One Person Wrote 80% of the Commits Behind 268,819 Stars
As of August 8, 2026, obra/superpowers has 268,819 stars and 38 contributors — but one contributor accounts for 80% of commits. The repository is 302 days old, pushed to today, and ships a test directory with no CI configuration to run it automatically. It's MIT licensed with no copyleft obligations from the license itself.
| Last push | 0 days ago |
| Repository age | 302 days |
| Stars / Forks | 268,819 / 24,002 |
| Open issues | 325 |
| License | MIT (not copyleft) |
| Contributors | 38 |
| Top contributor's commit share | 80% |
| Tests present | Yes |
| CI configured | No |
| Releases | 5 (latest v6.2.0) |
Ten months, a quarter-million stars
GitHub's created_at field puts obra/superpowers at 302 days old. In that window it accumulated 268,819 stars and 24,002 forks, and it was pushed to again today — last_push_days_ago is 0. That's not a project that grew slowly and then got discovered; the star count and the age together describe something that took off fast and is still being actively worked on right now, not a repo coasting on an old spike.
The language breakdown backs up what the description says it is: mostly Shell (55%) and JavaScript (38%), with TypeScript and HTML each at 2%. This reads as a framework of scripts and tooling glue rather than a compiled application — worth knowing before you assume it fits into a codebase the way a typical library would.
Who's actually writing this
38 people have contributed commits to this repository. Of those, one person authored 80% of them. The other 37 contributors split the remaining 20% between them. This is a measure of who has written the code that's currently in the repository — it says nothing about who reviews pull requests, who else understands the codebase well enough to maintain it, or what succession plan (if any) exists if that person moves on. GitHub's public data doesn't expose any of that, and this piece won't guess at it.
A quarter-million stars can make a project feel like an institution. The commit history says something narrower: most of the code currently in this repository was written by one person. If you're deciding whether to depend on this long-term, that's a data point worth weighing alongside the popularity, not instead of it.
Tests exist. Nothing runs them automatically.
The repository has a test directory. It does not have a CI configuration — no GitHub Actions workflow, no CircleCI config, nothing that would automatically run those tests on a push or pull request. That means a test suite exists, but there's no public signal that it's being run on every change, or that it's currently passing. There's also no Dockerfile, so there's no committed, reproducible way to spin the project up in a container. Neither of these is unusual for a young project, but they're both worth checking yourself before you build on top of this rather than assuming they're handled.
The license, precisely
obra/superpowers is MIT licensed, which is not a copyleft license — it doesn't require you to open-source code you build alongside it. That's what the license itself grants and requires; it says nothing about whether the project's own dependencies carry different terms, and this data doesn't cover that. If your legal review needs a clean answer, the license text is the place to start, not this article.
Questions people ask
How many people maintain obra/superpowers?
38 people have contributed commits, as of August 8, 2026. One contributor is responsible for 80% of those commits.
Is obra/superpowers still actively maintained?
Yes, as of the check on August 8, 2026 — the most recent push to the repository was 0 days before that date.
Does obra/superpowers have automated testing in CI?
A test directory is present in the repository, but no CI configuration file was found, so there is no public evidence the tests run automatically on each change.
What license does obra/superpowers use?
MIT License, which is not a copyleft license.
Who should use this, who should wait
If you're experimenting or building something where you can absorb breaking changes, the popularity, the daily commit activity, and the permissive MIT license make this a low-friction thing to try. If you're a team weighing this against building the equivalent in-house or paying a vendor, look past the star count at the two things GitHub's numbers can't paper over: most of the current codebase was written by one person, and there's no automated pipeline publicly confirming the test suite passes on every change. Neither fact means the project is unsound — but a team betting production infrastructure on it should verify both independently rather than take the star count as a proxy for either.
Source: obra/superpowers on GitHub
Comments
Post a Comment