The potential of LLM-generated reports in DevSecOps