By CrewWorkAI
Use GitHub Copilot CLI from Claude Code for code review, rescue, and tracked background workflows. Companion to openai/codex-plugin-cc, but with Copilot as the worker.
Run an adversarial-framed Copilot review on the current work.
Cancel a running GitHub Copilot delegation background job.
Delegate a coding task to GitHub Copilot CLI. Useful for second opinions or when Copilot's model lineup includes the right tool for the job.
Fetch the final output of a tracked GitHub Copilot delegation job.
Run a GitHub Copilot code review on the current work using the copilot-plugin-codex script.
Use when the user asks Copilot for a harsh review, red-team pass, challenged assumptions, or a sharper critique than `$copilot:review`.
Use when the user asks to cancel, kill, or stop a tracked background Copilot job. Do not trigger for foreground runs.
Use when the user mentions Copilot CLI generically, asks to delegate to Copilot, or asks for a Copilot second opinion without naming a specific `$copilot:<command>`.
Use when the user asks Copilot to perform a coding task, delegate work, or use a Copilot-only model. Do not trigger for code reviews or host-agent work.
Use when the user asks what a tracked Copilot job produced or wants the result transcript for a completed background job.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Delegate work from Codex or Claude Code to GitHub Copilot CLI.
A modified fork of sendbird/cc-plugin-codex, retargeted at GitHub Copilot CLI as the worker and extended to support both Codex and Claude Code as hosts from a single codebase. Layout, command surface, and the per-command skill structure are inherited from the upstream; the wrapper script, manifests, and skill bodies are rewritten for Copilot. See NOTICE for full attribution.
Companion to the existing cross-CLI delegation plugins:
| Plugin | Host | Worker |
|---|---|---|
openai/codex-plugin-cc | Claude Code | Codex |
sendbird/cc-plugin-codex | Codex | Claude Code |
CrewWorkAI/copilot-plugin-codex (this) | Codex or Claude Code | Copilot |
Copilot CLI is the only major coding agent CLI that lets you switch between Anthropic, OpenAI, and Google models mid-session (Claude Sonnet/Opus/Haiku 4.5–4.6, GPT-5.3-Codex, Gemini 3 Pro). From inside Codex or Claude Code, that turns Copilot into a useful "use the right model for this task" subprocess — including models the host doesn't natively expose.
codex plugin marketplace add CrewWorkAI/copilot-plugin-codex
Current Codex CLI versions register plugins through the marketplace command above. Start a new Codex thread after adding the marketplace, then use:
$copilot:setup
$copilot:review
$copilot:rescue <task description>
/plugin marketplace add CrewWorkAI/copilot-plugin-codex
/plugin install copilot-plugin-codex@copilot-plugin-codex
Then:
/copilot:setup
/copilot:review
/copilot:rescue <task description>
scripts/validate.sh and used as a JSON fallback by the runtime wrapper)jq recommended for job metadata reads; without it, the runtime wrapper falls back to Python 3npm install -g @github/copilot (run $copilot:setup to verify or get install instructions)gh auth login completed at least once| Codex | Claude Code | Purpose |
|---|---|---|
$copilot:setup | /copilot:setup | Verify Copilot CLI install & auth |
$copilot:review | /copilot:review | Copilot review on current branch |
$copilot:adversarial-review | /copilot:adversarial-review | Hostile-reviewer-framed review |
$copilot:rescue <task> | /copilot:rescue <task> | Delegate an arbitrary task |
$copilot:status [job-id] | /copilot:status [job-id] | List/inspect tracked jobs |
$copilot:result <job-id> | /copilot:result <job-id> | Get final output of a job |
$copilot:cancel <job-id> | /copilot:cancel <job-id> | Cancel a background job |
All commands accept --model <model-id> to override Copilot's default (currently Claude Sonnet 4.5). Full model list in skills/copilot/SKILL.md.
This plugin targets both Codex and Claude Code from one codebase:
copilot-plugin-codex/
├── .codex-plugin/plugin.json ← Codex reads this
├── .claude-plugin/
│ ├── plugin.json ← Claude Code reads this
│ └── marketplace.json ← Claude Code marketplace catalog (one-entry)
├── .agents/plugins/marketplace.json ← Codex marketplace catalog (one-entry)
├── skills/copilot/SKILL.md ← Canonical instructions (both hosts read this)
├── commands/copilot-*.md ← Claude Code slash command definitions
├── hooks/hooks.json ← Codex hooks (review gate, disabled by default)
└── scripts/
├── copilot-exec.sh ← The `copilot -p` wrapper
└── setup.sh ← Install/verify Copilot CLI
scripts/copilot-exec.sh reads HOST=codex or HOST=claude from the environment to pick the right state directory (~/.codex/plugins/... vs ~/.claude/plugins/...). Both hosts shell out to the same script.
copilot -p mode (github/copilot-cli#633). If a delegated task needs MCP tooling, the plugin will surface this and suggest an interactive Copilot session instead.$copilot:setup checks for this.npx claudepluginhub crewworkai/copilot-plugin-codex --plugin copilot-plugin-codexv9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Persistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development