Superpowers for GitHub Copilot CLI
Until obra/superpowers ships native GitHub Copilot CLI support, this repo packages the full Superpowers skills library using Copilot CLI's native plugin and skills system. All skills are Jesse Vincent's original work — this repo only adds the Copilot CLI wiring.
What is Superpowers?
Superpowers by Jesse Vincent is a complete software development workflow for AI coding agents. It transforms how your agent works:
- Brainstorming → Refines ideas through questions before writing code
- Test-Driven Development → Enforced RED-GREEN-REFACTOR, no exceptions
- Systematic Debugging → 4-phase root cause analysis, no random fixes
- Plan Writing → Bite-sized tasks with exact file paths and commands
- Subagent-Driven Development → Fresh subagent per task with two-stage review
- Code Review → Automated review against plan and coding standards
Superpowers already supports Claude Code, Cursor, Codex, Gemini CLI, and OpenCode. This repo adds GitHub Copilot CLI to that list.
Installation
Method 1 — One command (easiest)
macOS / Linux:
curl -fsSL https://raw.githubusercontent.com/nguyenducviet4/superpowers/main/install.sh | bash
Windows (CMD):
install.cmd
Windows (PowerShell):
powershell -ExecutionPolicy Bypass -File install.ps1
Method 2 — Native plugin system (direct install)
copilot plugin install nguyenducviet4/superpowers:plugins/superpowers
Method 3 — Via marketplace
copilot plugin marketplace add nguyenducviet4/superpowers
copilot plugin install superpowers@superpowers
Verify
Start a new Copilot CLI session and run:
/skills list
You should see all 14 Superpowers skills listed. To verify the agent:
/agents
What's Included
14 Skills
| Skill | When it activates |
|---|---|
| brainstorming | Before any creative work — creating features, building components |
| test-driven-development | When implementing any feature or bugfix, before writing code |
| systematic-debugging | When encountering any bug, test failure, or unexpected behavior |
| writing-plans | When you have specs for a multi-step task, before touching code |
| executing-plans | When executing an implementation plan with human checkpoints |
| subagent-driven-development | When executing plans with independent tasks via subagents |
| dispatching-parallel-agents | When facing 2+ independent tasks that can be worked on in parallel |
| requesting-code-review | Between tasks, to review against plan |
| receiving-code-review | When responding to review feedback |
| verification-before-completion | Before declaring any fix or feature done |
| using-git-worktrees | After design approval, to create isolated workspace |
| finishing-a-development-branch | When all tasks complete, to merge/PR/cleanup |
| writing-skills | When creating new skills |
| using-superpowers | Meta-skill that ensures all other skills are used when relevant |
1 Custom Agent
| Agent | Description |
|---|---|
| code-reviewer | Senior code reviewer that validates completed work against plans and coding standards |
Usage
Skills activate automatically when Copilot detects relevance. You can also invoke them explicitly:
Use the /brainstorming skill to explore this feature idea
Use the /test-driven-development skill to implement this
Use the /systematic-debugging skill to find the root cause
The Basic Workflow
- Brainstorming → "Let's build a notification system"
- Writing Plans → Agent creates detailed implementation plan
- Subagent-Driven Development → Agent dispatches subagents per task
- Test-Driven Development → Each task follows RED-GREEN-REFACTOR
- Code Review → Automated review after each step
- Verification → Confirm everything works before marking done
How It Works
This repo follows the exact same patterns as other Copilot CLI plugins: