assumptions, missing steps, and hidden dependencies. Plan content can be provided via --plan-file <path> or positional args. Use --focus <angle> with --plan-file to specify the adversarial angle.

Examples:

/copilot:adversarial-plan-review --plan-file plan.md
/copilot:adversarial-plan-review --plan-file plan.md --focus "what if the cache goes down"

/copilot:status

Shows running and recent Copilot jobs for the current repository.

Examples:

/copilot:status
/copilot:status task-abc123

Use it to:

• check progress on background work
• see the latest completed job
• confirm whether a task is still running

/copilot:result

Shows the final stored output for a finished job. When available, it also includes the Copilot session ID so you can resume that session with copilot --resume=<session-id>.

Examples:

/copilot:result
/copilot:result task-abc123

/copilot:cancel

Cancels an active background Copilot job.

Examples:

/copilot:cancel
/copilot:cancel task-abc123

Typical Flows

Review Before Shipping

/copilot:review

Hand A Problem To Copilot

/copilot:rescue investigate why the build is failing in CI

Start Something Long-Running

/copilot:adversarial-review --background
/copilot:rescue --background investigate the flaky test

Then check in with:

/copilot:status
/copilot:result

Quick Question

/copilot:ask what's the purpose of the retry logic in lib/http.mjs

Plan Before Building

/copilot:plan add rate limiting to the API endpoints

Delegate A Task

/copilot:task --background --write refactor the database connection pool
/copilot:status
/copilot:result

Architecture

copilot -p "<prompt>" --silent --no-ask-user --model gpt-5.4 --output-format json

No daemon, no WebSocket, no IPC. Each command spawns one copilot CLI process, parses JSONL output, and returns. Background jobs are managed via state files in ~/.claude/plugins/data/copilot/.

Flag	Purpose
-p	Non-interactive prompt mode
--silent	Suppress stats output
--no-ask-user	No interactive prompts
--model gpt-5.4	Model selection
--output-format json	JSONL event stream
--allow-all-tools	Write-capable tasks
--effort <level>	Reasoning effort (low/medium/high/xhigh)
--resume=<id>	Continue a previous session
--autopilot	Self-continue for multi-step tasks

FAQ

Do I need a separate account?

No. The plugin uses your local Copilot CLI authentication. If you are already signed into the Copilot CLI on this machine, it works immediately.

If you have not used the Copilot CLI before, run copilot login to authenticate with your GitHub account.

How many premium requests does this use?

Each review, adversarial review, ask, task, plan, review-plan, adversarial-plan-review, and rescue command uses exactly 1 Copilot premium request. Setup, status, result, and cancel use 0. The optional review gate uses 1 per stop.

License

Apache-2.0. See LICENSE.