Skip to content

superpowers
Enforce Best Practices for AI Coding Agents

Don't let AI be lazy. Superpowers provides 14 composable skills that enforce systematic design, TDD development, and code review for AI agents, ensuring production-quality delivery every time. Supports Claude Code, OpenCode, and Codex platforms.

Why Choose Superpowers?

Inject Software Engineering Discipline into AI Coding Agents

🎯 Enforce Design Dialogue

Before AI writes the first line of code, enforce requirement clarification and architecture design to avoid rework caused by over-rationalization.

🧪 TDD Discipline Enforcement

No code without tests, enforcing the RED-GREEN-REFACTOR cycle to ensure test coverage and code maintainability.

🔍 Systematic Debugging

Four-stage debugging workflow (Observe → Hypothesize → Verify → Fix), rejecting blind guesses and random fixes.

👥 Sub-Agent Driven Development

Specialized division of labor for rapid iteration. Utilize multiple sub-agents to execute tasks in parallel, with automatic code review for every step.

🔀 Git Worktree Isolation

Create isolated environments for each feature, keeping the main branch clean and enabling conflict-free parallel development.

⚡ Concurrent Agent Dispatch

When facing multiple independent tasks, dispatch multiple sub-agents in parallel to significantly improve development efficiency.

📋 Code Review Workflow

Automatically trigger code review after each task completion, maintaining code quality and catching issues early.

🛡 Evidence-First Verification

Require genuine verification before claiming completion (passing tests, successful builds, updated docs), eliminating half-finished deliveries.

🔌 Three Platform Support

Seamless integration with Claude Code, OpenCode, and Codex AI coding platforms, with automatic skill loading.

✨ Extensible Skill System

Write custom skills and create specialized workflows. Skills are stored as files, making them easy to version control and share.

Course Outline
4 Chapters, 17 Lessons