This version of GitHub Enterprise was discontinued on 2021-09-23. No patch releases will be made, even for critical security issues. For better performance, improved security, and new features, upgrade to the latest version of GitHub Enterprise. For help with the upgrade, contact GitHub Enterprise support.

Running CodeQL code scanning in your CI system

You can use the CodeQL runner to perform CodeQL code scanning in a third-party continuous integration system.

Code scanning is available if you have a license for GitHub Advanced Security.

Note: The CodeQL runner is being deprecated. Please use the CodeQL CLI version 2.6.2 or greater instead. GitHub Enterprise Server 3.3 will be the final release series that supports the CodeQL runner. On GitHub Enterprise Cloud, the CodeQL runner will be supported until March 2022. For more information, see the CodeQL runner deprecation.

Note: Code scanning is in beta in GitHub Enterprise Server 2.22. For the generally available release of code scanning, upgrade to the latest release of GitHub Enterprise Server.

Note: Your site administrator must enable code scanning for your GitHub Enterprise Server instance before you can use this feature. For more information, see "Configuring code scanning for your appliance."

Using CodeQL code scanning with your existing CI system

If you use a continuous integration or continuous delivery/deployment (CI/CD) system other than GitHub Actions, you can use your existing system to run GitHub's CodeQL analysis and upload the results to GitHub. To do this, use the CodeQL runner.

About the CodeQL runner

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 GitHub Enterprise Server. For information, see "About code scanning."

You can use the CodeQL runner to run code scanning on code that you're processing in a third-party continuous integration (CI) system. Alternatively, you can use GitHub Actions to run code scanning on GitHub Enterprise Server. For information, see "Setting up code scanning for a repository."

The CodeQL runner is a command-line tool that runs CodeQL analysis on a checkout of a GitHub repository. You add the runner to your third-party system, then call the runner to analyze code and upload the results to GitHub Enterprise Server. These results are displayed as code scanning alerts in the repository.

Notes:

  • The CodeQL runner is available to customers with an Advanced Security license.
  • The CodeQL runner shouldn't be confused with the CodeQL CLI. The CodeQL CLI is an interactive command-line interface that lets you create CodeQL databases for security research and run CodeQL queries. For more information, see "CodeQL CLI."

Downloading the CodeQL runner

You can download the CodeQL runner from https://github.com/github/codeql-action/releases. On some operating systems, you may need to change permissions for the downloaded file before you can run it.

On Linux:

chmod +x codeql-runner-linux

On macOS:

chmod +x codeql-runner-macos
sudo xattr -d com.apple.quarantine codeql-runner-macos

On Windows, the codeql-runner-win.exe file usually requires no change to permissions.

Adding the CodeQL runner to your CI system

Once you download the CodeQL runner and verify that it can be executed, you should make the runner available to each CI server that you intend to use for code scanning. For example, you might configure each server to copy the runner from a central, internal location. Alternatively, you could use the REST API to get the runner directly from GitHub, for example:

wget https://github.com/github/codeql-action/releases/latest/download/codeql-runner-linux
chmod +x codeql-runner-linux

In addition to this, each CI server also needs:

  • A GitHub App or personal access token for the CodeQL runner to use. You must use an access token with the repo scope, or a GitHub App with the security_events write permission, and metadata and contents read permissions. For information, see "Building GitHub Apps" and "Creating a personal access token."
  • Access to the CodeQL bundle associated with this release of the CodeQL runner. This package contains queries and libraries needed for CodeQL analysis, plus the CodeQL CLI, which is used internally by the runner. For information, see "CodeQL CLI."

The options for providing access to the CodeQL bundle are:

  1. Allow the CI servers access to https://github.com/github/codeql-action so that the CodeQL runner can download the bundle automatically.
  2. Mirror the github/codeql-action repository on GitHub Enterprise Server. Unless you specify the --codeql-path flag, the runner automatically checks for the bundle in this location and on GitHub.com.
  3. Manually download/extract the bundle, store it with other central resources, and use the --codeql-path flag to specify the location of the bundle in calls to initialize the CodeQL runner.

Calling the CodeQL runner

You should call the CodeQL runner from the checkout location of the repository you want to analyze. The two main commands are:

  1. init required to initialize the runner and create a CodeQL database for each language to be analyzed. These databases are populated and analyzed by subsequent commands.
  2. analyze required to populate the CodeQL databases, analyze them, and upload results to GitHub Enterprise Server.

For both commands, you must specify the URL of GitHub Enterprise Server, the repository OWNER/NAME, and the GitHub Apps or personal access token to use for authentication. You also need to specify the location of the CodeQL bundle, unless the CI server has access to download it directly from the github/codeql-action repository.

You can configure where the CodeQL runner stores the CodeQL bundle for future analysis on a server using the --tools-dir flag and where it stores temporary files during analysis using --temp-dir.

To view the command-line reference for the runner, use the -h flag. For example, to list all commands run: codeql-runner-OS -h, or to list all the flags available for the init command run: codeql-runner-OS init -h (where OS varies according to the executable that you are using). For more information, see "Configuring code scanning in your CI system."

Notes:

  • SARIF upload supports a maximum of 1000 results per upload. Any results over this limit are ignored. If a tool generates too many results, you should update the configuration to focus on results for the most important rules or queries.

