obra/superpowers provides an agentic skills framework that enforces test-driven development and subagent-driven development for AI coding agents. The methodology activates skills automatically to introduce planning, brainstorming, and structured debugging before code generation.
Project overview
The framework introduces a structured software development methodology that counteracts the tendency of AI coding agents to skip planning and testing by enforcing a RED-GREEN-REFACTOR cycle and fresh subagent dispatch per task.
Project type
AI Agent · AI Coding
Use cases
Coding & Development
Deployment
Refer to project documentation
License
MIT
Best for
Developers and enterprise teams who rely on AI coding agents and require enforced planning, test-driven development, and structured reviews in the development lifecycle.
Key capabilities
A Socratic design refinement skill that activates before writing code to tease out a spec and refine ideas.
Dispatches fresh subagents per engineering task with two-stage review for spec compliance and code quality.
Enforces RED-GREEN-REFACTOR cycle: writing failing tests first, implementing minimal code, and committing once tests pass.
A 4-phase root cause process for debugging with verification before completion.
Creates isolated workspaces on new branches for parallel development and clean branch finishing workflows.
Limitations and risks
The framework must be installed separately in each coding agent harness if multiple agents are used.
The project does not generally accept community contributions of new skills.
Getting started
Installation is rated easy and requires opening a supported coding agent, such as Claude Code, and running the single plugin install command provided for the target harness.
Alternatives and comparisons
Provides over 260 reusable skills, rules, and command shims across 12+ language ecosystems alongside a cross-harness hook system for multiple AI agent harnesses.
Offers small, composable engineering skills that enforce software engineering fundamentals like test-driven development, domain modeling, and code review.
An open source AI coding agent with built-in development and planning agents that can explore codebases, plan changes, and perform development tasks.
GitHub project description: An agentic skills framework & software development methodology that works.
GitHub repository metadata: skills
README: As soon as it sees that you're building something, it *doesn't* just jump into trying to write code. Instead, it steps back and asks you what you're really trying to do.
README: Your coding agent just has Superpowers.
README: Give your agent Superpowers: [Claude Code](#claude-code), [Antigravity](#antigravity), [Codex App](#codex-app), [Codex CLI](#codex-cli), [Cursor](#cursor), [Factory Droid](#factor…