Analysis updated 2026-07-05 · repo last pushed 2026-06-03
Stop your AI assistant from writing 500 lines of code in the wrong direction.
Make your AI coding tool follow test-driven development automatically.
Get AI to brainstorm and plan a design document before writing any code.
| joshuakgoldberg/superpowers | 0xhassaan/nn-from-scratch | 0xzgbot/hermes-comfyui-skills | |
|---|---|---|---|
| Stars | — | 0 | 0 |
| Language | — | Python | — |
| Last pushed | 2026-06-03 | — | — |
| Maintenance | Maintained | — | — |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 2/5 | 4/5 | 1/5 |
| Audience | developer | developer | designer |
Figures from each repo's GitHub metadata at analysis time.
Requires choosing the right installation method for your specific AI coding tool out of roughly eight supported options.
Superpowers is a plugin that makes AI coding assistants work in a more deliberate, structured way. Instead of letting your AI agent immediately jump into writing code the moment you describe a feature, it forces the agent to slow down, ask you clarifying questions, write up a design document, create a detailed plan, and only then start building, all while running tests at every step. The workflow happens in stages. First, the agent brainstorms with you to nail down what you actually want to build. Once you approve the design, it breaks the work into small tasks (each a few minutes long) with specific file paths and verification steps. Then it dispatches subagents to tackle each task one by one, reviewing their work for both spec compliance and code quality before moving on. Throughout implementation, it enforces test-driven development: write a failing test, write the minimum code to pass it, then refactor. The system wraps up by verifying everything works and helping you merge or discard the branch. Anyone using tools like Claude Code, Cursor, GitHub Copilot CLI, Gemini CLI, or similar AI coding assistants would benefit. If you've ever watched an AI agent confidently write 500 lines of code in the wrong direction because it misunderstood your goal, this project exists to prevent that. It's designed for founders, solo developers, or small teams who want their AI assistant to function more like a disciplined engineer than an over-eager intern, one that checks its work, asks questions before acting, and proves things actually function rather than just claiming they do. The project is notable for being cross-platform: it works across roughly eight different AI coding tools, each with its own installation method. It also operates entirely through automatically triggered skills rather than requiring you to remember specific commands. The philosophy leans heavily on simplicity, systematic process over guesswork, and evidence over claims, tradeoffs that favor reliability over speed when the two conflict.
A plugin for AI coding assistants that forces them to slow down, ask questions, plan, and run tests before writing code. It makes AI act like a disciplined engineer rather than an over-eager intern.
Maintained — commit in last 6 months (last push 2026-06-03).
The license is not specified in the repository explanation.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.