This page contains information on how each check works and provide remediation steps to fix the failure. All of these checks are basically "best-guesses" currently, and operate on a set of heuristics.
They are all subject to change, and have room for improvement! If you have ideas for things to add, or new ways to detect things, please contribute!
This check tries to determine if the project has generated executable (binary) artifacts in the source repository.
Including generated executables in the source repository increases user risks. Many programming language systems can generate executables from source code (e.g., C/C++ generated machine code, Java .class files, Python .pyc files, and minified JavaScript). Users will often directly use executables if they are included in the source repository, leading to many dangerous behaviors. One problem is that reviews become ineffective and thus enable the use of obsolete or maliciously-subverted executables. Reviews generally review source code, not executables, since it's difficult to audit executables to ensure that they correspond. Since it's hard to audit this, over time the included executables might not correspond to the source code. Another problem is that it allows the executable generation process to atrophy, which can lead to an inability to create working executables. These problems can be countered with verified reproducible builds, but it's easier to implement verified reproducible builds when executables are not included in the source repository (since the executable generation process is less likely to have atrophied).
It's fine to include, in the source repository, files that are simultaneously reviewable source code and executables, since these are reviewable. (Some interpretive systems, such as many operating system shells, don't have a mechanism for storing generated executables that are different from the source file.) Similarly, we allow including in the source repository source code generated by other tools (e.g., by bison, yacc, flex, and lex). There are potential downsides to generated source code, but generated source code tends to be much easier to review and thus presents a lower risk. Generated source code is also often difficult for external tools to detect.
We allow including generated documentation in source repositories. Generated documentation is intended for use by humans (not computers), and humans can take context into account. Thus, generated documentation doesn't pose the same level of risk.
A low score is therefore considered High risk.
Remediation steps
- Remove the generated executable artifacts from the repository.
- Build from source.
Branch protection can enable various rules to enforce certain workflows for branches, such as preventing rewriting of public history (e.g., a force push), requiring review before acceptance into a main branch, or passing certain status checks before acceptance into a main branch.
Branch protection can reduce the risk of unintentional or malicious code from entering the "main" branch. Branch protection rules that prevent rewriting of public history (e.g., preventing force push of public branches) prevent history from changing without external notice. Branch protection rules that require status checks ensure that at least those checks are met before a change is accepted. Branch protection rules that require at least one other reviewer reviews greatly reduces the risk that a compromise of a contributor's account will lead to injection of malicious code. Review also increases the likelihood that an unintentional vulnerability in a contribution will be detected and fixed before the change is accepted. A low score is therefore considered High risk.
Note, however, that requiring reviews for every proposed change is impractical for many projects, since many projects simply don't have that many active participants. For more discussion, see Code Reviews.
In some cases these rules will need to be suspended. For example, if a past commit includes illegal content (such as child pornography), it may be impractical to hide the commit and in such cases the history may need to be rewritten.
This check determines if the default and release branches are protected with GitHub's branch protection settings. The check only works when the token has Admin access to the repository. This check determines if the default and release branches are protected.
Remediation steps
- Enable branch protection settings in your source hosting provider to avoid force pushes or deletion of your important branches.
- For GitHub, check out the steps here.
This check tries to determine if the project runs tests before pull requests are merged.
Running tests helps developers catch mistakes early on. A low score is considered 'Low' risk.
The check works by looking for a set of CI-system names in GitHub CheckRuns and Statuses among the recent commits (~30). A CI-system is considered well-known if its name contains any of the following: appveyor, buildkite, circleci, e2e, github-actions, jenkins, mergeable, test, travis-ci. It does not currently support other source hosting repositories (forges). A project may meet this criterion yet have a failing scorecard report; there are many ways to implement this criterion, and it's challenging for an automated tool (like scorecard) to detect them all. If a project's system was not detected and you think it should be, please open an issue in the scorecard project.
Remediation steps
- Check-in scripts that run all the tests in your repository.
- Integrate those scripts with a CI/CD platform that runs it on every pull request (e.g. if hosted on GitHub, GitHub Actions, Prow, etc).
