Improving the quality of your repository's code

Introduction

This tutorial guides you through using GitHub Code Quality to review, prioritize, and remediate code health issues across your repository — helping you systematically reduce technical debt, improve reliability and maintainability, and communicate your impact to stakeholders.

Prerequisites

• Code Quality is enabled for your repository. See Enabling GitHub Code Quality.
• If you're enabling GitHub Code Quality for the first time, ensure you've waited a few minutes after enablement for a full CodeQL scan of the default branch to complete.

1. Assess your repository's overall code health

1. Navigate to the "Security" tab of your repository, then under "コードの品質", click 標準の結果.

2. The overview on the "標準の結果" dashboard gives you an immediate assessment of the state of your default branch today:

   • Maintainability rating reflects the presence and severity of findings for dead code, duplication, complexity, missing documentation, and failure to follow best practices.
   • Reliability rating reflects the presence and severity of findings for correctness, performance, error handling, concurrency, and accessibility of your code.
   

2. Identify and prioritize the most impactful findings

On the "標準の結果" view, you'll see the list of results from Code Quality's last scan of the default branch of the repository. These findings are:

• Grouped by rule, so you can see which types of problem most affect your codebase.
• Assigned a severity level ("Error", "Warning", "Note").

Focus on high severity findings

Use the dashboard filters to focus on the highest-severity results first ("Errors"), and review which rules generate the most issues.

To improve your repository's maintainability or reliability rating, you must resolve (fix or dismiss) all findings with the highest severity level for that metric.

For example, to improve your repository's "Reliability" metric from Needs improvement to Fair, you would need to address and resolve all error-level findings that impact reliability. If you have one or more error-level findings, your rating cannot be higher than "Needs improvement". See メトリックと評価のリファレンス.

3. Investigate a group of findings and understand context

Once you've identified a rule with multiple results that you want to address, you can investigate further to understand the underlying problems.

1. Click the rule name to be taken to a detailed view of all findings for that rule.

   

2. Click Show more, then review the explanation of the rule, what the recommended fix is, supporting code examples and references.

   

4. Choose remediation options

Evaluate all the highlighted findings for validity, impact, and risk. To improve your quality rating, you need to resolve each finding by either choosing to fix or dismiss it.

Generate an autofix

If the finding looks valid and relevant for your codebase, you can generate a suggested fix.

1. To the right of an individual finding, click Generate fix.

2. Review carefully the diff of the proposed change, and if you agree with it, click Open pull request.

3. In the "Commit autofix to branch" dialog box, select "Open a pull request", then click Commit change.

   ヒント

   It's not currently possible to generate autofixes for a group of findings in bulk.

   If you want to address multiple findings with a single pull request, repeat steps 1 and 2 above, then in the "Commit autofix to branch" dialog box, use the branch name you already created for the first autofix, then select "Open pull request" and Commit change.

   The fix will be added to the existing draft pull request for your branch.

4. When you're ready, change the pull request status from "Draft" to "Ready for review", and carefully review the proposed changes. Wait for any CI checks and automated tests to complete and pass before merging the pull request.

Dismiss a finding

コードベースのコンテキストで関連性がない場合、またはアクション可能でない場合は、結果を無視できます。 結果を無視する一般的な理由は次のとおりです。

• この結果は、管理されなくなったレガシ コードにあります。
• これは、チームのコーディング標準に対する既知の例外です。
• これは、実際の品質リスクをもたらさない偽陽性です。

無関係なアラートを無視すると、品質チェックが意味のある問題に焦点を当て続けます。

1. To dismiss a finding, click .
2. The finding will disappear from the list of open findings. You can still review and reopen dismissed findings from under the "Dismissed" tab at the top of the page.

5. Measure improvement and communicate impact

After remediation work is complete, return to the "標準の結果" dashboard to review the updated reliability and maintainability metrics.

When communicating your impact to stakeholders, highlight:

• Any reduction in the number of findings for "Reliability" or "Maintainability".
• Any change in rating for the Reliability or Maintainability ratings.
• The requirement(s) that has been met to achieve the change in rating. For example, the remediation of all "Warning"-level findings caused the rating to change from "Fair" to "Good".

Use the improvements in quality ratings and reduction in number of findings to demonstrate progress.

6. Enforce code quality standards for pull requests

If you haven't already, set up quality thresholds for pull requests, to block any changes to the codebase that will reduce the health of your codebase. See Setting code quality thresholds for pull requests.

Next steps

• Reduce technical debt further by fixing findings in recently changed files. See Improving the quality of recently merged code with AI.
• Provide feedback on GitHub Code Quality in the community discussion.