S

Superpowers

7-stage agentic methodology that stops AI from just winging it

PriceOpen Source / Free (MIT)Reviewed2026-04-29
Verdict — Ship
3 Ships1 Skips
Visit github.com

The Panel's Take

Superpowers is an open-source agentic skills framework by Jesse Vincent (obra) that enforces a structured 7-stage software development methodology for coding agents. Instead of having Claude or Codex immediately start writing code, Superpowers makes the agent pause, brainstorm, create git worktrees, plan bite-sized 2-5 minute tasks, dispatch sub-agents, enforce TDD, do code review, and then handle branch completion — all as a coherent orchestrated workflow. The seven stages are: Brainstorming (iterative requirement refinement), Git Worktrees (isolated dev environments per feature), Planning (task decomposition), Subagent Development (parallel task execution with review cycles), TDD (red-green-refactor enforcement), Code Review (spec validation), and Branch Completion (merge decisions and cleanup). It works across Claude Code, OpenAI Codex, Cursor, GitHub Copilot CLI, and Gemini CLI. Released under MIT, Superpowers trended on GitHub with 1,683 stars in a single day — unusually high for a methodology-first tool. It hits a real pain point: agents are often good at writing individual functions but terrible at sustained, coherent feature development. This framework is explicitly designed to fill that gap.

Share this verdict

Superpowers verdict: SHIP 🚀

3 ships · 1 skip from the expert panel

Full review: shiporskip.io/tool/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026

Weekly AI Tool Verdicts

Get the next verdict in your inbox

7 critics review a new AI tool every day. Weekly digest — free.

Embed this verdict

Tool makers can add a live ShipOrSkip badge to their site. Badge loads track impressions; clicks route back to this review.

Ship · 7.5/10
HTML badge
<a href="https://shiporskip.io/api/badge-click/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026" target="_blank" rel="noopener"><img src="https://shiporskip.io/api/badge/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026" alt="Superpowers Ship verdict on ShipOrSkip" width="360" height="90" /></a>
Markdown badge
[![Superpowers Ship verdict on ShipOrSkip](https://shiporskip.io/api/badge/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026)](https://shiporskip.io/api/badge-click/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026)
Iframe widget
<iframe src="https://shiporskip.io/embed/obra-superpowers-agentic-skills-framework-software-methodology-mit-2026" title="Superpowers ShipOrSkip verdict" width="360" height="260" style="border:0;border-radius:16px;max-width:100%;" loading="lazy"></iframe>

The reviews

The git worktrees per feature approach is something I wish I'd done from day one — isolated environments per task means agents can't accidentally clobber each other's work. The RED-GREEN-REFACTOR enforcement alone makes this worth the setup time.

Helpful?

Seven stages sounds great in a README but in practice agents still go off-rails mid-workflow — you're just adding structure around unreliable behavior. And the cross-platform support claim needs stress-testing; behavior in Claude Code vs Cursor vs Codex will differ significantly.

Helpful?

Superpowers is proof that the killer abstraction for the agent era isn't a new model — it's structured methodology. Agent orchestration frameworks at the prompt level are the 'Scrum for AI' moment; whoever codifies this best will define how software is built for the next decade.

Helpful?

The brainstorming phase that forces agents to ask clarifying questions before touching code is such an underrated feature. So many of my worst agent sessions started with me giving a vague prompt and the agent just confidently building the wrong thing for 20 minutes.

Helpful?

Bookmarks

Loading bookmarks...

No bookmarks yet

Bookmark tools to save them for later