QualiGuard: An AI-driven risk-based quality gate for defect prediction and code smell detection in DevOps pipelines


Kocakaplan B. M., BORANDAĞ E., ÖZÇEVİK Y., YÜCALAR F., ALTAY O.

SOFTWAREX, cilt.36, 2026 (SCI-Expanded, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 36
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.softx.2026.103090
  • Dergi Adı: SOFTWAREX
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Directory of Open Access Journals
  • Manisa Celal Bayar Üniversitesi Adresli: Evet

Özet

Modern DevOps pipelines rely on continuous quality monitoring to detect defects before release. We propose QualiGuard, an AI-driven risk-based quality gate that integrates code smell detection with file-level defect prediction. The tool mines GitHub repositories, software metrics, derives defect labels using the SZZ algorithm, and detects seven code smells through AST analysis. Evaluated on 62,207 files from 1000 Python projects, a hybrid stacking ensemble achieved an F1 score of 0.64 for defect prediction, while an optimized LightGBM model reached 0.69 for code smell prediction. Calibrated risk scores support PASS/REVIEW/BLOCK quality-gate decisions through a Flask-based web interface.