←Leaderboard/Software Engineering/superpowers
Last commit on November 3, 2026·Created on October 9, 2025

obra/superpowers

“Plan engineering work and coordinate parallel agents.”
Combined rank
#2
across all skills
In Software Engineering
#2
category rank
Stars
318k
+1.1% in last 7d
Forks
28.4k
+1.3% in last 7d
Watchers
1.2k
+0.5% in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
TL;DR

An opinionated development workflow that asks for a design before writing code, then moves through planning, test-driven implementation, and review. The process is the point.

WHO IT'S FOR
AI agent / automation builders
implementing subagent-driven development workflows
Technical leads / architects
standardizing AI-assisted SDLC methodologies
Platform / DevEx teams
optimizing AI coding assistant performance
Software engineers using AI assistants
managing complex feature implementation with AI

Use this skill collection

Follow the documented setup, then try a first task.

I’m using

Claude Code official marketplace

Installs: The Superpowers plugin bundle from Anthropic's official marketplace. · Claude Code

  1. Enter this command in a Claude Code session to install the plugin.

    /plugin install superpowers@claude-plugins-official
README.md ↗ · Checked Nov 14, 2026

Give it something to do.

Suggested first task

Investigate one failing test or error

Uses systematic-debugging

  • Access to the affected project
  • One failing test or error, including its full output and stack trace when available

Make sure systematic-debugging is installed in your agent before trying this task.

Use the systematic-debugging skill to begin investigating one failing test or error in this project. Read the full error message and stack trace, identify relevant file paths, line numbers, and error codes, and summarize the evidence and what remains unknown before proposing any fix.
Suggested first task

Check one code review suggestion

Uses receiving-code-review

  • The review suggestion
  • Access to the relevant code and tests
  • Any applicable requirements or prior architectural decisions

Make sure receiving-code-review is installed in your agent before trying this task.

Use the receiving-code-review skill to evaluate one review suggestion against this codebase before changing anything. Restate the requirement, check whether it is technically sound and preserves existing behavior, and give an evidence-based assessment or ask for clarification if the suggestion is unclear.
Suggested first task

Verify one completion claim

Uses verification-before-completion

  • Access to the project and current change
  • One specific claim to verify
  • The project's documented verification command or enough project context to identify it
  • A configured environment capable of running that check

Make sure verification-before-completion is installed in your agent before trying this task.

Use the verification-before-completion skill to check one claim about my current change, such as whether its tests pass or its build succeeds. Identify the full verification command, run it fresh, inspect the complete output and exit code, and report only the status supported by that evidence. Stop after reporting.