The front door to open source
Find the issue
you can fix.
Star count is a terrible way to choose a first contribution. Score a repository on starter issues, contributor docs, and real activity — then read the issues that are already scoped for you.
The score
Is this repo
ready for you?
A high score is not a promise of a merged PR. It means the repo has the signals that make a first contribution more likely to be useful and less likely to be ignored.
Reading public data
Opening the repository…
What the score reads
Six signals, no vibes.
Every point comes from something public and checkable. If a signal is missing, the breakdown says so instead of quietly guessing.
- 40Starter issues
Open issues labeled
good first issueorhelp wanted. - 15Contributor docs
A real CONTRIBUTING file, not a link to a wiki.
- 15Recent activity
Pushed within the last 120 days.
- 10Starter labels
The repo actually uses the labels it wants you to find.
- 10Manageable scale
Under 50k stars scores full marks; huge repos are harder to navigate.
- 10Clear license
An explicit license, so your work has a home.
Score breakdown
0 / 100Starter issues
0 foundYour first-PR checklist
Copy it into the issueThe repo board
Don't have a repo
in mind?
Search public repositories that already use starter labels, filtered by language and recency. Pick one, score it, then read the actual issues.
Repository search uses one request. Results are sorted by stars; the score is separate and always shown after you run it.
How it works
A small score
with a paper trail.
The browser calls the public GitHub API directly. No backend, no token, no account. Repository data, open issues, labels, and the community profile are read once per report.
The CLI uses the same formula and can rank a list of repositories in one table. It works with gh when available and falls back to the public REST API.
- 01
Read the repo, not the stars
Starter labels, a CONTRIBUTING file, recent pushes, license, and scale. The score is 100 points and every point is visible.
- 02
Open a real issue
The report lists the actual labeled issues, with comments and last activity. If there are none, it says so.
- 03
Keep the first PR small
One behavior, one test, one clear description. The checklist in the report is the same one the CLI prints.