This check tries to determine if the project has earned a CII Best Practices Badge. This badge tells us if the project is applying a particular set of security-focused best development practices for open source software. The CII Best Practices badge has 3 tiers (passing, silver, and gold); we give full credit if the passing criteria are met, as achieving a passing badge is a significant achievement for many projects. We give a little credit if the project is at least working to achieve a badge, and increasingly more as more criteria are met. For example, to earn the passing badge, the project MUST publish the process for reporting vulnerabilities on the project site, it MUST provide a working build system that can automatically rebuild the software from source code (where applicable), it MUST have a general policy that tests will be added to an automated test suite when major new functionality is added, it MUST meet various cryptography criteria where applicable, it MUST have at least one primary developer who knows how to design secure software, at least one of the project's primary developers MUST know of common kinds of errors that lead to vulnerabilities in this kind of software (as well as at least one method to counter or mitigate each of them), and at least one static code analysis tool (beyond compiler warnings and "safe" language modes) MUST be applied to any proposed major production release. A low score is considered 'Low' risk. The check uses the URL for the Git repo and the CII API.
Remediation steps
- Sign up for the CII Best Practices program.
This check tries to determine if the project requires code review before pull requests (aka merge requests) are merged.
Reviewing code improves the quality of code in general, because reviews may detect various unintentional problems that can be fixed immediately before they are merged. Such problems include unintentional vulnerabilities, so unintentional vulnerabilities can be reduced through review. Reviews also make it more difficult for an attacker to insert malicious code (either as a malicious contributor or as an attacker who has subverted a contributor's account), because a reviewer might detect the subversion. This increased difficulty can even deter attackers. A low score is therefore considered High risk.
However, requiring that all proposed changes be reviewed by someone else before merging them is impractical for some projects. A project with only one active participant cannot practically enforce multi-person review, and even a project with multiple active participants may not have enough active participation to be able to require review of all proposed changes. Such projects will not be able to practically enforce rules such as requiring a minimum number of reviewers, and administrators in practice will be exempt from reviews. Projects with a small number of active participants, but more than one, instead sometimes aim for a review of a percentage of proposals (e.g., "at least half of all proposed changes are reviewed").
In addition, note that requiring review does not eliminate all risks. The other reviewer(s) might fail to notice unintentional vulnerabilities or malicious code. They might be colluding with a malicious developer. The "other" reviewers might even be the same person (aka a "sock puppet").
In short, requiring others' review before accepting a change reduces risks to users, as those additional reviewers may detect problems early, but it is not a practical requirement for some projects.
The check first tries to detect if Branch-Protection is enabled on the default branch ,and the number of reviewers is at least 1. If this fails, it checks if the recent (~30) commits have a Github-approved review or if the merger is different from the committer (implicit review). It also performs similar check for reviews using Prow (labels "lgtm" or "approved") and Gerrit ("Reviewed-on" and "Reviewed-by").
Remediation steps
- If the project has only one contributor, or does not have enough reviewers to practically require that all contributions be reviewed, try to recruit more maintainers to the project who will be willing to review others' work. Ideally at least some of these people will be from different organizations. If the project has very limited utility, consider expanding its intended utility so more people will be interested in improving it, and make that larger scope clear to potential contributors.
- Follow security best practices by performing strict code reviews for every new pull request / merge request.
- Make "code reviews" mandatory in your repository configuration. E.g. GitHub.
- Enforce the rule for administrators / code owners as well. E.g. GitHub
This check tries to determine if the project has a set of contributors from multiple organizations (e.g., companies).
Low score has 'Low' risk.
Some projects cannot meet this requirement. Obviously projects with only one active participant cannot meet it. Many projects have a narrow scope and may not be able to attract the interest of multiple organizations. See Code Reviews for added notes about projects with a small number of participants.
The check works by looking at the authors of recent commits and checking the Company field on the GitHub user profile. A contributor must have at least 5 commits in the last 30 commits. The highest score is achieved when there are contributors from at least 3 different companies in the last 30 commits. It does not currently support other source hosting repositories (forges).
Remediation steps
- There is NO remediation work needed here. This is to provide some insights on which organization(s) have contributed to the project and making trust decisions based on that. But you can ask your contributors to join their respective organizations.
This check tries to determine if the project uses a dependency update tool.
Not updating dependencies makes a project vulnerable to known flaws and prone to attacks. A low score is therefore considered High risk.
