novasky-ai/skyrl
Grade a repo

0

/ 100

GradeC

Good community traction and solid fundamentals. Nearly in the top tier.

Top 33% of 5,156 graded repos

SkyRL: A Modular Full-stack RL Library for LLMs

Documentation88
Engineering68
Health96
Python2,017Apache-2.01mo ago

Top fixes

Highest-impact changes first, ranked by point weight

13 to address
  1. 1
    Tests18pt

    Wire your tests to a documented command (e.g. a test script in your build config) so the suite is reproducible.

  2. 2
    CI/CD14pt

    Add a step like `run: npm test`, `run: pytest`, or `run: tox` to your workflow file.

  3. 3
    CI/CD14pt

    Add a lint step to catch style issues automatically.

  4. 4
    CI/CD14pt

    Add `tsc --noEmit`, `mypy`, or `cargo check` to catch type errors before they merge.

Working through the fixes? Let every push regrade itself.

The free GitHub App rescans this repo on every push and posts the grade as a commit check, so the score climbs without coming back to rescan by hand.

Install the GitHub App

Scorecard

Every check, grouped by category and sorted worst-first

Documentation

88

Contributing guide5pt72

CONTRIBUTING guide found.

README12pt80

README is present.

Install and run instructions9pt100

README documents how to install the project.

License6pt100

Licensed under Apache-2.0.

Engineering

68

Issue and PR templates6pt0

No issue or PR templates found (−100 pts).

→ Add .github/ISSUE_TEMPLATE/ with bug_report.md and feature_request.md to guide contributors. It dramatically improves issue quality.

CI/CD14pt57

CI is configured (.github/workflows/gpu_ci_h100.yaml).

Reproducibility6pt80

Lockfile present (uv.lock). Installs are reproducible.

Tests18pt85

Test files detected (examples/train/gptoss/test_gptoss.py).

Linting and formatting5pt100

Linter or formatter configured ([tool.ruff] / [tool.black] in pyproject.toml).

Project health

96

Dependency manifest6pt86

Dependency manifest found (pyproject.toml).

Repository metadata5pt100

Repository has a description.

Activity5pt100

Actively maintained (pushed within the last month).

Housekeeping3pt100

.gitignore present.

Repository health signals

Activity, community, and responsiveness at scan time

Activity

  • -
    Commits (30d / 90d)
  • 359
    Forks
  • 6
    Releaseslatest 3mo ago

Community

  • -
    Community health
  • -
    authors own >50% of commits
  • 2,017
    Watchers

Responsiveness

  • 13h
    Median issue response
  • <1h
    Median PR merge time
  • 357
    Open issues
Repository files22 root entries
  • .claude
    Good: CONTRIBUTING guide found.
    Issue: CONTRIBUTING guide contents could not be read (−28 pts vs a readable file).Fix: Move the file to the repo root or docs/CONTRIBUTING.md so its setup, style, test, and PR sections can be graded.
  • .gemini
  • .github
    Good: CI is configured (.github/workflows/gpu_ci_h100.yaml).
  • ci
  • docker
    Good: Environment pinned via docker/Dockerfile.
  • docs
  • examples
    Good: Test files detected (examples/train/gptoss/test_gptoss.py).
  • skyrl
  • skyrl-agent
  • skyrl-gym
  • skyrl-train
  • skyrl-tx
  • tests
  • .gitignore
    Good: .gitignore present.
  • .pre-commit-config.yaml
  • .python-version
  • CLAUDE.md
  • format.sh
  • LICENSE
    Good: Licensed under Apache-2.0.
  • pyproject.toml
    Good: Dependency manifest found (pyproject.toml).
  • README.md
    Good: README is present.
    Good: README is well structured with multiple sections.
    Issue: No screenshots or images in the README (−20 pts).Fix: Add a GIF, screenshot, or logo image. It is the fastest way to show what your project does.
    Good: README has code examples.
    Good: README links to a live demo or deployed app.
    Good: README includes status badges.
    Good: README documents how to install the project.
    Good: README documents how to run the project.
  • uv.lock
    Good: Lockfile present (uv.lock). Installs are reproducible.
RepoGrade badge preview

Add this badge to your README

It updates automatically each time the repo is re-graded.

[![RepoGrade](https://www.repo-grade.com/api/badge/novasky-ai/skyrl)](https://www.repo-grade.com/report/novasky-ai/skyrl)

More graded Python repos