human-agent-society/coralNew scan
97/ 100 · A

Polished and well engineered. Punching above its star count.

CORAL is a robust, lightweight infrastructure for multi-agent autonomous self-evolution, built for autoresearch. Works with Claude Code, Codex, Cursor, OpenCode, Kiro, and more.

Python749 starsApache-2.0updated today

Outstanding work. A score of 97/100 puts this repo in a very small tier of truly well-engineered open source projects.

DocumentationREADME, setup, examples, license
95
EngineeringTests, CI, linting, lockfiles
98
Project healthDescription, activity, stars, deps
100

What to fix first

The highest-impact improvements for this repo.

  1. 1
    CI/CD
    EngineeringInfo

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

  2. 2
    CI/CD
    EngineeringInfo

    Upload coverage to Codecov, Coveralls, or report it with `--coverage` flags.

  3. 3
    Install and run instructions
    DocumentationIssue

    Add a .env.example listing all required environment variables so contributors know what to set up.

Detailed breakdown

Documentation

95
  • README100
    • README is present.
    • README is well structured with multiple sections.
    • README includes screenshots or visuals. Great for first impressions.
    • README has code examples.
    • README links to a live demo or deployed app.
    • README includes status badges.
  • Install and run instructions90
    • README documents how to install the project.
    • README documents how to run the project.
    • No .env.example found (−10 pts).Add a .env.example listing all required environment variables so contributors know what to set up.
  • License100
    • Licensed under Apache-2.0.
  • Contributing guide88
    • Contributing guide is detailed and thorough.
    • Contributing guide lacks a setup section (−12 pts).Show new contributors how to get a local dev environment running.
    • Contributing guide describes code style expectations.
    • Contributing guide explains how to run tests.
    • Contributing guide describes the PR/review workflow.
    • Contributing guide includes code examples.
    • Optional: add a Code of Conduct.A CODE_OF_CONDUCT.md signals that your project is welcoming. GitHub has a template you can add in one click.

Engineering

98
  • Tests100
    • Test files detected (examples/frontier_eng/WirelessChannelSimulation/HighReliableSimulation/seed/tests).
    • Pytest is fully configured in pyproject.toml with testpaths and test files detected.
  • CI/CD100

    Not applicable?

    • CI is configured (.github/workflows/ci.yml).
    • CI workflow runs tests.
    • CI runs on pull requests, not just on pushes to main.
    • CI workflow runs a lint or format check.
    • Optional: add type checking to CI.Add `tsc --noEmit`, `mypy`, or `cargo check` to catch type errors before they merge.
    • Optional: report test coverage in CI.Upload coverage to Codecov, Coveralls, or report it with `--coverage` flags.
    • CI tests across multiple environments or versions.
  • Linting and formatting100
    • Linter or formatter configured (web/eslint.config.js).
  • Reproducibility80
    • Lockfile present (uv.lock). Installs are reproducible.
    • Environment pinned via docker/claude/Dockerfile.
    • No Dependabot config (adding it earns up to +20 pts).Add .github/dependabot.yml with at least one package-ecosystem entry so dependencies are updated automatically.
  • Issue and PR templates100
    • Issue or PR templates present.
    • Optional: add a SECURITY.md.A SECURITY.md explains how to responsibly disclose vulnerabilities. Worth adding once the project has real users.

Project health

100
  • Dependency manifest100
    • Dependency manifest found (pyproject.toml).
    • pyproject.toml has a [project] table with package metadata.
    • pyproject.toml includes a description.
    • pyproject.toml specifies requires-python, preventing installs on incompatible versions.
    • pyproject.toml has a [build-system] table. The package can be built and published.
  • Repository metadata100
    • Repository has a description.
    • Primary language detected: Python.
    • pyproject.toml [project] metadata is complete (description, authors, urls).
  • Activity100
    • Actively maintained (pushed within the last month).
    • 749 stars.
  • Housekeeping100
    • .gitignore present.

Repository health signals

Activity, community, and responsiveness at scan time

Activity

  • —
    Commits (30d / 90d)
  • 96
    Forks
  • 6
    Releaseslatest 10d ago

Community

  • —
    Community health
  • —
    authors own >50% of commits
  • 749
    Watchers

Responsiveness

  • 2h
    Median issue response
  • <1h
    Median PR merge time
  • 11
    Open issues
Repository files25 root entries
  • .claude
  • .github
    Good: CI is configured (.github/workflows/ci.yml).
    Good: Issue or PR templates present.
  • assets
  • blog
  • coral
  • demos
  • docker
    Good: Environment pinned via docker/claude/Dockerfile.
  • docs
  • examples
    Good: Test files detected (examples/frontier_eng/WirelessChannelSimulation/HighReliableSimulation/seed/tests).
  • scripts
  • tests
  • web
    Good: Linter or formatter configured (web/eslint.config.js).
  • .dockerignore
  • .gitignore
    Good: .gitignore present.
  • .pre-commit-config.yaml
  • AGENTS.md
  • CLAUDE.md
  • CONTRIBUTING.md
    Good: Contributing guide is detailed and thorough.
    Issue: Contributing guide lacks a setup section (−12 pts).Fix: Show new contributors how to get a local dev environment running.
    Good: Contributing guide describes code style expectations.
    Good: Contributing guide explains how to run tests.
    Good: Contributing guide describes the PR/review workflow.
    Good: Contributing guide includes code examples.
  • install.sh
  • LICENSE
    Good: Licensed under Apache-2.0.
  • litellm_config.yaml
  • pyproject.toml
    Good: Dependency manifest found (pyproject.toml).
  • README_CN.md
    Good: README is present.
    Good: README is well structured with multiple sections.
    Good: README includes screenshots or visuals. Great for first impressions.
    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.
  • README.md
  • uv.lock
    Good: Lockfile present (uv.lock). Installs are reproducible.