GitHub Code Quality analyzes your code for quality and coverage issues and delivers Copilot-powered fixes you can apply in one click. It runs in two places:
- On pull requests, findings appear as inline comments before code is merged. If you upload a Cobertura XML coverage report, coverage metrics show whether a change maintains or reduces coverage. You can enforce quality and coverage thresholds with rulesets to block pull requests that don't meet your criteria, so new quality debt doesn't accumulate.
- On the default branch, scans identify existing quality debt across your codebase, with autofixes you can apply directly or assign to Copilot クラウドエージェント to resolve on your behalf.
Detection combines deterministic CodeQL rules for known anti-patterns with AI-powered analysis for issues that fall outside existing rule sets, including languages not yet covered by CodeQL queries.
Use cases
Here's what GitHub Code Quality looks like in practice.
For developers and teams:
- A developer opens a pull request that introduces a reliability or maintainability issue. Code Quality posts a comment explaining the issue and offers a one-click fix before the code is merged. The developer also sees a report of coverage metrics, and can tell at a glance whether the pull request improves or reduces coverage compared to the default branch.
- A team adopts AI coding assistants and needs assurance that generated code meets the same bar as hand-written code. AI-powered analysis catches issues that rule-based queries weren't written for, while CodeQL rules cover well-defined anti-patterns.
- A team inherits a large codebase with years of accumulated quality debt. Code Quality scans the default branch, surfaces findings with autofixes on a dashboard, and the team assigns remediation work to Copilot クラウドエージェント to open fix pull requests automatically.
For administrators and leads:
- An engineering lead sets coverage and quality thresholds using rulesets. Pull requests that don't meet the criteria are blocked from merging, so no new quality or coverage debt accumulates.
- An administrator needs visibility across repositories for audits or compliance reporting. Code Quality reports through the security overview alongside security tools, so they can see quality posture across the organization at a glance, identify which repositories need attention, and track improvement metrics using standard GitHub audit controls and policies.
Availability and billing
Usage costs are determined by:
- A per-seat license fee based on active committers.
- AI-powered detections and Copilot-powered autofixes, which consume GitHub AI Credits (no Copilot license required).
- GitHub Actions minutes for deterministic CodeQL scans, if you don't use self-hosted runners.
Optional features, such as delegating code quality remediation work to Copilot, require a Copilot license.
For more information, see GitHub Code Quality 請求.
Supported languages
Code Quality performs rule-based analysis of the following languages using CodeQL:
- C#
- Go
- Java
- JavaScript
- Python
- Ruby
- TypeScript
It also performs AI-powered analysis on pull requests and on your repository's recently changed code, including languages beyond those supported by rule-based queries.
Next steps
- For your enterprise: Ensure repositories in your enterprise can enable Code Quality. See GitHub Code Quality を企業で使用できるようにする.
- For your repository or organization: Turn on Code Quality to start generating results. See GitHub Code Quality の有効化.
- On your pull request: Learn how to fix code quality findings on your pull request. See Preventing code quality issues from reaching your default branch.