  • For each upload, SARIF upload supports a maximum size of 10 MB for the gzip-compressed SARIF file. Any uploads over this limit will be rejected. If your SARIF file is too large because it contains too many results, you should update the configuration to focus on results for the most important rules or queries.

Basic example

This example runs CodeQL analysis on a Linux CI server for the octo-org/example-repo repository hosted on https://github.example.com. The process is very simple because the repository contains only languages that can be analyzed by CodeQL directly, without being built (that is, Go, JavaScript, Python, and TypeScript).

In this example, the server has access to download the CodeQL bundle directly from the github/codeql-action repository, so there is no need to use the --codeql-path flag.

  1. Check out the repository to analyze.

  2. Move into the directory where the repository is checked out.

  3. Initialize the CodeQL runner and create CodeQL databases for the languages detected.

    $ /path/to-runner/codeql-runner-linux init --repository octo-org/example-repo
        --github-url https://github.example.com --github-auth TOKEN
    > Cleaning temp directory /srv/checkout/example-repo/codeql-runner
    > ...
    > Created CodeQL database at /srv/checkout/example-repo/codeql-runner/codeql_databases/javascript.
  4. Populate the CodeQL databases, analyze them, and upload the results to GitHub Enterprise Server. The results will appear in the Security tab for your repository.

    $ /path/to-runner/codeql-runner-linux analyze --repository octo-org/example-repo
        --github-url https://github.example.com --github-auth TOKEN
        --commit 5b6a3078b31dc346e5ce7b86837d6abbe7a18bbd --ref refs/heads/my-branch
    > Finalizing database creation
    > ...
    > POST /repos/octo-org/example-repo/code-scanning/sarifs - 202 in 786ms
    > Successfully uploaded results
  5. To upload code scanning results as pull request checks, specify the pull request using the --ref flag. We recommend setting up the CodeQL runner so that it runs on the pull_request webhook event.

    $ /path/to-runner/codeql-runner-linux analyze --repository octo-org/example-repo
        --github-url https://github.example.com --github-auth TOKEN
        --commit 1dc7a1346e5ce7b86835b68bbda3078b37d6abbe --ref refs/pull/123/merge
    > Finalizing database creation
    > ...
    > POST /repos/octo-org/example-repo/code-scanning/sarifs - 202 in 786ms
    > Successfully uploaded results

For more information about viewing code scanning alerts, see "Triaging code scanning alerts in pull requests" and "Managing code scanning alerts for your repository."

Compiled language example

This example is similar to the previous example, however this time the repository has code in C/C++, C#, or Java. To create a CodeQL database for these languages, the CLI needs to monitor the build. At the end of the initialization process, the runner reports the command you need to set up the environment before building the code. You need to run this command, before calling the normal CI build process, and then running the analyze command.

  1. Check out the repository to analyze.

  2. Move into the directory where the repository is checked out.

  3. Initialize the CodeQL runner and create CodeQL databases for the languages detected.

    $ /path/to-runner/codeql-runner-linux init --repository octo-org/example-repo-2
        --github-url https://github.example.com --github-auth TOKEN
    > Cleaning temp directory /srv/checkout/example-repo-2/codeql-runner
    > ...
    > CodeQL environment output to "/srv/checkout/example-repo-2/codeql-runner/codeql-env.json"
      and "/srv/checkout/example-repo-2/codeql-runner/codeql-env.sh".
      Please export these variables to future processes so that CodeQL can monitor the build, for example by running 
      ". /srv/checkout/example-repo-2/codeql-runner/codeql-env.sh".
  4. Source the script generated by the init action to set up the environment to monitor the build. Note the leading dot and space in the following code snippet.

    $ . /srv/checkout/example-repo-2/codeql-runner/codeql-env.sh
  5. Build the code. On macOS, you need to prefix the build command with the environment variable $CODEQL_RUNNER. For more information, see "Troubleshooting CodeQL code scanning in your CI system."

  6. Populate the CodeQL databases, analyze them, and upload the results to GitHub Enterprise Server. The results will appear in the Security tab for your repository.

    $ /path/to-runner/codeql-runner-linux analyze --repository octo-org/example-repo
        --github-url https://github.example.com --github-auth TOKEN
        --commit 5b6a3078b31dc346e5ce7b86837d6abbe7a18bbd --ref refs/heads/my-branch
    > Finalizing database creation
    > ...
    > POST /repos/octo-org/example-repo/code-scanning/sarifs - 202 in 786ms
    > Successfully uploaded results
  7. To upload code scanning results as pull request checks, specify the pull request using the --ref flag. We recommend setting up the CodeQL runner so that it runs on the pull_request webhook event.

    $ /path/to-runner/codeql-runner-linux analyze --repository octo-org/example-repo
        --github-url https://github.example.com --github-auth TOKEN
        --commit 1dc7a1346e5ce7b86835b68bbda3078b37d6abbe --ref refs/pull/123/merge
    > Finalizing database creation
    > ...
    > POST /repos/octo-org/example-repo/code-scanning/sarifs - 202 in 786ms
    > Successfully uploaded results

For more information about viewing code scanning alerts, see "Triaging code scanning alerts in pull requests" and "Managing code scanning alerts for your repository."

Note: If you use a containerized build, you need to run the CodeQL runner in the container where your build task takes place.

Further reading