第 2 阶段：准备大规模启用

Preparing to enable code scanning

Code scanning is a feature that you use to analyze the code in a GitHub repository to find security vulnerabilities and coding errors. Any problems identified by the analysis are shown in your repository. For more information, see About code scanning.

Rolling code scanning out across hundreds of repositories can be difficult, especially when done inefficiently. Following these steps will ensure your rollout is both efficient and successful. As part of your preparation, you will work with your teams, use automation to collect data about your repositories, and enable code scanning.

Preparing teams for code scanning

First, prepare your teams to use code scanning. The more teams that use code scanning, the more data you'll have to drive remediation plans and monitor progress on your rollout. During this phase, you should focus on leveraging APIs and running internal enablement events.

Your core focus should be preparing as many teams to use code scanning as possible. You can also encourage teams to remediate appropriately, but we recommend prioritizing enablement and use of code scanning over fixing issues during this phase.

Collecting information about your repositories

You can programmatically gather information about the different programming languages used in your repositories and use that data to enable code scanning on all repositories that use the same language, using GitHub Enterprise Server's GraphQL API.

If you want to gather information from repositories belonging to multiple organizations in your enterprise, you can use the query below to obtain the names of your organizations and then feed those into repository query. Replace OCTO-ENTERPRISE with your enterprise name.

query {
  enterprise(slug: "OCTO-ENTERPRISE") {
    organizations(first: 100) {
      totalCount
      nodes {
        name
      }
      pageInfo {
        endCursor
        hasNextPage
      }
    }
  }
}

You can identify which repositories use which languages by collating repositories by language at the organization level. You can modify the sample GraphQL query below, replacing OCTO-ORG with the organization name.

query {
  organization(login: "OCTO-ORG") {
    repositories(first: 100) {
      totalCount
      nodes {
        nameWithOwner
        languages(first: 100) {
          totalCount
          nodes {
            name
          }
        }
      }
      pageInfo {
        endCursor
        hasNextPage
      }
    }
  }
}

For more information about running GraphQL queries, see Forming calls with GraphQL.

Then, convert the data from the GraphQL query into a readable format, such as a table.

Language	Number of Repos	Name of Repos
JavaScript (TypeScript)	4212	org/repo org/repo
Python	2012	org/repo org/repo
Go	983	org/repo org/repo
Java	412	org/repo org/repo
Swift	111	org/repo org/repo
Kotlin	82	org/repo org/repo
C	12	org/repo org/repo

You can filter out the languages that are currently not supported by GitHub Advanced Security from this table.

If you have repositories with multiple languages, you can format the GraphQL results as shown in the table below. Filter out languages that are not supported, but retain all repositories with at least one supported language. You can enable code scanning on these repositories, and all supported languages will be scanned.

Language(s)	Number of Repos	Name of Repos
JavaScript/Python/Go	16	org/repo org/repo
Rust/TypeScript/Python	12	org/repo org/repo

An understanding of which repositories are using which languages will help you identify candidate repositories for pilot programs in phase 3, and prepares you to enable code scanning across all repositories, one language at a time, in phase 5.

Enabling code scanning for your appliance

Before you can proceed with pilot programs and rolling out code scanning across your enterprise, you must first enable code scanning for your appliance. For more information, see Configuring code scanning for your appliance.

Preparing to enable secret scanning

If a project communicates with an external service, it might use a token or private key for authentication. If you check a secret into a repository, anyone who has read access to the repository can use the secret to access the external service with your privileges. Secret scanning will scan your entire Git history on all branches present in your GitHub repositories for secrets and alert you or block the push containing the secret. For more information, see About secret scanning.

Considerations when enabling secret scanning

GitHub Enterprise Server’s secret scanning capability is slightly different from code scanning since it requires no specific configuration per programming language or per repository and less configuration overall to get started. This means enabling secret scanning at the organizational level can be easy, but clicking Enable All at the organization level and selecting the option Automatically enable secret scanning for every new repository has some downstream effects that you should be aware of:

License consumption

Enabling secret scanning for all repositories will maximize your use of GitHub Advanced Security licenses. This is fine if you have enough licenses for the current committers to all those repositories. If the number of active developers is likely to increase in the coming months, you may exceed your license limit and then be unable to use GitHub Advanced Security on newly created repositories.

Initial high volume of detected secrets

If you are enabling secret scanning on a large organization, be prepared to see a high number of secrets found. Sometimes this comes as a shock to organizations and the alarm is raised. If you would like to turn on secret scanning across all repositories at once, plan for how you will respond to multiple alerts across the organization.

Secret scanning can be enabled for individual repositories. For more information, see Enabling secret scanning for your repository. Secret scanning can also be enabled for all repositories in your organization, as described above. For more information on enabling for all repositories, see Managing security and analysis settings for your organization.

Custom patterns for secret scanning

Secret scanning detects a large number of default patterns but can also be configured to detect custom patterns, such as secret formats unique to your infrastructure or used by integrators that GitHub Enterprise Server's secret scanning does not currently detect. For more information about supported secrets for partner patterns, see Supported secret scanning patterns.

As you audit your repositories and speak to security and developer teams, build a list of the secret types that you will later use to configure custom patterns for secret scanning. For more information, see Defining custom patterns for secret scanning.

Push protection for secret scanning

Push protection for organizations and repositories instructs secret scanning to check pushes for supported secrets before secrets are committed to the codebase. For information on which secrets are supported, see Supported secret scanning patterns.

If a secret is detected in a push, that push is blocked. Secret scanning lists any secrets it detects so the author can review the secrets and remove them or, if needed, allow those secrets to be pushed. Secret scanning can also check pushes for custom patterns.

Developers have the option to bypass push protection by reporting that a secret is a false positive, that it is used in tests, or that it will be fixed later.

If a contributor bypasses a push protection block for a secret, GitHub:

• Creates an alert in the Security tab of the repository.
• Adds the bypass event to the audit log.
• Sends an email alert to organization or personal account owners, security managers, and repository administrators who are watching the repository, with a link to the secret and the reason why it was allowed.

Before enabling push protection, consider whether you need to create guidance for developer teams on the acceptable conditions for bypassing push protection. You can configure a link to this resource in the message that's displayed when a developer attempts to push a blocked secret.

Next, familiarize yourself with the different options for managing and monitoring alerts that are the result of a contributor bypassing push protection.

For more information, see About push protection.