Superpowers is an open-source plugin framework by obra that gives AI coding agents a complete, disciplined development workflow built on composable, auto-triggering skills. Rather than letting an agent jump straight into writing code, Superpowers enforces a structured pipeline: Socratic brainstorming to refine a spec, isolated git worktrees, granular implementation plans, subagent-driven development with two-phase code review (spec compliance, then code quality), strict red-green-refactor TDD, and clean branch finishing. The article walks through installation on Claude Code via its plugin marketplace, plus manual setup for Codex and OpenCode, and details the seven core workflows and skill categories (testing, debugging, collaboration, meta-skills). It also traces the project's evolution from v2.0.0—when skills were split into the separate obra/superpowers-skills repository—to v4.1.1, covering features like DOT flowcharts as executable specifications and slash-command restrictions. The underlying philosophy: test-driven development, systematic processes over improvisation, complexity reduction, and evidence over claims, transforming agents from聪明-but-unreliable coders into dependable engineers.
Superpowers: Turning AI Coding Agents from Ordinary to Legendary
Summary
Superpowers is an open-source plugin framework by obra that gives AI coding agents a complete, disciplined development workflow built on composable, auto-triggering skills. Rather than letting an agent jump straight into writing code, Superpowers enforces a structured pipeline: Socratic brainstorming to refine a spec, isolated git worktrees, granular implementation plans, subagent-driven development with two-phase code review (spec compliance, then code quality), strict red-green-refactor TDD, and clean branch finishing. The article walks through installation on Claude Code via its plugin marketplace, plus manual setup for Codex and OpenCode, and details the seven core workflows and skill categories (testing, debugging, collaboration, meta-skills). It also traces the project's evolution from v2.0.0—when skills were split into the separate obra/superpowers-skills repository—to v4.1.1, covering features like DOT flowcharts as executable specifications and slash-command restrictions. The underlying philosophy: test-driven development, systematic processes over improvisation, complexity reduction, and evidence over claims, transforming agents from聪明-but-unreliable coders into dependable engineers.
This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/176922612