This page answers frequently asked questions about Scorecard, including its purpose, usage, and checks. This page is continually updated. If you would like to add a question, please [contribute](../CONTRIBUTING.md)!
- [Fuzzing: Does Scorecard accept custom fuzzers?](#fuzzing-does-scorecard-accept-custom-fuzzers)
- [Pinned-Dependencies: Will Scorecard detect unpinned dependencies in tests with Dockerfiles?](#pinned-dependencies-will-scorecard-detect-unpinned-dependencies-in-tests-with-dockerfiles)
Over a million projects are automatically tracked by the Scorecard project. Use the webviewer to see these scores, replacing the placeholder text with the platform, user/org, and repository name: https://scorecard.dev/viewer/?uri=<github_or_gitlab>.com/<user_name_or_org>/<repository_name>.
Most code scanning tools are focused on detecting specific vulnerabilities already existing in your codebase. Scorecard, however, is focused on improving the project's overall security posture by helping it adopt best practices. The best solution for your project may well be to adopt Scorecard along with other tools!
### Wasn't this project called "Scorecards" (plural)?
Yes, kind of. The project was initially called "Security Scorecards" but that form wasn't used consistently. In particular, the repo was named "scorecard" and so was the program. Over time people started referring to either form (singular and plural) and the inconsistency became prevalent. To end this situation the decision was made to consolidate over the use of the singular form in keeping with the repo and program name, drop the "Security" part and use "OpenSSF" instead to ensure uniqueness. One should therefore refer to this project as "OpenSSF Scorecard" or "Scorecard" for short.
While it isn't currently possible to allowlist such binaries, the Scorecard team is working on this feature ([#1270](https://github.com/ossf/scorecard/issues/1270)).
This is quite a complex question. Right now, there is no way to do that. Here are some pros and cons of allowing users to set up an ignore-list for bots.
- Pros: Some bots run very frequently; for some projects, reviewing every change is therefore not feasible or reasonable.
- Cons: Bots can be compromised (their credentials can be compromised, for example). Or if commits are not signed, an attacker could easily send a commit spoofing the bot. This means that a bot having unsupervised write access to the repository could be a security risk.
Both Dependabot and Renovate bot won't update your dependencies immediately. They have some precautions to make sure a release is reasonable / won't break your build (see [Dependabot compatibility documentation](https://docs.github.com/en/code-security/dependabot/dependabot-security-updates/about-dependabot-security-updates#about-compatibility-scores)).
You can either configure the tools to only update your dependencies once a week or once a month. This way, if a malicious version is released, it's very likely that it'll be reported and removed before it even gets suggested to you. Besides, there's also the benefit that it gives you the chance to validate the new release before merging if you want to.
Another configuration possibility that would limit even more the release updates only to trusted releases is enabling to only perform Security Updates, which means you only be notified about releases that fixes a previous vulnerability you might be exposed to.
Scorecard can show the dependencies that are referred to in tests like Dockerfiles, so it could be a great way for you to fix those dependencies and avoid the vulnerabilities related to version pinning dependencies. To see more about the benefits of hash pinning instead of version pinning, please see the [Pinned-Dependencies check description](checks.md#pinned-dependencies)
Version pinning is a significant improvement over not pinning your dependencies. However, it still leaves your project vulnerable to tag-renaming attacks (where a dependency's tags are deleted and recreated to point to a malicious commit).
The OpenSSF therefore recommends hash pinning instead of version pinning, along with the use of dependency update tools such as Dependabot to keep your dependencies up-to-date.
Currently, the main benefit of [signed releases](checks.md#signed-releases) is the guarantee that a specific artifact was released by a source that you approve or attest is reliable.
However, there are already moves to make it even more relevant. For example, the OpenSSF is working on [implementing signature verification for NPM packages](https://github.blog/2022-08-08-new-request-for-comments-on-improving-npm-security-with-sigstore-is-now-open/) which would allow a consumer to automatically verify if the package they are downloading was generated through a reliable builder and if it is correctly signed.
### Branch-Protection: How to setup a 10/10 branch protection on GitHub?
To get a 10/10 score for Branch-Protection check using a non-admin token, you should have the following settings for your branches:
When using an admin token, Scorecard can verify if a few other important settings are ensured:
It's important to reiterate that Branch-Protection score is Tier-based. If a setting from Tier 1 is not satisfied, it does not matter that all other settings are met, the score will be truncated up the Tier's maximum. In this case, 3/10. The following table shows the relation between branch protection settings on GitHub and the score Tier:
| Name | Status | Required only for admin token | Tier |
