GitHub Adds AI-Based Scanning to Code Security

· Original article ↗

Summary

GitHub plans to add AI-based scanning to Code Security, supplementing CodeQL to cover more languages and frameworks. Public preview is expected in early Q2 2026; GitHub also reports internal testing and Copilot Autofix results.

Key points

  • AI detections will supplement CodeQL, which GitHub says will continue providing deep semantic analysis for supported languages.
  • The expanded coverage is expected to include Shell/Bash, Dockerfiles, Terraform, PHP, and other ecosystems.
  • GitHub expects the hybrid scanning model to enter public preview in early Q2 2026.
  • GitHub says internal testing processed more than 170,000 findings in 30 days, with 80% positive developer feedback.
  • GitHub reports that Copilot Autofix users resolved security alerts in an average of 0.66 hours, compared with 1.29 hours without Autofix.

Article Details

Event Type
GitHub announced AI-based vulnerability detection for Code Security, complementing CodeQL in a hybrid scanning model. Public preview is expected in early Q2 2026.
Impact
The system is intended to broaden code-scanning coverage to Shell/Bash, Dockerfiles, Terraform, PHP, and other ecosystems, with findings surfaced in pull requests alongside Copilot Autofix suggestions. GitHub reported that internal testing processed more than 170,000 findings in 30 days with 80% positive developer feedback. For 2025, GitHub reported average resolution times of 0.66 hours with Copilot Autofix versus 1.29 hours without it.

Vendors

Products

Related Articles