Superpowers — Agentic Skills Framework for AI Coding
obra/superpowersSuperpowers is a comprehensive framework that provides AI coding agents with a structured software development methodology and a suite of composable skills. It guides agents through a systematic workflow from brainstorming to code review and completion.
Superpowers offers a structured, TDD-focused methodology for AI coding agents, enhancing their systematic approach and autonomy in software development. It's suitable for developers aiming to improve agent reliability and workflow adherence.
The problem it solves
AI coding agents often lack structured development processes, leading to ad-hoc code generation without proper planning, design, or adherence to best practices like Test-Driven Development (TDD). This can result in unstructured, error-prone, or off-spec software.
What is it?
Superpowers is an agentic skills framework and software development methodology designed to equip AI coding agents with a systematic approach to software development. It integrates a library of skills that guide agents through phases like design, planning, implementation, and review, enforcing practices such as TDD and subagent-driven development across various AI coding platforms.
Why it's getting attention
With over 250,000 stars and its focus on AI, brainstorming, and software development lifecycle (SDLC) topics, Superpowers is a notable framework for enhancing the capabilities of AI coding agents.
Key features
- ✓Agentic skills framework for AI coding
- ✓Complete software development methodology for agents
- ✓Supports subagent-driven development processes
- ✓Enforces Test-Driven Development (TDD) and other best practices
- ✓Provides a structured workflow from brainstorming to code review
- ✓Integrates with multiple AI coding platforms (e.g., Claude, Cursor)
- ✓Includes a skills library for testing, debugging, and collaboration
Best use cases
- •Guiding AI agents through full software development cycles
- •Automating TDD practices within agent-generated code
- •Facilitating structured design and implementation planning by AI
- •Enabling parallel development with Git worktrees via agent orchestration
- •Enhancing AI agent adherence to engineering best practices
How to install / try
Installation instructions are provided for various AI coding harnesses, including Claude Code, Antigravity, Codex App/CLI, Cursor, Factory Droid, GitHub Copilot CLI, Kimi Code, OpenCode, and Pi. The process typically involves using plugin installation commands specific to each platform or selecting Superpowers from their respective marketplaces.
How to use
Once installed with a compatible coding agent, Superpowers activates automatically. It guides the agent through a systematic workflow starting with brainstorming to refine ideas, creating implementation plans, and executing tasks via subagent-driven development. The framework enforces TDD and review steps, with skills triggering automatically.
Strengths
- ✓Offers a comprehensive, structured methodology for AI agents
- ✓Enforces core engineering best practices like TDD, YAGNI, and DRY
- ✓Supports autonomous subagent-driven development for task execution
- ✓Integrates with a wide range of popular AI coding platforms
- ✓Provides a modular skills library for various development phases
- ✓Aims to reduce ad-hoc coding by AI agents through systematic processes
Limitations & risks
- △Requires separate installation for each AI coding harness used
- △The project generally does not accept contributions of new skills
- △Skill updates must ensure compatibility across all supported coding agents
- △Functionality is dependent on specific AI agent platforms, not a standalone tool
Alternatives
Who should try it — and who should skip
Developers and teams utilizing AI coding agents (e.g., Claude Code, Cursor, GitHub Copilot CLI) who want to enforce structured software development methodologies, Test-Driven Development, and a systematic workflow for their agents. Individuals seeking a standalone, agent-agnostic coding tool or who need to frequently contribute new custom skills to the core framework might find it less suitable.
Frequently asked questions
Superpowers is an agentic skills framework and software development methodology designed to give AI coding agents a structured, systematic approach to building software.
Superpowers supports integration with platforms such as Claude Code, Antigravity, Codex App/CLI, Cursor, Factory Droid, GitHub Copilot CLI, Kimi Code, OpenCode, and Pi.
It strongly emphasizes Test-Driven Development (TDD), systematic problem-solving over ad-hoc methods, complexity reduction, and evidence-based verification.
The project generally does not accept contributions of new skills, and any proposed updates must maintain compatibility across all supported coding agents.