The checks looks for dependabot or renovatebot. This check only looks if it is enabled and does not ensure that it is run and pull requests are merged. A project may meet this criterion yet have a failing scorecard report; there are many ways to implement this criterion, and it's challenging for an automated tool (like scorecard) to detect them all.
Remediation steps
- Signup for automatic dependency updates with dependabot or renovatebot and place the config file in the locations that are recommended by these tools.
This check tries to determine if the project uses fuzzing. Fuzzing is important to reduce the number of vulnerabilities in code. A low score is considered 'Medium' risk. The check currently works by checking if the repo name is in the OSS-Fuzz project list. A project may meet this criterion yet have a failing scorecard report; there are many ways to implement this criterion, and it's challenging for an automated tool (like scorecard) to detect them all.
Remediation steps
- Integrate the project with OSS-Fuzz by following the instructions here.
This check tries to determine if the project is "actively maintained".
A project which is not active might not be patched, might not have its dependencies patched, or might not be actively tested and used. A low score is therefore considered High risk.
A lack of active maintenance is not necessarily a problem. Some software, especially smaller utility functions, does not normally need to be maintained. For example, a library that determines if an integer is even would not normally need maintenance (it might only need to be modified if its underlying implementation language definition changed, e.g., by adding a new type). A lack of active maintenance simply suggests that potential users should investigate further.
The check currently works by looking whether the repo is archived or not. If it is archived, it returns the minimum score. If it is not, the check looks for commits within the last 90 days, and outputs the highest score if there are at least 1 commit/week during this period.
Remediation steps
- There is NO remediation work necessarily needed here from projects; this just indicates the project activity and maintenance commitment. External users should determine if this is the kind of software that would not normally need active maintenance.
This check tries to determine if the project is published as a package that others can easily download, install, easily update, and uninstall.
It's important that the project provide an easy way to download, install, update, and uninstall the software. It's particularly important to make it easy for users to receive security patches as updates.
This is often done by creating a "package" that is easy to install and uninstall by a package manager. Many program language ecosystems have a generally-used packaging format supported by a language-level package manager tool and public package repository. Many operating system platforms also have at least one package format, tool, and public repository (in some cases the source repository generates system-independent source packages, which are then used by others to generate system executable packages). Container images are yet another way to package software. In some situations packaging is not sensible, but it's wise to package software in so many circumstances that it's worth checking for it.
A low score is considered Medium risk.
The check currently looks for GitHub packaging workflows and language-specific GitHub Actions that upload the package to a corresponding hub, e.g., Npm. It does not currently support other source hosting repositories (forges). There is a plan to add better support to query package manager hubs directly in the future, e.g., for Npm, PyPi. A project may meet this criterion yet have a failing scorecard report; some widely-used tools are relatively easy to detect, but it's challenging for an automated tool (like scorecard) to detect them all. If scorecard fails to detect the way you publish a package and you think scorecard should support your use case, please open an issue in the scorecard project.
Remediation steps
- Publish your project as a downloadable package, e.g., if hosted on GitHub, use [GitHub's mechanisms for publishing a package]](https://docs.github.com/en/packages/learn-github-packages/publishing-a-package).
- If hosted on GitHub, use a GitHub action to release your package to language-specific hubs.
This check tries to determine if the project is an application that has declared and pinned its dependencies. A "pinned dependency" is a dependency that is explicitly set to a specific version instead of allowing a range of versions. If this project is a library (not an application), this check should automatically pass (but see limitations below).
It's important for applications to pin dependencies to ensure that checking and deployment are all done with the same software, reducing deployment risks, simplifying debugging, and enabling reproducibility. In some cases pinning dependencies can help mitigate compromised dependencies from undermining the security of the project (in the case where you've evaluated the pinned dependency, you are confident it's not compromised, and a later version is released that is compromised). In particular, pinning dependencies is one way to counter dependency confusion (aka substitution) attacks. In a dependency confusion attack, an application uses multiple feeds to acquire software packages (a "hybrid configuration"), and attackers can fool the user into using a malicious package via a feed that was not expected for that package. A risk of pinning dependencies is that it can inhibit software updates (e.g., because of a security vulnerability or because the pinned version is compromised); this can be mitigated by having applications and not libraries pin to specific versions, using automated tools to notify applications when their dependencies are outdated, and by applications that do pin dependencies update quickly. Low score is therefore considered Medium risk.
