Interpreting the code quality results for your repository

View GitHub Code Quality findings for your default branch.

Who can use this feature?

Users with write access

GitHub Team or GitHub Enterprise Cloud

In this article

Note

GitHub Code Quality is currently in public preview and subject to change. During public preview, Code Quality will not be billed, although Code Quality scans will consume GitHub Actions minutes.

Prerequisites

Viewing the full backlog of code quality results

  1. Navigate to the Security and quality tab of your repository.
  2. Click to expand Code quality, then click Standard findings.

Alternatively, if you want to view AI-powered findings for the most recently changed files, see Improving the quality of recently merged code with AI.

Exploring the backlog for your repository

The "Standard findings" dashboard shows all the results found by CodeQL analysis on the default branch of your repository. This view helps you visualize the full backlog of quality results and prioritize work to fix specific types of problems.

The overview, at the top of the page, summarizes the maintainability and reliability of the codebase.

Screenshot of the "Standard findings" dashboard for code quality results. The summary is outlined in dark orange.

Underneath the overview, the full list of results is shown with a header with filters that you can use to focus on a specific set of findings. The results are:

  • Grouped by the rule that detected each finding
  • Within each rule, ordered by file path alphabetically

Explore the results by expanding a rule to list the affected files and clicking on the name of a rule to see full details of the findings.

Screenshot of the Rules table on the "Standard findings" dashboard for code quality. The "Overwritten property" rule name is outlined in dark orange.

Interpreting ratings and metrics

Code quality results should always be interpreted in the context of your repository. For example:

  • Small repositories, or repositories with only a small amount of code written in supported languages, tend to have few results and good ratings.
  • Repositories with a lot of generated code may have many maintenance results, lowering the rating for maintainability. This is not a problem if the source code itself is maintainable.
  • Large repositories with a lot of code in a fully supported language often have many results even if the majority of the code has good maintainability and reliability standards.

To learn more about the metrics and how the ratings are calculated, see Metrics and ratings reference.

Viewing code coverage on pull requests

After code coverage is configured for your repository, the github-code-quality[bot] posts a coverage summary comment on each pull request. The comment includes:

  • Branch-vs-default-branch comparison: The aggregate coverage percentage for the pull request branch and the default branch (for example, "65% on pull request branch, 44% on default branch").
  • Most impacted files: The 10 files with the largest coverage changes between the default branch and the pull request branch. This list may include files not directly modified in the pull request. New or modified files are ranked above deleted files. Within each group, files are sorted by the absolute change in line coverage weighted by the number of lines in the file, with coverage magnitude and then filename as tiebreakers.
  • Per-file breakdown: An expandable section listing each file with its coverage percentage and delta value. A positive delta means the file gained coverage on this branch. A negative delta indicates coverage decreased, which may signal untested code paths introduced by the change.

Use the coverage summary to identify files with low or declining coverage and prioritize review attention on untested changes.

For more information about how the coverage percentage is calculated, see Metrics and ratings reference.

Next steps