Skip to content

Gives criticality score for an open source project

License

Notifications You must be signed in to change notification settings

nathannaveen/criticality_score

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Open Source Project Criticality Score (Beta)

OpenSSF Scorecard

This project is maintained by members of the Securing Critical Projects WG.

Goals

  1. Generate a criticality score for every open source project.

  2. Create a list of critical projects that the open source community depends on.

  3. Use this data to proactively improve the security posture of these critical projects.

Criticality Score

A project's criticality score defines the influence and importance of a project. It is a number between 0 (least-critical) and 1 (most-critical). It is based on the following algorithm by Rob Pike:

We use the following default parameters to derive the criticality score for an open source project:

Parameter (Si) Weight (αi) Max threshold (Ti) Description Reasoning
created_since 1 120 Time since the project was created (in months) Older project has higher chance of being widely used or being dependent upon.
updated_since -1 120 Time since the project was last updated (in months) Unmaintained projects with no recent commits have higher chance of being less relied upon.
contributor_count 2 5000 Count of project contributors (with commits) Different contributors involvement indicates project's importance.
org_count 1 10 Count of distinct organizations that contributors belong to Indicates cross-organization dependency.
commit_frequency 1 1000 Average number of commits per week in the last year Higher code churn has slight indication of project's importance. Also, higher susceptibility to vulnerabilities.
recent_releases_count 0.5 26 Number of releases in the last year Frequent releases indicates user dependency. Lower weight since this is not always used.
closed_issues_count 0.5 5000 Number of issues closed in the last 90 days Indicates high contributor involvement and focus on closing user issues. Lower weight since it is dependent on project contributors.
updated_issues_count 0.5 5000 Number of issues updated in the last 90 days Indicates high contributor involvement. Lower weight since it is dependent on project contributors.
comment_frequency 1 15 Average number of comments per issue in the last 90 days Indicates high user activity and dependence.
dependents_count 2 500000 Number of project mentions in the commit messages Indicates repository use, usually in version rolls. This parameter works across all languages, including C/C++ that don't have package dependency graphs (though hack-ish). Plan to add package dependency trees in the near future.

NOTE:

  • You can override those defaut values at runtime as described below.
  • We are looking for community ideas to improve upon these parameters.
  • There will always be exceptions to the individual reasoning rules.

Usage

The program only requires one argument to run, the name of the repo:

$ go install github.com/ossf/criticality_score/cmd/criticality_score

$ criticality_score github.com/kubernetes/kubernetes
repo.name: kubernetes
repo.url: https://github.com/kubernetes/kubernetes
repo.language: Go
repo.license: Apache License 2.0
legacy.created_since: 87
legacy.updated_since: 0
legacy.contributor_count: 3999
legacy.watchers_count: 79583
legacy.org_count: 5
legacy.commit_frequency: 97.2
legacy.recent_releases_count: 70
legacy.updated_issues_count: 5395
legacy.closed_issues_count: 3062
legacy.comment_frequency: 5.5
legacy.dependents_count: 454393
default_score: 0.99107

The score can be changed by using the -scoring-config parameter and supplying a different configuration file to specify how the score is calculated.

By default the original_pike.yml configuration is used to calculate the score. However, other config files can be supplied to produce different scores. See config/scorer for more.

Feel free to copy one of the configurations and adjust the weights and thresholds to suit your needs.

Authentication

Before running criticality score, you need to:

# For posix platforms, e.g. linux, mac:
export GITHUB_AUTH_TOKEN=<your access token>

# For windows:
set GITHUB_AUTH_TOKEN=<your access token>

Formatting Results

There are three formats currently: text, json, and csv. Others may be added in the future.

These may be specified with the -format flag.

Other Commands

The criticality score project also has other commands for generating and working with criticality score data.

  • enumerate_github: a tool for accurately collecting a set of GitHub repos with a minimum number of stars
  • collect_signals: a worker for collecting raw signals at scale by leveraging the Scorecard project's infrastructure.
  • scorer: a tool for recalculating criticality scores based on an input CSV file.

Public Data

If you're interested in seeing a list of critical projects with their criticality score, we publish them in csv format and a BigQuery dataset.

This data is generated using a production instance of the criticality score project running in GCP. Details for how this is deployed can be found in the infra directory.

NOTE: Currently, these lists are derived from projects hosted on GitHub ONLY. We do plan to expand them in near future to account for projects hosted on other source control systems.

CSV data

The data is available on Google Cloud Storage and can be downloaded via:

BigQuery Dataset

This data is available in the public BigQuery dataset.

With a GCP account you can run queries across the data. For example, here is a query returning the top 100 repos by score:

  SELECT repo.url, default_score
    FROM `openssf.criticality_score_cron.criticality-score-v0-latest`
ORDER BY default_score DESC
   LIMIT 100;

Contributing

If you want to get involved or have ideas you'd like to chat about, we discuss this project in the Securing Critical Projects WG meetings.

See the Community Calendar for the schedule and meeting invitations.

See the Contributing documentation for guidance on how to contribute.

About

Gives criticality score for an open source project

Resources

License

Code of conduct

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Languages

  • Go 84.2%
  • Python 12.3%
  • Dockerfile 1.5%
  • Shell 1.2%
  • Makefile 0.8%