About this project

Superpowers is a complete software development methodology designed to change how AI coding agents work. It is built as a set of composable skills plus initial instructions that cause the agent to use them automatically. Instead of jumping straight into writing code, an agent using Superpowers first clarifies what the user is trying to build, presents a design in readable sections, and gets sign-off before creating an implementation plan. The project describes a core workflow made of mandatory, automatically triggered stages: brainstorming refines the idea into a design document; git worktrees create an isolated environment; writing-plans breaks work into small tasks with exact file paths and verification steps; subagent-driven development or executing-plans dispatches tasks with two-stage review; test-driven development enforces a RED-GREEN-REFACTOR cycle; requesting-code-review checks work against the plan; and finishing-a-development-branch handles merging or cleanup. The skills library includes testing, systematic debugging, verification before completion, brainstorming, planning, parallel agent dispatch, code review, worktree management, and meta-skills for writing new skills. The project emphasizes test-first development, systematic process over guessing, complexity reduction, and evidence over claims. Superpowers can be installed across many coding agent harnesses, including Claude Code, Antigravity, Codex App/CLI, Cursor, Devin CLI, Factory Droid, Gemini CLI, GitHub Copilot CLI, Grok Build CLI, Kimi Code, OpenCode, Pi, and Hermes Agent, with instructions provided for each. The project is MIT-licensed. It includes an optional visual companion telemetry feature that loads the Prime Radiant logo to estimate usage and version spread; it can be disabled via the SUPERPOWERS_DISABLE_TELEMETRY environment variable or Claude Code telemetry opt-outs.