The checks works by (1) looking for the following files in the root directory: go.mod, go.sum (Golang), package-lock.json, npm-shrinkwrap.json (Javascript), requirements.txt, pipfile.lock (Python), gemfile.lock (Ruby), cargo.lock (Rust), yarn.lock (package manager), composer.lock (PHP), vendor/, third_party/, third-party/; (2) looks for unpinned dependencies in Dockerfiles, shell scripts and GitHub workflows. It does not currently support other source hosting repositories (forges) other than GitHub.
Limitations: This check should only apply to applications, as libraries shouldn't normally have enforced pinned dependencies. Unfortunately, Scorecard currently can't detect if a project is a library or application. Even when application detection is added, it's always possible for an automated tool like Scorecard to incorrectly categorize software (especially in projects that include both libraries and applications).
If the project is hosted on GitHub, You can learn more about dependencies for projects using GitHub dependency graph.
Remediation steps
- First, determine if your project is producing a library or application. If it is a library, then generally you don't want to pin dependencies of library users. If your project is producing an application, consider the remediations below.
- Declare all your dependencies with specific versions in your package format file (e.g.
package.jsonfor npm,requirements.txtfor python). For C/C++, check in the code from a trusted source and add aREADMEon the specific version used (and the archive SHA hashes). - If the package manager supports lock files (e.g.
package-lock.jsonfor npm), make sure to check these in the source code as well. These files maintain signatures for the entire dependency tree and saves from future exploitation in case the package is compromised. - For Dockerfiles and GitHub workflows, pin dependencies by hash. See example main.yaml and Dockerfile examples.
- To help update your dependencies after pinning them, use tools such as Github's dependabot or renovate bot.
This check tries to determine if the project uses static code analysis.
SAST tool may prevent known classes of bugs to be inadvertently introduced in the codebase. A low score is considered Medium risk.
The checks currently looks for known Github apps such as github-code-scanning (codeql) and sonarcloud in the recent (~30) merged PRs. The check also looks for the use of "github/codeql-action" in a GitHub workflow. It does not currently support other source hosting repositories (forges) other than GitHub. A project may meet this criterion yet have a failing scorecard report; there are many ways to implement this criterion and it's especially difficult for an automated tool (like scorecard) to detect them all.
Remediation steps
- Run CodeQL checks in your CI/CD by following the instructions here.
This check tries to determine if the project has published a security policy. It works by looking for a file named SECURITY.md (case-insensitive) in a few well-known directories.
Remediation steps
- Place a security policy file
SECURITY.mdin the root directory of your repository. This makes it easily discoverable by a vulnerability reporter. - The file should contain information on what constitutes a vulnerability and a way to report it securely (e.g. issue tracker with private issue support, encrypted email with a published public key).
This check tries to determine if the project cryptographically signs release artifacts. Signed releases attest to the provenance of the artifact. A low score is considered 'High' risk. It works by looking for filenames: *.minisig (https://github.com/jedisct1/minisign), *.asc (pgp), *.sign. for the last 5 releases if it's hosted on GitHub. The check does not verify the signatures. It does not currently support other source hosting repositories (forges) other than GitHub.
Remediation steps
- Publish the release.
- Generate a signing key.
- Download the release as an archive locally.
- Sign the release archive with this key (should output a signature file).
- Attach the signature file next to the release archive.
- If the source is hosted on GitHub, check out the steps here.
This check tries to determine if the project's automated workflows follow the principle of least privilege, i.e. if the tokens are set read-only by default.
Attackers may use a compromised token with write access to push malicious code into the project. A low score is therefore considered High risk.
The check looks at the GitHub workflows. For each workflow yaml file, for the permission definitions. To obtain the highest score, the permissions should be set as read-only at the top level and the required write permissions should be declared at the run-level. The check cannot detect if the "read-only" GitHub permission settings is enabled, as there is no API available. It does not currently support other source hosting repositories (forges) other than GitHub.
Remediation steps
- Set permissions as
read-allorcontents: readas described in GitHub's documentation.
This check determines if the project has open, unfixed vulnerabilities using the OSV service.
An existing vulnerability is can readily be used by attackers, so a low score is considered High risk.
Remediation steps
- Fix the vulnerabilities. The details of each vulnerability can be found on https://osv.dev.